Award Winners and Abstracts of the 30Th Anniversary Symposium of the Protein Society, Baltimore, MD, July 16-19, 2016

Award Winners and Abstracts of the 30th Anniversary Symposium of The Protein Society, Baltimore, MD, July 16-19, 2016

4 ABSTRACT

2016 Award Winners BALTIMORE, MD – The Protein Society, the premiere international society dedicated to supporting protein research, announces the winners of The 2016 Protein Society Awards. The awards were conferred at the 30th Anniversary Symposium of The Protein Society (July 16-19, 2016, Bal- timore, Maryland USA). Plenary talks from each recipient were scheduled throughout the 3.5 day event.

The Carl Brand€ en Award The Carl Brand€ en Award, sponsored by Rigaku Corporation, honors an outstanding protein scientist who has also made exceptional contributions in the areas of education and/or service to the field. The 2016 recipient of this award is Dr. Gary Pielak (University of North Carolina at Chapel Hill). Dr. Pielak has advanced the field of protein chemistry through pioneering research in unraveling protein biophysics in living cells. Dr. Pielak and his students developed innovative quantitative techniques to measure protein stability and diffusion in crowded samples that interfere with standard measurement techniques. He has also made major advances in elucidating how the intracellular environment impacts both globular and intrinsically disordered proteins in surprising ways. Dr. Pielak has revolutionized our understanding of how proteins work where they actually function – inside cells – and not in the artificial environment of the test tube.

The Dorothy Crowfoot Hodgkin Award The Dorothy Crowfoot Hodgkin Award, sponsored by Genentech, is granted in recognition of exceptional contributions in protein science which profoundly influence our understanding of biology. The 2016 award will be presented to Dr. Rachel Klevit (University of Washington). Dr. Klevit’s research contributions have made a profound impact on the way we understand very important aspects of biological chemistry including how phosphorylation alters protein activity, regulation of transcription, and ubiquitylation. From the first structures of a zinc-finger and the RING E3 ubiquitin ligase BRCA1, Dr. Klevit has pushed NMR spectroscopy to establish new paradigms. Her research has been instrumental in understanding the mechanism of disease of two scourges, breast cancer and Parkinson’s Disease. Moreover she has changed the way research in this area is done. Additionally, she is an exceptional mentor of younger scientists and wonderful role model for other scientists and educators at all stages of their careers.

The Hans Neurath Award The Hans Neurath Award, sponsored by The Neurath Foundation, seeks to honor individuals who have made a recent contribution of exceptional merit to basic protein research. In 2016, the Hans Neurath winner is Dr. H. Eric Xu (Van Andel Research Institute). Dr. Xu established and served as the distinguished Director of the VARI-SIMM Center for Drug Discovery at Shanghai Institute of Materia (SIMM) of Chinese Academy of Sciences. Research in his group has previously been supported by four NIH R01 grants, one Senior Investigator Award from American Asthma Foundation, and one past DOD prostate cancer idea development award, which cover structures and drug discovery of nuclear hormone receptors, hepatocyte growth factor and its receptor Met tyrosine kinase, G-protein coupled receptors, and plant hormones. Two of his research papers on plant hormones were selected as top 10 breakthroughs by Science in 2009 and by Chinese Academy of Sciences in 2014, and his recent X-ray laser structure of the first GPCR-arrestin complex was also selected as a top 10 breakthrough by Chinese Academy of Sci- ences in 2016.

5 ABSTRACT

The Christian B. Anfinsen Award The Christian B. Anfinsen Award, sponsored by The Protein Society, recognizes technological achievement or significant methodological advances in the field of protein science. The recipient of this award in 2016 is Dr. Andreas Pluckthun€ (University of Zurich). Dr. Pluckthun€ is a pioneer of protein engineering. By combining rigorous biophysical studies with the invention of new combinatorial and evolutionary technologies, he has advanced both basic and applied science. His research greatly contributed to enabling the emergence of antibody engineering, by the use of E. coli as an engineering platform and studies on synthetic antibodies which led to the first fully synthetic antibody library. To create a true in vitro protein evolution technology he developed ribosome display of whole proteins. Through his work, designed ankyrin repeat proteins (DARPins) were created as a robust alternative scaffold for binding proteins. Innovative directed evolution technologies have led to highly stable G protein-coupled receptors that can be used for structural studies and in drug screening. Several engineered therapeutics, developed on the basis of his research, are now in late phase clinical development.

The Emil Thomas Kaiser Award The Emil Thomas Kaiser Award recognizes a recent, highly significant contribution to the application of chemistry in the study of proteins. The 2016 recipient is Dr. Charles S. Craik (University of California, San Francisco). Dr. Craik the founder and director of the Chemistry and Chemical Biology Graduate Program. He received his education and training at Allegheny College (BS), Columbia University (Ph.D.) and UCSF (Postdoctoral). He joined the UCSF faculty in 1985 where his research interests focus on defining the roles and the mechanisms of enzymes in complex biological processes and on developing technologies to facilitate these studies. He is also founder of Catalyst Biosciences, a biotechnology company focused on therapeutic proteases. Craik is a Fellow of the American Association for the Advancement of Science (AAAS) and the National Academy of Inventors (NAI). The current research in the Craik lab focuses on the chemical biology of proteolytic enzymes, their receptors and their natural inhibitors. A particular emphasis of his work is on identifying the roles and regulating the activity of proteases and degradative enzyme complexes associated with infectious diseases and cancer. These studies coupled with his global substrate profiling and noninvasive imaging efforts are providing a better understanding of both the chemical make-up and the biological importance of these critical proteins to aid in the rapid detection, monitoring and control of infectious disease and cancer.

The Stein and Moore Award The Stein and Moore Award is named for Nobel laureates Dr. William Stein and Dr. Stanford Moore. The award venerates eminent leaders in protein science who have made sustained, high impact research contributions to the field. The 2016 recipient is Dr. Jane Clarke (University of Cambridge, UK). Dr. Clarke is Professor of Molecular Biophysics in the Chemistry Department of the University of Cambridge. Her research is multidisciplinary, combining single molecule and ensemble biophysical techniques with protein engineering and simulations to investigate protein folding, misfolding and assembly. In her role as Deputy Head of the Chemistry Department in Cambridge, Dr. Clarke became involved in mentoring, career development, and leadership training for scientists at all stages in their careers.

The Protein Science Young Investigator Award The Protein Science Young Investigator Award, named for the academic journal of the Society, Recog- nizes a scientist generally within the first 8 years of an independent career who has made an important contribution to the study of proteins. The 2016 winner is Dr. Benjamin Garcia (University of Pennsylvania Perelman School of Medicine). Dr. Garcia has been developing analytical and computational tools to

6 ABSTRACT understand the combinatorial complexity of simultaneously occurring histone modifications, identifying thousands of uniquely modified histone H3 forms, the significance of which is the focus of current research interest (e.g. combinatorial Histone Code). He also been involved in development of advanced mass spectrometry instrumental approaches using electron transfer dissociation and data-independent acquisition to increase the accuracy and precision for protein and proteome characterization. The Garcia lab has been developing and applying novel mass spectrometry based proteomic approaches for interrogating protein post-translational modifications (PTMs), especially those involved in epigenetic mechanisms such as histones, publishing over 170 publications.

The 2016 Protein Science Best Paper Award At the beginning of each year, two “best papers” are selected from articles published in “Protein Scien- ce” during the preceding 12 months. A junior author (typically the first author) is designated as the award winner and invited to give a talk at the following Annual Protein Society Symposium. Tracy Clinton is an Air Force biochemist who earned her Ph.D. at the University of Utah through the Air Force graduate education civilian institute program. As Tracy puts it, “I am passionate about science and earning my Ph.D. was a wonderful way to further my scientific education while working in an area of biological relevance to my career. The research I had the privilege to take part in was both exciting and challenging and allowed me to bring many new skills and knowledge back to my professional life. I am very thankful for the opportunity to be a small part of the great things that Dr. Kay’s lab has and will continue to achieve.” Michael Thompson grew up in the San Fernando Valley area, outside of Los Angeles, and was an undergraduate at UC-Berkeley (degree 2007), in the Department of Molecular and Cell Biology. While at Berkeley, he worked as a research assistant in Tom Alber’s laboratory, where he developed interests in protein crystallography and in understanding how conformational changes control the functions of proteins. He then attended graduate school at UCLA (degree 2014), under the mentorship of Todd Yeates. For more on the 2016 Protein Science Best Paper Award winners and their research, read the article by Protein Science Editor-in-Chief Brian W. Matthews “Protein Science best paper awards to Tracy Clin- ton and Michael Thompson” in the May 2016 Issue.

7 ABSTRACT

Abstract Grouping by Topics Pages PA – Amyloid and Aggregation 9 PB – Bioinformatics 19 PC – Chaperones 22 PD – Chemical Biology 26 PE – Computational Modeling/Simulation 34 PF – Design/Engineering 49 PG – Dynamics and Allostery 76 PH – Enzymology 89 PI – Evolution 99 PJ – Folding 101 PK – Intrinsically Disordered Proteins 109 PL – Membrane Proteins 113 PM – Motors and Machines 121 PN – Peptides 121 PO – Protein in Cells 126 PP – Protein Interactions and Assemblies 128 PQ – Proteomics 145 PR – Proteostasis and Quality Control 148 PS – Single Molecule Studies 149 PT – Structure (X-Ray/NMR/EM) 150 PU – Synthetic Biology 165 PV – Therapeutics and Antibodies 166 PW – Transcription/Translation/Post-Translational Medications 171 PX – Metabolic Engineering/Energy Applications 174 Abstracts of the 2016 Award Winners 174

8 ABSTRACT

PA - AMYLOID AND AGGREGATION A new mechanism of pancreatic b-cell toxicity in type 2 diabetes Andisheh Abedini1, Annette Plesner, Ping Cao, Jinghua Zhang, Chris T. Middleton, Daniel Sartori, Julia Derk, Rosa Rosario, Fei Song, Jacqueline Lonier, Martin T. Zanni, Daniel P. Raleigh and Ann Marie Schmidt 1New York University, School of Medicine Progressive loss of insulin-producing pancreatic b-cells leads to hyperglycemia. Islet amyloidosis by the hormone islet amyloid polypeptide (IAPP or amylin) causes b-cell stress/death in type 2 diabetes (T2D), exacerbating hyperglycemia and islet transplant failure. But the molecular nature of toxic IAPP species and their mechanisms of toxicity remain elusive; no therapeutic strategies exist to treat/prevent this disorder. Our recent advances fill this critical knowledge gap. We define the properties of toxic species produced during IAPP amyloid formation, link their characteristics to induction of b-cell death, and show that they have both distinguishing properties and share common features with toxic species reported for other amyloidogenic polypeptides. This is highly significant because it provides the first target for rational drug development to treat IAPP-induced b-cell death and overturns a decades old paradigm that toxic species produced by different amyloidogenic proteins are similar. Indeed, propelled by these key insights, we identify a new receptor-mediated mechanism of islet amyloidosis toxicity. The receptor for advanced glycation endproducts (RAGE) selectively binds to toxic intermediates, but not to non-toxic forms of IAPP, including IAPP amyloid fibrils. We demonstrate that the isolated soluble extracellular ligand binding domain of RAGE (sRAGE) blocks both IAPP toxicity and amyloid formation. Inhibi- tion of IAPP/RAGE interaction by sRAGE, RAGE-blocking antibodies, or by genetic deletion of RAGE protects pancreatic islets and cells from IAPP-induced toxicity. Treatment of transgenic mice with sRAGE prevents islet amyloid formation, loss of insulin and b-cell inflammation and apoptosis. In human diabetic pancreas and in in vitro and in vivo models of islet amyloidosis, increased expression of RAGE accompanies IAPP induced b-cell/islet stress. These findings define the nature of toxic IAPP species and establish a new receptor-mediated mechanism of islet amyloidosis toxicity, identifying RAGE as a therapeutic target, and sRAGE as a therapeutic agent, to mitigate this source of b-cell stress in T2D.

Amyloidogenicity and Cytotoxicity of Bovine Amylin; Implications for Xenobiotic Transplantation and the Design of Non-toxic Amylin Variants Rehana Akter1, Andisheh Abedini2, Rebekah L. Bower3, Ann Marie Schmidt2, Debbie L. Hay3 and Daniel P. Raleigh1,4 1Department of Chemistry, Stony Brook University, 2Diabetes Research Program, NYU School of Medi- cine, 3School of Biological Sciences, University of Auckland, 4Graduate Program in Biochemistry and Structural Biology, Stony Brook University Islet amyloid formation contributes to b-cell death and dysfunction in type-2 diabetes and to the failure of islet transplants. Amylin (Islet amyloid polypeptide, IAPP), a normally soluble 37 residue polypeptide hormone, is responsible for amyloid formation in type-2 diabetes. The peptide is produced in the b- cells and is thus absent in type-1 diabetes. Amylin normally plays an adaptive role in metabolism and the development of non-toxic, non-amyloidogenic, but bioactive variants of human amylin are of interest for use as an adjunct to insulin therapy. The discovery of naturally occurring non-amyloidogenic variants is also of potential interest for xenobiotic transplantation, and because they can produce clues towards understanding the amyloidogenicity of human amylin. The sequence of amylin is well conserved among species, but sequence differences strongly correlate with in vitro amyloidogenicity and with the ability to form islet amyloid in vivo. Bovine amylin differs from the human peptide at 10 positions and is one of the most divergent among known amylin sequences. We show that bovine amylin is not toxic to cultured b-cells and is considerably less amyloidogenic than the human polypeptide, but does not bind human amylin receptors. The bovine sequence contains a number of non-conservative

9 ABSTRACT substitutions relative to human amylin including His to Pro, Ser to Pro and Asn to Lys replacements. The effect of these substitutions is analyzed in the context of wild type human amylin and the results provide insight into the mode of assembly of human amylin and into the design of soluble amylin analogs.

Aggregation/fibrillation of bovine serum albumin and its inhibition by osmolytes Moumita Das Gupta1, Nand Kishore1 1Department of Chemistry, Indian Institute of Technology Bombay, Powai, Mumbai 400 076, India Amyloid fibrils formed due to aggregation of misfolded or partially folded protein molecules, are associated with neurodegenerative diseases. Therefore, understanding the mechanism of fibrillation/aggregation of protein and design of suitable inhibitor molecules for stabilizing the native conformation of protein, are of utmost importance. In the current study, an approach has been taken to explore and understand the heat induced aggregation of Bovine Serum Albumin (BSA) at physiological pH (pH 7.4), in absence and presence of osmolytes (proline, hydroxyproline, glycine betaine, sarcosine and sorbitol), employing a combination of fluorescence spectroscopy, Rayleigh scattering, isothermal titration calorimetry (ITC), dynamic light scattering (DLS) and transmission electron microscopy (TEM). Formation of fibrils by BSA under the given conditions was confirmed from increase in fluorescence emission intensities of Thioflavin T over a time period of 600 minutes and TEM images. Absence of change in fluorescence emission intensities of 8-Anilinonaphthalene-1-sulfonic acid (ANS) in presence of native and aggregated BSA, signify the absence of any amorphous aggregates. ITC results have provided important insights on nature of interaction of these osmolytes with different stages of the fibrillar aggregates of BSA. Among the osmolytes used here, we found glycine betaine to be supporting and promoting the aggregation process while hydroxyproline to be maximally efficient in suppressing the fibrillation process of BSA, followed by sorbitol, sarcosine and proline in the following order of their decreasing potency: Hydroxyproline> Sorbitol> Sarcosine> Proline> Glycine betaine.

Typical and Atypical Prion-like Propagation of Neurotoxic Amyloid-b Oligomers Dexter N. Dean1, Kayla M. Pate2, Pradipta K. Das1, Sarah E. Morgan1, Melissa A. Moss2, and Vijayaraghavan Rangachari1 1The University of Southern Mississippi, 2The University of South Carolina Soluble oligomers of the amyloid-b (Ab) peptide have emerged as the primary neurotoxic agents in Alzheimer disease (AD). Recent evidence from animal models also implicates aggregates of Ab to undergo prion-like propagation towards seed-specific fibril deposition. However, dearth in a molecular understanding of Ab oligomers has confounded the insights into propagation and dissemination of toxic amyloid aggregates. Our recent reports on a distinct 12-24mer oligomer of Ab, called large fatty acid-derived oligomers (LFAOs), have opened doors in investigating this elusive mechanism. We have previously established that LFAOs undergo replication upon interacting with monomers to form quantitative amounts of identical oligomers. Here, we sought to investigate the concentration-dependent dynamics of LFAOs to reveal how such transitions manifest in their ability to replicate and induce neuronal apoptosis. We have also investigated the ability of LFAOs to undergo prototypical prion-like propagation, where LFAOs grow as distinct repeating units leading to morphologically unique fibrils. We discovered that LFAOs undergo a concentration-dependent transition between 12mers and disperse 12- 24mers, which correlates with their ability to replicate and induce apoptosis. At low concentrations (sub-mM), LFAOs exist as 12mers and undergo atypical prion-like propagation (replication) upon interaction with Ab monomers. While at high concentrations (> 10 mM), LFAOs exist as disperse 12-24mers and propagate towards morphologically unique fibrils in typical prion-like fashion. The observations

10 ABSTRACT reported here may have profound significance in deciphering the emerging roles of Ab oligomer phenotypes in prion-like propagation and dissemination of toxicity in AD.

Lipid interaction and membrane perturbation of different protofibrillar Ab9-40 trimers: an atomistic simulation study Xuewei Dong1, Yunxiang Sun1, Buyong Ma2, Ruth Nussinov2,3, and Guanghong Weia 1State Key Laboratory of Surface Physics, Key Laboratory for Computational Physical Sciences (MOE), and Department of Physics, Fudan University, Shanghai, P. R. China, 2Basic Science Program, Leidos Biomedical Research, Inc. Cancer and Inflammation Program, National Cancer Institute, 3Sackler Inst. of Molecular Medicine Department of Human Genetics and Molecular Medicine Sackler School of Medicine, Tel Aviv University Alzheimer’s disease (AD), a common protein misfolding and neurodegenerative disease, is characterized by the aggregates of amyloid beta (Ab) peptides. The direct interaction of Ab with membrane lipids is believed to be responsible for its neurotoxicity and could accelerate the growth of fibrils (1). Experimen- tal studies reported that Ab fibers could cause changes in the membrane morphology and permeation (2). However, the mechanism of Ab-lipid interaction and consequent membrane perturbation at atomic level is still not clear. In this study, we have investigated the interactions of three different solid-state NMR derived protofibrillar Ab9-40 trimers with an anionic palmitoyloleoyl-phosphatidylglycerol (POPG) bilayer by using atomistic molecular dynamics (MD) simulations. Our simulation results show that the N-terminal residues of all Ab protofibrils preferentially bind to the membrane surface due to electrostatic interaction. We also find that the lipid membranes enhance the structural stability of protofibrils by increasing the b-sheet content and salt-bridge number. Furthermore, the interaction of Ab protofibril with POPG bilayers decreases the local thickness of lipid membranes. This work provides insights into the Ab-lipid interaction and the mechanism of lipid perturbation by Ab protofibrils at atomic level. References: 1. Capone, R.; Jang, H.; Kotler, S. A.; Kagan, B. L.; Nussinov, R.; Lal, R., Probing structural features of Alzheimer’s amyloid-beta pores in bilayers using site-specific amino acid substitutions. Biochemistry 2012, 51 (3), 776-85. 2. Williams, T. L.; Day, I. J.; Serpell, L. C., The effect of Alzheimer’s Abeta aggregation state on the permeation of biomimetic lipid vesicles. Langmuir: the ACS journal of surfaces and colloids 2010, 26 (22), 17260-8.

Decreasing amyloidogenicity of immunogloublin light chain variable domain by mutation of surface exposed residues and peptide inhibitor targeting partially unfolded state Daizo Hamada1 1Graduate School of Engineering and Center for Applied Structural Science (CASS) Kobe University AL amyloidosis is a protein-misfolding disease characterized by accumulation of amyloid fibrils by immunoglobulin light chains (LCs) that induce tissue damage and multiple organ failures. Destabiliza- tion of the native state of the variable domain of the LC (VL) is one of the critical factors in promoting the amyloid fibril formation. However, determining the key residues involved in this destabilization remains challenging, because of the existence of a number of sequence variations within VL. We identified the key residues for destabilization of the native state of amyloidogenic VL in the LC of BRE by analyzing the stability of chimeric mutants of BRE and REI VL; the latter is not associated with AL amyloidosis. The results suggest that the surface-exposed residues N45 and D50 are the key residues in the destabilization of the native state of BRE VL. The K45N, E50D, and K45N/E50D mutants of REI VL destabilized the native state and increased amyloidogenicity. However, the reverse mutations in BRE VL (N45K, D50E, and N45K/D50E) re-established the native state and decreased amyloidogenicity. X-ray crystallography and NMR analyses indicated the regions with increased structural fluctuations in the

11 ABSTRACT native state of K45N/E50D REI VL. Based on this information, we could designed peptide fragments that prevent the amyloid formation by binding to partially unfolded states of VL domains. These results suggest that the modulation of surface properties of VL may improve the stability and prevent the formation of amyloid fibrils, and that the analysis on molecular fluctuations provides the information for the design of amyloid inhibitors that target the partially folded states.

The Role of Micelle-like Oligomers in the Aggregation of Human Calcitonin Kian Kamgar-Parsi1, Ayyalusamy Ramamoorthy2 1Applied Physics Program, University of Michigan, 2Department of Chemistry and Biophysics, Univer- sity of Michigan Calcitonin is a peptide hormone involved in skeletal regulation, and as such is used as a therapy for osteoporosis. However, calcitonin’s aggregation along the canonical amyloid pathway limits its therapeutic efficacy. In several amyloids, micelle-like oligomers have been shown to have strong effects on aggregation kinetics, and thus present a possible target species for kinetics modulation. In this paper, we report the formation of micelle-like oligomers in human calcitonin, and characterize their effects on aggregation kinetics and morphology. In contrast to other amyloids, micelle-like oligomers of human calcitonin did not affect aggregation kinetics as determined via thioflavin-T binding assay, indicating that the micelle-like oligomer in human calcitonin plays a significantly different role than in other amyloids. Micelle-like oligomers in human calcitonin were not found to significantly alter the morphology and secondary structure of intermediate or end-stage aggregates by transmission electron microscopy and circular dichroism spectroscopy. Interestingly, human calcitonin presents a direct relationship between peptide concentration and aggregation kinetics previously unseen in amyloids. Despite the micelle independence of this trend, the slowed aggregation appears associated with the formation of a long persisting oligomeric species. Further elucidation of this behavior could prove informative in improving calcitonin therapeutics.

Sialylation of the prion protein controls prion infectivity Elizaveta Katorcha1 1University of Maryland, Baltimore Prion diseases are a family of lethal, transmissible, neurodegenerative disorders that can be sporadic, inheritable or transmissible in origin. Prion diseases include Creutzfeldt-Jakob disease in humans and Bovine spongiform encephalopathy in cattle, they are 100% lethal and have no treatment. Prions or PrPSc lack nucleic acids bur consist of a self-replicating state of a sialoglycoprotein called prion protein or PrPC. While prions are not conventional pathogens, they might use a similar strategy for invading and colonizing their host. We hypothesized that PrPSc, like some microbes, use a carbohydrate post- translational modification - sialic acid - to camouflage themselves from the host immune system upon invasion. We found that eliminating sialic acids from PrPSc glycans abolishes its infectivity, whereas restoring sialylation restores prion infectivity. Moreover, sialylation was found to be critical for effective trafficking of PrPSc to secondary lymphoid organs, which is the first step of prion invasion. We also showed that upon colonization of secondary lymphoid organs, the sialylation status of PrPSc is enhanced. We proposed that sialylation is responsible for camouflaging prions from a host innate immune system. Our work suggests new glycobiology-driven approaches for developing therapeutic strategies against prion diseases. References: 1. Katorcha et al. PLoS Pathog. (2014) Sep 11;10(9):e1004366. 2. Srivastava et al. Proc Natl Acad Sci U S A. (2015) Dec 1;112(48):E6654-62.

12 ABSTRACT

Identification of the Sequence Cleavage Preference for the Protease BACE2 Using Proteomics, Mass Spectrometry, and Bioinformatics Hyojung Kim1 and Joseph L. Johnson1 1Department of Chemistry and Biochemistry, University of Minnesota Duluth Proteases play critical roles in vital biological processes including the regulation, breakdown, and recy- cling of proteins. Understanding the preferred amino acid sequence for a given protease can elucidate its substrates and therefore its function. The information from this characterization can be used to develop inhibitors of a protease to treat disease or to suggest potential undesired effects resulting from its inhibition. b-secretase (BACE) membrane associated aspartyl protease family consists of two homologs, BACE1 and BACE2. BACE1 has been studied in depth as the primary member of this family leading to Alzhei- mer’s disease (AD) while its homolog of BACE2 has been largely ignored because BACE2 mRNA is mini- mally expressed in the brain. However, recent studies suggest that both BACE1 and BACE2 are involved in the generation of beta-amyloid (Ab) which aggregates to form the plaques that are a hallmark pathology of AD. Thus, characterization of the substrate cleavage sequence preference of BACE2 should help to identify putative substrates and its native and AD-related functions. Proteomic identification of cleavage sites (PICS) is a combined proteomic and mass spectrometric method for characterizing the cleavage site preferences of proteases. It is less biased than previous methods requiring the synthesis of a small number of peptides designed around a specific amino acid sequence. Briefly, peptide fragments generated by the BACE2 cleavage of semi-random peptide libraries are sequenced using mass spectrometry. The remaining portion of recognition and cleavage sequence is determined using bioinformatics and database searching. With the preferred sequence(s) determined, we will then search the human proteome for novel putative substrates of BACE2. In summary, a first step towards a better understanding of BACE2 is to analyze its substrate specificity and then use that information to identify its native substrates which will enable us to elucidate its native and AD-related functions.

DLPC Liposomes Inhibit Aß Fibrillation and Remodel Preformed Fibrils Through a Detergent-like Mechanism Kyle Korshavn1, Cristina Satriano2, Rongchun Zhang3, Mark Dulchavsky4, Anirban Bhunia5, Magdalena Ivanova4, Carmelo La Rosa2, Mi Hee Lim6, Ayyalusamy Ramamoorthy1,3 1Department of Chemistry, University of Michigan, Ann Arbor, MI 48109, USA, 2Department of Chem- ical Sciences, University of Catania, Viale A. Doria 6, 95125 Catania, Italy, 3Biophysics Program, Uni- versity of Michigan, Ann Arbor, MI 48109, USA, 4Department of Neurology, University of Michigan, Ann Arbor, MI 48109, USA, 5Department of Biophysics, Bose Institute, Kolkata 700 054, India, 6Department of Chemistry, Ulsan National Institute of Science and Technology (UNIST), Ulsan, 44919, Republic of Korea. The aggregation of amyloid-b (Ab) on neuronal lipid bilayers is implicated in the neurotoxicity associated with Alzheimer’s disease. It is suggested that the aggregation on lipid bilayers is accelerated. This aggregation, in turn, is capable of disrupting bilayer integrity which induces cell death. We have found, however, that lipid bilayers composed of dilauroyl phosphatidylcholine (DLPC) undergo a unique mechanism of disruption. Through a combination of fluorescence measurements, microscopy, CD spectroscopy, and solid state NMR we have found that DLPC liposomes are rapidly disrupted through interactions with monomeric and early oligomeric forms of Ab. This disruption generates free lipid which successfully impedes amyloid formation through a detergent-like mechanism and stabilizes non- toxic, off pathway oligomers. DLPC liposomes are also capable of remodeling preformed fibrils of Ab into on-pathway, toxic protofibrillar species. All previous evidence of membrane disruption by Ab has

13 ABSTRACT been the result of fibrillation, pointing to a unique interaction occurring at the surface of DLPC membranes which is capable of halting amyloid formation in the early stages.

Supermetallization of Peptides and Protiens Studied by High Resolution Mass Spectrometry Eugeny Kukaev1, 3, 4, Yury Kostyukevich 1, 2, Alexey Kononikhin1, 3, Maria Indeykina3, 4, Igor Popov1,4, Eugene Nikolaev 1, 2, 3, 4 1Moscow Institute of Physics and Technology, Moscow, 2Skolkovo Institute of Science and Technology, Skolkovo, 3Institute for Energy Problems of Chemical Physics, Moscow, 4Emanuel Institute of Bio- chemical Physics, Moscow, Russia Transition metal ions are involved in neurodegenerative diseases. Recently we have reported the supermetallization phenomenon, which is the formation of complex ions of peptide–metal in the gas phase when the peptide accepts an unexpectedly large number of metal atoms. This supermetallization takes place during electrospray ionization when charged droplets are evaporating at relatively high temperature (ca 4008C). The effect has been demonstrated for the divalent Zn21 atom and several peptides and proteins. It is observed that small protein ubiquitin can incorporate up to 20 Zn atoms and the attachment of each Zn requires the removal of two hydrogen atoms. Here we report the first observation, to our knowledge, of gas-phase complexes of peptides and proteins with tetravalentmetals. Supermetallized complexes were studied by FTICR MS in combination with CID/ECD fragmentation. It was found that during ECD complexes with many metal atoms do not dissociate and only decrease their charge by capturing the electron. Using high resolution FTICR it was observed that with the increase of the number of metal atoms in the supermetallized complex there is a trend toward forming the z11 fragments instead of the z fragments in ECD experiments. The effect of hydrogen rearrangement for c fragments was also observed but was found to be very weak. This work was supported by the project of Russian Science Foundation 16-14-00181. References: 1. Y. Kostyukevich et al. J. Mass Spectrom. 50, 1079–1087 (2015). 2. Y. Kostyukevich et al. Eur. J. Mass Spectrom. 22,39–42 (2016).

Key asparagine and glutamine residues promote cross-species prion conversion Timothy D. Kurt1, Lin Jiang2, Nazilla Alderson1, Jun Liu1, David Eisenberg2, and Christina J. Sigurdson1,3 1Departments of Pathology and Medicine, UC San Diego, La Jolla, CA 92093, USA, 2UCLA-DOE Insti- tute, Howard Hughes Medical Institute, and Molecular Biology Institute, UCLA, Los Angeles, California 90095, USA, 3Department of Pathology, Immunology, and Microbiology, UC Davis, Davis, CA 95616, USA The central event underlying prion disease transmission is the conversion of the cellular prion protein, PrPC, into a misfolded, self-templating conformer called PrPSc. Sequence similarity between PrPSc and PrPC is critical for efficient prion templating, yet the specific interacting residues and the mechanism by which a residue promotes or inhibits conversion remain unclear. Previous studies have shown that single residue differences between PrPSc and PrPC may dramatically impede conversion, as illustrated by the G127V variant of human PrPC, which confers protection against transmission of human prions (1). However, the transmission of prions between species suggests that many PrPC residue differences are not protective, or may even promote prion cross-seeding. Recent studies have found that the PrPC sequence of bank voles, a European rodent, is permissive to conversion by a wide array of prions from many species (2, 3) despite residue differences between the PrPSc and vole PrPC sequences. We recently demonstrated that human PrPC having only two bank vole amino acid substitutions (E168Q

14 ABSTRACT and S170N) was converted efficiently by chronic wasting disease (CWD) prions from elk (4). Here we show that key asparagine and glutamine residues promote prion conversion by diverse prions in a strain-specific manner. Further, it is possible to use these results to engineer a “universal” PrPC sequence that is permissive to templating by PrPSc having diverse sequences and conformations. These findings suggest that the molecular mechanism of PrPC to PrPSc conversion involves formation of asparagine/glutamine ladders via side chain hydrogen bonding along the length of the fibril. Identifying the heterologous asparagine/glutamine residues that promote prion cross-seeding also provides important insight into cross-species prion transmission and zoonotic disease risk. References: 1. E. A. Asante, M. Smidak, A. Grimshaw, R. Houghton, A. Tomlinson, A. Jeelani, T. Jakubcova, S. Ham- dan, A. Richard-Londt, J. M. Linehan, S. Brandner, M. Alpers, J. Whitfield, S. Mead, J. D. Wadsworth, J. Collinge, A naturally occurring variant of the human prion protein completely prevents prion disease. Nature 522, 478-481 (2015); published online EpubJun 25 (10.1038/nature14510). 2. C. D. Orru, B. R. Groveman, L. D. Raymond, A. G. Hughson, R. Nonno, W. Zou, B. Ghetti, P. Gambetti, B. Caughey, Bank Vole Prion Protein As an Apparently Universal Substrate for RTQuIC-Based Detection and Discrimination of Prion Strains. PLoS pathogens 11, e1004983 (2015); published online EpubJun (10.1371/journal.ppat.1004983). 3. J. C. Watts, K. Giles, S. Patel, A. Oehler, S. J. DeArmond, S. B. Prusiner, Evidence that bank vole PrP is a universal acceptor for prions. PLoS pathogens 10, e1003990 (2014); published online EpubApr (10.1371/journal.ppat.1003990). 4. T. D. Kurt, L. Jiang, N. Fernandez-Borges, C. Bett, J. Liu, T. Yang, T. R. Spraker, J. Castilla, D. Eisen- berg, Q. Kong, C. J. Sigurdson, Human prion protein sequence elements impede crossspecies chronic wasting disease transmission. The Journal of clinical investigation, (2015); published online EpubFeb 23 (10.1172/JCI79408).

Understanding the structure and self-assembly of the hydrophobin protein RodA from Aspergil- lus fumigatus and development of novel nanocarriers Jennifer I-Chun Lai1, Victor Lo1, Ivan Cheung2, Matthew Hampsey2, Ann H. Kwan2, Chi Pham1,Inaki~ Guijarro3, Ariane Pille3, Jake A. Campbell1, Margaret Sunde1 1Discipline of Pharmacology, School of Medical Sciences, 2School of Life and Environmental Sciences, 3The University of Sydney, NSW 2006, Australia; Institut Pasteur, Paris, France Hydrophobins are small proteins produced by filamentous fungi which self-assemble into amphipathic monolayers at hydrophobic:hydrophilic interfaces. They are characterised by a pattern of eight conserved cysteine residues that form four disulphide bonds. These proteins play a number of different functional roles, including reducing surface tension, facilitating attachment to surfaces and mediating fungal-host interactions. RodA has been shown to form a robust, protein monolayer on the surface of the Aspergillus fumigatus conidia. The assembled protein is arranged in filamentous structures known as rodlets, which share many structural characteristics with amyloid fibrils. This layer provides a mechanism for evasion of host innate immune response through masking of the pathogen-associated molecular patterns (PAMPs) expressed on the fungal cell wall [1]. We have carried out a program of mutagenesis to identify the regions of the protein that are important in the mechanism of self-assembly and structure of the functional rodlet monolayers. Four single- glycine mutations, a double-glycine mutation and a chimera in which the Cys7-Cys8 region was replaced with a non-rodlet-forming sequence were successfully produced and compared to recombinant wild-type RodA for changes in kinetics and morphology of self-assembly. In the kinetics assay all mutants displayed a general trend of increase in lag phase, implicating the involvement of these residues in self-assembly. Previous work on the hydrophobin EAS suggested that only the Cys7-Cys8 region is critically involved in hydrophobin self-assembly [2]. However, our work demonstrates that the region

15 ABSTRACT bound by Cys4-Cys5 in RodA is a secondary point of contact for self-assembly of RodA. The understanding generated through the mutagenesis study unlocked the possibility of engineering RodA to incorporate non-native but useful amino acid sequences, without loss of the ability to form amyloid structures. We have introduced non-native sequences that should target the hydrophobin constructs to specific molecular markers expressed only during disease states or to markers that are specific for a particular cell type. The amphipathic nature of hydrophobins allows them to act as solubilising agents for hydrophobic drugs. This work will lead to the development of engineered hydrophobins as novel targeted therapeutic nanocarriers. References: 1. Aimanianda, V., Bayry, J., Bozza, S., Kniemeyer, O,. Perruccio, K., Elluru, S.R., Clavaud, C., Paris, S., Brakhage, A.A., Kaveri, S.V., Romani, L., and Latge, J.P. (2009) Surface hydrophobin prevents immune recognition of airborne fungal spores. Nature 460, 1117-1121. 2. Macindoe, I., Kwan A.H., Ren, Q., Morris, V.K., Yang, W., Mackay, J.P., and Sunde, M. (2012) Self- assembly of functional, amphipathic amyloid monolayers by the fungal hydrophobin EAS. Proceedings of the National Academy of Sciences of the United States of America 109, 6951-6956.

Characterizing Pre-Amyloid Oligomers of hIAPP with a Photoinduced Cross-linking Technique Zachary Ridgway1, Andisheh Abedini2, Ann Marie Schmidt2, and Daniel P. Raleigh1 1Stony Brook University, 2Diabetes Research Program, NYU School of Medicine Human islet amyloid polypeptide (hIAPP) is a 37 residue peptide hormone implicated in the pathology of type II diabetes. Secreted from b-cells, it fulfills numerous roles in metabolic processes. Although it is soluble under physiological conditions, it can also misfold to form insoluble aggregates. Analogues that do not form amyloid have been utilized as adjunct treatments with insulin therapy, but their biophysical properties have not been well characterized. The correlation between hIAPP aggregates and the reduction of b-cell mass remains unclear. Recent studies have shown that the predominant cytotoxic species of hIAPP is not the mature aggregate, but rather low-order oligomers. In this work, a rapid and nonspecific crosslinking technique is paired with gel electrophoresis in order to take a “snapshot” of oligomeric populations in solution. With this system, we characterize oligomeric species of hIAPP and soluble analogues to better understand the interactions that govern oligomerization and cytotoxicity.

Predicting Aggregation and Cross-Seeding Activity of Yeast Prion-Like Proteins Jenifer Shattuck1, Aubrey Waechter1 and Eric Ross1 1Biochemistry and Molecular Biology Department, Colorado State University Prions are infectious proteins that generally result from the formation of self-propagating amyloid aggregates. Several yeast proteins can form amyloid-based prions. These proteins provide a model system to investigate the formation and propagating of prions and amyloids. Various prion prediction algorithms have been developed to predict yeast proteins that have the propensity to form prions, including the Prion Aggregation Prediction Algorithm, PAPA. We used PAPA to scan the yeast proteome to identify proteins that contain domains predicted to have prion activity. These prion-like domains (PrLDs) were tested for aggregation and prion activity in vivo. While PAPA was highly effective at identifying aggregation-prone PrLDs with a range of Q/N content, the ability to form the detergent-insoluble aggregates that typically characterize amyloid-based prions was limited to domains with high Q/N content. This suggests that high Q/N content may specifically promote conversion to the amyloid state. We then tested these PrLDs for the ability to promote prion formation by the prion protein Sup35. This ability was highly correlated with the ability to form detergent-insoluble aggregates, suggesting a specific characteristic required for promoting prion formation. Together we reveal more information about

16 ABSTRACT the parameters necessary for prion and prion-like formation, all contributing to understanding the functional role prion-like proteins play in cellular processes.

Disulfide Bond Scrambling at Acidic pH is Key to Insulin Aggregation and Toxicity Ashutosh Tiwari1 and Colina Dutta1 1Department of Chemistry, Michigan Technological University Protein aggregates are hallmark of several neurodegenerative diseases such as Amyotrophic Lateral Sclerosis, Alzheimer’s, and Parkinson’s disease. To better understand the relationship between distinct aggregated forms of proteins and its associated toxicity, we used insulin as a model protein. Insulin aggregates were prepared by incubating the protein at acidic, neutral, or basic pH at 378Cor658Cin presence/absence of disulfide reducing agents. The morphology and nature of the aggregates and its associated kinetics were characterized using spectroscopy and electron microscopy techniques. At conditions close to physiological, insulin aggregates rapidly (10-20 minutes) in the presence of reducing agent. In absence of reducing agent at acidic pH it forms rod-like aggregates at 378C and fibrils at 658C. In the presence of reducing agent, insulin forms amorphous aggregates with distinct morphology at acidic pH and filamentous structures of varying lengths at basic pH. The cytotoxicity of these distinct aggregates were tested using human neuroblastoma (SH-SY5Y) cells. Overall, the results show that aggregates formed under disulfide reducing conditions at acidic pH are more toxic.

Rosmarinic acid, a catechol-containing natural product, potently inhibits amylin amyloidosis. Paul Velander1, Ling Wu1, Anne Brown1, Keith Ray1, Rich Helm1, David Bevan1, Bin Xu1,2 1Department of Biochemistry, Virginia Tech, USA, 2Center for Drug Discovery, Virginia Tech, USA Epidemiological studies show significant association between obesity-related type 2 diabetes (T2D) and risk for cerebrovascular disease and dementia (including Alzheimer’s disease, AD). Amylin is a peptide hormone secreted by pancreatic beta cells, whose normal function involves maintaining glucose homeostasis. However, its high propensity to form amyloid in the pancreas contributes to T2D pathology. Recently, heterologous amyloid composed of amylin and amyloid beta (the causative agent and amyloid forming species in AD) has also been found in the brain issue of individuals with AD and T2D but not age matched healthy controls. Collectively these data suggest that amylin aggregation may represent a molecular link between AD and T2D. Our cell-based studies demonstrated that amylin amyloid is highly toxic to human and mouse neuronal cell lines SH-SY5Y and Neuro2A, respectively. From a targeted screening of a collection of natural compounds used in complementary medicine, we identified that rosmarinic acid (RA) is a highly potent inhibitor against amylin amyloid formation (estimated to be 200 nM IC50,). We also show that RA and its metabolites, caffeic acid (CA) and salvianic acid A (SAA), exert additive anti amylin aggregation and cell rescue effects against amylin-induced neurotoxicity. Our data suggests that the mechanism of RA mediated amylin amyloid inhibition may in part be due to site-specific binding to the amine groups in the peptide, a mechanism that is congruent with other data that suggests the catechol moieties present in RA, CA and SAA may play an essential role in their observed anti-amylin aggregation activity. Consistent with these experimental results, the inhibition effect by RA is demonstrated in computational molecular simulation analyses, providing an additional mechanism of non-covalent interactions between RA and the peptide as a way to block amylin oligomer and amyloid formation. Acknowledgement: This project was supported in part by Virginia Tech new faculty start-up funds, Alzheimer’s and Related Disease Award Fund from Virginia Center on Aging, and the Diabetes Action Research and Edu- cation Foundation (DAREF).

17 ABSTRACT

RNA binding and subcellular localisation of TIA protein in the formation of stress granules Saboora Waris1, Menachem Gunzburg1, Kylie Wagstaff1, Belinda Maher1, David Jans1, Matthew Wilce1, Jackie Wilce1 1Department of Biochemisty and Molecular biology, School of Biomedical sciences, Monash Univer- sity, Australia, Victoria, VIC 3800, Australia Eukaryotic cells are able to respond promptly to environmental stresses at the level of mRNA translation. In response to heat, starvation, oxidative stress and viral infection cytoplasmic granules known as stress granules (SGs) are formed by quick sequestration of target mRNAs bound by their associated proteins. T-cell intracellular antigen-1 (TIA-1) is an RNA binding protein, which continuously shuttles between the nucleus and the cytoplasm. Upon stress TIA protein accumulate in the cytoplasm along with its bound mRNA. In particular, TIA-1 binds to mRNA containing A/U-rich elements in their 3’UTR and then escorts the translationally stalled mRNA into stress granule through the aggregation of its prion related domain (PRD). TIA-1 possesses three RNA recognition motifs RRM1, RRM2, RRM3 and a PRD, among the domains, RRM2 is the major RNA recognition domain for U and A/U-rich sequences. Here we report the first crystal structure of RRM2 engaged with nucleic acid. We also investigate the affect of TIA serine phosphorylation and the other components of stress granule on the amyloid-like properties of TIA-1 aggregation and cytoplasmic accumulation.

Discovery and Mechanisms of Small Molecule Inhibitors against Amylin Amyloidosis in the pancreas and the brain Ling Wu1, Paul Velander1, Keith Ray1, Anne Brown1, Rich Helm1, David Bevan1, Bin Xu1,2 1Department of Biochemistry, Virginia Tech, 2Center for Drug Discovery, Virginia Tech Epidemiological and clinical studies showed significant association between type 2 diabetes (T2D) and the risk for neurodegeneration (including Alzheimer’s disease). Amylin is a 37-residue peptide hormone that is co- secreted with insulin from pancreatic-cells. It is highly amyloidogenic and amylin amyloid deposition in the pancreas are hallmark features of T2D. Recent clinical studies reported that amylin plaques were deposited in the brain of diabetic patients, but not in healthy controls. We performed cell- based studies, demonstrating that amylin amyloid is highly toxic to multiple cell lines, including pancreatic -cells INS-1 and neuronal cells SH-SY5Y and Neuro2A. From a 384-well fluorescence-based screening of a collection of natural compounds used in complementary and alternative medicine, we identified multiple potent inhibitors, including rosmarinic acid (RA), baicalein, and 7,8-dihydroxyflavone (estimated 200 nM– 2 mM in apparent IC50). These lead compounds disaggregate amylin fibrils from transmission electron microscopic observations and significantly reduce amylin-induced cytotoxicity. Dissecting the functional groups of these natural compounds, we identified that the catechol groups contributed significantly to amyloid inhibition in more than a dozen catechol- containing compounds. Compounds with multiple catechol groups exhibited additive effect. Analog analyses demonstrated key roles played by vicinal hydroxyl groups of the catechol groups. We provided mass spectrometric evidence, for the first time to our knowledge, that incubating several of these catechol-containing inhibitors with amylin leads to (1) covalent adducts consistent with Schiff bases as a proposed mechanism of amyloid inhibition,; (2) loss of the amino-terminal lysine on peptide as a potentially novel mechanism of inhibition. The inhibition effects by these compounds were demonstrated in computational simulation analyses, providing additional non-covalent mechanisms. (Supported in part by grants from the Alzheimer’s and Related Diseases Research Award Fund, Diabetes Action Research and Education Foundation).

EGCG binds to different intermediates populated during Human lysozyme fibrillation and modulates them towards less toxic off-pathway aggregates Fatima Kamal Zaidi1, Rajiv Bhat1 1School of Biotechnology, Jawaharlal Nehru University, New Delhi, India-110067. Protein aggregation disorders are a group of devastating pathologies, characterized by the deposition of toxic protein aggregates in various organs of the body. It is therefore important, to design effective

18 ABSTRACT therapeutic strategies against these disorders, and naturally occurring polyphenols are emerging as promising candidates for such studies. The effect of EGCG (green tea polyphenol) was analyzed on amyloid fibril formation of recombinant Human lysozyme (HuL), which is a protein associated with the amyloidogenic disorder called as Lysozyme systemic amyloidosis (Alys). EGCG inhibited fibril formation of HuL in a dose-dependent manner with light scattering studies and TEM analysis showing large aggregates in the fibrillation pathway in presence of EGCG and ANS binding analysis revealing the less hydrophobic nature of these aggregates compared to amyloid fibrils. Cytotoxicity studies showed that these aggregates were also remarkably less cytotoxic compared to amyloid fibrils, and seeding experiments indicated that they were off-pathway species having seeding capabilities. Fluorescence titration and ITC measurements revealed weak to moderate binding affinity (Ka 104 M-1) between EGCG and the native state of HuL and FRET analysis indicated that EGCG was binding at a distance of 2.9 nm from the Trp- 109 located in the active site of HuL and that energy transfer was taking place between them. Addition of EGCG at different time-points during fibrillation indicated that EGCG was mediating inhibition of fibrillation not only by interacting with the native state of HuL but was also capable of interacting with other intermediates populated during HuL fibrillation, and was thus modulating the fibrillation pathway of HuL towards formation of these less toxic, off-pathway aggregates.

Unveiling the inhibitory behavior of n-acetylneuraminic acid against fibrillation of amyloidogenic proteins-a biophysical insight Nida Zaidi1, Rizwan Hasan Khan1 1Interdiscilplinary Biotechnology Unit, Aligarh Muslim University, Aligarh – 202002, India. A variety of neurodegenerative disorders including Parkinson’s disease, are due to fibrillation in amyloidogenic proteins. The development of plausible curative therapeutics for these disorders is topic of extensive research; however effective treatments are still unavailable. In the present study, n- acetylneuraminic acid (Neu5Ac), a non-reducing monosaccharide inhibits the amyloid fibrillation of a- synuclein (SYN) and hen egg white lysozyme (HEWL) by decelerating the nucleation phase as observed from various spectroscopic and microscopic techniques. Besides, Neu5Ac also attenuates the cytotoxic nature of amyloid fibrils as evaluated by cell cytoxicity assay. The Neu5Ac stabilizes the native state of SYN and HEWL, and hampers the amyloid fibrillation process as evident from differential scanning calorimetry. Additionally, molecular docking and isothermal titration calorimetry results collectively suggest that Neu5Ac binds with low affinity to hydrophobic patches responsible for formation of fibrils in both proteins.

PB - BIOINFORMATICS Automating TULIP, a Protein Clustering Method Nick Biffis1 1University of Richmond As the number of sequenced proteins continues to grow at a rapid rate, it is critical to develop a process to accurately and efficiently identify protein function. A previously created process, TuLIP (Two- Level Iterative clustering Process), is an iterative process that utilizes active site information to cluster proteins from a given superfamily into hypothesized functional groups; it was initially validated using the enolase superfamily, a gold-standard superfamily in the SFLD that has been well studied by experts in the field. Though TuLIP has demonstrated the ability to accurately cluster proteins into proposed functionally relevant groups, it is time consuming and tedious to execute as most steps involve manual data manipulation at every iteration of the process. Therefore, a comprehensive automation of TuLIP was paramount. Automation of TuLIP has allowed the clustering of any given superfamily into functionally relevant groups with minimal manual data manipulation throughout the process. Comprehensive testing of TuLIP with gold-standard SFLD superfamilies demonstrates the automated method produces

19 ABSTRACT results identical to a manual implementation of TuLIP with significantly improved efficiency. Ultimately, we have developed a product capable of clustering protein structures with both high accuracy and efficiency.

Manual Curation in the Conserved Domain Database Gonzales NR1, Chitsaz F, Derbyshire MK, Geer L, Gwadz M, Han L, He J, Hurwitz DI, Lanczycki CJ, Lu F, Marchler GH, Song JS, Thanki N, Wang Z, Yamashita RA, Zheng C, Bryant SH, Marchler-Bauer A. 1NCBI The Conserved Domain Database (CDD) is a collection of multiple sequence alignments that represent ancient conserved domains. CDD includes domain models from external sources including Pfam and TIGRFAMs, as well as conserved domain models (accessions with a “cd” prefix) manually-curated by NCBI. Most curated models contain annotation of features that are conserved across the domain family, supported by evidence obtained from 3D structures as well as the published literature. Hierarchical clas- sification and curation of protein domains, using our in-house tools CDTree (hierarchy viewer) and Cn3D (structure viewer and multiple alignment editor), have been the focus of our manual curation efforts. We also develop structural motif models (accessions with an “sd” prefix) to represent protein sequence segments such as short repeats, coiled coils, and transmembrane regions. We manually validate superfamily clusters (accessions with a “cl” prefix), formed by an automated clustering procedure as sets of models that generate overlapping annotation on the same protein sequences. Superfamily clustering allows the organization of data within CDD in a non-redundant way; it is aided by using Cytoscape as a visualization tool for the degree of overlap between conserved domain models. More recently, our manual curation efforts are focused on providing functional labels for domain architectures, using an in-house procedure called SPARCLE (“Specific ARChitecture Labeling Engine”). While we are able to assign functional labels to a large fraction of proteins, we have also identified areas of insufficient coverage and resolution of the current domain models that comprise CDD. The need for manual curation work always exceeds available resources and we hope to automate hierarchical classifications to some degree in the near future. Acknowledgement: This research was supported by the Intramural Research Program of the National Library of Medicine, NIH.

Functional Clustering of the Amidohydrolase Julia Hayden1 1University of Richmond High throughput methods of protein sequencing have rapidly increased the number of sequenced proteins. However, analyzing this influx of protein data remains a challenge as experimentally determining function has high cost and time requirements. Sequence similarity analysis was thought to be a solution to this challenge. However, subsequent work has demonstrated sequence similarity methods cannot accurately classify proteins based on functional detail. Thus, there is rampant protein misannotation in protein databases relying on these methods. We have developed two processes to functionally cluster proteins. The first process, TuLIP (Two-Level Iterative clustering Process), clusters structures based on expert-defined active site motifs. The second, MISST (Multi-level Iterative Sequence Searching Tech- nique), utilizes active site sequence motifs of TuLIP-defined protein groups to identify hypothesized functionally similar proteins in GenBank. The Amidohydrolase superfamily is comprised of metal- dependent proteins involved in a wide range hydrolysis reactions involving amide or ester functional groups at carbon and phosphorus centers. Protein structures of the Amidohydroase superfamily were clustered by TuLIP and resulted in thirty-two hypothesized functional groups. The overall number of

20 ABSTRACT

Amidohydrolase proteins identified in the initial MISST search to the final MISST search has increased by over 1.3-fold, indicating the importance of iterative searching to identify and cluster the entire superfamily. Furthermore, the MISST process produced protein clusters sharing a more detailed level of function than protein groupings in databases such as Pfam. Thus, the processes TuLIP and MISST are able to group known Amidohydrolase proteins in hypothesized functional groups, but also identify many proteins novel to the Amidohydrolase superfamily that hypothetically share function.

A proteome view of structural, functional, and taxonomic characteristics of major protein domain clusters Chia-Tsen Sun1, Austin W.T. Chiang1, Ming-Jing Hwang1 1Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan With the number of completely sequenced genomes increasing rapidly, studies on various aspects of protein structure, function, and evolution are increasingly amenable to study at the level of the proteome. Here, we present results from a bi-clustering analysis on presence/absence data for 6,580 unique protein domains (PDs) in 2,134 species with a sequenced genome, thus covering a complete set of proteins, for the three superkingdoms of life, Bacteria, Archaea, and Eukarya. Our analysis revealed eight distinctive clusters of PDs, which, following an analysis of enrichment of Gene Ontology functions and CATH classifi- cation of protein structures, were shown to exhibit structural and functional properties that are taxa- characteristic. The largest of the eight clusters constituted a set of PDs created early in evolution and retained in living organisms, as evidenced by their being present in the majority of species in the three superkingdoms, and characterized by basic cellular functions and ancient structural architectures. The identification of these taxa-characteristic PDs may prove pivotal in furthering our understanding of the intricate relationship linking proteins and the evolution of the species in which they are used.

Mapping side chain interactions at protein helix termini Nicholas Newell1 1Newell, NE. BMC Bioinformatics 2015 16:231 Side-chain interactions near the ends of protein helices stabilize helix termini and shape the geometry of the adjacent loops. Known N-terminal side-chain motifs include the Asx/ST N-caps, the capping box, and hydrophobic and electrostatic interactions. At the C-terminus, capping is often achieved with main- chain polar groups, (e.g. the Schellman loop), but here also particular side-chain motifs clearly favor specific loop geometries. Key questions that remain concerning side-chain interactions at helix termini include: 1) To what extent are observed helix-terminal interactions that include multiple amino acids likely to represent genuine cooperative interactions between side-chains, rather than chance alignments? 2) Which particular helix-terminal loop geometries are favored by each side-chain interaction? 3) Can an exhaustive statistical scan of a large, recent dataset identify new side-chain interactions at helix termini? In this work, three analytical tools are applied to answer the above questions. First, a new 3D clustering algorithm is applied to partition helix terminal structures by loop backbone geometry. Next, a motif detection algorithm (Cascade Detection, Newell, Bioinformatics, 2011) is applied to each cluster to determine which motifs are most important in each loop geometry. Finally, the results for each motif are displayed in a CapMap, a 3D conformational heatmap that depicts the distribution of the motif’s abundance and overrepresentation across all loop geometries. This work identifies a “toolkit” of side-chain motifs which are good candidates for use in the design of synthetic helix-terminal loops with specific desired geometries, because they are used in nature to support these geometries. Highlights of the analysis include determinations of the favored loop geometries for the Asx/ST motifs, capping boxes, big boxes, and other previously known and unknown hydrophobic, electrostatic, H-bond, and pi-stacking interactions.

21 ABSTRACT

Functional Labeling of Protein Domain Architectures with SPARCLE Thanki N1, Han L1, Lanczycki CJ1,HeJ1,LuS1, Chitsaz F1, Derbyshire MK1, Gonzales NR1, Gwadz M1, Lu F1, Marchler GH1, Song JS1, Yamashita RA1, Zheng C1, Bryant SH1, Geer L1, Marchler-Bauer A1 1National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health SPARCLE, the SPecific ARChitecture Labeling Engine, is a curation interface developed for the Conserved Domain Database (CDD) team at the National Center for Biotechnology Information (NCBI). SPARCLE is used to associate conserved domain architectures with suggested protein names and brief functional descriptions or labels, as well as selected corresponding evidence. The specific or sub-family domain architectures result from annotating protein sequences with domain footprints provided by the CDD, a resource that consists of a collection of well-annotated multiple sequence alignment models for ancient domains and full-length proteins. This resource includes NCBI-curated domains, which use 3D-structure information to explicitly define domain boundaries and provide insights into sequence/structure/function relationships, as well as domain models imported from external source databases such as Pfam, SMART, and TIGRFAMs. The protein names and labels recommended in SPARCLE range from generic to very specific, reflecting the status quo of the underlying protein and domain model collections. They may, however, provide concise functional annotations that are easier to interpret than raw representations of domain architecture.

PC – CHAPERONES Design and Characterization of Small Heat Shock Protein Dimers Caley Butler1 1Mount Holyoke College Small heat shock proteins (sHsps) are a ubiquitous part of the machinery that maintains cellular homeostasis by acting as molecular chaperones. When cells experience stress conditions such as heat, proteins

22 ABSTRACT denature and expose hydrophobic surfaces to the aqueous cellular environment. These exposed hydrophobic surfaces can form damaging aggregates that are hallmarks of many diseases such as Alzheimer’s and cataracts. sHsps bind to, prevent the aggregation, and can promote the refolding or degradation of these unfolded client proteins in an ATP-independent manner (1). sHsps are found in vivo as polydisperse oligomers. The dynamic nature of these oligomers has prevented the elucidation of a specific mechanism of action for how s Hsps recgnize and bind their client proteins. Structural studies of sHsps reveal a highly conserved a-crystallin domain flanked by variable N and C terminal domains. Despite the large gaps in understanding of the mechanism of action, several regions within the Nterminal domain are suggested to facilitate client binding. Additionally, based on crystal structures it is proposed that the dimer is the smallest active chaperone unit, and larger oligomers may act as storage depots for sHsps (2,3). HspB1 and HspB5 are both widely produced in human tissues as a stress response to prevent client protein aggregation. In order to pin down the role of the dimer in chaperone activity, glutathione-s-transferase (GST) was genetically linked as a fusion protein to the N-terminus regions of both proteins, in order to constrain HspB1 and HspB5 forms to a dimer. My studies suggest that the fusion proteins form dimers and they function as active molecular chaperones. Furthermore, the two different fusion proteins demonstrate different chaperone activity in relation to multiple different substrate proteins. References: 1. Basha, E., O’Neill, H., and Vierling, E. (2012) Small Heat shock proteins and a-crystallins: dynamic proteins with flexible functions. TIBS, 37, 1-12. 2. McDonald E. T., Bortolus, M., Koteiche, H.A., and McHaourah, H. S. (2012) Sequence, Structure, and Dynamic Determinants of Hsp27 (HspB1) and Equilibrium Dissociation are Encoded by the N-terminal Domain. Biochemistry. 3. Delbecq, S.P., Jehle, S., and Klevit, R. (2012) Binding determinants of the small heat shock protein, aB crystalline: recognition of the “IxI” motif. EMBO J. 31, 4587-4594.

Identification and Characterization of Small Heat Shock Protein Interacting Domains Elizabeth De Leon1 1Department of Chemistry, Mount Holyoke College The small heat shock proteins (sHsp), ubiquitous cellular homeostasis machinery, are oligomeric chaperone proteins. sHsps play an important role in maintaining cell function and survival under stress conditions such as high temperatures. By binding to non-native proteins in an ATP- independent manner, sHsps effectively prevent harmful aggregation of denatured proteins, promote proper protein folding, or facilitate protein degradation. Defects in sHsps can result in abnormal accumulation of proteins or aberrant protein folding, which lead to many pathologies such as cata- ract, tumor and neurodegenerative diseases including Parkinson’s disease and Alzheimer’s disease(1). Structural studies have demonstrated that sHsps contain a highly conserved a-crystallin domain flanked by variable N-terminal and C-terminal regions (NTR and CTR, respectively). sHsps usually exist as large, polydisperse oligomers and therefore, due to their dynamic organization, the specific mechanism(s) of how sHsp bind to substrate and function are not well understood. Nonetheless, it is suggested that the variable NTR of the protein contributes to substrate binding and oligomerization(2, 3). We have purified the NTR of HspB1 and determined it has chaperone activity for at least two model substrates. Furthermore, we have attached the NTR to gold nanoparticles, creating multivalent peptide- nanoparticle conjugates (artificial sHsps), which also demonstrate chaperone activity. Further experiments will include probing additional regions of HspB1 for chaperone activity and to determine the therapeutic potential of “artificial” sHsps.

23 ABSTRACT

References: 1. Arrigo, A. P., Simon, S., Gibert, B., Kretz-Remy, C., and Nivon, M. (2007) Hsp27 (HspB1) and aB- crystallin (HspB5) as therapeutic targets, FEBS Lett. 2. McDonald, E. T., Bortolus, M., Koteiche, H. A., and McHaourab, H. S. (2012) Sequence, Structure, and Dynamic Determinants of Hsp27 (HspB1) Equilibrium Dissociation are Encoded by the N-terminal Domain., Biochemistry. 3. Delbecq, S. P., Jehle, S., and Klevit, R. (2012) Binding determinantsof the small heat shock protein, aB-crystallin: recognition of the “IxI” motif., EMBO J. 31, 4587–4594.

Molecular mechanism of protein kinase recognition and sorting by the Hsp90 kinome-specific cochaperone Cdc37 Dimitra Keramisanou1, Adam Aboalroub1, Ziming Zhang1, Ralf Landgraf2 and Ioannis Gelis1 1Department of Chemistry, University of South Florida, 2Department of Biochemistry and Molecular Biology, University of Miami, Miller School of Medicine The Hsp90 chaperone machinery is a ubiquitous mediator of protein homeostasis and activity. Despite its essential roles, little is known about the molecular mechanism that controls substrate entry into its chaperone cycle. Here we show that the functional role of the kinome-specific cochaperone, Cdc37, reaches beyond that of an adapter protein and find that it actively participates in the selective recruitment of only client kinases. The cochaperone recognizes kinase specificity determinants in both clients and non-clients and thus it acts as a general kinase scanning-factor. Kinase sorting within the client to non-client continuum relies on the ability of Cdc37 to challenge the conformational stability of client kinases by locally unfolding them. This metastable conformational state has high affinity for Cdc37 and

24 ABSTRACT forms stable complexes through an extended, multi-domain cochaperone interface. On the other hand, the interaction with non-clients is not accompanied by conformational changes of the substrate and results in substrate dissociation. Collectively, Cdc37 performs a quality control of protein kinases, where induced conformational instability acts as a “flag” for Hsp90 dependence and thus stable cochaperone association.

Prion Propagation and Curing by AAA1 Chaperone Proteins in the Saccharomyces cerevisiae model system PSI1 Shannon May1, Jodi L. Camberg1,2 1Interdisciplinary Neurosciences Program, University of Rhode Island, 2Department of Cell and Molec- ular Biology, University of Rhode Island All living cells have a vast protein quality control network that maintains protein homeostasis. Molecular chaperone proteins are an integral component of this network, helping to fold nascent polypeptides, refold misfolded or partially unfolded proteins, and target irreversibly misfolded proteins for degradation by the proteasome. The AAA1 (ATPases associated with diverse cellular activities) superfamily of proteins includes a diverse set of ATPases that share similar domain organization. This protein family also includes Hsp100 proteins, such as the chaperone Hsp104, which can disassemble amyloid fibers along with co-chaperones, and Lon, which has both protein unfolding and proteolytic activity. Hsp104 is required for prion propagation from mother to daughter cells and overexpression of Hsp104 cures yeast cells of mature amyloid fibers by promoting their disassembly. In the yeast model system [PSI1], we are studying amyloid formation by Sup35, a translation termination factor that aggregates to form amyloid fibers. We have developed a quantitative colorimetric assay, based on the accumulation of the metabolite 5-aminoimidazole ribotide, to distinguish between cells that contain amyloid and cells that have been cured of amyloid. To investigate substrate targeting and amyloid disassembly by Hsp100 family members, we constructed a plasmid expression library of hsp104 and lon genes containing random N-domain mutations. We are currently screening mutants for gain- and loss-of-function mutations in the yeast [PSI1] model system that cause altered prion propagation and amyloid clearance. This study provides new insight into chaperone-mediated disassembly of amyloids, which have been implicated in several neurodegenerative diseases, such as Alzheimer’s disease, Parkinson’s disease, and Hun- tington’s disease.

Design and Characterization of the Chaperone Activity of “Artificial” Small Heat Shock Proteins Kathryn McMenimen1 1Mount Holyoke College Small heat shock proteins (sHsps) are a family of molecular chaperones broadly employed across many organisms and contain the canonical “a-crystallin domain” (ACD). This domain is flanked by an N- and C- terminal region (NTR and CTR, respectively) of varying length and sequence, believed to participate in substrate and quaternary interactions. In addition, a defining feature of sHsps is their ability to form dynamic, polydisperse oligomers that exhibit subunit exchange under certain conditions. The physiological relevance of oligomerization and mode(s) of chaperone function remain undetermined. Previous studies have identified “mini-chaperones”, which are fragments of the ACD that demonstrate chaperone activity by interacting with substrate proteins. In order to identify other functional regions of sHsps, we purified the relatively unstructured N-terminal sequence from human HspB1. Using two model substrates, malate dehydrogenase (MDH) and citrate synthase (CS) we have identified that this 88-residue sequence exhibits chaperone activity in solution. In further studies to determine the importance of oligomeric organization (multivalent interactions) in chaperone activity we constructed gold nanoparticles (AuNPs) appended with the HspB1 NTR. These sHsp-AuNPs exhibit chaperone activity, although their activity varies depending on the substrate. In exciting results, we observe that sHsp-AuNPs are more active

25 ABSTRACT chaperones for citrate synthase than unconjugated sHsp NTRs. Finally, our results indicate the importance of the NTRs in chaperone activity and demonstrate the therapeutic potential of sHsp-AuNPs.

PD – CHEMICAL BIOLOGY The oxidation of hydroxilamines with H2O2 mediated by Myoglobin Lucıa Alvarez1, Sebastian Suarez1, Fabio Doctorovich1, Marcela Martı2 1Departamento de Quımica Inorganica, Analıtica y Quımica Fısica/INQUIMAE-CONICET, 2Departa- mento de Quımica Biologica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Ciudad Universitaria, Pabellon 2, Buenos Aires, C1428EHA. Argentina Nitroxyl (HNO) has biological and pharmacological properties compared to those of other nitrogen oxides. It regulates calcium channels in both the cardiovascular and nervous systems. In addition, HNO donors increase cardiac function in an additive and independent manner to b-adrenergic signaling. Such findings have led recently to extensive investigation of the chemical, biochemical and pharmacological properties of HNO. Many heme proteins function as peroxidases in the catalyzed oxidation of a wide range of substrates. Despite the protective and functional utility of peroxidases, an increase in peroxidase activity has been implicated in the pathology of a number of diseases. In this work, we examined the ability of Myoglobin to produce HNO from peroxidation of hydroxyl- amines such as NH2OH and CH3NHOH with H2O2 using both experimental and computational approach. In previous works1 it has been proposed that this reaction mediated by different hemeproteins ocurrs by the following mechanism: According to this mechanism we studied the reaction kinetic of these two different steps following spectral changes that occurred during the reaction. In addition, the HNO production was confirmed using an electrochemical method with a Cobalt Porphyrin bound to gold developed in our group.2 In order to have an insight in the reaction mechanism, we performed QM/MM experiments calculating the energy barrier for the HNO production considering different spin states. It can be concluded that the oxidation of hydroxilamines with H2O2 mediated by Myoglobin produced HNO and it could be a possible endogenous source of HNO. References: 1. Donizelli, S., Free Radical Biology and Medicine; 2008, 45, 578-584. 2. Suarez, S., Inorganic Chemistry; 2010, 49, 6955-6966.

Probing cation-pi interactions of Heterochromatin Protein 1 using in vivo unnatural amino acid mutagenesis Stefanie Baril1 1University of North Carolina, Chapel Hill Regulation of gene expression is critical to cell development and function. Lysine methylation of histone proteins is a well-recognized means of epigenetic control and dysregulation of lysine methylation has

26 ABSTRACT been implicated in multiple cancers. Proteins that interpret methylation states of lysine are called reader proteins. In humans, these readers utilize a cage of aromatic residues that interact with the N-methyl groups of methylated lysine (MenK) residues. Not surprisingly, cation-pi interactions have been implicated as major contributors to MenK recognition. Dougherty and coworkers were the first to study cation-pi interactions using unnatural amino acids (UAAs). Here we expand on Dougherty’s approach by using in vivo UAA mutagenesis of a trimethyl lysine (Me3K) reader protein. Two tyrosine residues in the aromatic cage of Heterochromatin Protein 1 (HP1) of Drosophila melanogaster were mutated to UAAs with various electron-withdrawing and –donating groups. Me:3:K recognition was monitored by ITC and a correlation between binding and electrostatic potential of the UAAs’ R-group was observed at both residues. Protein crystallography confirmed that changes in binding were not due to changes in structure. One position (Y24) appeared to exhibit more of a pronounced effect on binding than the other (Y48). This is likely due to the orientation of the Me3K peptide within the pocket. This provides, to our knowledge, the first example of differential, residue-specific cation-pi effects observed in a reader protein.

The Influence of Distal Residues on Catalysis through Alpha-helical Motion Shanadeen C. Begay1, Penny J. Beuning1, Mary Jo Ondrechen1 1Department of Chemistry and Biochemistry, Northeastern University The human enzyme phosphoglucose isomerase (hPGI) is important in glycolysis. It has been shown that distal amino acids, residues 11-13 Å away from the site of reaction, contribute to catalysis, such that single-site, conservative mutations result in significant loss of catalytic activity. For instance, the catalytic efficiency of the D511N variant is 230-fold less than that of wild type and the K362A variant shows no detectable activity. Molecular dynamics simulations of wild-type and variant hPGI in solution will be discussed. The K362-D511 salt bridge is postulated to be important in dynamic processes that may be critical to catalysis. Preliminary studies of the motion of alpha-helix 23, and its possible role in catalysis, will be reported. These dynamic studies may reveal how distal residues modulate conformational changes that are important for catalysis.

Biochemical characterization of Med25 and its protein-protein interaction network Matthew Beyersdorf1,2, Mahmoud Abedaljawad1,2, Paul A Bruno2, Kevon Stanford2, Steven M Sturlis2, Ashootosh Tripathi2, David H Sherman1,2, Anna K Mapp1,2 1Program in Chemical Biology, University of Michigan, 2Life Sciences Institute, University of Michigan The transcriptional coactivator Mediator subunit 25 (Med25) is a component of the Mediator complex (1.2 MDa in size), a coactivator complex that is required for activated expression of most eukaryotic genes. The activator interaction domain (ACID) of Med25 is a discrete component which has been shown to interact with a number of transcriptional activators including VP16 (herpes simplex viral infections), ATF6a (endoplasmic reticulum oxidative stress response); and ERM (cancer metastasis; key member of ETS-domain transcription factors), leading to the recruitment of the Mediator complex to specific gene targets. Published structures of Med25 ACID have demonstrated that this motif contains two dis- parate binding sites however binding modes and specificities of the various interacting partners for each individual site have not been described. In this work, we have used HSQC NMR techniques and site-directed mutagenesis in order to determine modes of molecular recognition between Med25 ACID and its native interaction partners. Importantly, these experiments have defined which peptide ligands bind to each site of Med25 ACID. Additionally, we have been focused on the deconvolution and structural elucidation of natural product small molecules following a high-throughput screening effort of a library of natural product extracts maintained by the University of Michigan Center for Chemical Genomics. Through these efforts, we have demonstrated that Med25 ACID can be specifically targeted

27 ABSTRACT and is amenable to small molecule modulation. We anticipate that the use of small molecule inhibitors as well as native peptide ligands to selectively target Med25 ACID in cellulo will serve as useful tools towards elucidating functional significance of Med25 and its protein-protein interaction network.

Design of Modular Switches for Allosteric Control over Protein Kinases and Protein Phosphatases Matthew Bienick1 1Department of Chemistry and Biochemistry, University of Arizona Selectively modulating the activity of a desired enzyme in vivo is a major goal in protein design and can aid in the development of methods for understanding and rewiring cell-signaling pathways. Protein kinases and phosphatases are complementary enzymes that catalyze the addition and removal of phosphate groups upon substrate proteins, respectively. Kinases and phosphatases are implicated in almost every signaling pathway and their deregulation is implicated in many diseases such as cancer and neurodegeneration. The high structural homology of kinases and phosphatases presents a challenge in designing selective inhibitors for understanding their cellular roles. Though powerful genetic knockdown or knockout tools exist, they are susceptible to compensatory cellular mechanisms and do not allow for titratable activity. We have addressed this problem by designing a potentially general allosteric approach for gating kinase and phosphatase activity. We have utilized the well-studied protein-protein interactions between Bcl-2 and BH3-only peptides and their small molecule inhibitors. We have designed a system where specific BH3- only peptides, 20 to 25-residues, are inserted into an enzyme, at predetermined non-homologous positions. BH3-only peptides, such as Bad, are unstructured but adopt a rigid, a-helical conformation upon the addition of a protein binding partner, such as, Bcl-xL. Thus in our system, Bcl-xL acts as a poison and allosterically inhibits the function of the Bad-inserted-enzyme. Subsequently, the addition of a small molecule inhibitor, ABT-737, binds to and displaces Bcl-xL, acting as an antidote, thus restoring enzymatic activity. We have shown that this method allows for controlling the activity of both kinases and phosphatases with a small molecule in a dose dependent fashion both in vitro and in cellulo. We are currently optimizing these allosteric enzyme systems in order to study cell signaling and to redesign new pathways.

Structural Basis of Chemokine CXCL7 Function in Neutrophil-Platelet Crosstalk: Role of Dimer Formation and Glycosaminoglycan Interactions Aaron Brown1 and Krishna Rajarathnam1,2,3 1Department of Biochemistry and Molecular Biology, 2Sealy Center for Structural Biology, 3Depart- ment of Microbiology and Immunology at the University of Texas Medical Branch, Galveston, TX Chemokines are important signaling proteins involved in a variety of biological processes. One such process is a subset of chemokines that play a critical role in inflammation by recruiting and activating

28 ABSTRACT neutrophils to eliminate microbial infections. In particular, recruitment of neutrophils into the thrombus during bacterial infection or tissue injury, specifically, is regulated by CXCL7 (NAP-2), a neutrophil activating chemokine released in high concentration from platelets. Chemokine function must be tightly regulated and involves activating seven transmembrane G-protein coupled receptors and binding to glycosaminoglycans that regulate the receptor-binding process. Glycosaminoglycans (GAG) are highly sulfated polysaccharides that are present on the endothelium and the extracellular matrix. Furthermore, CXCL7 function is also intricately linked to dimer and potentially heterodimer formation. Our recent work on CXCL7-glycosaminoglycan interactions and heterodimer formation using engineered proteins and NMR spectroscopy revealed several unique features of CXCL7. Among these features are its ability to form heterodimer, unique GAG binding profiles, and a unique monomer-dimertetramer equilibrium. GAG titrations on 15N-labeled CXCL7 also provide molecular-level insight into interactions between the various states of CXCL7 and GAG. Our work shows for the first time the formation of heterodimer between CXCL7 and CXCL1 and to our knowledge is the first report of chemokine heterodimer interactions with GAG. These studies on the molecular-level interactions between CXCL7 and CXCL1 and GAG provide novel insight into the mechanisms of neutrophil recruitment and the link between oligomerization, GAG-binding, and function that may lead to development of new therapeutics for thrombus- related diseases.

Characterization of Novel Pth-like Nucleotide Binding Protein PTRHD1 Geordan Burks1 1University of Alabama in Huntsville Peptidyl-tRNA hydrolases (Pths) are essential enzymes ubiquitous across all living organisms. Based on sequence and structural prediction, PTRHD1 was hypothesized to be a human Pth, belonging to the Pth2 superfamily. However, the absence of key catalytic residues prompted further investigation into Pth activity. Herein presented are the recombinant expression, purification, and storage buffer optimization for PTRHD1. Also presented is characterization of the PTRHD1 interaction with oligonucleotides, but inability to remove the peptide moiety from peptidyl-tRNA. Thus, PTRHD1 is not a peptidyl-tRNA hydrolase, but a nucleotide binding protein with yet unknown function.

29 ABSTRACT

Controlling the Phosphoproteome: Ligand-Gated Split-Kinases and Split-Phosphatases Javier Castillo-Montoya1, Karla Camacho-Soto2, Blake Tye3, Luca Ogunleye4, and Indraneel Ghosh1 1Department of Chemistry and Biochemistry, University of Arizona, Tucson, AZ, 2Luceome Biotechnol- ogies, Tucson, AZ, 3Wolfe Laboratories, Woburn, MA, 4Department of Chemistry and Chemical Biol- ogy, Harvard University, Cambridge, MA. The activity and function of most proteins are regulated by chemical perturbations, such as phosphorylation, acetylation, and methylation. A third, if not more, of the proteome is phosphorylated and phosphorylation dependent signaling is of paramount importance in cells. Two large enzyme families, protein kinases and phosphatases, catalyze phosphorylation and dephosphorylation. Aberrant regulation of phosphorylation is implicated in many diseases, including cancer and neurological disorders. Thus, understanding the specific role of particular kinases or phosphatases is important for understanding their biology and designing new therapies. The structural similarity of both enzyme families makes it challenging to selectively turn-on or turn-off a specific enzyme using activators or inhibitors. To address this problem, we developed a sequence dissimilarity approach to identify sites in kinases and phosphatases tolerant to fragmentation. Subsequently, we dissected the kinase or phosphatase at these sites into two inactive fragments and attached them to protein pairs, such as FKBP and FRB, which can heter- odimerize only upon addition of a small molecule chemical inducer of dimerization (CID). The addition of a CID allows for ligand-gated reassembly of the splitenzyme and concomitant activity. We have demonstrated the feasibility of this approach in vitro with members of the tyrosine kinase group, such as Lyn and Src, and with tyrosine phosphatases, such as PTP1B and SHP1. In order to control several split- enzymes simultaneously, we adopted three orthogonal CIDs: rapamycin, abscisic acid, and gibberellic acid. We are currently developing methods to selectively turn-on desired kinases and phosphatases in live cells, and methods to develop a novel CID system that can be rapidly turned-off upon irradiation with UV light. Thus together, we provide a methodology that potentially allows for posttranslational orthogonal small molecule control over the activity of user defined split-kinases and split-phosphatases for interrogating and redesigning signaling pathways.

30 ABSTRACT

Semisynthesis of human selenoenzyme cSelS through native chemical ligation Rujin Cheng1, Zhengqi Zhang1, Sharon Rozovsky1 1University of Delaware Department of Chemistry and Biochemistry Selenoprotein S (SelS) is an intrinsically disordered membrane enzyme that is a member of the ER- associated protein degradation (ERAD) pathway. The ERAD governs the extraction of misfolded proteins or misassembled protein complexes from the ER’s membrane and lumen and their transport to the cytoplasm where they are degraded by the proteasome. Selenocysteine (Sec) is a genetically encoded amino acid but its incorporation into proteins is distinctly different than that of the 20 canonical amino acids. Uniquely, Sec insertion requires a dedicated suite of proteins to reprogram the opal codon UGA to encode for Sec using a structural element in the protein mRNA. Sec incorporation efficiency remains low due to the opal codon being decoded as a termination signal. Hence it remains challenging to prepare selenoproteins. In order to characterize SelS’s role in the ERAD membrane complex, we have developed a preparation method by native chemical ligation – a technique that relies upon an amide- forming reaction to generate proteins from their respective fragments. We describe the preparation, characterization and C terminus SelS enzymatic activity as a disulfide reductase.

Dissecting conformationally dynamic transcriptional activator-coactivator complexes via selective covalent ligands Andrew Henderson1,2, Matthew Henley1,2, Nicholas Foster1,2 Zachary Hill3, James Wells3, Anna Mapp1,2 1Program in Chemical Biology, University of Michigan, 2Life Sciences Institute, University of Michigan, 3Department of Pharmaceutical Chemistry, University of California San Francisco Protein-protein interaction (PPI) networks between transcriptional activators and coactivators are essential for the maintenance of regulated gene expression. The coactivator Med25, a member of the Media- tor complex, has emerged as a critical linchpin linking several disease-associated transcriptional activators and the transcriptional machinery. Med25 makes critical contacts with the amphipathic activators VP16 (viral infection), ATF6a (oxidative stress response), and PEA3 family members (tumor progression and metastasis). Activators interact directly with Med25 through two separate surfaces of the Activator Interaction Domain (AcID). Despite the importance of the AcID motif, the molecular basis of interactions between activators and Med25 is poorly understood. Using several classes of covalent chemical probes, the complexes between activators and Med25 AcID were studied to dissect the role that conformational dynamics and allostery play in mediating binary and ternary complex formation. A series of activator-derived covalent peptides were developed that display a striking degree of selectivity for distinct surfaces of AcID depending upon the activator sequence. These covalent activators have been used to detect an allosteric network linking the two binding surfaces of AcID and as inhibitors of PPIs between activators and Med25. A panel of covalent small molecules was also discovered through the use of a Tethering screening campaign. These molecules are being assessed for their ability to recapitulate the effects to AcID structure and activator binding that were observed with the covalent activators. While both sets of covalent probes have enabled the ability to study Med25-activator complexes in a site-specific manner, the covalent small-molecules represent the most attractive candidates for use in cellular studies and as eventual therapeutics.

Utilizing computational and experimental chemistry to characterize the functions of Structural Genomics proteins. Caitlyn Mills1, Penny J. Beuning1, Mary Jo Ondrechen1 1Department of Chemistry and Chemical Biology, Northeastern University There are now over 13,700 Structural Genomics (SG) protein structures deposited in the Protein Data Bank, but many are of unknown biochemical function or have putative functional assignments that are

31 ABSTRACT often incorrect. The accumulated structural information from SG represents a tremendous contribution to structural biology and genomics. However, the addition of better functional predictions for SG proteins would add substantial value to this structural information. While the majority of current practices in protein function prediction is almost purely informatics based, our more powerful approach incorporates structure-based computed chemical properties. Improved, verified computational methods can yield more reliable predictions of function that can lead to the identification of new drug targets and of proteins with other useful applications, from biomass conversion to disease control. Our approach starts with Partial Order Optimum Likelihood (POOL), a machine learning method, to computationally predict the residues important for catalysis in a protein structure. Next, Structurally Aligned Local Sites of Activity (SALSA) uses POOL-predicted residues and multiple structure alignments of proteins of known function to develop unique, spatially-localized consensus signatures for each functional family within a given superfamily. We then compare the POOL-predicted residues for the SG proteins to the consensus signatures by aligning the residues and scoring the alignment to determine the degree of similarity between the local structures. Finally, this score is used to determine the best functional assignment for each SG protein. This research focuses on the Crotonase Superfamily, which we have determined contains many mis- annotated SG proteins. In some instances, we can provide better putative functional annotations for the SG proteins and have acquired experimental data supporting our predictions. The goal is to provide a validated approach to functional annotation for wider application by the community. Aknowledgement: Funded by NSF-CHE-1305655.

Rapid Bioorthogonal Protein Conjugation Reactions via ortho-Formylphenylboronic Acid-Based Coupling Chemistry Kamalika Mukherjee1, Tak Ian Chio2, Saptarshi Ghosh2, Han Gu2, Zhen Lei2, Samantha L. Grieco2, Susan Bane2 1Department of Medicine, Division of Nephrology, Massachusetts General Hospital/Harvard Medical School, Boston, MA, 2Department of Chemistry, State University of New York at Binghamton, Bing- hamton, NY Modern chemical methods for modification of proteins are increasingly focused on bioorthogonal reactions, which are chemical reactions that occur between two exogenous reactive groups in the presence of the naturally occurring functional groups. One of the oldest methods used for bioorthogonal reactions is that of “a-effect amines” (hydrazine, hydrazide, and other amine-nitrogen compounds, aminooxy reagents, etc.) as nucleophiles and carbonyl functionalities (aldehydes or ketones) as electrophiles. A great advantage to this system is that the functional groups are synthetically accessible and may be incorporated into biomolecules. A significant disadvantage, which is shared by most bioorthogonal reactions, is slow reaction rate at physiological pH. We have discovered that addition of a boronic acid functional group ortho- to an aldehyde or ketone functionality dramatically increases the reaction rate and alters the outcome of the reaction, forming a boron-nitrogen heterocycle as the final product. In this work, the structure-reactivity relationship of the bioorthogonal pair is explored with respect to its utility for protein bioconjugations. Both hydrazines and hydrazides react rapidly with the boronic acid- containing electrophile 2-formylphenyboronic acid (2fPBA) to form the heterocyclic product. Fluorescent labeling of a protein can be accomplished in minutes at micromolar protein and fluorophore concentrations with these reagents. Products of the reactions are very stable in aqueous solution, and the conjugates remain intact after SDS-PAGE. Application of this conjugation chemistry to protein crosslinking is also demonstrated.

32 ABSTRACT

DNA binding and unwinding mechanisms of archaeal and human DNA repair helicase homologues Hel308 and HelQ. Sarah Northall2, Rebecca Lever1, Nathan Jones1, Panos Soultanas2, Edward Bolt1 1School of Life Sciences, The University of Nottingham, 2School of Chemistry, The University of Nottingham. Hel308 and HelQ are DNA translocase homologues implicated in recombination and repair of stalled replication forks in archaea and metazoans respectively. Although atomic structures of archaeal Hel308 have been obtained, some domains remain to be assigned a functional role, and mechanisms coupling core ATPase (RecA) domains with DNA translocation are unknown. We have used biochemical analyses of purified archaeal Hel308 and human HelQ to evaluate and compare their DNA binding and translocation mechanisms. This showed that purified non-canonical winged helix (WH) domains from Hel308 and HelQ bind to DNA, which we propose is important for targeting of branched substrates typical of DNA forks and recombination intermediates, possibly analogous to PriA helicases. Further biochemical analyses identified a motif conserved in Hel308 and HelQ is required to transduce RecA domain ATP hydrolysis into active DNA translocation. We also present new data on interplay of human HelQ with Rad51 and Rad51 paralogs that are essential for controlling DNA repair by formation of branched DNA intermediates triggering homologous recombination.

Next-generation assays to define and manipulate lysine acetyltransferase function Jonathan Shrimp1 1National Cancer Institute, NIH Lysine acetyltransferase (KAT) enzymes are key regulators of gene expression and signaling in many diseases, including cancer. However, a substantial limitation to our understanding of this enzyme superfamily is a lack of high quality chemical probes to study KAT activity in living cells and organisms. Here, we describe a multi-faceted assay approach to identify and evaluate small molecule KAT inhibitors. First, we detail the development of a microfluidic mobility shift assay used for high-throughput screening and identification of novel chemotypes targeting KAT activity. Second, we describe the use of chemo- proteomics to evaluate the global selectivity of known KAT inhibitors. Thus far, our findings provide guidance on the application of the currently known inhibitors along with revealing the need for improved KAT inhibitors. Finally, we report preliminary studies on the application of a Cas9-KAT fusion protein to serve as a live-cell reporter of KAT activity. These next-generation assay technologies are a powerful set of tools to identify novel KAT inhibitors and will facilitate the ultimate goal of defining the role of KATs in genome function and cancer pathogenesis.

From Synthesis to Pathology: Identifying Glucosepane’s Role in Diabetes and Aging Matthew Streeter1, Nam Kim1, Jonathan Clark1, Jason Crawford1, David Spiegel1 1Yale University Advanced glycation end-products (AGEs) are non-enzymatic protein modifications formed from the reaction of sugars with basic amino acid residues. Of the AGEs present in the extracellular matrix, a lysine-arginine crosslink called glucosepane appears to be of particular importance, being two orders of magnitude more prevalent than other AGEs and being strongly correlated to several diabetic complica- tions, atherosclerosis, and aging. Despite evidence suggesting an important role in human health, studies on glucosepane have proven difficult given its molecular complexity. Our lab has leveraged our synthetic capabilities towards a bottom-up approach to studying this molecule, publishing the first total synthesis last year. Currently, we are working to incorporate the molecule into more physiologically relevant systems, and developing tools to better probe its effects on human health and biology.

33 ABSTRACT

Novel Constrained Peptides Targeting Protein-Protein Interfaces Haifan Wu1, Matthew Bienick2, Lijun Liu3, Yibing Wu1, Indraneel Ghosh2, William DeGrado1 1Department of Pharmaceutical Chemistry, University of California, San Francisco, CA, 2Department of Chemistry and Biochemistry, University of Arizona, AZ, 3Cardiovascular Research Institute, Univer- sity of California, San Francisco, CA Protein-protein interactions (PPIs) are involved in virtually all kinds of biological processes. The modulation of these interactions by synthetic molecules or biologics presents a strategy to study these interactions, as well as a way to treat a variety of diseases. One feature of PPIs is the large interface, which makes modulation by small molecules challenging. Peptides derived from fragments of the protein interface can be used for inhibiting these interactions. However, short peptides are less likely to adopt well-defined secondary structures, and they generally have poor pharmacological properties, such as in- vivo stability and cell permeability. These drawbacks have significantly limited the use of peptides as therapeutic agents. To overcome this issue, macrocyclization, such as peptide stapling, is used. However, this method leaves the two termini of the helices exposed, which decreases the stability and permeability. To address this issue, we designed head-to-tail cyclized “cap-strapped peptides” using linkers that can incorporate helix N- and C-capping motifs. The design of a 14-residue cap-strapped peptide has been optimized to give almost 100% helicity. Furthermore, in a model study, cap-strapped peptides derived from a p53 segment showed potent inhibition of MDM2/p53 interaction. The results demonstrate that cap-strapped peptide can be used as a general tool for highly potent and specific inhibition of PPIs.

PE – COMPUTATIONAL MODELING/SIMULATION Quantitative description of pH dependent chemical shifts in model peptides Efrosini Artikis1, Charles L. Brooks III1,2 1Biophysics Program, University of Michigan, 930 North University, 2Department of Chemistry, Univer- sity of Michigan, 930 North University NMR chemical shifts (CSs) are important in the characterization of protein structure and dynamics. How- ever, pH effects influencing these observables are not well described, as it is difficult to disentangle the contributions of conformation and of pH on experimental CSs. Alternatively, empirical computation of CSs may be achieved when a reference set of experimental CS is available for parameterization. Still, reference CSs may be incomplete or unavailable for proteins in extreme pH conditions, and other methods must be employed. In this study, we explore the influence of charge effects by utilizing DFT calculations for the ab initio computation of NMR CSs of model tri–peptides. Preliminary data suggests that combined simulation and quantum mechanical modeling are able to reproduce trends in chemical shift changes due to changes in pH of titratable residues. Furthermore, the modular nature of these peptides enables the investigation of protonation equilibria in terms of nearest neighbor effects. Scaling the CS trends seen in the model peptides to protein systems, will allow for more accurate interpretation of experimental CSs and subsequently enable more informed structure determination.

Coevolutionary analysis and structural prediction of the bacterial divisome Samson Condon1, Gladys Diaz-Vazquez1, Deena-al Mahbuba1, Alessandro Senes1 1University of Wisconsin-Madison One of the most fundamental processes in the bacterial life cycle is division. This process is mediated by the divisome, a multi-protein complex whose core subunits are conserved across bacterial phyla. Though many protein-protein interactions between divisomal subunits are known, there are no experimental structures of complexes showing how different proteins in the divisome bind to one another. However, it may be possible to computationally model these interfaces by taking advantage of the vast

34 ABSTRACT evolutionary sequence record for divisomal proteins. Using an algorithm to predict evolutionary couplings between pairs of coevolving residues across homologous subunits of the divisome, I have begun to model their structural interactions. For example, the intermediate divisomal proteins FtsB and FtsL exhibit a large number of strongly coevolving positions that map to a single interface. With this information, the FtsB-FtsL subcomplex was successfully modeled as a stable fourhelix bundle that is consistent with both previous knowledge of the subcomplex as well as subsequent experimental analyses. These findings indicate that it may be possible to model protein-protein interfaces within the divisome using the same methods, which will be useful for understanding the role of different components of the divisome and for designing inhibitors of bacterial cell division.

An Improved Search Algorithm for Protein-Ligand Docking Using FFTs Xinqiang Ding1 1Department of Bioinformatics, University of Michigan Protein-ligand docking is a computational methodology to predict how a ligand interacts with a protein. The method plays an important role in the early stages of the drug discovery processes, such as in early lead identification and optimization. Overall, protein-ligand docking usually consists of two key components: an energy function by which to delineate good potential binders from poor potential binders and a search algorithm, which allows one to explore the protein-ligand binding pose space and identify the poses with significant binding energy. One of the key challenges in protein-ligand docking is to develop a fast search algorithm. This is a challenge because of the rough binding energy landscape comprised of a high dimensional search space including translation, rotation and ligand conformation or in some cases protein conformation. To address this challenge many stochastic optimization algorithms have been used, such as simulated annealing and genetic algorithms, but there is still significant room for improvement in terms of the search algorithm. In the present work we describe an improved search algorithm using spatial FFTs methods for the commercially distributed, but in-house developed proteinligand docking method CDOCKER. The algorithm takes advantage of two facts. One is that the interaction energy between the protein and the ligand can be approximately calculated as a cross- correlation between two sets of grids. The other is that the cross-correlation can be calculated using an FFT-based algorithm that performs an exhaustive translational search of the putative proteinligand dis- tances given a conformation and orientation of each. These two facts enable us to do an exhaustive translational search in H(NlogN) time instead of H(N2). Combining this fast exhaustive translational search method with traditional simulated annealing methods, we describe our improved protein-ligand search algorithm.

Differences on the conformational substates visited by native and mutants versions of the LAO- BP obtained by Accelerated Molecular Dynamics. Diego S. Granados1, Jesus BandaVazquez1, Alejandro SosaPeinado1 1Department of Biochemistry, School of Medicine, UNAM, Mexico City, 04510. Mexico. Protein engineering in the realm of binding has been focused historically on mutating residues directly involved in molecular recognition. However, redesigning of binding sites has proven to be even more complex that the binding site itself. Despite efforts, the rate of success of this approach is very limited and we are still lacking for optimal design principles. Additionally, many examples show that is possible to change binding patterns by mutations outside the binding site. This fact suggest we are leaving out important details while designing, since the native structure is a dynamical ensemble. In our laboratory, we use a periplasmic binding protein called LAO which binds positive amino acids (LLysine, Larginine and Lornithine) with nanomolar affinity as a model for understanding the interplay between dynamics and recognition. In previous works, we have designed punctual mutants for understanding the role of specific residues on its function. We took two of those: P16A and L117K. The first one shows a

35 ABSTRACT reduction on the rate of binding for arginine and the later shows affinity for glutamine, a ligand which the native version ignores. Both mutations are not in residues of the binding site turning the rationale of this functional differences an intriguing question. For gaining insight in the details of dynamical ensembles which were modified on these mutants we used accelerated molecular dynamics (aMD). aMD is an enhancedsampling method implemented on amber14. This methodology allow us to compare differences on substates visited by native and mutants forms of LAOBP.

A new definition of inter-residue interaction provides insight into sequence-structure relationships Jack Holland1 1Dartmouth College The native structure of a protein depends on the interactions between its residues, but quantifying these interactions and how they affect structural stability is an open problem. Simple distance and orientation based descriptions often fail to capture important details of inter-residue interactions. A more structurally informative measure of interaction may thus enable a more nuanced description of sequence-structure relationships. Here we propose such a measure, termed contact degree, and go on to show a contact degree-based statistical potential very accurately discriminates between native and decoy protein structures. Contact degree is roughly the fraction of all amino acid rotamer pairs that clash at a given pair of positions. This provides measure a of pseudo-distance that incorporates more information about the structural circumstances around a potentially interacting pair of residues than a straightforward measure of distance or orientation, enabling a more informative and discriminating interaction potential. In fact, this potential exhibits high accuracy in decoy discrimination, often outper- forming state-of-the-art detailed atomistic methods, such as DFIRE and FOLDX, on standard decoy sets. Unlike with these atomistic methods, however, side-chain coordinates are not used to calculate contact degrees, so that given backbone coordinates alone, a contact degree-based pseudo-energy can be computed rapidly for any sequence. Thus, rather than interpreting the input structure explicitly as a precise all-atom conformation, the potential implicitly integrates over side-chain degrees of freedom, giving it some tolerance to local structural errors and enabling a more robust identification of near-native conformations. Given these properties, and the results of our benchmarks, we believe the contact degree and its associated interaction potential could be of great utility for both protein design and structure prediction applications.

Generating in silico Mutations to Infer the Effect of Multiple Amino Acid Substitutions on Protein Stability Rebecca Hsieh1 1Western Washington University Proteins are comprised of amino acids that are joined end-to-end to form a polypeptide chain. A protein’s 3D shape facilitates its biological function which may be altered due to one or more amino acid substitutions. Performing wet lab experiments by engineering a protein with specific mutations and directly assessing the effect of those amino acid substitutions is time intensive and often prohibitive. We are developing an in silico technique to generate multiple amino acid substitutions in protein structure files. No such tool currently exists. We study the mutants using rigidity analysis, which is an efficient method for determining the flexibility of a molecule. Our in silico approach systematically generates multiple mutations in a protein. We process a Protein Data Bank (PDB) file that contains the x, y, and z coordinates of the atoms in a protein. Our rMutant algorithm edits the ATOM entries of a PDB file to simulate a mutation from a larger to smaller amino acid. The rigidity analysis is quick, and identifies rigid clusters of atoms in a protein. We compare the rigid cluster sizes of the wild type and in silico mutant, to infer the effect of the amino acid

36 ABSTRACT substitutions. We find that our stability predictions based on rigidity analysis correlate well with experimentally derived DDG data for mutations performed in physical proteins.

Unique allosteric mechanism regulating protein-protein interaction through phosphorylation: a case study of the conformational changes in the Syk tandem SH2 protein Duy P. Hua1, Carol Beth Post1 1Department of Medicinal Chemistry and Molecular Pharmacology, Purdue University Spleen tYrosine Kinase (Syk) is one of the crucial signaling proteins involved in the development of immune cells and the initiation of inflammatory responses and an attractive target for therapeutic treatments of chronic autoimmune and allergic diseases such as asthma, rheumatoid arthritis and lupus. Syk, a 72-kDa protein, comprises a tandem SH2 (tSH2) – two SH2 domains connected by linker A - and a kinase domain. The activity of Syk is regulated by the association of its tandem SH2 to a short, highly conserved, doubly tyrosine-phosphorylated motif (dpITAM) on the immunoreceptor. tSH2-dpITAM association initiates signaling, and some time thereafter, Tyr130 on linker A of tSH2 is phosphorylated, which leads to a lower binding affinity between the tSH2 and dpITAM and the dissociation of the tightly bound Sykimmunoreceptor complex. The molecular mechanism for the allosteric regulation of protein- protein interaction via phosphorylation of Tyr130, a site far from the dpITAM binding pockets, is unknown. Results obtained with long MD simulations of the unphosphorylated and Tyr130- phosphorylated tandem SH2 indicate that the domain-domain structure of the two species differ (see Fig. 1), and provide a molecular interpretation for previous biophysical measurements. In addition, the simulations reveal that phosphorylation at Tyr130 alters electrostatic interactions between charged residues at the interface of the two SH2 domains as well as between solvent ions and an arginine residue, and thus present mechanistic insight into the allosteric mechanism regulating protein-protein interaction. Distance-based correlations and correlated network analysis will elucidate how Tyr130 phosphorylation leads to changes in interfacial ionic interactions and subsequently, domain-domain structure of the tandem SH2 protein. Together with experimental data, results from MD simulations provide a detailed molecular picture of the effect of Tyr130 phosphorylation on the structure and dynamics of the tandem SH2 that enables a better understanding of an unusual regulation of tSH2-dpITAM association.

The slow to fast transition in cytochrome c oxidase catalysis is facilitated by a loop flip in subunit ii of the enzyme Trevor Alleyne1, Diane N. Ignacio1, Damian Ashe1, Valerie B. Sampson2 1Faculty of Medical Sciences, The University of the West Indies, St. Augustine Campus, Trinidad and Tobago, 2Nemours Center for Cancer and Blood Disorders, Alfred I. duPont Hospital for Children, Wil- mington, DE. USA The enzyme cytochrome c oxidase (COX) also referred to as complex IV of the electron transport chain, catalyses the reduction of molecular oxygen to water in a process that is coupled to the synthesis of ATP. The ATP synthesized provides energy for cells to carry out their various functions. Several studies have led to the conclusion that COX exists in at least two conformations; a slow acting form referred to as ‘Resting’ and a much more active form referred to as ‘Pulsed’. Little is known presently about the structural features that distinguishes the Pulsed conformation from the Resting. Here we used computer modelling to attempt to address this issue. Structures of COX and its substrate, cytochrome c, were obtained from the Protein Data Bank. The Insight II program was then employed to simulate a conformational change in the region of the enzyme’s substrate (cytochrome c) binding site. The conformational change involved flipping an 8 residue loop of subunit-II through 908. Finally a stable COX-cytochrome c enzyme-substrate complex was simulated and the complex evaluated for probable electron transfer pathways within COX.

37 ABSTRACT

Similar to our earlier energy minimized ES-complex which had 7 salt bridges and 24 hydrogen bonds, the new complex formed subsequent to the conformational change was stabilized by 7 salt bridges and 22 hydrogen bonds. However, compared to complexes formed in the absence of the conformational change, the distance between the redox centres of the two proteins was reduced by half. Furthermore, only four COX residues (Trp-104, Try-105, His-102 and Glu-198) rather than the original nine were found along the axis linking the redox metal centres of the two proteins. We accept that the binding and reduction of oxygen at the heme a3 center is probably the trigger for the Resting to Pulsed trans- formation, but propose that the initial activity at heme a3 is communicated via the latter’s long hydrophobic side chain to the substrate binding site. We further propose that intramolecular electron transfer in fast acting COX occurs via the resultant shortened charge/hydrogen relay system that is created near to the substrate binding site.

Atomistic simulations of unfolding and translocation of the Immunoglobulin domain I27 in repetitive cycles of the ClpY Biological Nanomachines Abdolreza Javidialesaadi1 and George Stan1 1Department of Chemistry, University of Cincinnati Cellular protein quality control plays a critical role in maintaining cell viability by recognizing and degrading misfolded proteins to prevent toxic protein aggregation, which is a common feature of many neurodegenerative diseases such as Alzheimer’s and Parkinson’s diseases. Powerful AAA1 ATPases, such as ClpY, are hexameric biological nanomachines which selectively process abnormal proteins targeted for degradation by unfolding and threading them through a narrow central channel. The molecular details of unfolding and translocation of substrate proteins by these nanomachines have not yet been fully understood. We performed targeted molecular dynamics simulations coupled with repetitive pulling forces of an implicit solvent model of ClpY-mediated unfolding and translocation of an Immu- noglobulin protein domain, Titin I27. Cyclical opening and closing of the ClpY ring are modeled through sequential conformational changes of ClpY subunits. To elucidate the dependence of I27 unfolding and translocation pathways on the direction of force applied by ClpY, we performed simulations in which the I27 N-terminus is either restrained, which reproduces the setup in single molecule experiments, or unrestrained to mimic in vivo geometries. Results of these simulations were compared with bulk mechanical unfolding of I27 by pulling along the N-C direction. We find unrestrained I27 is reoriented by ClpY during the threading, which allows I27 unfolding to proceed by unzipping the C-terminal strand. This unfolding mechanism requires smaller force than the shearing mechanism utilized in the one-directional force application onto the restrained I27 by ClpY or bulk unfolding. However, restrained I27 was found to experience less reorientation during unfolding and translocation.

The degradation and inhibition mechanism of Alzheimer’s Ab fiber by dihydrochalcone molecules Yibo Jin1, Yunxiang Sun1, Guanghong Wei* 1State Key Laboratory of Surface Physics, Key Laboratory for Computational Physical Sciences (Minis- try of Education), Department of Physics, Fudan University Alzheimer’s disease (AD) is associated with the aggregation of amyloid-b (Ab) peptides into toxic prefi- brillar aggregates[1,2]. To destabilize and dissolve Ab fibrils, a number of candidate molecules have been proposed. Recent experiments reported that small molecule dihydrochalcone could effectively reduce the cytotoxicity induced by Ab[3]. Dihydrochalcone is a compound extracted from daemonorops draco tree[3], so it’s a potential candidate drug for treating AD. However, the atomic-level details of dihydrochalcone2Ab-oligomer interaction are largely unknown. In this work, We found that there were three possible stable binding sites including two sites in hydrophobic grooves on surface of Ab protofibril that made no significant changes in Ab structures and one site in the interior that caused destabilization of the protofibril. In this site, dihydrochalcone was disrupted the D232K28 salt bridges, and

38 ABSTRACT partially opened the tightly compacted interlink states Ab chains. We also found how dihydrochalcone molecules get into amyloid beta fiber and how it destroy the Ab fiber. The molecular mechanism of this novel drug candidate dihydrochalcone to disaggregate Ab protofibril may provide some insight into the strategy of structure-based drug design for AD. References: 1. Soto C, Sigurdsson E M, Morelli L, et al. b-sheet breaker peptides inhibit fibrillogenesis in a rat brain model of amyloidosis: implications for Alzheimer’s therapy[J]. Nature medicine, 1998, 4(7): 822- 826. 2. Laird F M, Cai H, Savonenko A V, et al. BACE1, a major determinant of selective vulnerability of the brain to amyloid-b amyloidogenesis, is essential for cognitive, emotional, and synaptic functions[J]. The Journal of Neuroscience, 2005, 25(50): 11693-11709. 3. Viet M H, Chen C Y, Hu C K, et al. Discovery of dihydrochalcone as potential lead for Alzheimer’s disease: in silico and in vitro study[J]. PloS one, 2013, 8(11): e79151.

A Computational Study on Aquaporin-embedded Water Purification System Youngjin Kim1, Sangjae Seo1, Eugene Ham2, Jong Sung Kim2, Moon Ki Kim1 1School of Mechanical Engineering, Sungkyunkwan University, 2Stevenson High School, 1 Stevenson Drive Aquaporins are transmembrane proteins that form pores in the cell membrane and selectively transport water molecules across the cell membrane. By utilizing this unique feature of aquaporin, many attempts have been carried out to develop aquaporin-embedded or aquaporin-mimicking water purification system. In this study, several computational approaches including normal mode analysis (NMA) and molecular dynamics (MD) simulation were applied to investigate not only dynamics characteristics of aquaporin itself, but also interaction between aquaporin and various lipid layers. First, an elastic network model was applied to construct a simulation model of aquaporin structure and then NMA was performed. The calculated normal modes successfully demonstrate intrinsic vibration features of aquaporin, which are strongly related with water transportation. Second, MD simulation of aquaporin-embedded membranes reveals that different lipid layers slightly alter topological arrangement of aquaporin tetramer resulting in change of water permeability. The MD simulation results have a good agreement with experimental results already reported elsewhere. Consequently, this basic understanding of water transportation mechanism at atomic level, as well as lipid-dependent permeability change at system level would be a cornerstone for us to make a breakthrough in optimal design of aquaporin based water purification system.

Mechanism of Improved Doripenem Binding in Several Clinical Mutations in OXA-66 b-Lactamase Zachary Klamer1, Emma Schroder2, Troy Wymore3, David A. Leonard2, Agnieszka Szarecka1 1Department of Cell and Molecular Biology, Grand Valley State University, Allendale MI, 2Department of Chemistry, Grand Valley State University, Allendale MI, 3Department of Chemistry, University of Michigan, Ann Arbor, MI Antibiotic resistance is a serious and global public health concern. One of the major mechanisms of resistance involves beta-lactamases. In particular, the class D beta-lactamases have been evolving rapidly and gaining extended spectrum and/or carbapenemase profiles. OXA-66 beta-lactamase is a parent enzyme to one of the subfamilies of class D carbapenemases – proteins that provide resistance to the newest line of beta-lactam antibiotics. Several of its clinically observed single-amino acid variants show significantly increased activity against carbapenems but the mechanism of this gain of function is not clear. In this study, we investigate the structure and dynamics of OXA-66 wild type and the same protein with P130Q, I129L, and L167V substitutions to explain the mutants’ more favorable interactions

39 ABSTRACT with doripenem. We have analyzed 250-nsec Molecular Dynamics simulations of each fully hydrated protein model. Our data indicate that the rotational state of I129 plays an important role in relieving a steric clash between this residue and doripenem. In all the mutants, the distance between residue 129 and doripenem increases, but the P130!Q mutation has a particularly strong effect on the conformational freedom of the I129 side chain. In addition, based on the OXA66 structure, doripenem binding may also be hindered by W222. Our results indicate that the dynamics of the W222 residue changes in the I!L and P!Q mutations - in a way that is consistent with easier doripenem binding. Both effects explain increased hydrolytic activity of the variants and provide insight into the plasticity of the OXA-66 binding pocket.

The Role of Electrostatic Dominated Cation-? Interactions for CREBBP Bromodomain Inhibition Kiran Kumar1, Wilian A. Cortopassi1, Prof. Robert S. Paton1 1Chemistry Research Laboratory, University of Oxford CREBBP bromodomain containing proteins are epigenetic readers that have received increasing atten- tion in recent years as promising cancer drug targets. To aid in inhibitor design, we utilized computational chemistry techniques such as classical molecular dynamics (MD) and quantum mechanics (QM) to analyse key interactions that drive selectivity for CREBBP by a series of 15 5-isoxazolyl-benzimidazole inhibitors (IBIs). In this study we explore the formation of a key cation-p interaction between 15 known IBIs and a conserved arginine residue unique to CREBBP. 100 ns MD simulations with explicit solvation were performed with an initial inhibitor template bound to a CREBBP protein. Cation-p interactions were critical for binding and present for more than 70% of simulation time, although not initially observed in crystal structures. Given the importance of this interaction for the stability of the inhibitor, we performed additional binding free energy calculations on

40 ABSTRACT

15 IBIs using molecular mechanics/Poisson-Boltzmann (PB) and generalized Born (GB) surface area (MM-PBSA, MM-GBSA) scoring functions. We also performed QM-complexation energies for an accurate prediction of cation-p interactions. A third technique analysing the Electrostatic Potential Surface Area (ESPs) of the substituted benzenes demonstrated that consideration of only the surface above the p system is enough for a quantitative (R2 5 0.84, n515) and qualitative prediction of these interactions. Results from our three methodologies correlated well with experimental binding affinities for prediction of cation-p interactions with CREBBP inhibitors. However, ESP is significantly faster without compromising accuracy. Applications of these MM, QM, and ESP calculations can be easily extended to other small molecule protein complexes in which cation-p interactions are necessary for driving selectivity.

A Power Flex in Hsp70: The nucleotides impact on the actin-like ATPase domain Daniela Bauer1, Dale R. Merz2, Benjamin Pelz1, Kelly E. Theisen2, Gail Yacyshyn2, Dejana Mokranjac3, Ruxandra I. Dima2, Matthias Rief1,4, Gabriel Zoldak1 1Physik Department, Technische Universitat Munchen, 2Department of Chemistry, University of Cin- cinnati, 3Department of Physiological Chemistry, Medical Faculty, University of Munich, 4Munich Cen- ter for Integrated Protein Science The regulation of protein function through ligand-induced conformational changes is crucial for many signal transduction processes. Heat shock protein (70 kDa) is one example of a protein undergoing conformational changes upon ligand hydrolysis. It consists of a nucleotide binding domain (NBD) covalently linked to a substrate binding domain (SBD) via an interdomain linker [1]. We studied the nucleotide binding domain (NBD) of Hsp70 using a combination of coarse-grained pulling molecular dynamics

41 ABSTRACT simulations [2] along with corresponding single molecule experiments to uncover the changes in mechanical stability between the two lobes of the NBD during binding, hydrolysis, and exchange of ATP [3]. We found that nucleotide binding and hydrolysis played a significant part in regulating the mechanical balance between the two lobes of the NBD. ADP binds strongly to lobe II increasing its stability relative to lobe I, which is the more stable lobe in the absence of a nucleotide. Extraction of the C- Terminal helices occurs first independent of the nucleotide status. The helix extraction and shift in the mechanical hierarchy between the lobes of the NBD offer insight into its signal transduction mechanism with the substrate binding domain and can be used for a general interpretation of signal transduction within the actin/sugar kinase superfamily. References: [1] Swain, J. F. J., Dinler, G., Sivendran, R., Montgomery, D. L., Stotz, M., & Gierasch, L. M. (2007). Hsp70 chaperone ligands control domain association via an allosteric mechanism mediated by the inter- domain linker. Molecular Cell, 26, 27–39. [2] Zhmurov, A., Dima, R. I., Kholodov, Y., & Barsegov, V. (2010). Sop-GPU: accelerating biomolecular simulations in the centisecond timescale using graphics processors. Proteins, 78, 2984–99. [3] Bauer, D., Merz, D. R., Pelz, B., Theisen, K. E., Yacyshyn, G., Mokranjac, D., ... Zold ak, G. (2015). Nucleotides regulate the mechanical hierarchy between subdomains of the nucleotide binding domain of the Hsp70 chaperone DnaK. PNAS, 112(33), 10389–10394.

An Investigation of the Thermostability of DNA Polymerase using Molecular Dynamics Simulations Erica Modeste1, Lily Mawby1, Bill Miller III1, Eugene Wu1, and Carol Parish1 1Department of Chemistry, University of Richmond Since the invention of the polymerase chain reaction (PCR), DNA polymerase has become a critical tool in biotechnology. DNA polymerase I from Thermus aquaticus (Taq DNA polymerase) is an ideal prototype for PCR due to its thermostability and activity at high temperatures. Despite the ability of Taq DNA polymerase to function at high temperatures, there are disadvantages. For example, Taq DNA polymerase often begins DNA replication at room temperature, resulting in mispriming and nonspecific products. To improve functionality as a PCR agent, Kermechiev and colleagues discovered that a single point mutation from isoleucine to leucine at residue 707 reduced Taq DNA polymerase’s activity at low temperatures. Unrestrained Molecular Dynamics (MD), Steered Molecular Dynamics (SMD), and Umbrella Molecular Dynamics (UMD) were performed on the mutated and wild-type Taq DNA polymerase at both 341K and 298K to determine how the mutation affects activity. The Weighted Histogram Analysis Method (WHAM) determined the potential of mean force (PMF) for the displacement of the DNA strand within the polymerase. The results suggest that the mutated Taq DNA polymerase remained relatively immobile at room temperature and became more flexible at the higher temperature while the wild- type demonstrated no noticeable differences in dynamics at the low and high temperatures. SMD, UMD, and WHAM suggest that a high energy barrier separates important conformations of the protein.

Carbohydrate and protein effects on antibody–receptor binding Morgan L. Nance1,2, Jason W. Labonte2, Jeffrey J. Gray2 1College of Biological Sciences, University of California, Davis, Davis, California, 2Department of Chemical & Biomolecular Engineering, Johns Hopkins University, Baltimore, Maryland It is well understood that differential glycosylation can be used to modulate an antibody’s therapeutic potential. Unfortunately, since an antibody’s carbohydrate composition is dependent on which enzymes are expressed in the eukaryotic cell line used for production, the testing of different glycoforms is challenging, time-consuming, and expensive. An alternative method of design is thus required to create new and improved therapeutic antibodies more efficiently. We have developed a framework within the

42 ABSTRACT

Rosetta structure-prediction and design software suite for modeling saccharides and glycoconjugates, including glycoproteins. Updates to the Rosetta code have allowed access to all of the relevant torsion angles, (u, w, x, m, and v), enabling the ability to capture the energetic contribution of the conformational variability that is intrinsic to carbohydrates, such as flexibility, stereochemistry, and branching. The addition of a library of low-energy ring conformers and a glycosidic bond scoring method has facilitated the incorporation of carbohydrates into Rosetta’s core functions and protocols. We will present the beginning framework of an algorithm that will rationally predict amino acid and carbohydrate mutations to improve antibody–receptor binding and therapeutic potential. We will demonstrate Rosetta’s modeling capabilities by presenting images of a representative antibody and antibody–receptor structure. Lastly, we will highlight Rosetta’s ability to recapitulate experimental binding affinity data, as well as our plans to further test and improve this functionality. Our work will advance the efforts of antibody engineering for increased therapeutic potential, as well as expand the computational tools available to scientists working with antibodies and other glycoproteins. Financial support: EAGER award, NSF Division of Biological Infrastructure (1541278)

On the ability of molecular dynamics force fields to recapitulate NMR derived protein side chain NMR order parameters Evan S. O’Brien1, A. Joshua Wand1, Kim A. Sharp1 1Department of Biochemistry & Biophysics, Perelman School of Medicine, University of Pennsylvania Molecular dynamics (MD) simulations have become a central tool for investigating various biophysical questions with atomistic detail. While many different proxies are used to qualify molecular dynamics force fields, most are based on structural parameters such as the root mean square deviation from experimental coordinates or NMR chemical shifts and residual dipolar couplings. NMR derived Lipari- Szabo generalized order parameter values (O2) values of amide N-H bond vectors of the polypeptide chain are also often employed for refinement and validation. However, with a few exceptions, side chain methyl axis order parameters have not been incorporated into experimental reference sets. Using a test set of five diverse proteins, we examine the performance of several force fields implemented in the NAMDD simulation package. We find that simulations employing explicit water implemented using the TIP3 model generally perform significantly better than those using implicit water in reproducing experimental methyl symmetry axis O2 values. Overall the CHARMM27 force field performs nominally better

43 ABSTRACT than two implementations of the Amber force field. It appears that recent quantum mechanics modifications to side chain torsional angles of leucine and isoleucine in the context of the Amber ff12 force field significantly hinder proper motional modeling for these residues. There remains significant room for improvement as even the best correlations of experimental and simulated methyl group Lipari- Szabo generalized order parameters fall below an R2 of 0.8.

Electrostatic interactions and the multi-layered local structure of active sites: Key features in natural and designed enzymes Timothy A. Coulther1, Lisa Ngu1, Penny J. Beuning1, Mary Jo Ondrechen1 1Department of Chemistry & Chemical Biology, Northeastern University Catalytic sites in natural enzymes are characterized by networks of strongly coupled protonation states. These networks impart the necessary electrostatic and proton-transfer properties to the catalytic residues in the first layer around the reacting substrate molecule(s). Typically these networks include not only first-layer residues but also residues in the second and third layers. Our machine learning methodology, Partial Order Optimum Likelihood (POOL), predicts that distal residues, typically in the second and third layers around the reacting substrate molecule(s), participate in catalysis in most enzymes. POOL-predicted, multi-layer active sites, with significant participation by distal residues, have been verified experimentally by single-point site-directed mutagenesis and kinetics assays for nitrile hydratase, phosphoglucose isomerase, ornithine transcarbamoylase, and the DNA polymerase DinB. POOL predicts the amino acids in a protein 3D structure that contribute to catalysis using input features from Theoret- ical Microscopic Anomalous Titration Curve Shapes (THEMATICS). The THEMATICS metrics, based on computed electrostatic and chemical properties, are effectively measures of the strength of coupling between protonation events. In natural enzymes, THEMATICS metrics are high for all of the catalytic residues; mutations to coupled residues in the second and third layers typically result in observed reduction in catalytic efficiency. In designed enzymes that have not been fully evolved, these networks have not been optimized and it is possible to identify mutations in distal positions that increase THEMATICS metrics for all of the ionizable first-layer residues, to obtain a stronger network of coupled protonation states. The importance of coupled protonation states and of distal residues in enzyme catalysis are features that should be considered in enzyme design. Supported by NSF MCB-1517290 and a National Institute of Justice Predoctoral Fellowship awarded to TAC.

Analysis of Multi-domain Protein Dynamics Amitava Roy1 1National Institutes of Health The biological function of multi-domain proteins, or the regulation of their activity, depends on the variation in orientation and separation of the domains. Careful characterization of motion and correlated fluctuations within and between the domains is relevant for understanding the functional behavior of multi-domain proteins. Molecular dynamics (MD) simulations can provide atomic details of these motions. Nevertheless, the common procedure for analyzing fluctuations from MD simulations after rigid-body alignment greatly overestimates correlated positional fluctuations in the presence of relative domain motion. We show here that expressing the atomic motions of a multi-domain protein as a combination of displacement within the domain reference frame and motion of the relative domains correctly separates the internal motions to allow a useful description of correlated fluctuations. The approach makes it possible to calculate the proper correlations in fluctuations internal to a domain as well as between domains.

44 ABSTRACT

Computational prediction and functional annotation of enzymes in the Haloacid Dehalogenase Superfamily for Bioremediation Lydia A. Ruffner1, Mong Mary Touch1, Penny J. Beuning1, Mary Jo Ondrechen1 1Department of Chemistry & Chemical Biology, Northeastern University Halogenated organic compounds are serious environmental pollutants that are difficult to eliminate from the soil and groundwater. Some enzymes in the Haloacid Dehalogenase (HAD) superfamily possess the ability to detoxify and degrade halogenated compounds. Unfortunately, the majority of the members of this superfamily, along with over 13,000 Structural Genomics protein structures in the Pro- tein Data Bank (PDB), have unknown biochemical function or an incorrect putative function. In order to fully utilize the functions of these proteins, reliable methods must be developed to functionally anno- tate these proteins in order to identify potential applications, such as bioremediation. Computational methods, including the machine learning method Partial Order Optimum Likelihood (POOL) and the local-structure-based function predictor Structurally Aligned Local Sites of Activity (SALSA), developed at Northeastern, will be used to predict the biochemical function of proteins in the HAD superfamily that lack validated function by using the functionally important residues designated by POOL. These predictions will then be experimentally validated by direct biochemical assay to establish the function of each protein and to verify our computational approach to function prediction. This project is funded by NSF CHE-1305655.

Computational Studies of Green Fluorescent Protein Unfolding and Translocation by the ClpY ATPase during Protein Degradation Yu-Hsuan Shih1, George Stan1 1Department of Chemistry, University of Cincinnati Clp ATPases are heat-shock proteins that belong to the AAA1 (ATPases Associated with diverse cellular Activity) superfamily. They play critical roles in the protein quality control system by dispatching harmful proteins via the degradation process. The regulatory mechanism of Clp ATPases is controlled by ATP- driven large conformational changes and dynamic interactions to the substrate proteins (SPs). Without protease assistance, protein misfolding can result in deleterious pathways, which can lead to neurodegenerative diseases, such as Alzheimer’s, Parkinson’s diseases. The purpose of this study is to understand the unfolding and translocation mechanisms of substrate proteins (SPs) by the Clp ATPase. We

45 ABSTRACT apply coarse-grained molecular dynamics simulations using the CHARMM program to pinpoint the mechanical response of the green fluorescent protein (GFP) to the action of the ClpY ATPase. Results of our computational simulations are compared with single-molecule experiments. We observed that unraveling GFP from N-terminus is about 50% more efficient than from C-terminus. The intermediate without central a helix decomposes unfolds to 2-state mechanism. Interestingly, some minor intermediates can unravel either from C- or N-terminus. This finding indicates that mechanical resistance of partially unfolded intermediates is correlated with the topology of local domains, rather than the global SP fold. Simulations of 6-fold and 2-fold conformational transitions of ClpY subunits yield similar unfolding pathways. Restraining the distal SP terminus results in specific direction dependent unfolding pathways. This project serves to help bridge the gap between experiments and simulations providing vivid microscopic details on the degradation capability of Clp proteases.

GeoFold and InteractiveROSETTA: Providing improved tools for computational protein design Benjamin Walcott1,2, Christian Schenkelberg1,2, Oluwadamilola Lawal1,2,3, Wenyin San2,4, William Hooper1,2, Quadis Evans1,2, Alexander McDonald1,2,4, Donna E. Crone1,2, Christopher Bystroff1,2,4 1Department of Biology, Rensselaer Polytechnic Institute, Troy, NY, 2Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, NY, 3Department of Biology, Medgar Evers College, Brooklyn, NY, 4Department of Computer Science, Rensselaer Polytechnic Institute, Troy, NY Knowledge of the pathways along which proteins fold and unfold provides invaluable insight into protein stability and design. However, existing methods for folding pathway elucidation, such as U-value analysis, hydrogen-deuterium exchange (HX), and molecular dynamics (MD) are too resource intensive for practical application in computational protein design. GeoFold is a fast computational method for predicting protein folding/unfolding pathways and rates. GeoFold unfolds proteins through recursive hierarchical breaking of a protein’s native tertiary contacts via four types of moves: separation of chains (break), single point revolution (pivots), rotation around two points (hinge), and the opening of a b- barrel between two strands (seam). Forward and reverse rates for each move are calculated based on

46 ABSTRACT changes in solvation free energy and configurational entropy. A finite differences calculation on the directed acyclic graph (DAG) of all moves then simulates the kinetics, giving an unfolding pathway represented as a time-ordered contact map. Several modifications to GeoFold, including incorporation of a new solvation model and changes in the criteria for the pivot move, have improved the correlation between calculated and experimental unfolding rates for a database of 86 proteins spanning various protein families and kinetic types. To visualize GeoFold output, we use InteractiveROSETTA, a graphical interface for the PyRosetta suite of protein engineering software. Using InteractiveROSETTA, key steps identified via U-value analysis for two proteins in the database, barnase and Im7, can be readily identified from GeoFold simulations. InteractiveROSETTA is being further developed to include utilities such as constrained energy minimization, Ramachandran plots, DNA docking, HMMSTR local structure prediction, and improved remote server support. Together, GeoFold and InteractiveROSETTA provide powerful tools for studying thermodynamic properties and structure-function relationship within proteins. Insights gained from these studies will aid greatly in engineering proteins with novel properties.

Modeling Protein Folding/aggregation on Nanoparticle—based Biosensors in Complex Solvent Environments by a Coarse—grained Simulation System Shuai Wei1, Charles L. Brooks III1 1Department of Chemistry and Biophysics, University of Michigan Understand protein behavior in a complex environment is the key in rationalizing the design of many novel techniques such as protein biosensors based on novel nanoparticle (NP) materials. Protein folding and aggregation properties can be largely affected by the environments such as solid abiotic interfaces and complex co-solvents, which would directly affect the expected functionality or performance of the biosensors. Difficulties remain in applying current experimental methods to obtain a detailed picture of protein structures on a solid surface, so our ability is limited by these factors in the design of biosensor surface chemistry a priori for desired protein behavior. To address this point, in this work we present a coarse-grained simulation system that is able to model protein folding, misfolding, and aggregation on various NP surfaces and in the presence of co-solvents based on the Karanicolas and Brooks (KB) Go-like protein model. A flat-surface potential was first included and carefully parameterized based on a large experimental data set. This surface model has been shown in many studies to be able to reproduce and predict protein behavior at different types of surfaces as verified by experimental tests. Further effort in developing the surface potential was applied in describing surface curvature effects on protein behavior as a means of understanding nano-sized biosensor materials. Tanford’s transfer free energy model was recently added to this simulation system as a descriptor of co-solvent effects. The energy of each part of the underlying folding/surface/solvation potential has been well balanced so that each element works collaboratively as a single complex environment for the protein. Together, this coarse-grained modeling system is expected to accurately predict the protein adsorption, folding/unfolding, aggregation, and protein-protein interactions in a complex environment as it would experience in differing cellular or solvent environments as well as in the context of a NP-based biosensor.

Studying protein conformational transitions using adaptive biased sampling Heng Wu1, Carol B. Post1 1Purdue University Conformational transitions are fundamental to function of many proteins such as signaling proteins, membrane receptors and molecular machines. Studying the molecular mechanisms of protein conformational transitions is critical for understanding the biological function of these proteins. Computational methods have been valuable in elucidating such transitions; however, specific computational methods

47 ABSTRACT are usually required to overcome the free energy barrier between different protein conformational states. Previously our group developed Adaptive Biased Path Optimization (ABPO), which is a transition path sampling and optimization method that works in collective variable space. ABPO was applied to Src kinase activation using a Go-model. How ABPO works in an all-atom protein transition is not yet explored. Here we report application of ABPO on three proteins that undergo relatively simple transitions. Triose phosphate isomerase (TIM) has a flexible loop that moves when the protein is bound to a ligand. Dihydrfolate reductase (DHFR) has a loop that can adopt open and occluded forms in different states. Estrogen receptor helix-12 has different conformations when the protein is bound to an agonist or an antagonist. For each protein, distance-based or torsion- angle collective variables were identified from equilibrium trajectories. Lastly, ABPO was launched to obtain the transition path between the two states. Our results suggest that torsion angles can be good collective variables for transitions like loop movement, while distance-based CVs are required for rigid body movement. The all- atom transition paths for the three systems identified from the simulations will be described.

Self-Assembling Nano-Architectures Created from a Protein Nano-Building Block Using an Inter- molecularly Folded Dimeric de Novo Protein Naoya Kobayashi1, Keiichi Yanase1, Takaaki Sato1, Satoru Unzai2, Michael H. Hecht3, and Ryoichi Arai1 1Fac. of Text. Sci. & Tech., Shinshu Univ., Ueda, Nagano 386-8567, Japan, 2Fac. of Biosci. & Appl. Chem., Hosei Univ., Koganei, Tokyo 184-8584, Japan, 3Dept. of Chemistry, Princeton University The design of novel proteins that self-assemble into supramolecular complexes is an important step in the development of synthetic biology and nanotechnology. Recently, we described the three- dimensional structure of WA20, a de novo protein that forms an intermolecularly folded dimeric 4- helix bundle (Arai, R. et al., J. Phys. Chem. B 2012, 116, 6789). To harness the unusual intertwined structure of WA20 for the self-assembly of supramolecular nanostructures, we created a protein nano-

48 ABSTRACT building block (PN-Block), called WA20-foldon, by fusing the dimeric structure of WA20 to the trimeric foldon domain of fibritin from bacteriophage T4. The WA20-foldon fusion protein was expressed in the soluble fraction in Escherichia coli, purified, and shown to form several homooligomeric forms. The stable oligomeric forms were further purified and characterized by a range of biophysical techniques. Size exclusion chromatography, multi-angle light scattering, analytical ultracentrifugation, and small-angle X-ray scattering (SAXS) analyses indicate that the small form (S form), middle form (M form), and large form (L form) of the WA20-foldon oligomers exist as hexamer (6-mer), dodecamer (12-mer), and octadecamer (18-mer), respectively. These findings suggest that the oligomers in multiples of 6-mer are stably formed by fusing the interdigitated dimer of WA20 with the trimer of foldon domain. Pair-distance distribution functions obtained from the Fourier inversion of the SAXS data suggest that the S and M forms have barrel- and tetrahedron-like shapes, respectively. These results demonstrate that the de novo WA20-foldon is an effective building block for the creation of self-assembling artificial nano-architectures.(Kobayashi,N.etal.,J.Am.Chem.Soc.2015, 137, 11285)

PF – DESIGN/ENGINEERING A General, Symmetry-Based Methodology in Protein Cage Assembly Aaron Sciore1, Somayesadat Badieyan1, Min Su1, Philipp Koldewey1, Kelsey A. Diffley1, Brian M. Linhares1, James C. A. Bardwell1, Georgios Skiniotis1 and E. Neil G. Marsh1 1Department of Chemistry, University of Michigan Using natural proteins as building blocks for the construction of symmetry-based self-assembled architectures has introduced exciting new possibilities for a broad range of bio-nanotechnology applications. Currently, the dominant paradigm for synthesizing these self-assembled protein superstructures involve computational alignment of protein subunits followed by design of rigid protein-protein interfaces. However, this approach suffers from a lack of generality and requires significant experimental characterization to identify robustly assembled systems. Here we demonstrate a general method for generating symmetrical protein assemblies by genetically fusing two symmetrical proteins together via a flexible linker of variable length. By attaching either a trimeric or a tetrameric coiled- coil motif to the C-terminus of a trimeric esterase building block, we could assemble, purify, and characterize a tetrahedral or an octahedral assembly, respectively. These assembled superstructures retained esterase activity, representing a potential step forward in the targeted design of functional protein assemblies.

49 ABSTRACT

Engineering a Surface Immobilized Enzyme to Obtain High Levels of Water-free Activity Somayesadat Badieyan1, Qiuming Wang1, Xingquan Zou1, Yaoxin Li1, Maggie Herron2, Nicholas L. Abbott2, Zhan Chen1, E. Neil G. Marsh1 1Department of Chemistry, University of Michigan, 2Department of Chemical and Biological Engineer- ing, University of Wisconsin Water-free biologics introduce exciting new possibilities for precisely control of all the thermodynamic and kinetics parameters of biological reaction. However, biomolecules tend to lose majority of their functionality upon removal of the bulk water. Here, enzymatic dehalogenation of gas-phase substrate was selected as a model system to investigate how engineered abiotic/biotic interface may affect protein performance in its solid-state. We reported almost 40 times increase in dry-state activity of haloalkane dehalogenase when covalently conjugated to support matrix alongside poly-sorbitol methacrylate (PSMA). The PSMA aptitude to stabilize water-free state was found to be far significant than the accu- mulative effect of the same amount of sugar physically adsorbed to the protein and independent of surface water activity.

Engineering a Universal Dengue Vaccine Using a Virus-Like Particle Scaffold Danielle Basore1,2, Thomas Jordan1,2, Carolyn Barcellona3, Eric Carlin3, Emilie Mausser1, Kanthi Bommareddy1, Donna Crone1, Sharon Isern3, Scott Michael3, Chris Bystroff1,2,4 1Biological Sciences, Rensselaer Polytechnic Institute, 2CBIS, Rensselaer Polytechnic Institute, 3Biologi- cal Science, Florida Gulf Coast University, 4Computer Science, Rensselaer Polytechnic Institute Dengue virus is an emerging tropical disease, affecting over 300 million people each year. Dengue Fever, the disease caused by this virus, is flu like and severe. There is no cure or preventative other than avoid- ing the carrying mosquitoes. Dengue presents a challenge for vaccine development however, because it exhibits an antibody dependent enhancement of infection. Infection with another Dengue serotype after an initial infection can cause severe symptoms due to weakly binding antibodies that enhance uptake of virions of the subsequent serotype. This condition must be avoided when developing a vaccine for Den- gue virus. The antigen must be conserved across all Dengue serotypes, and generate only neutralizing antibodies. The Fusion Loop of the Envelope Protein of the virus (L65-F119, PDB 1OAN) is highly conserved and binds monoclonal antibodies that are known to neutralize dengue infection. This segment of the protein has been designed into the b4b5 loop of the major capsid protein L1 of Human Papilloma Virus (HPV). This protein spontaneously forms a 55nm virus-like particle (VLP) in solution, and in the cell. A locking disulfide bond is incorporated into the design to add stability to the insert; short linkers (CPG- insert-GPC) connect the Fusion Loop insert to the L1 monomer. A similar construct without the locking bond showed low expression by Western blot, suggesting that a large insert interferes with folding of the monomer or assembly of the VLP. The protein is expressed and matured into VLPs, which are then analyzed by transmission electron microscopy. Using an anti-L1 antibody, VLPs of the correct size have been identified by gradient ultracentrifugation and Western blotting. These chimeric VLPs will also be used to generate antibodies for immunogenicity assays, and in mice to assess efficacy of the vaccine.

Computational design and screening of leave-one-out green fluorescent protein biosensors for viral targets Keith Fraser, Shounak Banerjee, Danielle Chan, Julia Reimertz, Nailah Wade, Emily Crone, Monica Miles, Mason Gentner, Alex Stevens, Megan Jenkins, Eric Carlin, Lauren Paul, Angela Choi, Rachel Atschul, Casey Thornton, Colleen Lamberson, Christian D. Schenkelberg, Yao-ming Huang, Derek J. Pitman, Donna E. Crone, Jonathan S. Dordick, Scott Michael, Christopher Bystroff1 1Rensselaer Polytechnic Institute We present Leave-one-out Green Fluorescent Protein (LOOn-GFP), a circularly permuted and truncated GFP in which the nth b-strand has been ‘left out’ of the protein, making it unable to glow. Fluorescence

50 ABSTRACT

can be reconstituted with the addition of the left out peptide. We used InteractiveRosetta, an graphical interface for the Rosetta molecular modeling suite, to design mutations that accommodate a virus peptide in place of the nth b-strand. The quality of the computationally designed libraries was assessed by plate screening for green fluorescence in bacteria. On the way to eventual success, we discovered many ways to fail. Designed LOO-GFPs variously did not express, were not soluble, oligomerized in solution, gave false positive fluorescence, did not carry out the auto-catalytic maturation of the intrinsic chromophore, or glowed very weakly. Genetic libraries that were computationally designed to accomodate a 12-14 residue target peptide, did not produce glowing LOO-GFP•target complexes when the sequence complexity was high (>1e9 sequences, >20 variable positions) or when it was low (1000 or less, <8 variable positions), even when the target peptide was fused with the biosensor. On the other hand, genetic libraries designed to focus on 1 to 3 residues of the target peptide, with 8-10 variable positions and moderate sequence complexity (1e3 to 1e6) succeeded, giving 1 to 10% fluorescent bacterial colonies when co-expressed with or fused to the target peptide. Improvements were seen when a second round of design was carried out in the context of a first round “hit”. Here we report the improved binding of a LOO7-GFP biosensor to influenza hemagglutinin and the binding of LOO11-GFP to a tagged dengue virus envelope protein.

Mispacking and the fitness landscape of green fluorescent protein (GFP). Shounak Banerjee, Keith Fraser, Danielle Chan1, Julia Reimertz, Emily Crone, Mason Gentner, Robert Langan, Priscilla Hempel, Oluwadamilola Lawal, Kangning Zhang, Colleen Lamberson, Angela Choi, Rachel Alschul, Christian D. Schenkelberg, Yao-ming Huang, David Rosenman, Benjamin Walcott, Derek J. Pitman, Donna E. Crone, John Karanicolas, Jonathan S. Dordick, Christopher Bystroff 1Rensselaer Polytechnic Institute In order to fluoresce, green fluorescent protein (GFP) must fold efficiently, catalyze the spontaneous maturation of its intrinsic chromophore, and stabilize the excited state against rotational isomerization. The GFP fitness landscape, defined as relative fluorescence with respect to sequence, depends on on the chromophore maturation pathway, the flexibility of the chromophore, and on folding kinetics. An exhaustive study of mutations in seven positions surrounding the chromophore showed a surprising degree of mutational tolerance, producing bright green fluorescence in 120 out of 133 point mutants. Only four point mutations effectively blocked chromophore maturation, but even these eventually formed the chromophore, on a timescale of months. In contrast, double mutants of contacting side chains glowed less frequently than expected, and mutating additional positions continues the trend of non-additive loss of fluorescence. Both positive and negative non-additive effects

51 ABSTRACT

of mutation were observed. This paints a picture of the fitness landscape of GFP as rough and disjoint. In case studies where structural data is available, the thermodynamic stability of folding (-DGf) correlates with both the chromophore maturation rate and the quantum yield, even when mutations are distal to the chromophore. Rotational isomerization of the chromophore is the most favored route to fluorescence quenching by non-radiative decay, but the nearly identical positions of the proximal atoms does not explain the significant differences in quantum yield and chromophore maturation rate. The size of the buried cavities in the vicinity of the chromophore correlates with chromophore maturation rate, quantum yield, and the unfolding rate. All of these phenomena can be explained by side chain mispacking in the core upon folding, which is relieved only upon refolding. We propose that the thermodynamic stability and the propensity to form trapped, mispacked states are closely linked.

De novo design of protein homo-oligomers with modular hydrogen bond network-mediated specificity Scott E. Boyken, Zibo Chen1, Benjamin Groves, Robert A. Langan, Gustav Oberdorfer, Alex Ford, Jason M. Gilmore, Chunfu Xu, Frank DiMaio, Jose Henrique Pereira, Banumathi Sankaran, Georg Seelig, Peter H. Zwart, David Baker 1University of Washington In nature, structural specificity in DNA and proteins is encoded quite differently: in DNA, specificity arises from modular hydrogen bonds in the core of the double helix, whereas in proteins, specificity arises largely from buried hydrophobic packing complemented by irregular peripheral polar interactions. Here we describe a general approach for designing a wide range of protein homo-oligomers with specificity determined by modular arrays of central hydrogen bond networks. We use the approach to design dimers, trimers, and tetramers consisting of two concentric rings of helices, including previously not seen triangular, square, and supercoiled topologies. X-ray crystallography confirms that the structures overall, and the hydrogen bond networks in particular, are nearly identical to the design models, and the networks confer interaction specificity in vivo. The ability to design extensive hydrogen bond networks with atomic accuracy is a milestone for protein design and enables the programming of protein interaction specificity for a broad range of synthetic biology applications.

52 ABSTRACT

De Novo Design of Protein Scaffolds Suited for Ligand-Binding Tamuka Chidyausiku1, Benjamin Basanta1, Enrique Marcos PhD1, David Baker PhD1 1Biochemistry Department, University of Washington The Holy Grail of biochemistry lies in the understanding of the structure-function relationship of proteins. By folding into structures with pockets, proteins enable small-molecule binding events associated with molecular recognition, metabolic pathways or cell signaling. Likewise, computational design of

53 ABSTRACT ligand-binding proteins and enzymes relies on finding an existing protein structure with a pocket satis- fying the binding site requirements of the targeted small-molecule - while remaining stable after incorporating the necessary mutations. Despite the successes of this approach for repurposing native proteins, this dependence on existing protein structures can be a limitation for certain applications and, instead, building proteins de novo with custom-made binding sites should be more effective. However, de novo design of proteins supporting pockets remains an outstanding challenge. Among them, b- sheet curvature is ubiquitous in native ligand-binding protein structures, like the b-barrel, NTF2-like and jelly-roll fold domains; and the control of its geometry by design may open up new possibilities for cus- tomizing ligand-binding sites. However, two factors make b-sheets challenging to design: their non- local nature, which slows the folding rate and can cause misfolding; and their solvent-exposed strand edges, which are prone to aggregation – explaining the slower progress in de novo design of b-sheets compared with helical proteins. Here we describe an approach to design quasi-ideal protein folds with curved b-sheets supporting pockets. We used this method to de novo design a panel of protein folds of incremental complexity, ranging from cystatin-like to NTF2-like folds, at atomic-level accuracy and explored stabilization strategies to balance the stability-activity tradeoff, while preserving the pocket accessibility, for potential functional applications.

Lessons in Protein Electrostatics from Natural Enzymes Applied to Enzyme Design Tim Coulther1 1Department of Chemistry & Chemical Biology, Northeastern University Enzymes have applications in a variety of industries due to their ability to catalyze reactions at high rates under mild conditions. Enzymatic catalysis has advantages over conventional catalytic processes, namely less energy consumption and fewer unwanted byproducts. However, a natural enzyme that can catalyze the desired reaction does not always exist. While there have been successes in the protein design field, including de novo enzyme design for a few reactions that have no known natural counter- part, these enzymes have low activity compared to natural enzymes. Laboratory techniques such as directed evolution have been used to improve activity levels of designed enzymes, but these methods are costly and time consuming, limiting their widespread use. In order to further develop new enzymes successfully, a better understanding of natural enzymes is needed, specifically the interactions between residues that facilitate catalysis. The computational method THEMATICS identifies active site residues through their perturbed proton transfer behavior. In addition to the residues that interact directly with the substrate molecule, distal residues, or those farther away from the substrate, are also predicted and have been verified to contribute to catalysis through biochemical assays on variants. Computationally, every possible single-site variant can be examined, deconstructing the active site to understand the contributions of each residue to the chemical properties needed for catalysis. Many designed enzymes lack the electrostatic properties of their natural counterparts; while natural enzymes can have extensive networks of coupled proton transfer processes, designed enzymes are typically less connected, with fewer residues seemingly contributing to catalysis. These differences may be one reason for the low activity observed for designed enzymes. Since these electrostatic features are computationally identifiable, design processes incorporating these metrics.

Computational Design of Dynamics into a Stable Globular Protein Adam M. Damry1, James A. Davey1, Natalie K. Goto1, and Roberto A. Chica1 1Department of Chemistry and Biomolecular Sciences, University of Ottawa Proteins are widely used in research, industry, and medicine for their ability to carry out complex molecular processes with high precision and efficiency. It is generally thought that the three-dimensional

54 ABSTRACT structure of proteins dictates their function, but increasing evidence demonstrates that complex functions are equally mediated by dynamics. Traditional rational design methodologies however do not con- sider this important property when engineering proteins. This is in part due to the lack of a framework for the structure-based rational design of protein dynamics. To address this issue, we have developed a methodology based on multistate design (MSD), an emerging methodology in computational protein design that optimizes sequences in the context of multiple structural states. As a proof-of-concept for our framework, we predicted and experimentally validated sequences that stabilize and facilitate exchange between two non-native conformations of Trp43 in the streptococcal protein G domain b1 (Gb1) fold. Four candidate sequences predicted to exchange between core and solvent-exposed conformations of Trp43 were identified by MSD across a sequence space of 1296 possible mutants evaluated on an ensemble of 12,648 unique Gb1 backbones. 15N-HSQC and ZZ-exchange NMR spectroscopy confirmed that all four candidate sequences are dynamic, with two of them exchanging between two distinct conformations on the 10 to 100 millisecond time scale. Solution structures of two Gb1 mutants displaying decreased dynamics, whose 15N-HSQC spectra showed peaks with highly similar chemical shifts to those from a single conformer of each dynamic variant, were solved. The structures of these mutants, coupled with an analysis of NNOE correlations from the Trp43 He1, confirmed that Trp43 adopts conformations corresponding to those designed in the exchanging mutants. Overall, the successful application of our MSD strategy paves the way to the rational design of protein dynamics as we move towards increasingly complex protein functions.

Gas Vesicle Nanoparticles: Composition, Biogenesis, and Therapeutic Applications Shiladitya DasSarma1, Priya DasSarma1, Wolf Pecher1, Ram Karan1, and Jong-Myoung Kim1 1University of Maryland School of Medicine, Baltimore Gas vesicle nanoparticles (called GVNPs or gas vesicles) are extremely stable, hollow structures bounded by a protein membrane [1]. One halophilic microbe Halobacterium sp. NRC-1 produces these buoyant organelles in abundance for buoyancy and is being used for studies and applications of GVNPs. More than a dozen proteins encoded by a large gene cluster in a NRC-1 plasmid have been implicated in GVNP formation [2,3]. GvpA, an amyloid-like protein, constitutes the major component of the nanoparticles. GvpA is not easily solubilized, but was identified in an oligomeric form using a highly denaturing phenol-acetic acid-urea (PAU)-polyacrylamide gel [4]. Two small pfam741 polypeptides similar to GvpA, GvpJ and GvpM, were also identified using immunoblotting and confirmed by mass spectrometry [5,6]. Spectroscopic and de novo modeling studies of GvpA have suggested that this protein adopts a coil-a- b-b-a-coil fold with anti-parallel b-strands and forms a hydrophobic surface on the internal face and a hydrophilic surface on the external side of the nanoparticles [7-9]. The GvpA family of proteins vary considerably between species and have been proposed to function in shape-determination of the nanoparticles. Another protein, GvpC, readily detectable on the exterior surface of the GVNPs, strengthens the nanoparticles and may be released from GVNPs under mildly denaturing conditions [4,10]. This pfam1304 protein family contains a motif that is repeated eight times in NRC-1, and the repeats are thought to bind to multiple copies of GvpA on the external surface of the GVNP membrane [4,9]. An acidic tail is present in the NRC-1 GvpC protein for stability under the highly saline conditions in which they are found. GVNPs have been produced in NRC-1 with the acidic tail of GvpC fused to foreign proteins, which results in their display on the nanoparticles and allows their use for antigen delivery, vaccine development, and potential therapeutic agents [11-15]. Three additional proteins, GvpF, GvpG, and GvpL, have been observed in NRC-1 GVNPs using antisera against unique peptides from these proteins. GvpF and GvpL, members of pfam6386, contain coiled-coil domains involved in laddering, self- association, and nucleation [4]. This presentation will review the current state of knowledge on GVNP proteins, their structure and biogenesis, and summarize recent studies on immunogenicity of pathogenic viral, bacterial, and eukaryotic proteins and peptides displayed on the nanoparticles.

55 ABSTRACT

References: 1. DasSarma, S. and P. DasSarma. Vacccines 3:686-702, 2015. 2. DasSarma S. et al. J. Bacteriol. 176:7646-7652, 1994. 3. Ng, W.V. et al. Genome Res. 8:1131-1141, 1998. 4. Halladay, J.T. et al. J. Bacteriol. 175:684-692, 1993. 5. Shukla, H.D. and S. DasSarma. J. Bacteriol. 186:3182-3186, 2004. 6. Chu, L.J. et al. J. Proteome. Res. 10:1170-1178, 2011. 7. Sivertsen, A.C. Biophys. J. 99:1932-1939, 2010. 8. Ezzeldin, H.M. et al. J. Struct. Biol. 179:18-28, 2012. 9. Jones, J.G. et al. Gene 102:117-122, 1991. 10. Hayes, P.K. et al. Arch. Microbiol. 157:229-234, 1992. 11. Stuart, E.S. et al. J. Biotechnol. 88:119-128, 2001. 12. Stuart, E.S. et al. J. Biotechnol. 114:225-237, 2004. 13. DasSarma, S. et al. BMC Biotechnol. 13:112, 2013.

Controlling the Phosphoproteome: Ligand Activated Split-Kinases and SplitPhosphatases Mahsa M.E. Ghaffari1, Karla Camacho-Soto1, Javier Castillo- Montoya1, Indraneel Ghosh1 1Department of Chemistry and Biochemistry, University of Arizona Reversible phosphorylation controls both temporal and spatial activity of proteins within almost all signaling cascades. The human complement of >500 protein kinases and 147 protein phosphatases regulate various cellular events from cell division to cell death. The aberrant function of these proteins is implicated in diseases such as cancer and neurodegeneration. Thus, understanding the function of a specific kinase of phosphatase, by turning them on or off, is important for understanding their biology and for designing new therapies. Toward this goal, a few emerging methods have made significant advances towards turning-on the activity of a specific kinase. However, current methods do not neces- sarily allow for orthogonal control over two or more kinases or phosphatases in living cells. We hypothesized that if we could design fragmented or split-kinases that can be turned on by different ligands then two or more kinases or phosphatases could be temporally controlled. We first developed a sequence dissimilarity based approach to identify sites in the catalytic domain of kinases tolerant to a 25-residue loop insertion. The successful loop insertion sites, guided the fragmentation of the kinases at these sites into two inactive fragments, which were subsequently attached to two proteins, FKBP and FRB that dimerize in the presence of the small molecule, rapamycin. We demonstrated that the addition of rapamycin to the designed split-kinases could selectively turn-on enzymatic activity. We have

56 ABSTRACT successfully tested this approach for members of the tyrosine kinase such as Src, Lyn and Fak as well as tyrosine phosphatases, TCPTP and HePTP. Currently, we are both improving the design of the split- enzymes for achieving orthogonal control and studying ligand dependent activation of signal transduction pathways.

Developing a soluble bifunctional receptor/co-receptor mimetic for structural characterization of the HIV envelope glycoprotein Agnes Hajduczki1 1NIH/NIAID The HIV envelope glycoprotein (Env) mediates virus entry by initiating fusion of the viral envelope with the cell membrane. The surface-exposed gp120 subunit undergoes stepwise conformational changes upon interactions with the primary receptor CD4, and co-receptors, CCR5 or CXCR4, both of which are membrane-anchored G-protein-coupled receptors. Obtaining structural information on the intermediates during viral entry is a key focus of antiviral and vaccine research and could open the doors for more effective treatment and prevention. Due to the inherent insolubility of membrane proteins, working with the intact co-receptors outside the context of the membrane is not an option. This project aims to develop recombinant soluble co-receptor mimetics featuring one or both of the critical determinants of CCR5, the N-terminus (Nt) and extracellular loop 2 (E2), fused to soluble CD4 (sCD) by flexible polypeptide linkers. We have successfully overexpressed and purified the recombinant proteins from mammalian cells, and demonstrated that the Nt-bearing variants contain sulfated tyrosines, shown to be critical for co-receptor activity. Characterization of the gp120-binding properties of the variants is underway using a vaccinia-based cell fusion assay where sCD4 has been shown to induce membrane fusion between Env-expressing effector cells and target cells bearing CCR5, but no CD4. When the sCD4-fused co-receptor mimetic is added to the reaction, there is strongly impaired activation of membrane fusion compared with sCD4 alone or an unsulfated control, suggesting that the CCR5-derived portion of the protein competes with cell-surface CCR5 for binding to gp120. Conversely, the mimetic displays stronger neutralization activity than sCD4 or the unsulfated control in pseudovirus neutralization assays. The soluble variants will be used to elucidate the conformational changes in gp120 that immediately precede membrane fusion, and potentially for collaborative structural analyses of the gp120-coreceptor complex.

Antigen clasping: novel antibody-antigen recognition mechanism enabling extraordinarily high specificity Takamitsu Hattori1,2, Darson Lai2, Irina Dementieva2, Sherwin Montano2, Kohei Kurosawa2, Akiko Koide2, Alexander J. Ruthenburg2,3, Shohei Koide2 1Department of Biomolecular Engineering, Graduate School of Engineering, Tohoku University, 2Department of Biochemistry and Molecular biology, The University of Chicago, 3Department of Molecular Genetics and Cell Biology, The University of Chicago The antigen-binding fragment (Fab or Fv) of immunoglobulins containing an antigen-binding site is viewed as an independent and single unit for antigen recognition, with one Fab binding to one antigen. Here, we report novel antibody-antigen recognition mechanism where two Fab cooperatively recognize one antigen [1]. We isolated highly specific antibodies to trimethylated Lys9 and Lys4, respectively, on histone H3 (H3K9me3 and H3K4me3), which are critical epigenetic modifications and challenging targets for molecular recognition. Unexpectedly, the crystal structures of peptide-antibody complexes revealed that two antigen-binding sites of these antibodies form a head-to-head dimer and sandwich the antigen in the dimer interface. This antigen clasping created extensive antigen recognition interface where trimethylated Lys interacted with an unusual aromatic cage in one Fab and the histone N- terminus with another pocket in the other Fab. The distance between two binding pockets were

57 ABSTRACT different in each antibody and fine-controlled by antigen clasping, thereby achieving exquisite specificity. A new antibody format designed for promoting antigen clasping showed extraordinarily high specificity and potency. Together, antigen clasping expands capacity for molecular recognition and guide the design of the next generation of recombinant antibodies to challenging targets including histone modifications and small molecules. Acknowledgement: This work was supported by NIH grants DA025725 and DA028779. [1] Hattori T. et al., Proc. Natl. Acad. Sci. U. S. A., 113, 2092-2097 (2016).

Novel Proteins Provide Life Sustaining Activities In Vivo Michael Hecht1 1Princeton University One of the key goals of synthetic biology is to design novel proteins that fold and function in vivo. A particularly challenging objective would be to produce non—natural proteins that don’t merely generate interesting phenotypes, but which actually provide essential functions necessary for the growth of living cells. The successful design of such life—sustaining proteins would represent a first step toward constructing artificial “proteomes” of non—natural sequences. In initial work toward this goal, we designed large libraries of novel proteins encoded by millions of synthetic genes. Many of these novel proteins fold into stable 3—dimensional structures, and many bind biologically relevant metals, metabolites, and cofactors. Several of the novel proteins function in vivo providing essential activities necessary to sustain the growth of E. coli cells. In some cases, these novel proteins function by providing specific catalytic activities, while in other cases, the non—natural protein sustains cell growth by providing a novel regulatory function that alters the expression of endogenous genes. These results suggest that (i) the molecular toolkit for life need not be limited to proteins that already exist in nature, and (ii) artificial genomes and proteomes might be built from non—natural sequences.

A Directed Evolution Approach to Engineer Caspase Specificity Maureen E. Hill1, Derek J. MacPherson1, Peng Wu, Ph.D.1, Olivier Julien, Ph.D.1, James A. Wells, Ph.D.1 and Jeanne A. Hardy, Ph.D.1 1University of Massachusetts Amherst Proteases are valuable enzymes that provide indispensable tools for applications in industry, academia and pharmaceutical research. When applicable, proteases may even be engineered for enhanced properties that allow for improved activity and stability. With recent advancements in directed evolutionary approaches, proteases may be exploited for even more sophisticated purposes. For instance, approaches targeting the active site would allow for precise control over cleavage of potential therapeutic targets. Our lab has developed a directed evolution method using a dark-to-bright GFP reporter that enables us to tune specificity into intracellular proteases. In our first application of this technology, we sought to change the specificity of the apoptotic caspase-7 (DEVD) into that of caspase-6 (VEID). Saturation mutagenesis at chosen active site residues produced evolved-specificity caspase-7 (esCasp-7) variants with caspase-6-like activity after a single round of directed evolution. Crystal structures of the esCasp-7 variants show the restructuring of one active site loop to accommodate the caspase-6 substrate mimic. Further investigation of esCasp-7 specificity using N-terminomics with natural proteins revealed very similar substrate preferences to caspase-6 across the entire human proteome. With the body of caspase-7 intact, we sought to decipher proteins that may rely on non-active site interactions (exosites) for recognition. Based on esCasp-7 activity, we predict that the caspase-6 substrate lamin C relies on an exosite for cleavage. Such reprogrammed caspases represent a new method to decipher exosite driven substrates and should be broadly applicable to many other proteases.

58 ABSTRACT

Artificially designed peptides that degrade amyloid fibrils Yoshihiro Iida1 and Atsuo Tamura1 1Kobe University, Graduate School of Science, Department of Chemistry Amyloid fibrils are misfolded and self-assembled aggregates of proteins approximately 10 nm in diameter and several micrometers in length. Beta-strands of the fibrils are arranged perpendicularly to the fibril long axis to form the characteristic cross-beta structure. Amyloidosis is the extracellular deposition of these insoluble protein fibrils, leading to tissue damage and disease such as Alzheimer’s disease, prion disease, type II diabetes, and dialysis-related amyloidosis. To remedy amyloidosis, one of the most effective ways is to hydrolyze amyloid fibril directly. We thus tried to design short peptides having hydrolysis activity against amyloid fibrils. As a strategy for the design, we made peptides to have the catalytic triad composed of histidine, aspartic acid and serine, and synthesized 9 peptides named me1- 9. We employed thioflavin T assay, circular dichroism spectroscopy, atomic force microscopy and mass spectrometry for analyses. It has been shown that me5 takes the alpha-helical conformation and can hydrolyze amyloid fibrils comprising of various proteins, i.e., beta lactoglobulin, insulin, amyloid beta40, amyloid beta42 and beta 2 microglobulin. It is concluded that the designed peptide me5 can be regarded as a hydrolase which is capable of hydrolyzing amyloid fibrils specifically.

Computational design of allosteric antibody Olga Khersonsky1, Sarel J. Fleishman1 1Department of Biological Chemistry, Weizmann Institute of Science, Rehovot, Israel Allosteric regulation is central to the control of many metabolic and signaltransduction pathways, and thus it was dubbed “the second secret of life”. The mechanisms of allostery, however, are largely remaining enigmatic. De novo design of allosteric proteins is particularly challenging due to the need to model the backbone degrees of freedom, and to define the multiple protein states. Recently, AbDe- sign algorithm was developed in our lab for de novo design of antibodies. It is guided by natural conformations and sequences, and exploits the modular nature of antibodies to generate an immense

59 ABSTRACT space of conformations, which can be used as scaffolds for design of stable high-affinity binders. We have used AbDesign to design an allosteric antibody, in which Zn21 is an allosteric effector of fluorescein binding. We introduced Zn21 binding site into the H3 CDR of fluorescein-binding antibody, with the assumption that without Zn21 chelation, the floppiness of H3 backbone would prevent fluorescein from binding. We used as a template the high-affinity fluorescein-binding antibody (PDB ID 1x9q), and spliced into it various H3 segments to obtain a variety of scaffolds. Zn21 binding site was introduced into these scaffolds with RosettaMatch, and the sequence was subsequently optimized by design. 100 designs were experimentally tested by yeast display for fluorescein binding in the presence or absence of Zn21. One of the designs demonstrated significantly higher fluorescein binding in the presence of Zn21, and its Zn21 binding was specific and reversible. Mutational analysis and SPR measurements confirmed the allosteric nature of the designed antibody.

Super WA20 (SUWA), an Ultra-Stabilized Dimeric de Novo Protein for Self-Assembling Protein Nanobuilding Blocks Naoya Kimura1, Naoya Kobayashi1, Ryoichi Arai1 1Grad. Sch. of Sci. & Tech., Shinshu Univ., Ueda, Nagano 386-8567, Japan Recently, we designed and created a protein nanobuilding block (PN-Block), WA20-foldon,by fusing an intermolecularly folded dimeric de novo protein WA20 and a trimeric foldon domain from bacteriophage T4 fibritin (Kobayashi, N., et al., 2015, J. Am. Chem. Soc., 137, 11285). The WA20-foldon, as a simple and versatile nano-building block, self-assembled into several oligomeric nanoarchitectures (6-mer, 12-mer, 18-mer, 24-mer) in multiples of 6-mer. To further stabilize the de novo protein WA20 for extensive application of PN-Blocks in nanotechnology, we designed and created a variant, called Super WA20 (SUWA), by several mutations for stabilization of helices and hydrophobic cores. Thermal denaturation experiments by circular dichroism show that denaturation midpoint temperature (Tm) for SUWA is extremely high, 1228C, whereas Tm for WA20 is 758C. In addition, chemical denaturation experiments with guanidine hydrochloride show that denaturation midpoint concentration for SUWA is very high, 6 M, whereas that for WA20 is 3.5 M. Furthermore, we constructed a PN-Block, SUWA-foldon, and its native PAGE analysis suggests that the PN-Block self- assembled into several nanostructures. Thermal denaturation experiment of the SUWA-foldon nanostructures suggests very high structural stability. These results demonstrate that we successfully created the ultra-stabilized dimeric de novo protein, SUWA, which expands the possibilities of PN- Block approach.

Self-Assembling Supramolecular Nanostructures Created by de Novo Extender Protein Nanobuilding Blocks Naoya Kobayashi1, Kouichi Inano2, Takaaki Sato2, Keisuke Miyazawa3, Takeshi Fukuma3, Michael H. Hecht4, Ryoichi Arai1 1Dept. of Appl. Biol., Fac. of Text. Sci. & Tech., Shinshu Univ., Ueda 386-8567, Japan, 2Dept. of Chem. & Mater., Fac. of Text. Sci. & Tech., Shinshu Univ., Ueda 386-8567, Japan, 3Div. of Electric. Eng. & Comput. Sci., Kanazawa Univ., Kanazawa 920-1192, Japan, 4Dept. of Chem., Princeton Univ., Prince- ton, NJ 08544 The design of novel proteins that self-assemble into supramolecular complexes is an important step in the development of synthetic biology and nanobiotechnology. Recently, we designed and created a protein nanobuilding block (PN-Block), called WA20-foldon, by fusing an intermolecularly folded dimeric de novo protein WA20 and a trimeric foldon domain of T4 phage fibritin (Kobayashi, N., et al., J. Am. Chem. Soc. 2015, 137, 11285). The WA20-foldon formed several distinctive types of self- assembling nanoarchitectures in multiples of 6-mer (6-, 12-, 18-, 24-mer) because of the combination of dimer and trimer. In this study, to construct self-assembling extended chain nanostructures, we

60 ABSTRACT designed de novo extender protein nanobuilding blocks (ePN-Blocks) as the second series of PN- Blocks. The ePN-Blocks were constructed by fusing tandemly two de novo WA20 proteins with various linkers. The ePN-Block proteins with the long helical linkers or flexible linkers were expressed mainly in soluble fractions in Escherichia coli. The purified ePN-Blocks migrated as ladder bands in native PAGE, suggesting that the ePN-Blocks form several homooligomeric states, probably circular chain structures, in the soluble fraction. Then, we reconstructed heteromeric complexes from extender and stopper PN- Blocks by denaturation and refolding. Size exclusion chromatography-multiangle light scattering and small-angle X-ray scattering analyses suggest that extender and stopper PN-Block (esPN-Block) complexes formed different types of extended chain structures depending on their linker types. Moreover, observation by atomic force microscopy in liquid revealed the esPN-Block complexes with metal ion further self-assembled into supramolecular nanostructures on mica surface. These results suggest that the PN-Block approach using the de novo proteins is a powerful strategy to create novel self- assembling supramolecular nanostructures.

Substitutions affecting the flexibility of the N-terminus of the cold adapted subtilase, VPR, are important for its temperature adaptive properties. Kristinn R. Oskarsson 1, Magnus M. Kristjansson1 1Department of Biochemistry, Science Institute, University of Iceland, Reykjavık, Iceland. Protein function relies upon a fine balance between two opposing structural factors; molecular flexibility and stability. It is proposed that molecular mechanisms of temperature adaptation of proteins involve the adjustment of their molecular flexibility. Thus, the higher catalytic activity usually observed for cold adapted enzymes compared to homologues from thermophiles is believed to reflect their higher, global or local, molecular flexibility. VPR is a subtilisin-like serine proteinase (subtilase) from a psychrophilic Vibrio sp. Comparative studies of the cold adapted VPR and related subtilases, especially aqualysin I (AQUI), a close homologue from the thermophile Thermus aquaticus, have guided us in selecting mutations which we have used to test our hypotheses regarding the molecular origins of temperature adaptation of these enzymes to extreme conditions. Observations made in such studies have indicated that substitutions near the N-terminus may be important for the temperature adaptive properties of these subtilases. A variant of VPR has two proline residues inserted at corresponding sites to AQUI, near the N-terminus of the enzyme. The thermal stability of the VPR_N3P/I5P variant was shown to be significantly increased, but it displayed a concomitant loss of catalytic efficiency. We propose that the insertion of Pro3 and Pro5 may constrain the configuration of the N-terminal region of the enzyme. We show that a highly conserved proximal tryptophan residue, Trp6, is a major contributor to the intrinsic fluorescence of VPR. The accessibility of Trp residues may be assessed by measuring the quenching of their intrinsic fluorescence by small molecules such as acrylamide. The degree of the quenching effect reflects how readily the quencher can penetrate the protein and hence the flexibility at this site in the protein structure. We measured the Trp fluorescence quenching of VPR, AQUI and the VPR_N3P/I5P and VPR_W6F variants under the same set of conditions in order to assess whether their different properties could be related to flexibility of the N-terminal region, as monitored by quenching of the fluorescence of Trp6. From Stern-Volmer plots we observe a significantly decreased quenching effect for the VPR_N3P/I5P variant as compared to VPR. The fluorescence of the double Pro variant was indeed similar to that of the thermophilic AQUI, containing prolines at the corresponding sites, as well as Trp6. These results suggest that the N3P and I5P substitutions appear to constrain the N-terminal region of enzyme, which may contribute to higher thermal stability of the thermophilic AQUI. To further emphasize the structural role of the N-terminal region for these enzymes we observe a 14-158C decrease in thermal stability for the VPR_W6F variant.

61 ABSTRACT

Harnessing the Reactivity of Selenocysteine for Expressed Protein Ligation Jun Liu1, Qingqing Chen, Sharon Rozovsky1 1Brown lab 138, Department of Chemistry & Biochemistry, University of Delaware Protein semisynthesis is expanded by selenocysteine-mediated chemical ligation using heterologously expressed protein fragments. The method incorporates selenocysteine at any position with no constraints on the fragment size and properties and exploits its high reactivity to facilitate challenging ligation reactions. Furthermore, following ligation the selenocysteine can be readily and selectively deselenized to an alanine or serine or used as a chemical handle for site-specific tagging with chemical probes.

Directed Evolution of a LOV-Trap for Deciphering Intracellular Signaling Pathways Rihe Liu1, Hui Wang1 1Eshelman School of Pharmacy, Carolina Center for Genome Sciences, Department of Pharmacology, University of North Carolina We developed a versatile optogenetic approach that allows for repeated and reversible control of protein activity under precise kinetics (Wang et. al 2016). This system is based upon a small trapping protein that specifically bind to the dark but not the lit state of LOV2 protein, a photo sensor that is used by a wide variety of higher plants to sense environmental conditions. We started with a protein domain library based on the Z subunit of protein A, in which 13 surface-exposed residues were totally randomized. Trapping molecules that selectively bind to the closed form of LOV2 with low nanomolar affinities were evolved from an mRNA-displayed Z domain library with a diversity of 50-trillion unique sequences. The availability of this LOV-TRAP makes it convenient to sequester a protein of interest away from its site of action by interaction in the dark with the LOV2 engineered to being anchored on desired intracellular membranes. This sequestered protein of interest can be easily released from membrane-bound LOV2 upon irradiation by blue light, and move to its site of action to turn on or off the signaling pathways. One unique feature of this system is that it provides diffusion limited activation kinetics with tunable deactivation rates. This broadly applicable approach can be used to decipher the regulation of numerous intracellular signaling pathways in a spatio-temporal manner. References: Wang, H., Vilela, M., Winkler, A., Tarnawski, M., Hartmann, E., Schlichting, I., Yumerefendi, H., Kuhlman, B., Liu, R.*, Danuser, G.*, Hahn, K.* (2016) “LOVTRAP, A Versatile Optogenetic System, Reveals Resonator Motifs in Mammalian Mechano-chemical Signaling Pathways”, Nature Methods, Under minor revision.

62 ABSTRACT

Induced domain swapping (INDOS): a modular design for controlling protein function with a small molecule Jeung-Hoi Ha1, Joshua M. Karchin1, Stewart N. Loh1 1Department of Biochemistry & Molecular Biology, State University of New York Upstate Medical University Domain swapping is the process by which identical proteins exchange segments in a reciprocal fashion. We are developing engineered swapping as a mechanism for controlling protein function as well as for creating self-assembling biomaterials that retain and combine the functions of their constituent proteins. We recently demonstrated that a target protein can be induced to domain swap by inserting a ‘lever’ protein into a surface loop of the target. If the target is more stable than the lever, the fusion protein is a monomer and consists of a folded target domain and an unfolded lever domain (functional OFF state; see below). If the lever is more stable than the target, the lever folds and splits the target, forcing the latter to refold in trans to generate domain-swapped dimers and oligomers (functional ON state). Here we design a modular system in which domain swapping of an arbitrary target protein can be triggered by the cell-permeable molecule FK506. The lever, FK506 binding protein (FKBP), is inserted into one of multiple surface loops in two targets: ribose binding protein (RBP) and staphylococcal nuclease (SNase). The biological function of each target is turned off by one of two point mutations, rendering the RBP-FKBP and SNase-FKBP fusion proteins inactive in their monomeric forms. Addition of FK506 induces folding of the FKBP domain. Subsequent swapping of the RBP and SNase domains yields dimers in which one of the swapped copies is composed of the non-mutated, functional sequence. We observe 10- to 50-fold activation of ribose binding and DNAse enzymatic activities upon FK506 addition. This design is intended to switch on and off the function of an arbitrary protein in vivo. We present guidelines for creating protein switches based on this induced swapping mechanism.

Caspase-7 with Reprogrammed Specificity Allows Identification of Exosites for Substrate Recognition Derek MacPherson1, Maureen Hill1, Peng Wu1, Olivier Julien2, James A. Wells2, Jeanne A. Hardy1 1Department of Chemistry, University of Massachusetts Amherst, 2Department of Pharmaceutical Chemistry and Cellular & Molecular Pharmacology, University of California at San Francisco Caspases are proteases essential for apoptosis, a process necessary for maintaining cellular homeostasis and differentiation. Disruption of caspase function has been implicated in cancer, neurodegeneration and inflammation, emphasizing the need to deconvolute individual caspase function, regulation and molecular substrates. Caspases have similar active site composed of four highly flexible loops and recognize similar peptide-length substrates. Nevertheless, each of the caspases exhibits extraordinary distinguishing power for a particular subset of intracellular targets. One hypothesis is that caspases utilize distal interactions via unique exosites to differentiate their substrates. Due to the implications of caspase-6 in neurodegenerative disorders (i.e. Alzheimer’s, Huntington’s) we aimed to generate an evolved specificity caspase (esCasp) that would allow us to distinguish the role of exosites in caspase-6 substrate recognition. esCasp proteins were generated by saturation mutagenesis at critical substrate binding residues in the active site of caspase-7. Utilizing a caged GFP reporter, the library of caspase-7 variants was sorted by their ability to cleave a new recognition sequence (VEID) encoded in the reporter. The end result enabled us to generate esCasp-7 variants that maintained the body of caspase- 7 but had the specificity of the caspase-6. Using N-terminomics to profile the human proteome, we found esCasp-7 displayed the same specificity as caspase-6, indicating the effectiveness of our evolutionary screen. We predicted that proteins requiring exosites for recognition by caspase-6 would failto be hydrolyzed by esCasp-7, enabling the first known approach for directly assessing exosite contributions to substrate recognition. We identified lamin A/C, as a substrate relying on putative exosite interactions for caspase-6 recognition. The most compelling promise of identifying exosites is their potential

63 ABSTRACT

for precision medicine, allowing us to exploit exosite-directed inhibitors to prevent recognition of one therapeutic substrate, while leaving interactions with all other substrates unchanged.

Dynamic combinatorial libraries from designed armadillo repeat protein fragments Erich Michel1, Andreas Pluckthun€ 2, Oliver Zerbe1 1Department of Chemistry, University of Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzer- land, 2Department of Biochemistry, University of Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzerland Molecular reagents for the specific detection of marker proteins in cellular states are of key importance in diagnostics and the understanding of cellular signaling and regulation in healthy and diseased cells. A fundamentally new approach based on designed armadillo repeat proteins (dArmRP) has recently been initiated for the development of modular reagents for specific detection of regular and post- translationally modified polypeptide sequences. A key feature of dArmRPs is the specific modular recognition of two amino acids per repeat module. In addition, complementary fragments of split dArmRPs spontaneously self-assemble into a native-like conformation that retains the ability to bind its target peptide. Using biochemical and biophysical data, we present the development and characterization of conditional protein fragment complementation exclusively in presence of target peptide ligands. We further outline the application of the obtained knowledge to develop novel dynamic combinatorial polypeptide libraries and split in-vivo protein reporters for the detection of cellular states.

Construction of Zn-SO4 Cluster-encapsulating Protein Nanocage by Domain Swapping Takaaki Miyamoto1, Mai Kuribayashi1, Satoshi Nagao1, Yasuhito Shomura2, Yoshiki Higuchi3,4, Shun Hirota1 1Graduate School of Materials Science, Nara Institute of Science and Technology, 8916-5 Takayama, Ikoma, Nara 630-0192, Japan, 2Graduate School of Science and Engineering, Ibaraki University, 4-12-1 Nakanarusawa, Hitachi, Ibaraki 316-8511, Japan, 3Department of Life Science, Graduate School of Life Science, University of Hyogo, 3-2-1 Koto, Kamigori-cho, Ako-gun, Hyogo 678-1297, Japan, 4RIKEN SPring-8 Center, 1-1-1 Koto, Sayo-cho, Sayo-gun, Hyogo 679-5148, Japan Protein nanostructures have been of great interest, because they have potential for various applications in bionanotechnology. To date, protein nanostructures have been constructed by several methods,

64 ABSTRACT including chemical modification, computational design, metal coordination, and protein fusion. How- ever, development of other methods, such as domain swapping, is required for increase in protein nanostructure variety. Domain swapping is a mechanism of protein oligomerization, where a secondary structural region or a domain of one protein molecule is replaced with the corresponding region or domain of another protein molecule. We have previously shown that small spherical heme proteins, c- type cytochrome (cyt) proteins and myoglobin, form oligomers by domain swapping. In this study, we have succeeded in constructing a unique protein nanocage using domain swapping.1 We used a four- helix bundle heme protein, cytochrome cb562 (cyt cb562). Oligomeric cyt cb562 was obtained by treatment with acetic acid. According to X-ray crystallographic analysis, dimeric cyt cb562 exhibited a domain-swapped structure, where the two helices in the N-terminal region of one protomer interacted with the other two helices in the C-terminal region of the other protomer. In the crystal, three domain- swapped cyt cb562 dimers formed a unique cage structure with a Zn-SO4 cluster inside the cavity. The Zn-SO4 cluster consisted of fifteen Zn21 and seven SO42– ions, whereas six additional Zn21 ions were detected inside the cavity. The cage structure was stabilized by coordination of the amino acid side chains of the dimers to the Zn21 ions and connection of two four-helix bundle units through the conformation-adjustable hinge loop (Lys51–Asp54). These results show that domain swapping can be applied in the construction of unique protein nanostructures. References: 1. T. Miyamoto, et al., Chem. Sci, 2015, 6, 7336–7342.

Chemical synthesis of homogeneous antifreeze glycoprotein having uniform GalNAc modification and its antifreeze activity Ryo Okamoto1, Ryo Orii1, Masayuki Izumi1, Daichi Fukami2, Sakae Tsuda2, Yasuhiro Kajihara1 1Department of Chemistry, Graduate School of Science, Osaka University, 2Bioproduction Research Institute, National Institute of Advanced Industrial Science and Technology (AIST) Antifreeze glycoprotein (AFGP) is a unique mucin type O-glycoprotein consists of Ala-Thr-Ala triglyco- peptide repeat (repeating number: n 5 450) where the Thr residue is glycosylated with a D-Galb1-3- D-GalNAca1. The functional analyses of the mixture of short AFGP (e.g. n < 7) have suggested that the GalNAc moiety is essential modification for the antifreeze activity of AFGP. We have been interested in the precise roll of such simple protein modification that can confer antifreeze activity on the simple polypeptide chain; However, so far the precise structural analysis at atomic level have been hampering due to the inherent heterogeneity of AFGP obtained from natural sources as well as its unique simple primary structure. In this context, we set out to establish the chemical synthesis of AFGP having uniform GalNAc modification. Chemical synthesis can provide not only highly pure protein derivative with defined chemical structure but also unnatural derivative such as selectively isotope labeled protein molecules. We rea- soned the structure-activity relationship study using such structurally-defined AFGPs would give insight into the detailed roll of the O-GalNAcylation, which is a ubiquitous posttranslational modification in eukaryote. The unique repeating structure was successfully assembled by a site selective peptide coupling reaction using glycopeptide-athioester and glycopeptidyl-N-acylguanidine that is an emerging peptide derivative. This synthetic method allowed us to obtain a range of AFGPs (n 5 5, 10, 20, 40) having uniform O-GalNAc modification. The antifreeze activity of the synthetic AFGP was evaluated by thermal hysteresis assay. This revealed that the O-GalNAcylated AFGPs had antifreeze activity. These results suggested that GalNAc modification could be generally essential modifications to exhibit the antifreeze activity of AFGP. The synthetic basis would give further opportunity for the elucidation of the antifreeze mechanism of AFGP at atomic level.

65 ABSTRACT

Statistical and combinatorial approaches to designing repeat proteins as recognition elements in microbial sensors Rachael N. Parker1, Ana Mercedes, Tijana Z. Grove1 1Dept. of Chemistry, Virginia Tech The efficiency of a biosensor relies highly on the sensitivity and selectivity of the molecular recognition event. High affinity molecular recognition elements are therefore a necessity for specific and effective biosensors, as well as diagnostic devices. We implement rational and combinatorial protein design approaches to engineer repeat-protein scaffolds as high affinity probes for whole cell pathogen sensing. Repeat- proteins are especially well-suited alternative binding scaffolds due to their modular architecture and biophysical properties. Here we present the design of an affinity probe based on the consensus sequence of the leucine-rich repeat (LRR) domain of the Nod-like receptor (NLR) family of cytoplasmic innate immune system receptors. Defined by their tripartite domain architecture, NLRs contain an N-terminal effector domain, nucleotide-binding oligomerization (NACHT) domain, and a C-terminal LRR domain, which is responsible for pathogen sensing. We have designed a consensus leucine-rich repeat protein (CLRR2) that is a stable, monomeric, and cysteine free scaffold. Without any affinity maturation, CLRR2 displays micromolar binding affinity to muramyl dipeptide, a gram negative bacterial cell wall fragment. We are using CLRR2 as a starting point for development of a synthetic module that specifically binds glycolylated muramyl dipeptide, a cell-wall fragment specific to Mycobacteria, with the goal of developing inexpensive point-of-care diagnostic devices based on direct recognition of Mycobacteria cell wall.

Computational design of novel repeat protein families with atomic level accuracy Fabio Parmeggiani1, TJ Brunette1, Po-Ssu Huang1, Damian Ekiert2, Gira Bhabha2, Susan Tsutakawa3, Greg L. Hura3, John A. Tainer3, David Baker1 1University of Washington, Seattle, WA, USA, 2UCSF, San Francisco, CA, USA, 3Lawrence Berkeley National Laboratory, Berkeley, CA, USA Computational design is a promising tool to create custom made proteins with three-dimensional structures able to support specific molecular recognition and catalysis or capable to assemble into nanoparticles and protein arrays. However, the complexity of most proteins found in Nature limits the possibility of modifying their structures. Here we present our approach for the design of new protein structures, using modular repeating units as building blocks, and we explore the range of achievable geometries. We focused on building blocks formed by two helices of variable length and we developed a computational approach to design and validate sequence using the Rosetta macromolecular modeling suite. Sampling of backbone conformations and side chain identities was performed simultaneously in all repeat units, reducing dramatically the complexity of the search space [1]. Out of 83 proteins selected for experimental characterization, 54% possessed the expected secondary structure, were stable at more than 958C and 3M guanidinium hydrochloride (some above 7M), were monomeric in solution with the structure confirmed by Small Angle X-ray Scattering. 15 crystal structures revealed the high accuracy of the designs, with root mean square deviations between 0.6 Å and 2.4 Å. The lack of sequence and structural similarity with existing repeat proteins establishes our designs as novel repeat protein families characterized by a broad range of geometries [2]. The designs described above can be used as scaffold for protein, nucleic acid, small molecule binders and enzymes, and new structures can be designed for specific applications. Additionally, the stability and modular structure of designed repeat proteins make them the ideal candidates for the production of protein fibers, arrays and nanoscale architectures. References: 1. Parmeggiani F, Huang PS, et al. J Mol Biol (2015), 427(2), 563-575 2. Brunette T, Parmeggiani F, Huang PS, et al. Nature (2015), 528, 580-584

66 ABSTRACT

Protein engineering of Caf1 from the plague bacterium Yersinia pestis for tissue engineering applications Daniel Peters1, Yakup Ulusu1,2, Helen Waller1, Jeremy Lakey1 1Institute of Cellular and Molecular Bioscience, Medical School, University of Newcastle, UK, 2Depart- ment of Bioengineering, Faculty of Engineering, Karamanoglu-Mehmetbey University, Karaman, Turkey The capsular antigen F1 (Caf1) protein of Y. pestis forms a gel-like, non-stick coat, allowing the bacteria to resist phagocytosis by macrophages. As cells cannot adhere to Caf1, new functions can be engineered in to control cell adhesion, differentiation and proliferation, through the mutation of the protein at key sites. Previously, a mutant Caf1 polymer containing an insertion mutant corresponding to the integrin binding motif (RGD) was produced, which reversed the non-stick phenotype and facilitated the adhesion of cells. Caf1 can also be made to form a hydrogel, highlighting the potential for this protein in tissue engineering applications. Building on this work, we test Caf1’s ability to retain its thermostability under different chemical conditions, and demonstrate its resistance to common proteases. We then show that several regions of the protein can be modified to contain new functional mutations such as growth factor peptides, cell adhesion motifs and protease recognition sites which allow for specific polymer cleavage. Finally, we show the engineered proteins can be combined to form mixed Caf1 polymers with multiple properties, similar to extracellular matrix proteins. The production of defined Caf1 polymers with different functionalities will greatly expand its use as a material in regenerative medicine, for example as a wound care product.

Engineering affinity reagents for chaperone enabled structural studies Katarzyna Radziwon1, Somnath Mukherjee1, Lucas J. Bailey1, Anthony A. Kossiakoff1 1Department of Biochemistry and Molecular Biology, The University of Chicago Recent advances in cryo-electron microscopy have enabled near-atomic resolution structural determination of a number of macromolecules. Despite the growing number of structures, several challenges remain. Notably, the technique is limited to larger molecules where proteins less than 200 kDa do not contain suitable features necessary for image alignment. One promising strategy is the use of affinity reagents to both significantly increase the size of macromolecules subjected to cryo-EM. Antibody fragments (Fabs) can serve as fiducial markers, where their large size (50 kDa) and defined shape aid in image alignment. Here, we report a strategy for developing a general reagent we call “Fab on Fab” that increases the size of the target protein by 100 kDa. Using phage display antibody panning technology, we developed Fabs against a complex of human IgG1 Fab and an engineered Protein-G. A number of selected Fabs require both protein components for efficient protein complex formation as indicated by immunoassays. Furthermore, engineering of the bacterial immunoglobulin binding protein, Protein-G, should enable high-affinity ternary complex formation necessary for homogenous sample preparation. Development of these powerful reagents should greatly increase the potential number of proteins whose structures can be elucidated by Cryo-EM.

Designing highly specific protein-based small molecule biosensors Srivatsan Raman1 1Department of Biochemistry, University of Wisconsin, Madison Allosteric transcription factors (aTF) act as switches that transduce small molecule binding into transcriptional actuation to regulate various cellular processes. This switch-like behavior makes aTFs a cornerstone in synthetic biology applications. In metabolic engineering, they act as in vivo sensors to report on the level of target metabolites, allowing directed evolution of production pathways by revealing rare, high-producing cells. aTFs also play a pivotal role as switches to control information flow and

67 ABSTRACT feedback regulation in synthetic gene networks. Expanding aTFs beyond naturally-occurring aTF-small molecule pairs would greatly increase their utility. For example, new chemical recognition capabilities would facilitate the evolution of new biosynthetic routes for natural products and synthetic chemicals useful as therapeutics, novel materials and fuels. Additional switches with orthogonal chemical specificity would allow the engineering of higher-order synthetic circuits that function more robustly for applications outside the laboratory. However, redesigning an aTF to bind to a new molecule is challenging. Mutations in the ligand-binding site often disrupt allosteric communication with the DNA-binding domain, destroying the switch-like behaviour. Here, we present a general strategy to engineer an aTF to respond to new inducer molecules using the E. coli LacI protein as a test case. We evaluate thousands of candidate designs, derived from computational design to identify those that are both active as a switch and responsive to a target molecule. We enhance the activity of the initial hits toward greater specificity and stronger induction. We demonstrate the utility of this approach by engineering LacI variants to respond to gentiobiose, fucose, lactitol or sucralose with response comparable or superior to the wild-type LacI response to its synthetic inducer, IPTG. This technology enables us to build highly specific intracellular biosensors for small molecules toward many applications in synthetic biology and cellular engineering. Reference: 1. Taylor ND, Garruss AS, Moretti R, Chan S, Arbing MA, Cascio D, Rogers JK, Isaacs FJ, Kosuri S, Baker D, Fields S, Church GM, Raman S (2016): Engineering an allosteric transcription factor to respond to new ligands. Nature Methods 13, 177-83.

Flipping the Switch: Engineering Alternate Function in the Lactose Repressor David H Richards1, Corey J Wilson1 1Yale University The lactose repressor (LacI) is a classical regulatory protein that represses gene expression and then reduces repression upon binding the allosteric inducer IPTG. Several recent works have been able to invert the function of LacI to generate Anti-Lacs, which allow gene expression and then increase repression after binding IPTG. In both studies, reversing function was contingent upon first blocking allosteric

68 ABSTRACT communication via a “super-repressor” (IS) mutation and then providing additional compensatory mutations through random mutagenesis. Here we investigated the generality of this method by using a different set of IS mutations to initially block allostery. From there we generated a new series of Anti-Lacs using error-prone PCR. We analyzed the function of the Anti-Lacs with classical b-galactosidase screens as well as more standardized flow cytometry experiments. Interestingly, the Anti-Lacs showcase both different mutational trends as well as different dynamic ranges of gene repression. Together, these new Anti-Lacs serve as a versatile tool for synthetic biology by allowing precise tuning of gene expression. They may also act as useful model systems for examining the molecular basis of allostery in the lactose repressor.

Controlling protein structure and function using engineered allosteric effectors Matthew Krusen1, Akiko Koide2, Shohei Koide2, Anthony A. Kossiakoff2, Kasturi Haldar1, Robert V. Stahelin1,3, and Shahir S. Rizk4 1Boler—Parseghian Center for Rare and Neglected Disease, University of Notre Dame, 2Department of Biochemistry and Molecular Biology, University of Chicago, 3Indiana University School of Medicine, South Bend, 4Department of Chemistry and Biochemistry, Indiana University South Bend Regulation of protein function is an essential feature of biological systems. One strategy employed by nature is to take advantage of the conformational diversity of protein structure, whereby allosteric molecules can influence protein function by conformational switching. Our work is aimed to mimic natural control of protein conformation using engineered antibody fragments (Fabs) that can act as allosteric effectors. These Fabs are generated by directed evolution using a high diversity phage display library of humanized antibody fragments. Unlike traditional methods of antibody generation, phage display selection allows exquisite control over the target protein conformation during the selection process. We have utilized this technique to generate Fabs that can distinguish between the ligand—bound and the apo forms of a binding protein. This conformational selectivity produces Fabs that dynamically modulate the affinity of the protein for its ligand by acting as allosteric effectors in vitro and in vivo. Addi- tionally, Fabs that recognize different conformations of the prolactin receptor can influence its hormone specificity in situ and reprogram its downstream signaling pathway. Further, Fabs have been generated to recognize different complexes of the Ebola viral protein VP40, which is able to form dimers, hexamers and octamers, each playing an important role in the viral life cycle. We show that complex—specific Fabs can modulate the equilibrium between the different VP40 complexes and can be used as a powerful tool in understanding this process and for developing therapies. Our recent work has focused on modulating enzyme function, specifically by generating Fabs that can act as allosteric activators. We show that Fabs that recognize the active form of an enzyme can enhance activity and importantly can restore function to a mutant enzyme. Our work will serve as a first step towards developing therapies for a large number of genetic disorders where mutations have disrupted protein function.

Engineering the FabA and FabZ fatty acid dehydratase domains from Escherichia coli into dimeric artificial constructs Carlos Rullan1 1University of Puerto Rico, Medical Sciences Campus Modulating the biosynthesis of microbially derived fatty acids is an attractive strategy towards generating precursors for biodiesel mixtures. Escherichia coli produces fatty acids using a set of independent, stand-alone enzymes, while several deep-sea bacteria are capable of producing higher fatty acid yields using a multidomain type I fatty acid synthase. In an effort to mimic this naturally linked protein architecture and increase fatty acid yields, we have created artificially linked gene constructs of two E. coli dehydratases that are essential for fatty acid biosynthesis, fabA and fabZ. While these enzymes function as homodimers in E. coli, the effect of covalent linkage is unknown. Using overlap PCR, we have

69 ABSTRACT generated 2 hybrid gene constructs encoding for dimers separated by a short amino acid linker: fabA/ fabA, fabZ/fabZ. Each construct has been cloned into pET200, expressed in E. coli BL21 cultures, and the soluble proteins purified by NiNTA affinity chromatography. Using size exclusion chromatography, we elucidated the oligomeric states of the enzymes in solution, where FabA/FabA forms an intramolecular dimer, and FabZ/FabZ forms a trimer of dimers. Enzymatic activity assays demonstrated that the artificially linked enzymes are active towards crotonyl-CoA. Our kinetics data shows that FabZ/FabZ is 12 times more active than the FabZ monomer. By contrast, the activity of FabA/FabA is not enhanced. Lastly, the fatty acid profiles measured by GC/MS reveal that E. coli cultures overexpressing FabZ/FabZ produce a higher proportion of saturated fatty acids: the more stable components of biodiesel preparations. In conclusion, we have generated artificially linked E. coli dehydratase genes, purified their soluble protein products, and confirmed their activity towards acyl-CoA substrates. This is the first time that these enzymes have been covalently linked or structurally modified to make them useful tools to enhance the production of fatty acids in bacterial cultures.

An Engineered High Affinity Fbs1 Carbohydrate Binding Protein for Selective Capture of N- glycans and N-glycopeptides Minyong Chen, Xiaofeng Shi, Rebecca Duke, Cristian Ruse, Nan Dai, Christopher H. Taron, James C. Samuelson1 1New England Biolabs The objective of this study was to develop a method for selective and comprehensive enrichment of N- glycopeptides to facilitate biomarker discovery. The method takes advantage of the inherent properties of Fbs1, which functions within the ubiquitin-mediated degradation system to recognize the common core pentasaccharide motif (Man3GlcNAc2) of N-linked glycoproteins. We show that Fbs1 is able to bind diverse types of N-linked glycomolecules, however, wild-type Fbs1 preferentially binds high mannose containing glycans. Using a plasmid display selection method we identified Fbs1 variants, which possess higher affinity and improved recovery of complex N-glycomolecules. In particular, we demonstrate that the Fbs1 GYR variant may be employed for substantially unbiased enrichment of N- glycopeptides. Using the N-glyco-FASP method, Fbs1 GYR was applied to enrich and analyze N- glycopeptides from human serum samples. After enrichment, 65%-70% of the peptides in the Fbs1 GYR

70 ABSTRACT enrichment sample were identified as N-glycopeptides. Furthermore, the N-glycan profiles of the enriched and pre-enrichment samples were remarkably similar. This substantially unbiased N- glycopeptide enrichment method enables the simultaneous determination of N-glycosites and N-glycan composition.

Engineering of a modular “Split-enzyme” protein G-based sandwich immunoassay for Ebola-virus Nucleoprotein detection Tomasz Slezak1, Mateusz Jaskolowski, Lucas J. Bailey, Zachary P. Schaefer, Elena K. Davydova, Anthony A. Kossiakoff 1Department of Biochemistryand Molecular Biology The University of Chicago, USA. The 2014 massive outbreak of Ebola virus disease in West Africa demonstrated an urgent need for the point-of-care diagnostics for early-stage Ebola-virus infection. To that end, we used two recently generated Fabs, MJ6 and MJ20, which simultaneously bind to non-overlapping epitopes on the C-terminal domain of Ebola Zaire Nucleoprotein (NPCT), to serve as the basis for a “Split-Enzyme” protein G-based sandwich Immunoassay (SEGIA). In order to potentiate the assay sensitivity, we grafted MJ6 and MJ20 with a new Fab Lc scaffold possessing 20-fold improvement in off-rate of binding to an engineered high-affinity variant A1 of immunoglobulin-binding protein G. This new scaffold was selected from M13 phage library randomized at aa 123-127 of Fab LC that show a close contact with Protein G “helical cap” in a crystal structure. Two protein-fragment complementation reagents for this assay were constructed by fusion of Protein G-A1 to each of two TEM-1 b-lactamase (BL) complementary fragments: BLF1(M182T) and BLF2, via a 30 aa-long Gly-Ser linker. We hypothesized that simultaneous binding of MJ6 and MJ20 in complexes with the complementary G-A1_BLF fusions to NPC would result in BL refolding and reconstitution of its activity detectable using a fluorogenic BL substrate. Indeed, we were able to detect up to 10-fold increase in fluorescent signal upon addition of the NPCT at the concentration range of 10-200 nM. The modular format of SEGIA allows the assembly of a desired diagnostic assay from the G-A1_BLF fusions and a pair of specific independently-binding Fabs instantaneously. Potentially, SEGIA can be applied for ultra-sensitive point-of-care diagnostics of wide–range of infections, disease and disorders.

Engineering Porous Protein Crystals as Scaffolds for Programmed Assembly Thaddaus R. Huber, Luke F. Hartje, Ann E. Kowalski, Lucas B. Johnson, Jacob C. Sebesta, Christopher D. Snow1 1Colorado State University A key motivation for nano-biotechnology efforts is the creation of designer materials in which the assembly acts to organize functional domains in three dimensions. Crystalline materials are ideal from the validation perspective because X-ray diffraction can elucidate the atomic structure. Relatively little work has focused on engineering protein crystals as scaffolds for nanotechnology, due to the technical challenges of coaxing typical proteins into crystallizing, and the likelihood of disrupting the crystallization process if changes are made to the monomers. We have circumvented these limitations by instal- ling guest protein domains within engineered porous crystals (13 nm pore diameter) that have been rendered robust using covalent crosslinks. The retention of the scaffold structure despite changes to the solution conditions and macromolecule uptake can be validated through X-ray diffraction. We have engineered scaffold crystals for the non-covalent and covalent capture of guest macromolecules. By controlling the reversible loading and release, we can prepare “integrated” crystals with spatially segre- gated guest loading patterns. As assessed using confocal microscopy, such host-guest crystals are highly stable. Ultimately, the resulting crystals may serve as a robust alternative to DNA assemblies for the programmed placement of macromolecules within materials.

71 ABSTRACT

Development of single chain Fv of neutralizing antibody to measles virus targeting the receptor binding site of hemagglutinin Takashi Tadokoro1, Lubna Mst Jahan1, Atsutoshi Imai1, Natsumi Sugimura1, Koki Yoshida1, Mizuki Saito1, Yuri Ito1, Surui Chen1, Takao Hashiguchi2, Yusuke Yanagi2, Maino Tahara3, Makoto Takeda3, Hideo Fukuhara1, Katsumi Maneaka1 1Faculty of Pharmaceutical Sciences, Hokkaido University, 2Department of Virology, Faculty of Medicine, Kyushu University, 3Department of Virology, National Institute of Infectious Diseases Measles Virus (MV) is a major cause of childhood morbidity and mortality worldwide although an effective vaccine is available. Most of the MV neutralizing antibodies developed at present target to hemagglutinin (H) protein of MV, a surface glycoprotein responsible for the host cell entry. We have previously prepared the neutralizing mouse monoclonal antibody (MAb) 2F4 using a cell line expressing the H protein as an antigen. The infection assays demonstrated that the 2F4 MAb shows high neutralizing titers against eight different recombinant MVs. The results suggest that the epitope of MAb 2F4 is overlapped with the binding site of the receptors although the epitopes have not been fully determined yet. The single chain variable antibody fragment (scFv) strategy is one of the most popular methods in antibody engineering because of its lower immunogenicity, and its small molecular size allowing better tissue penetration. To reveal molecular basis for the inhibition of the interaction between MV-H and cellular receptors, we have prepared the recombinant scFv from 2F4 MAb. The kinetic analysis using surface plasmon resonance indicated that the 2F4 scFv interacts with MV-H protein at lM level of KD. Further competitive analysis revealed that the scFv is able to inhibit the binding of MV-H to its receptors, SLAM and nectin-4. Fusion assay as well as virus infection assay using receptor expressing cells showed that the scFv effectively inhibits the syncytia formations and virus infections mediated by the interaction between MV-H and cellular receptors. Collectively, 2F4 scFv has sufficient inhibitory activity for the binding of MV-H and cellular receptors in a competitive manner. Our results may contribute to the development of an effective vaccine and antiviral drugs to measles.

72 ABSTRACT

Drug design against trichomoniasis Vique Sanchez JL1, Brieba Luis2, Rossana Arroyo3, Jaime Ortega4, Arturo Rojo4, Benıtez Cardoza C1. 1Laboratorio de Investigacion Bioquımica, ENMyH, Instituto Politecnico Nacional, 2LANGEBIO, CINVESTAV-Irapuato; 3Departamento de Infectomica y Patogenesis Molecular; CINVESTAV-Zacatenco, Departamento de Biotecnologıa, CINVESTAV-Zacatenco; 4UAM-Cuajimalpa. Mexico Trichomonas vaginalis is a protozoan, the causal agent of trichomoniasis, the most common non-viral sexually transmitted infection (STI) spread worldwide. Trichomoniasis is associated with perinatal compli- cations and infections in the genitourinary tract in both sexes. For over 40 years, the treatment against trichomoniasis is the provision of nitroimidazoles, commonly metronidazole and tinidazole. However, 5 to 20% of the patients show no improvement by this treatment. This highlights the need for new therapeutic regimens against trichomoniasis. Carbohydrates are the main nutrient source for T. vaginalis. Therefore, the enzymes in the glycolytic pathway on T. vaginalis like triose phosphate isomerase (TIM) are potential therapeutic targets. We performed molecular interaction simulations between a set of compounds obtained from libraries and triose phosphate isomerase from T. vaginalis. Subsequently, the compounds with higher probability of interaction were assayed in their ability to inhibit or destabilize the mentioned glycolytic enzyme. Some compounds selected by docking strategies were able to reduce the replication and viability of T. vaginalis cultures. These findings have important implications in the development of new therapeutic strategies against trichomoniasis.

Comprehensive evaluation of protein sequence-function landscapes using deep sequencing Klesmith JR1, Wrenbeck EE2, Kowalsky CA2, Whitehead TA2,3 1Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, Michi- gan, 48824;, 2Department of Chemical Engineering and Materials Science, Michigan State University, East Lansing, Michigan, 48824;, 3Department of Biosystems and Agricultural Engineering, Michigan State University, East Lansing, Michigan, 48824 Mutational studies have been used for over six decades to probe protein sequence-function relationships. Classical experiments involve site directed mutagenesis of a single residue, followed by assessment of the function of the mutant with an in vitro assay. By contrast, deep mutational scanning has emerged as a method to assess the effect of thousands of mutations on function using massively parallel functional screens and DNA counting via deep sequencing. In this talk I will first describe several technical advances developed by my lab in order to resolve sequence-function landscapes for complete protein sequences. Next, I will describe results from six deep mutational scanning experiments involving protein binders and enzymes. For protein binders I will show how to reconstruct the change of binding affinity upon mutation (interface DDG) from sequencing counts, and will demonstrate how deep mutational scanning can be used for conformational epitope mapping for antibody/antigen interactions. I will also present comprehensive single point mutant fitness landscapes for two enzymes involved in biomass conversion. In both enzymes gain of function mutations are spread throughout the protein tertiary structure, with no statistical difference in abundance between the first shell surrounding the active site, the protein core, or the protein surface. For one enzyme, an aliphatic amidase, we report on the comprehensive sequence determinants to specificity. We find that the distribution of fitness effects varies remarkably between substrates, with 20.7% of possible mutations showing enhanced fitness over wild- type. Many of the sequence determinants to function can be understood from biophysical principles of protein stability, aggregation, catalytic rate enhancement, and transition state binding stabilization. The advantageous mutations are exponentially distributed, consistent with modern theories of adaptation. The presented comprehensive fitness landscapes contribute to our fundamental understanding of how protein sequence imparts function.

73 ABSTRACT

Experimental Characterization of Computationally Predicted “Metamorphic” Proteins James O. Wrabl1, Jordan Hoffmann1,3, Mark Sowers1, and Vincent J. Hilser1,2 1Department of Biology, 2T. C. Jenkins Department of Biophysics, Johns Hopkins University, 3Paulson School of Engineering and Applied Sciences, Harvard University The emerging biological phenomenon of “metamorphic” proteins, single amino acid sequences that adopt two physiologically distinct structures and functions, challenges current prediction methods largely reliant on sequence similarity. To address this problem, we develop a novel metric for sequence structure compatibility, using energetic information derived from an experimentally validated ensemble-based description of protein thermodynamics. The compatibility metric takes both native and denatured state energetic information into account, with separate calibration of Gaussian probability models for background sequence-structure scores in each state. High-identity sequences, previously demonstrated to adopt either Streptococcus protein GA or GB folds, were correctly recapitulated, suggesting that the simple compatibility metric indeed reflected the energetic determinants of fold. To test this hypothesis, arbitrarily chosen uncharacterized members of the high-identity sequence space were expressed and purified; most were found to be consistent with their predicted folds as assessed by circular dichroism spectroscopy. Strikingly, several single Glycine substitution mutants appeared to promote a fold-switch. Because our ensemble-based framework is general and could be applied to any fold, it may be useful for the future targeted design, or large-scale proteomic detection, of novel metamorphic proteins.

Nicking Mutagenesis: A Plasmid-Based Single-Pot Saturation Mutagenesis Method Emily E. Wrenbeck1, Justin R. Klesmith2, James A. Stapleton1, Timothy A. Whitehead1,3 1Department of Chemical Engineering and Materials Science, Michigan State University, 2Department of Biochemistry and Molecular Biology, Michigan State University, 3Department of Biosystems and Agricultural Engineering, Michigan State University A core component of directed evolution and deep mutational scanning experiments is the generation of a mutational library1,2. Error prone PCR and cassette mutagenesis are the most common strategies but suffer from limitations in scalability, deficient codon sampling, and imprecise control over the number of mutations introduced. Based on these limitations, we developed Nicking Mutagenesis3,a robust and accessible method for the construction of high quality, user-defined mutational libraries. Nicking Mutagenesis is a single day, single pot method using routinely prepped plasmid dsDNA as an input substrate. This poster will describe the overall method and demonstrate the efficacy on multiple systems. Data will be presented for saturation mutagenesis at multiple positions and single-site saturation mutagenesis libraries containing all possible single codon substitutions for full-length genes. Effi- cacy of the method will be presented by assessing library coverage using deep sequencing, revealing 100% coverage of all possible single non-synonymous mutations (2840 total) and 100% of all possible programmed codon mutations (8946 total), with over 60% of the library containing exactly one mutation. Resources for portability of the method, including plasmid sharing and protocol capture will also be discussed. References: 1. Fowler, D. M. & Fields, S. Deep mutational scanning: a new style of protein science. Nat. Methods 11, 801– 807 (2014). 2. Currin, A., Swainston, N., Day, P. J. & Kell, D. B. Synthetic biology for the directed evolution of protein biocatalysts: navigating sequence space intelligently. Chem. Soc. Rev. 44, 1172–1239 (2015). 3. Wrenbeck, E. E. et al. Plasmid-based single-pot saturation mutagenesis. Nat. Methods, in revision (2016).

74 ABSTRACT

De Novo Design of Multi-nuclear Clusters in Helical Bundles Shao-Qing Zhang1, Lijun Liu1, Youzhi Tang1, Yibing Wu1, William F. DeGrado1 1Department of Pharmaceutical Chemistry, University of California at San Francisco Metalloorganic complexes contain versatile functional groups that coordinate complicated and novel metal clusters. These metal clusters carry out challenging chemical reactions, however usually in organic solvent. In comparison, natural metalloproteins have incorporated limited sets of metal clusters, even with diverse moiety of 20 amino acids. Here we focus on building unprecedented multi-nuclear clusters in artificial metalloproteins through de novo design, with the aim to pave the way for their application in synthetic biology as new enzymes. The residues Asp and His, which always reside at the active site of transition-metal enzymes, are employed to form multi-nuclear clusters at the core of four-helix bundles. The crystal structures shows that we have successfully de novo designed the FIRST protein to bind four transition metal Zn(II) ions as a cluster in a cubane-like manner, with carboxylate and imidazole as the ligating residues, in a biological scaffold. Analytical ultracentrifugation reveals well-behaved four-helix bundles in aqueous solution. NMR demonstrates interesting sequential binding phenomena of Zn(II) ions with the bundles. We have also finely modulated coordination geometry of ligands and metal ions, by tuning the interhelical distance. The crystal structures uncover subtle change in the orientation of the sidechains of the ligating residues, when the residues at helix-helix interface are modified. In summary, we have designed novel multi-nuclear clusters, previously existing only in metalloorganic complexes, in biological protein scaffolds, and we can exquisitely modulate the coordination environment of metal clusters by adjusting the global geometry of the scaffolds.

A Tertiary Alphabet for the Observable Protein Structural Universe Captures Sequence-Structure Relationships Craig O. Mackenzie1, Jianfu Zhou2, Fan Zheng3,4, Gevorg Grigoryan1,2,3,4 1Institute for Quantitative Biomedical Sciences, Dartmouth College, Hanover, NH, USA, 2Department of Computer Science, Dartmouth College, Hanover, NH, USA, 3Molecular and Cellular Biology Pro- gram, Dartmouth College, Hanover, NH, USA, 4Department of Biological Sciences, Dartmouth College, Hanover, NH, USA We systematically decompose the known protein structural universe into its basic tertiary motifs, which we call TERMs. A TERM is a compact structural fragment that captures the secondary, tertiary, and quaternary environments around a given residue. We seek the set of universal TERMs that capture all structural environments observed in the PDB, finding remarkable degeneracy. Strikingly, only 600 TERMs are sufficient to describe 50% of the PDB at sub-Angstrom resolution. Further, the PDB appears to be close to converged with respect to TERM usage. On the other hand, more rare geometries also exist and the overall structural coverage grows logarithmically with the number of TERMs. We go on to show that universal TERMs provide an effective means of describing structure-sequence relationships. We demonstrate that TERM-based statistics alone are sufficient to recapitulate close-to-native sequences given either NMR or X-ray backbones, with little difference between the two classes (28% and 24% sequence identity, respectively). Furthermore, sequence variability predicted from TERM data agrees closely with evolutionary variation (42% identity between predicted and evolutionary consensus sequences). Finally, we show that TERM statistics can discriminate between accurate and erroneous structural models, further providing information on poorly predicted regions. In fact, considering submissions from recent Critical Assessment of Structure Prediction (CASP) experiments, we find a strong correlation (R 5 0.69) between TERM-based scoring and model accuracy. Though all TERMs recur in unrelated proteins, some appear to be specialized for certain functions, such as interface formation, metal coordination, or even water binding, providing us with simple design principles towards engineering such functions. Structural biology has benefited greatly from previously observed degeneracies in structure space, particularly on secondary and super-secondary structural levels. The decomposition of the known

75 ABSTRACT

structural universe into a finite set of compact TERMs offers exciting opportunities towards better understanding, design, and prediction of protein structure.

Resurfacing Proteins Using a Structural Search Engine Jianfu Zhou1, Gevorg Grigoryan1,2 1Department of 1Computer Science, 2Biological Sciences, Dartmouth College Searching the Protein Data Bank (PDB) for backbone substructures that match an arbitrary tertiary structural motif, comprising multiple disjoint segments, is an important task in structural biology. Previous approaches to this problem have generally imposed additional simplifying restrictions to limit search times. Here, we describe a novel search approach, called MASTER, which is both fast and provably correct, returning all matches within a user-specified root-mean-square deviation cutoff. We show that MASTER is fast in practice, with queries against the PDB returning in a matter of seconds even for motifs with many disjoint segments. We go on to demonstrate that this rapid search capability enables a novel approach to protein surface design, based on mining motif matches for sequence-structure relationships. Redesigned surfaces are shown to be native-like by a number of metrics, including surface charge, desolvation penalty, secondary-structure propensity, and hydrogen bonding, despite the lack of any molecular mechanics in the design model. We provide MASTER in two forms: as a standalone program and a C11 library. With the latter, efficient structural queries can be easily incorporated into arbitrary applications. Given the broad utility of the protein motif structural search problem, we believe our open-source package can be used towards a number of applications in structural biology.

PG – DYNAMICS AND ALLOSTERY Probing the Domain Architecture and Dynamics of Caspase-6 Reveal Mechanisms for its Regulation Kevin B. Dagbay1, Nicolas Bolik-Coulon2, and Jeanne A. Hardy1 1Department of Chemistry, University of Massachusetts Amherst, 104 Lederle Graduate Research Tower, 710 North Pleasant Street, Amherst, MA 01003, USA, 2Department of Chemistry, Ecole Nor- male Superieure, 45th rue d’Ulm, Paris, France Caspases are cysteine aspartate proteases that are major players in key cellular processes, including apoptosis and inflammation. Recently caspase-6 has been implicated to be crucial in the maturation of

76 ABSTRACT

Hungtingtin protein and amyloid precursor protein leading to neurodegeneration. Thus the goal of our work is to understand the molecular basis of caspase-6 function and regulation that accounts for the diverse functional roles of caspase-6 in the cell and which should enable caspase-6-specific inhibition for the treatment of neurodegenerative diseases including Alzheimer’s and Hungtington’s. The prodomain and intersubunit linker are the most distinctive regions of caspases and thus may hold the key to the unique roles of caspase-6 relative to all other caspases. We have explored the role of the prodomain and the intersubunit linker in caspase-6 function, structural dynamics, and stability. The results of CD and intrinsic fluorescence spectroscopy together with mutagenesis studies suggest that a specific region in the prodomain is critical for caspase-6 stability where charged interactions are found to be important for the observed stability. In addition to the canonical caspase conformation, caspase-6 can also exist in a helical conformation in its 130’s region that is not observed in any caspases. Results from the hydrogen-deuterium exchange mass spectrometry (HDX-MS) revealed the unique conformational ensemble of caspase-6 and that the prodomain and linker are highly flexible, which influenced the overall global and local structural dynamics of caspase-6. The interesting implications for caspases in neurodegeneration and other related disorders amplify the need to better understand how these proteases work and how they can be controlled in search for improved therapeutic design.

Direct Measurements of the Long-Range Collective Vibrations of Photoactive Yellow Protein Yanting Deng1, Mengyang Xu1, Katherine A. Niessen1, Marius Schimidt2, Andrea G. Markelz1 1Department of Physics, University at Buffalo, SUNY, Buffalo, NY, USA, 2University of Wisconsin, Mil- waukee, WI, USA Long-range collective vibrations are thought to be crucial to protein functions. In the case of photoactive protein family, modeling suggests the intramolecular vibrations provide an efficient means of energy relaxation[1], feedback for enhancement of chromophore vibrations that promote structural transitions[2] and can assist in charge energy transfer[3]. As a paradigm of this family, photoactive yellow protein (PYP) is a cytoplasmic photocycling protein related to negative phototactic response to blue light in purple photosynthetic bacteria. PYP has a p-coumaric acid chromophore binding to the cysteine residue via a thioester bond, whose vibrations were found to overlap calculated vibrations of the protein scaffold. Using our unique technique of anisotropic terahertz microscopy(ATM)[4], we measure the intramolecular vibrations for PYP for the first time, including cycling between ground and blue

77 ABSTRACT shift (pB) states. Room temperature ATM measurements are performed in the dark and with continuous wave illumination at 488nm, resulting in a steady pB state with approximately 5% population conversion. In pB state, we find an overall decrease in the strength of resonant band in frequency range of 30-60 cm-1. Our calculated spectra using quasi-harmonic analysis indicate that our measurements are dominated by the protein vibrations, rather than the pCA chromophore, allowing us to characterize how the scaffold dynamics changes with functional states and mutations. Acknowledgement: This work is supported by NSF-DBI-1556359. References: 1. Levantino, M., et al. Nat Commun, 2015. 6. 2. Mataga, N., et al. Chem. Phys. Lett., 2002. 352(3-4): p. 220-225. 3. Fokas, A.S., et al. Photosynth. Res., 2014. 122

Heme induced allostery drives the interaction of pseudomonas aeruginosa cytoplasmic heme binding protein (phus) with heme oxygenase Daniel Deredge1, Weiliang Huang1, Colleen Hui2, Hirotoshi Matsumura2, Pierre Moenne-Locoz2, Patrick Wintrode1, Angela Wilks1 1School of Pharmacy, Department of Pharmaceutical Sciences, University of Maryland, Baltimore, MD;, 2Institute of Environmental Health, Oregon Health and Science University, Portland, OR. Integral to its essential role in the metabolism of extracellular heme, Pseudomonas Aeruginosa’s Phus binds incoming heme and transfers it to the iron-regulated heme oxygenase (HemO). Previously, heme- induced conformational changes were shown to be required to interact with HemO. Using mass spectrometry and spectroscopic techniques, we probed heme binding, characterized the heme-induced conformational changes in Phus and mapped the holo-Phus-HemO interaction. Hydrogen-Deuterium exchange (HDX-MS) of apo- and holo-PhuS reveals that heme binding results in significant rearrangements of the heme-binding pocket. Notably, C-terminal proximal helices a6/a7/a8 which are highly labile in the apo-Phus become largely protected from exchange in holo-Phus only to subsequently undergo EX1-type cooperative unfolding. Also, N-terminal a1/a2 helices undergo cooperative unfolding in apo-Phus which is significantly slowed in holo-Phus. Remarkably, mapping the interface of the holo- PhuS-HemO complex using chemical crosslinking and MALDI-TOF-MS revealed that these same regions, the a7/a8 helices and a1 helix contact HemO. Together, HDX-MS and crosslinking data is consistent with a heme-binding induced conformational rearrangement of the C-terminal domain driving the initial PhuS-HemO interaction. A binding-competent yet transfer-incompetent mutant, H212R-Phus was used to further interrogate the conformational changes of Phus in response to heme binding. Spectroscopic studies show H212R-Phus binds heme with approximately 2-fold faster kinetics than wt-Phus. In accord- ance, HDX-MS shows, in the apo form, the C-terminal binding pocket helices a6/a7/a8 of H212R-Phus are slightly but significantly more protected from exchange than wt-Phus. In holo-H212R-Phus, whereas the large conformational rearrangements in the heme binding pocket is still observable, the cooperative unfolding of the C-terminal and N-terminal helices is largely abolished, suggesting a possible role for such cooperative unfolding in heme transfer.

Deciphering the Dual Regulatory Mechanisms of Phosphorylation on Caspase–?7 Scott J. Eron, Dr. Jeanne A. Hardy 1University of Massachusetts Amherst The family of death—inducing caspases play a dominant role in the apoptotic cascade. Specifically, the executioner caspases cleave a multitude of downstream targets, resulting in a structured extinction of the cell. Due to their death potential, each caspase is regulated uniquely. Understanding individual regulation would be instrumental in harnessing the ability to control cell death. In fact, particular cancers

78 ABSTRACT

have done this, finding ways to evade the cell death signal and avoid the executioner caspases. Certain cancers manipulate caspase activity by exploiting phosphorylation as a means to keep the cell alive. Our focus centers around the effects of phosphorylation on the executioner caspase—7 by the kinase PAK2, which is upregulated in many cancers. Through a variety of biophysical and biochemical techniques we have deciphered a dual inhibitory mechanism of caspase inactivation by phosphorylation. Kinetic analysis of phosphomimetic variants pinpointed a critical phosphorylation site, which decimated caspase activity. The crystal structure of this phosphomimetic at 2.2 Å illustrates the molecular details behind protein inactivation. A clear loop rearrangement below the active site alters the loop dynamics and significantly slows substrate binding. In addition, our investigation revealed a regulatory role of an orthogonal phosphorylation site near the N—terminus. Phosphorylation at this site, though distal from the cleavage event in sequence, dramatically slows zymogen processing and thus caspase activation. Further analysis confirmed that this phosphorylation interrupts the protein—protein interaction necessary for upstream caspases to activate the executioner caspase—7. Our investigation of caspase—7 phosphorylation has elucidated a dual inhibitory role imposed by the kinase PAK2. Phosphorylation at two different sites modulates protein function by two completely different mechanisms. Nature has again proven dexterous by targeting the protein activity directly, as well as slowing processing necessary for activation. Together this data propels us towards therapeutically relevant means to manage programmed cell death.

Allostery in Trp-Dependent RNA Remodeling by TRAP Elihu C. Ihms, Ian R. Kleckner, Craig A. McElroy, Melody L. Holmquist, Aparna Unnikrishnan, Vicki Wysocki, Paul Gollnick, Mark P. Foster 1Department of Chemistry and Biochemistry, The Ohio State University, Columbus, OH The ring-forming oligomeric Bacillus trp RNA binding attenuation protein (TRAP), defines a paradigm for gene regulation by ligand-mediated alteration of the structure of non-coding RNA, and for mechanisms of both homotropic and heterotropic allostery. Undecameric (11-mer) TRAP serves as a sensor for intracellular tryptophan (Trp), which occupy its 11 identical sites, and thereby activates the protein for binding to specific RNA sequences in the 5’ untranslated regions of messenger RNAs. RNA binding by TRAP can be prevented by another homo-oligomeric protein, Anti-TRAP (AT). TRAP regulates gene expression through its participation in several linked equilibria: (1) NMR revealed that exchange between different protein conformations is linked to Trp binding; (2) stopped-flow fluorescence showed that Trp binding kinetics are affected by the occupancy of other sites in in homo-oligomeric TRAP; (3) the affinity of RNA for TRAP is coupled to the binding of Trp to TRAP, and (4) is negatively coupled to TRAP binding by AT, which is in turn (5) linked to pH via protonation-coupled oligomerization; (6) and lastly, transcription and translation of the trp operon are coupled to TRAP binding via its effect on the secondary structure of the 5’ leader. We show that by globally fitting temperature-dependent

79 ABSTRACT isothermal titration calorimetry data to mechanistic nearest-neighbor interaction models, we are able to distinguish competing statistical thermodynamics models of cooperativity. This analysis reveals positive cooperativity, and quantifies on a microscopic scale the contribution of neighboring ligands to individual binding sites. Such mechanistic analysis of ITC data has the potential to reveal the microscopic origins of cooperativity in a wide range of ligand-regulated homo-oligomeric assemblies.

Trapping Catalytic Conformations by Crosslinking the Swinging-Arm Domain of Pyruvate Carboxylase Joshua Hakala1 1Marquette University Pyruvate carboxylase (PC) catalyzes the ATP-dependent production of oxaloacetate from pyruvate and bicarbonate. This serves as an important anaplerotic reaction to replenish citric acid cycle intermediates. PC is a multi-domain enzyme with a swinging-arm domain that coordinates reactions between two remote active sites. The biotin cofactor on the biotin carrier domain (BCCP) is carboxylated through a MgATP-dependent cleavage reaction in the biotin carboxylase (BC) domain using bicarbonate as the CO2 donor. The BCCP domain then swings to the carboxyltransferase (CT) domain where the carboxyl group is transferred from the biotin cofactor to pyruvate, generating oxaloacetate. These reactions, which take place in distantly located active sites, are tightly coordinated under optimal reaction conditions. The mechanism for this coordination between active sites is not well understood. Equilibrium positioning of the BCCP domain under a variety of conditions was studied by introducing cysteine residues at a specific location in the so called “exo-binding site”, located just outside of the CT domain, and in the catalytic CT domain. These were paired with an additional cysteine residue introduced on the BCCP carrier domain. These probes permitted a detailed analysis of how the BCCP domain is positioned on the enzyme in response to various substrates and allosteric effectors. Crosslink- ing studies revealed that intermediate analog binding in the BC domain shifts the equilibrium positioning of the BCCP domain away from the exo-binding site. Conversely, the allosteric activator, acetyl CoA, shifts the equilibrium positioning towards the exo-binding site. However, the presence of both acetyl- CoA and pyruvate shifts the BCCP positioning to the CT domain. Using this methodology along with inactivation kinetics to compare relative rates of inactivation, we can assess the impact of substrates and allosteric effectors on the positioning of the BCCP domain.

Single Molecule Analysis of Allosteric Interactions in Protein Kinase A Yuxin Hao1, Jeneffer England1, Susan S. Taylor2, Rodrigo A Maillard1. 1Department of Chemistry, Georgetown University, Washington, DC, 2Department of Pharmacology, University of California, San Diego, La Jolla, CA Protein Kinase A (PKA) is composed of Catalytic (C) and Regulatory (R) subunits that assemble as an inactive holoenzyme. The allosteric activation of PKA is triggered by the cooperative binding of cAMP to two cyclic-nucleotide binding domains (CNB-A and CNB-B) located in the R subunit. The goal of this study is to establish a quantitative relationship between protein stability, conformation and the observed cooperativity between CNB domains. We used optical tweezers force spectroscopy to probe at the single molecule level the stability and conformation of the CNB domains, either as isolated structures or together as found in the R subunit. Moreover, we investigated the effect of cAMP binding on the stability of each CNB domain and on the interaction energy between them. We found that in the absence of cAMP, the CNB-A unfolds at forces higher than those observed for the CNB-B (10pN and 6pN, respectively). Interestingly, however, the CNB domains as part of the R subunit unfold at similar forces (10pN), indicating that the CNB-B is stabilized by the presence of CNB-A due to inter-domain interactions. Binding of cAMP further stabilizes the interaction between the CNB domains, as evidenced by higher unfolding forces near 16pN. The mutant R241A, a hotspot residue position for allosteric

80 ABSTRACT

interactions, reduces the folding probability and the mean unfolding force of the CNB-B, and severely disrupts the interaction between the CNB domains in the R subunit. Altogether, our experimental strategy allowed us to trace and dissect the energetic bases associated with the allosteric interactions triggered by cAMP binding and by inter-domain interactions between CNB domains. Moreover, this study enabled the identification and quantitative characterization of aberrant allosteric events associated with mutations observed in disease states.

Hemi-methylated DNA regulates DNA methylation inheritance through allosteric activation of H3 ubiquitylation by UHRF1 Joseph S. Harrison1,2, Evan M. Cornett3, Dennis Goldfarb4, Paul A. DaRosa5, Zimeng M. Li6, Feng Yan7, Bradley M. Dickson3, Angela H. Guo1, Daniel V. Cantu1, Lilia Kaustov8, Peter J. Brown8, Cheryl H. Arrowsmith8, Rachel E. Klevit5, Dorothy A. Erie9, Michael B. Major4,7, Krzysztof Krajewski1, Brian Kuhlman1,2, Brian D. Strahl1,2, Scott B. Rothbart3 1Department of Biochemistry and Biophysics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA, 2Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA, 3Center for Epigenetics, Van Andel Research Institute, Grand Rapids, Michigan 49503, USA, 4Department of Computer Science, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA, 5Department of Biochemistry, Univer- sity of Washington, Seattle, Washington 98195, USA, 6Department of Physics and Astronomy, Univer- sity of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA, 7Department of Cell Biology and Physiology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA, 8Structural Genomics Consortium, University of Toronto, Toronto, Ontario M5G 1L7, Canada, 9Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA. DNA methylation is a heritable chromatin modification essential to mammalian development that functions with histone post-translational modifications to regulate chromatin structure and gene expression

81 ABSTRACT

programs. Inheritance of DNA methylation requires UHRF1, a histone- and DNA-binding RING E3 ubiquitin ligase that facilitates recruitment of DNA methyltransferase 1 (DNMT1) to sites of newly replicated DNA through the ubiquitylation of histone H3. UHRF1 binds DNA with modest selectivity towards hemi-methylated CpGs (HeDNA); however, the contribution of HeDNA sensing to UHRF1 function remains elusive. Here, we reveal that the interaction of UHRF1 with HeDNA is required for DNA methylation but is dispensable for chromatin interaction, which is governed by reciprocal positive cooperativity between the UHRF1 histone- and DNA-binding domains. We further show that HeDNA allosterically activates UHRF1 ubiquitylation towards multiple lysines on the H3 tail adjacent to the UHRF1 histone- binding site. Collectively, our studies demonstrate the first example of a DNA-protein interaction and an epigenetic modification directly regulating E3 ubiquitin ligase activity and define a highly orchestrated epigenetic control mechanism involving modifications both to histones and DNA that facilitate UHRF1 chromatin targeting, H3 ubiquitylation, and DNA methylation inheritance.

First Structure of Full-Length Mammalian Phenylalanine Hydroxylase Reveals the Architecture of the Resting-state Tetramer Emilia C. Arturo1,2, Kushol Gupta3, Annie Heroux4, Linda Stith1, Penelope J. Cross5, Emily J. Parker5, Patrick J. Loll2, Eileen K. Jaffe1 1Fox Chase Cancer Center-Temple Health, Philadelphia PA 19111, 2Drexel University College of Medi- cine, Philadelphia PA 19102, 3Perelman School of Medicine, University of Pennsylvania, Philadelphia PA 19104, 4Brookhaven National Laboratory, Upton, NY 11973, 5University of Canterbury Christchurch 8041, New Zealand Improved understanding of the relationship among structure, dynamics, and function for phenylalanine hydroxylase (PAH) can lead to needed new therapies for phenylketonuria, the most common inborn error of amino acid metabolism. PAH is a multi-domain homo-multimeric protein whose conformation and multimerization properties respond to allosteric activation by the substrate phenylalanine (Phe); allosteric regulation is necessary to maintain Phe below neurotoxic levels. A recent model for allosteric regulation of PAH involves major domain motions and architecturally distinct PAH tetramers (Jaffe, E.K., et al., 2013, Arch Biochem Biophys, 530:73-82). We present the first X-ray crystal structure for full-length rat PAH in an auto-inhibited conformation. Chromatographic isolation of a monodisperse tetrameric PAH, in the absence of Phe, facilitated determination of the 2.9 Å crystal structure. The new structure supersedes a composite homology model that had been used extensively to rationalize phenylketonuria genotype-phenotype relationships. Small-angle X-ray scattering (SAXS) confirms that this tetramer is different from a Phe-stabilized allosterically activated PAH tetramer. The lack of structural detail for

82 ABSTRACT activated PAH remains a barrier to complete understanding of phenylketonuria genotype-phenotype relationships. Nevertheless, the use of SAXS and X-ray crystallography together to inspect PAH structure provides the first complete view of the enzyme in a tetrameric form that was not possible with prior partial crystal structures, and facilitates interpretation of a wealth of biochemical and structural data that was hitherto impossible to evaluate.

15N Direct-Detect NMR Spectroscopy Experiments to Examine Determinants of Anomalous pKa Values of Lys Residues in Hydrophobic Environments Christos M. Kougentakis1, Emily M. Grasso1, Ananya Majumdar2, Bertrand Garcıa-Moreno1 1Department of Biophysics and Biomolecular NMR Center, 2Johns Hopkins University Ionizable groups buried in the hydrophobic core of proteins play important roles in biological energy transduction, including catalysis, H1 and e- transfer, and regulation of ion homeostasis. To examine molecular determinants of buried ionizable groups in detail this lab previously generated a set of variants of staphylococcal nuclease (SNase) with Lys, Glu, Asp, and Arg at 25 different internal positions. The pKa values of most naturally occurring buried ionizable groups and of most of the buried residues in SNase are highly anomalous relative to the normal pKa values in water, shifted in the direction that favors the neutral state. In the set of internal Lys variants it has been shown that ionization of the internal Lys can lead to structural reorganization. The amplitude and timescale of reorganization varies greatly depending on the identity of the internal Lys residue. To examine in more detail the molecular determinants of anomalous pKa values, including the conformational reorganization that is coupled to the ionization of Lys side chains buried in the hydrophobic core of a protein, we have developed a set of NMR spectroscopy experiments to directly detect the Nf amine of the internal Lys side chain. pKa values determined by tracking the chemical shift of the internal Nf versus pH are in good agreement with pKa values determined through thermodynamic linkage analysis of the pH dependence of stability of Lys-containing variants. Importantly, we have found that the behavior of the internal Lys Nf resonance varies considerably for Lys residues at different internal positions. A few cases will be presented to illustrate how the combination of these 15N detect experiments to report on the Lys side chains, and 15N 1H-HSQC pH titrations that report on the backbone, will allow an unprecedented description of the conformational response of a protein to the ionization of groups buried in hydrophobic environments.

Extended Impact of Catalytic Loop Phosphorylation in Human Pin1 Brendan J. Mahoney, Meiling Zhang, and Jeffrey W. Peng 1Department of Chemistry & Biochemistry, University of Notre Dame, Notre Dame, IN 46556 Pin1 is two-domain human protein that catalyzes the cis-trans isomerization of pSer/pThr-Pro (pS/T-P) motifs in numerous cell cycle regulatory proteins. These pS/T-P motifs bind to Pin1’s peptidyl-prolyl isomerase (PPIase) domain in a catalytic pocket, between an extended catalytic loop (20 residues) and the PPIase domain core. Previous studies have shown that post-translational phosphorylation of the loop at serine-71 (S71) decreases substrate binding affinity and isomerase activity. To define the basis for these phosphorylation-induced effects, we have begun studies of a phosphomimetic mutant, S71E, using solution NMR. Our results indicate that S71E perturbs not only its host loop, but also the nearby PPIase core. The perturbations imply a local network of hydrogen bonds and salt bridges that is more extended than previously suspected, and include interactions between the binding loop and the a2/a3 turn on the PPIase core. Explicit-solvent molecular dynamics simulations and evolutionary trace analysis suggest these interactions act as conserved “latches” between the loop and PPIase core that may enhance favorable substrate binding, as they are not found in homologous PPIases that lack pS/T-P specificity. The results suggest that S71 phosphorylation perturbs a hub residue (S71) that belongs to an extended electrostatic network important for Pin1 activity.

83 ABSTRACT

A role for the heme propionates in in hemoglobins: Dictating the nature of the iron distal ligand Dillon Nye, Jaime Martinez, Matthew Preimesberger, Ananya Majumdar, and Juliette Lecomte 1Johns Hopkins University “Hexacoordinate” hemoglobins are proteins in which a pair of residues, the conserved proximal histidine and a distal histidine or lysine, form ligation bonds to the iron of the heme cofactor. Functionally relevant exogenous ligands, such as O2, are able to bind to the iron only upon dissociation of the endogenous distal residue. As a result, the affinity of these proteins for exogenous ligands is linked to the affinity of the distal residue for the heme iron. In the X-ray structure of many different hemoglobins, whether hexacoordinate or not, the distal histidine or lysine interacts with one of the heme propionate side chains. We explored the influence of the propionates on distal ligation with two hexacoordinate hemoglobins: THB1, from the green alga Chlamydomonas reinhardtii (distal lysine) and GlbN, from the cyanobacterium Synechococcus PCC 7002 (distal histidine). These well-characterized proteins were modified so as to disrupt specific protein-propionate interactions either by mutagenesis or by esterifica- tion of the propionate groups. Optical absorption and NMR spectroscopies revealed that native hexacoordination was destabilized in the variant proteins, favoring a water molecule as the distal ligand in ferric THB1, and an unexpected protein residue in GlbN. These results demonstrated that endogenous hexacoordination of the heme may be modulated by protein-propionate interactions and illustrated the flexibility of the hemoglobin scaffold. Acknowledgement: Supported by NSF MCB-1330488, NIH T32 GM-008403, and NIH T32 GM-080189

Tuning Catalytic Activity by Perturbing Amino Acid Networks in a (b/a)8 Barrel Enzyme Kathleen O’Rourke1 1Pennsylvania State University Amino acid networks describe the web of noncovalent interactions between residues spanning an enzyme. These networks may be responsible for the propagation of regulatory signals across the protein that influence conformation, binding of substrate(s), and catalysis. Tryptophan synthase (TS), the final enzyme in the tryptophan biosynthetic pathway, is a tetramer consisting of a pair of alpha and beta heterodimers arranged in a linear conformation. The alpha and beta subunits are connected by a 25Å intramolecular tunnel that channels indole, a product from the alpha reaction, to the active site in the beta subunit. In addition to this tunnel, the conformational states of these subunits are highly coordinated making TS an ideal and heavily studied model for substrate channeling and enzyme-enzyme interactions. We used nuclear magnetic resonance chemical shift covariance analysis to delineate amino acid networks in the alpha subunit, a (b/a)8 barrel enzyme. These networks were perturbed by making a variety of amino acid substitutions of surface residues correlated to the catalytic Glu49. These modifications resulted in modest changes to the catalytic rate, although they are 25 Å away from the active site. Amino acid networks are important for the function of an enzyme and may be manipulated to tune its function and enhance protein engineering.

Regulation, activation and deactivation of guanylate cyclase Olga Petrova, Isabelle Lamarre, Michel Negrerie 1Laboratory for optics and biosciences, Ecole Polytechnique, France Soluble guanylate cyclase (sGC) is an enzyme involved in signal transduction which catalyzes the formation of cGMP from GTP and is activated by the binding of NO to its heme group. While sGC is present in many mammalian cells, its homologous domain (H-NOX) in bacteria is involved in NO detection and metabolism regulation. An important objective is to find sGC activators (to cure hypertension) or sGC inhibitors (to stop tumor progression or fight bacterial biofilm formation). We studied the reactivity of

84 ABSTRACT bacterial H-NOX sensors towards NO using time-resolved spectroscopy (0.5-500 ps). The photodissocia- tion of NO from H-NOX demonstrated diverse heme coordination changes as a function of temperature. The three bacterial H-NOX reacted differently with NO, despite their homologous sequences and same structure. Two functions of H-NOX are revealed: redox sensing and NOsensing, the mechanism of NO-sensing appears to be different from that of sGC. Consecutively, we investigated the inhibition mechanism of human sGC and measured the effect of two potential inhibitors, hypericin and hypocrellin, on the activity of sGC. The inhibition constant for cGMP synthesis measured in vitro on purified sGC (hypericin: 0.3 lM; hypocrellin: 0.6 lM) was found to be 2.5 time lower than measured in vivo on HUVEC (hypericin: 0.8 lM; hypocrellin: 1.6 lM). In vitro, the presence of the NO-independent activator BAY41-2272 of sGC didn’t change the inhibition induced by hypericin. In vivo, BAY41-2272 activated sGC in the absence of NO, and reduced the inhibition of sGC by the hypericin and hypocrellin. Finally, using surface plasmon resonance we have shown irreversible binding of hypericin to sGC. These experiments demonstrate the complex allosteric regulation of NO-sensors through both molecular dynamics and interactions with pharmacological compounds.

Subunit exchange and activation of human CaMKII variants Ana Pamela Torres Ocampo, Brendan Page, Margaret M. Stratton 1Department of Biochemistry and Molecular Biology, University of Massachusetts, Amherst, MA, USA Ca21-calmodulin dependent protein kinase II (CaMKII) assembles into a dodecameric or tetradecameric ring in which the kinase domains are organized around a central hub. Notably, the stimulated activity of CaMKII persists even after the withdrawal of a calcium stimulus. CaMKII acquires this Ca21- independent activity at a threshold frequency and this property is implicated in long-term potentiation (LTP). Indeed, transgenic mice expressing mutant versions of CaMKII have limited LTP and defects in learning and memory. We have previously shown that CaMKII has a remarkable property, which is that activation of CaMKII triggers the exchange of subunits between holoenzymes, including inactive ones, enabling the Ca21-independent activation of neighboring subunits. Our results have implications for an earlier idea that subunit exchange in CaMKII may have relevance for long-term memory formation. These studies were done using primarily human CaMKIIa, isoform 2. There are four human CaMKII genes, CaMKIIa and b are found in the brain, CaMKIId is in the heart, and CaMKIIg is found throughout the body. Each of these genes has several splice variants encoding 20 different isoforms in total. The primary difference between these isoforms is in the composition and length of the variable linker domain that connects the kinase to the hub. Previous studies have shown that the length of this linker determines the threshold frequency for activation. A comprehensive biochemical study of existing human CaMKII isoforms has not been completed. We have expanded our study of frequency activation and subunit exchange to the remaining isoforms of CaMKII in order to investigate whether these properties are ubiquitous and why specific isoforms are selectively expressed in different cell types. Our new data show that as you lengthen the variable linker domain, less CaM is needed for activation. However, above a certain linker length, there is no added effect.

Localizing Stimulator Binding to the b1 H-NOX Domain of Soluble Guanylate Cyclase Jessica A. Wales, Cheng-Yu Chen, Linda Breci, Andrzej Weichsel, Sylvie G. Bernier, James E. Sheppeck II, Joon Jung, Paul A. Renhowe, William R. Montfort 1Department of Chemistry and Biochemistry, University of Arizona, Tucson, AZ, 8572, USA ‡Ironwood Pharmaceuticals, Boston, MA, 02142 Nitric oxide (NO) signaling regulates numerous physiological processes, including vascular function, wound healing, and neurotransmission. Central to NO signaling is soluble guanylate cyclase (sGC), the primary NO receptor. Upon binding NO, sGC catalyzes the conversion of GTP to cGMP, thereby

85 ABSTRACT initiating the subsequent signaling cascade. Dysregulation of NO/sGC/cGMP signaling is implicated in numerous forms of vascular pathology, making sGC a highly sought after therapeutic target. Stimulators are a family of compounds that allosterically enhance sGC catalytic activity both independently and syn- ergistically with NO. Stimulators are a promising step forward for the treatment of vascular pharmacology, with the most advanced compound recently gaining FDA approval to treat pulmonary hypertension. However, where these compounds bind and their mechanism of action remains unknown. We localized drug binding by photoaffinity labelling sGC with a diazirine-containing stimulator known as IWP-854. IWP-854 selectively crosslinked to the b1 subunit of sGC, and modified residues were identified by high resolution mass spectrometry. Residues labelled by IWP-854 converge around a single location at the base of the b1 H-NOX domain. Photoaffinity labelling of a bacterial b1 H-NOX homolog confirmed that the b1 H-NOX domain is sufficient for stimulator binding. Characterization of candidate binding sites within the b1 H-NOX domain with respect to drug binding and protein allostery is underway. Understanding the mechanisms that underlie stimulator binding and catalytic enhancement of sGC will substantially aid in drug development, as well as provide novel insights into the allosteric regulation of this key therapeutic target.

Internal motion and conformational entropy in protein function A. Joshua Wand 1Department of Biochemistry & Biophysics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104 At a fundamental level, biological processes are most often controlled using molecular recognition by proteins. Protein-ligand interactions impact critical events ranging from the catalytic action of enzymes, the assembly of macromolecular structures, complex signaling and allostery, transport phenomena, force generation and so on. The physical origin of high affinity interactions involving proteins continues to be the subject of intense investigation. Conformational entropy represents perhaps the last piece of the thermodynamic puzzle that governs protein structure, stability, dynamics and function. The presence and importance of internal conformational entropy in proteins has been debated for decades but has resisted experimental quantification. Over the past few years we have introduced, developed and validated an NMR-based approach that uses a dynamical proxy to determine changes in conformational entropy. This new approach, which we term the NMR “entropy meter,” requires few assumptions, is empirically calibrated and is apparently robust and universal. Using this “entropy meter,” it can now be quantitatively shown that proteins retain considerable conformational entropy in their native functional states and that this conformational entropy can play a decisive role in the thermodynamics of molecular recognition by proteins. Recent results show that changes conformational entropy of a protein upon binding a high affinity ligand is highly system specific and can vary from strongly inhibiting to even strongly promoting binding and everything in between. Thus one cannot possibly understand comprehensively how proteins work without knowledge of the breadth and underlying principles of the role of conformational entropy in protein function. Supported by the NIH and the Mathers foundation.

Reciprocal allosteric communication in E.coli BirA Jingheng Wang, Dorothy Beckett 1Department of Chemistry and Biochemistry, University of Maryland, College Park Allostery regulates many cellular functions including gene expression, signal transduction and enzyme catalysis. However, the mechanism of allosteric communication remains unclear. The E. coli biotin protein repressor (BirA) is an allosteric protein to which binding of the effector molecule, bio- 5’-AMP, enhances homodimerization by 24.0 6 0.3kcal/mol. The effector binding and dimerization surfaces are separated by approximately 30 Å. Previous studies have demonstrated unidirectional allosteric

86 ABSTRACT communication from the bio-5’-AMP binding surface to the dimerization surfaces, as well as from the dimerization surface to the bio-5’-AMP binding surface. In order to provide direct evidence for reciprocal communication between the two functional surfaces, double-mutant cycle analysis with combined single alanine substitutions on each surface was employed. Functional analysis by isothermal titration calorimetry (ITC) and sedimentation equilibrium demonstrates reciprocal coupling between residues on the two BirA surfaces.

Understanding the role of distal residues in the activity of ornithine transcarbamoylase using small angle x-ray solution scattering Jenifer N. Winters, Lisa Ngu, Lee Makowski, Penny J. Beuning, Mary Jo Ondrechen Partial Order Optimum Likelihood (POOL) is a machine learning method that predicts residues that are important for catalytic activity based on the protein tertiary structure and computed electrostatic properties. For many enzymes, POOL has been able to predict spatially extended active sites, where residues that are not in direct contact with the substrate still contribute to catalysis. An example of such an enzyme is ornithine transcarbamoylase (OTC). OTC is an enzyme that is involved in the urea cycle and arginine biosynthesis pathway. Using single-site directed mutagenesis, E.coli OTC variants were constructed, expressed, and purified based on these predictions. Kinetics assays have shown that these distal variant positions predicted by POOL contribute to catalysis whereas negative controls - variants with mutations at distal positions not predicted by POOL - do not contribute to catalysis. This work uses small angle x-ray solution scattering (SAXS) to understand if the predicted distal residues play a role in dynamical structural changes in the protein, thus influencing catalysis from a distance. Three-dimensional reconstructions of solution scattering data for wild-type OTC and variants were generated using x-ray solution scattering programs GNOM and GASBOR. Variants with mutations in distal positions suggest a structural rearrangement based on these SAXS reconstructions. In addition to electrostatic effects, these distal residues may a play a key role in modulating dynamics and thus eludidate the catalytic mechanism of OTC. Acknowledgement: Supported by NSF MCB-1517290.

The Role of Dynamical Transition in Protein Function: Coupling of Protein Collective Vibrations and Water Dynamics Mengyang Xu1, Katherine Niessen1, Yanting Deng1, Nigel Michki1, Edward Snell2, Andrea Markelz1 1Department of Physics, University at Buffalo, SUNY, Buffalo, NY, USA, 2Hauptman-Woodward Medical Research Institute & Department of Structural Biology, University at Buffalo, SUNY, Buffalo, NY, USA Computational simulations have revealed protein collective vibrations prompt structural rearrangements to accomplish biological function. However, the biological importance of collective vibrations has not been experimentally demonstrated. The attempts have been hampered by the inability to distinguish localized water or side-chain relaxational motions from protein long-range vibrations using conventional techniques. The dynamical transition (DT), extensively observed using X-ray, neutron scattering, NMR and terahertz techniques [1,2], describes a rapid increase in the temperature-dependent dynamics of critically hydrated proteins above 220 K, and has been attributed to thermally activated solvent motions. While some proteins lose function below the specific temperature, others do not. We suggest the difference arises from the nature of the required motions for function. Specifically, functional motions enabled by long-range vibrations will be vulnerable to DT, which require surrounding solvent to be sufficiently mobile. We explored the coupling of protein vibrations to solvent dynamics by applying a recently developed technique, anisotropy terahertz microscopy [3], to directly measure the collective vibrations for lysozyme and investigate the temperature dependence in 150-300 K range. We find long-range intramolecular vibrations occur at 220K and rapidly increase in strength with increasing temperature,

87 ABSTRACT

consistent with enhanced access above the DT. The results suggest collective vibrations are slaved to DT, and those proteins with function reliant on these motions will cease function below DT. Acknowledgement: This work was supported by NSF (DBI-1556359). References: 1. Doster,W., et al. Phys.Rev.Lett., 2010.104(9):098101. 2. Niessen,K., et al. Biophys.Rev., 2015.7,201. 3. Acbas,G., et al. Nat.Commun., 2014.5,3076.

Effector-linked high-frequency thermal fluctuations of hemoglobin regulate its O2-affinity, cooperativity, and Bohr effects Takashi Yonetani1, Kenji Kanaori2 1Univ. Pennsylvania, Philadelphia, 2Kyoto Inst. Tech, Kyoto, The O2-affinitiy of hemoglobin (Hb) is substantially higher than that of free protoheme complex with an axial N-base (P50 > 103 mmHg). In Hb, globin interferes with the dissociation process of O2 from heme Fe to solvent by physically blocking with protein matrix (“Caged” state), forcing O2 geminately to rebind back to heme Fe and forming H-bonds between O2 and distal His to increase [R(oxy)-Hb], resulted in substantially increased apparent O2-affinity (P50 5 2 mmHg) in stripped Hb. This O2-affinity of stripped Hb is too strong to be useful as a physiological O2-transporter. Binding of heterotropic effectors to Hb reduces its O2-affinity to more physiologically appropriate levels. Binding of the effectors enhances high- frequency thermal fluctuations of Hb, leading to breaking the distal H-bonds and increasing the transpar- ency of globin matrix toward intra-globin migrations of diatomic ligands such as O2, CO, and NO and concomitantly reducing/eliminating the geminate rebinding process, resulted in increases in {[T(deoxy)- Hb] 1 [free O2]} and decreases in [R(oxy)-Hb] or a lower O2-affinity. pH-dependent differential bindings of the effectors to T(deoxy)- and R(oxy)-Hb cause changes the cooperativity and Bohr effect.

Studying protein–protein binding through T-jump induced dissociation: Transient 2D IR spectroscopy of insulin dimer Xin-Xing Zhang, Kevin C. Jones, Andrei Tokmakoff Insulin homodimer associates through the coupled folding and binding of two partially disordered monomers. We aim to understand this dynamics by instead observing insulin dimer dissociation initiated with a nanosecond temperature-jump using transient two-dimensional infrared spectroscopy (2D IR) of amide I vibrations. With the help of equilibrium FTIR and 2D IR spectra, and through a systematic study of the dependence of dissociation kinetics on temperature and insulin concentration, we are able to decompose and analyze the spectral evolution associated with different secondary structures. We find that the dissociation under all conditions is characterized by two processes whose influence on

88 ABSTRACT the kinetics varies with temperature: the unfolding of the b sheet at the dimer interface observed as exponential kinetics between 250-1000 ls, and non-exponential kinetics between 5-150 ls that we attribute to monomer disordering. Microscopic reversibility arguments lead us to conclude that dimer association requires significant conformational changes within the monomer in concert with the folding of the interfacial b sheet. While our data indicates a more complex kinetics, we apply a two-state model to the b-sheet unfolding kinetics to extract thermodynamic parameters and kinetic rate constants. The association rate constant, ka (238C) 5 8.8 3 105 M-1s-1 (pH 0, 20% EtOD), is approximately three orders of magnitude slower than the calculated diffusion limited association rate, which is explained by the significant destabilizing effect of ethanol on the dimer state and the highly positive charge of the monomers at this pH.

PH - ENZYMOLOGY Promiscuous hydrolysis of p-nitrophenyl sulfate by Vibrio alkaline phosphatase coincides with loss of active-site magnesium ions and loss of phosphatase activity. Bjarni Asgeirsson1, Jens G. Hjorleifsson€ 1, Tinna Palmadottir 1, Sandeep Chakraborty2 1Science Institute, Department of Biochemistry, University of Iceland, Dunhaga 3, 107 Reykjavik, 2Department of Plant Sciences, UC-Davis, CA, USA Although modern enzymes are highly specific, many of them have promiscuous activity which is usually several magnitudes lower than the native activity (1, 2, 3, 4). The enzymes in the alkaline phosphatase family offer some of the best examples of such catalytic promiscuity, i.e. the ability of specialized enzymes to catalyse other reactions that they are not optimized to facilitate. We were curious to see how a cold-active Vibrio AP (VAP) would compare with E. coli (ECAP) as an aryl sulfatase (2). Initial results showed no, or very low, sulfatase activity with (VAP) compared with ECAP when assayed in buffer containing 500 mM NaCl. However, without the added salt, aryl sulfatase activity appeared after 24 hours at room temperature with kcat/Km 3-5 times higher than that of ECAP and a concomitant fall in phosphatase activity. Metal ion analysis (Zn/Mg) suggested that Mg ions were gradually leaving VAP and played a part in determining this selective action of VAP. We also made a variant of VAP in order to simulate the electrostatic potential in the active site of ECAP based on computational calculations. The quadruple VAP variant, T112A/R113E/H116D/W274K, had a similar sulfatase activity to ECAP. References: 1. Jensen, R. A. (1976) Ann Rev Microbiol 30, 409-425 2. O’Brien, P. J., and Herschlag, D. (1999) Chem & Biology 6, R91-R105 3. Khersonsky, O., and Tawfik, D. S. (2010 Ann Rev Biochem 79, 471-505 4. Pabis, A. and S. C. L. Kamerlin (2016). Curr Opin Structural Biol 37: 14-21.

In crystallo Phosphorylation of Tobramycin by the Antibiotic Kinase APH(2’’)-Ia Shane J. Caldwell, Albert M. Berghuis 1McGill University, Department of Biochemistry- Montreal, Canada Phosphorylation of aminoglycoside antibiotics by the kinase enzyme APH(2’’)-Ia confers resistance by blocking the compounds from interacting with their site of action on the bacterial ribosome. Recent studies have determined an activation mechanism of the enzyme induced by binding of the aminoglycoside substrate. This activation mechanism involves the closure of an important loop, the gly-loop, over the nucleoside triphosphate substrate to facilitate phosphotransfer. We introduced the aminoglycoside tobramycin to a crystal of APH(2’’)-Ia grown with the phosphotransfer-resistant analogue GMPPNP. In this crystal, productive phosphotransfer has occurred, indicating that the enzyme is active, and capable of transferring phosphate from a relatively inactive GTP analogue to the antibiotic substrate. This structure validates mechanistic inferences about the activity of this enzyme, confirms the critical role of the gly-loop in this kinase enzyme, and provides new insights to help understand the

89 ABSTRACT

mechanism of this enzyme and to help direct anti-resistance strategies toward this enzymatic resistance factor.

Mechanism of a cytosolic O-glycosyltransferase essential for the synthesis of a bacterial adhesion protein Yu Chen1, Ravin Seepersaud2, Barbara A. Bensing2, Paul M. Sullam2, and Tom A. Rapoport1 1Howard Hughes Medical Institute and Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston, MA 02115, USA, 2San Francisco Veteran Affairs Medical Center and Uni- versity of California, San Francisco, CA 94121, USA. Protein O-glycosylation is an important process in all organisms. Substrates are often modified at multiple Ser/Thr residues, but how a glycosyltransferase can act on a continuously changing substrate is unknown. In many species of Grampositive bacteria, such modification is essential for the biogenesis of

90 ABSTRACT adhesin proteins, which mediate bacterial attachment to the eukaryotic host cell and play major roles in bacterial pathogenicity. Here, we have analyzed the mechanism by which the cytosolic O- glycosyltransferase GtfA/B of Streptococcus gordonii modifies the Ser/Thr-rich repeats of adhesin. Crys- tal structures indicated that the enzyme is a tetramer containing two molecules each of GtfA and GtfB. The two subunits have the same glycosyltransferase fold, but only GtfA contains an active site. Both subunits are required for performing glycosylation reaction. Biochemical studies suggested that GtfB provides the primary binding site for adhesin, in both un-glycosylated and glycosylated states. During a first phase of glycosylation, the conformation of GtfB is restrained by GtfA to bind substrate with unmodified Ser/Thr residues. In a slow second phase, GtfB recognizes residues that are already modified with N-acetylglucosamine, likely by converting into a relaxed conformation in which one interface with GtfA is broken. The restoration of this interface facilitates the release of glycosylated adhesin. These results provide mechanistic insight into bacterial adhesin biogenesis, and explain how glycosyltransferase.

IpdAB, a key virulence determinant in Mycobacterium tuberculosis, is a cholesterol ring-opening hydrolase Adam M. Crowe, Nobu Watanabe, Isra€el Casabon, Kirstin Brown, Jason Rogalski, Timothy Hurst, Victor Snieckus, Leonard Foster, Natalie C. Strynadka, Lindsay D. Eltis. 1The University of British Columbia, Vancouver, Canada IpdAB, a predicted type I CoA transferase (CoT), is a key virulence factor in Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis, and a promising therapeutic target. We have recently demonstrated that the enzyme plays a role in degrading the final two rings of cholesterol, a key nutrient for Mtb during infection. An DipdAB mutant of Mtb accumulated the metabolite 2-(2-carboxyethyl)23- methyl-6-oxocyclohex-1-ene-1-carboxyl-CoA (COCHEA-CoA) when incubated with cholesterol. In vitro incubation of COCHEA-CoA with IpdAB, a CoA thiolase, FadA6, and CoASH yielded methyl-5-oxo- octanedioyl-CoA (MOODA-CoA) and acetyl-CoA, suggesting IpdAB catalyzes the final steroid ring- opening reaction. Despite IpdAB’s homology with type I CoTs, the enzyme showed no detectable CoA transferase activity towards small CoA thioesters. The crystal structure of the IpdAB heterotetramer was solved to 1.8 Å. A structure-based alignment of IpdAB with glutaconate CoT (PDB:1POI, Ca rmsd52.1 Å) indicates that IpdAB lacks the catalytic glutamate residue that is conserved in known CoA transferases. Moreover, differences in the orientation of the active site loop harbouring this glutamate in CoT are consistent with IpdAB not being a CoA Transferase. Lastly, the IpdAB-mediated incorporation of deuterium from D2O into COCHEA-CoA is inconsistent with the ping-pong reaction mechanism typical of CoA transferases. Overall, we propose that IpdAB is not a CoT but catalyzes a retro-Claisen opening of cholesterol ring C. Acknowledgement: This research was funded in part by a CIHR grant to LDE

Kinetic characterization of Trypanosoma cruzi His10-b-hydroxybutyrate dehydrogenase (bHBDH) and functional exploration of Trypanosoma brucei bHBDH via RNA interference. William Escboar-Arrillaga1, Linh Nguyen1, Jennifer Palenchar1 1Villanova University, USA Trypanosomes are single-celled eukaryotic parasites that are the causative agents of Human African Try- panosomiasis (Trypanosoma brucei), Chagas Disease (Trypanosoma cruzi), and Leishmaniasis (Leishma- nia). The parasites are replete with unique biochemistry, including a b-hydroxybutyrate dehydrogenase (bHBDH) present only in the Trypanosoma species. The physiological role of Trypanosoma bHBDH, which has distinct cofactor preference among bHBDHs, is unknown. The in vitro characterization of

91 ABSTRACT

Trypansoma cruzi His10-bHBDH will be presented, including its unique ability to utilize NADP(H). Site directed mutagenesis was utilized to generate a parasite bHBDH in which cofactor specificity was altered to more closely resemble that of the NAD(H)-dependent bacterial homologs. Finally, the physiological role of the enzyme in Trypanosoma brucei is being explored through a bHBDH RNA interference cell line, the results of which will be presented.

Rapid detection of single-stranded DNA-specific 3’ exonucleases in human serum Simin Fang, Meiping Zhao 1College of Chemistry and Molecular Engineering, Peking University, Beijing100871, China Accurate repair of genomic DNA is vital to maintain genomic integrity. Many different repair enzymes cooperate to achieve this goal, among which 3’-5’ exonucleases that specifically degrade single stranded DNA (ssDNA) play important roles. They can interact with certain polymerase to enhance ligation efficiency and increase replication accuracy in the BER pathway [1]. Misfunction of these enzymes could lead to hereditary and sporadic human diseases such as cancer and accelerated aging, implying the high potential of exonucleases in disease diagnosis. At present, there are only a few available approaches for detecting ssDNA-specific 3’ exonuclease (3’-ssExos), but they are subject to interference from various nontarget proteins in real samples [2]. Therefore, it is meaningful to develop a simple and rapid method for direct and accurate detection of 3’-ssExos in biological samples without the need of complex sample pretreatment. To solve this problem, we first systematically studied the interactions among 3’-ssExos, DNA substrates and coexisting proteins. Taking advantages of the novel interactions that we found in our experiments, we successfully developed an ideal fluorescent assay for rapid detection of 3’-ssExos with a liner working range from 0.1-2 U/mL and a detection limit of 0.1 U/mL. The method can be directly applied to detect the target enzymes in human serum samples with good recovery and reproducibility. The newly developed method provides a useful tool for high-throughput measurement of the levels of 3’-ssExos in a large number of biological samples. It is simple, low-cost and easy to use, which may hold great potential for disease diagnostics. References: 1. M.Hoss, et al., EMBO J.,1999,18,3868–3875 2. C.Song, et al., J.Mater.Chem.B.,2014,2,1549-1556 Acknowledgement: The work was financially supported by the NSFC (21375004).

Active Site Binding is not Sufficient for Reductive Deiodination by Iodotyrosine Deiodinase Nattha Ingavat1, Jennifer M. Kavran2, Zuodong Sun1, Steven E. Rokita1 1Department of Chemistry, Johns Hopkins University, Baltimore, MD (USA), 2Department of Biochem- istry and Molecular Biology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD (USA) Iodotyrosine deiodinase (IYD) is a flavoprotein that is necessary for iodide homeostasis in mammals. It generates free iodide from 3-iodotyrosine (I-Tyr) and 3,5-diiodotyrosine (I2-Tyr), the byproducts of thy- roid hormone biosynthesis. How this enzyme may avoid deiodinating thyroxine and its precursors was not initially apparent. We now compare the nature of I-Tyr and 2-iodophenol (2IP) with IYD to learn the basis for substrate selectivity. Our studies began with human IYD for which binding of 2IP (Kd 5 1,340 mM) is 13,400-fold weaker than I-Tyr (Kd 5 0.1 mM). Binding may determine specificity for human IYD but not for IYD from bacterium (Haliscomenobacter hydrossis), hhIYD that also exhibits a high selectivity for dehalogenation of I-Tyr over 2IP. Low turnover of 2IP was observed even though its affinity for hhIYD (Kd 5 67 mM) is only eight-fold less than that of I-Tyr (Kd 5 8.8 mM). Dehalogenation of 2IP by

92 ABSTRACT hhIYD (kcat/Km 5 4.4 x 10-5 mM-1 min-1) is 105-fold less efficient than that of I-Tyr (kcat/Km 5 1.2 mM- 1 min-1). Thus, binding affinity is not the sole determinant of specificity. Crystallographic data indicates 2IP is unable to stabilize the active site lid that is formed in the presence of I-Tyr. Furthermore, association of 2IP to hhIYD does not share the ability of halotyrosines to stabilize a FMN semiquinone intermediate that is associated with effective catalysis. Hence, IYD appears to use a substrate-induced conformational change to prevent consumption of substrates lacking the amino acid zwitterion such as thyroxinyl and iodotyrosyl residues within thyroglobulin.

Calpain/calpastatin proteolytic system as a target of geomagnetic field disturbances Kantserova N.P.1, Krylov V.V.2, Lysenko L.A.1, Tushina E.D.3, Nemova N.N.1 1Institute of Biology, Karelian Research Centre of Russian Academy of Sciences, 2I.D. Papanin Institute for Biology of Inland Waters of Russian Academy of Sciences, 3Petrozavodsk State University Calcium-dependent proteases, or calpains, and their proteinaceous inhibitor, calpastatin, function in all eukaryotic cells; both composition and regulation of calpain/calpastatin system are similar in all verte- brates. Calpain upregulation threatens cell to excessive hydrolysis of functional and structural proteins resulting in histopathologies, such as muscular dystrophy, cancer, or neurodegeneration. Because of absolute calcium dependency calpain/calpastatine system is a suitable experimental model to study biological effects of magnetic fields, both artificial and natural. We have found that low-frequency magnetic fields led to significant decrease in calpain activity in model organisms. Geomagnetic disturbances, or geomagnetic storm, constitute of the phase of slow changes of the geomagnetic field (frequencies up to 0.001 Hz) and various geomagnetic pulsations (0.001–5 Hz). The effect of the slow changes of geomagnetic fields on calpain proteolytic activity occurs more significant as compared with geomagnetic pulsations indicating biologically effective constituent of a storm. Our findings contribute to the understanding on calpain physiology and promote a search of non-pharmaceutical therapeutic approaches targeting calpain activity. Acknowledgement: This work was carried out using IB KarRC RAS Sharing Equipment Centre facilities and financially supported by budget donation no. 0221-2014-0003 and President Program grant no. 4737.2016.4.

Analysis of a polyextremophilic? -galactosidase from an Antarctic Haloarchaeon: Mutagenic analysis of residues important for cold activity Victoria Laye, Priya DasSarma, Wolf Pecher, Shiladitya DasSarma 1Department of Microbiology & Immunology, and Institute of Marine and Environmental Technology, University of Maryland, Baltimore, MD, U.S.A. A combination of bioinformatic and experimental approaches is being used to study a salt and cold- active family 42 b-galactosidase enzyme from Halorubrum lacusprofundi, an Antarctic haloarchaeal microorganism. The enzyme was compared bioinformatically to 12 related family 42 enzymes to identify residues likely important for cold-activity. Three putative cold-adapted residues were targeted for site- directed mutagenesis and over-expression in the related Haloarchaeon Halobacterium sp. NRC-1: serine-189 was mutated to aspartic acid (S189D) and phenylalanine-387 (F387L) and valine-434 (V434L) were mutated to leucine. The pDRK42 expression vector employed also added a His-6 tag to the N- terminus. The wild-type and mutated enzymes were purified using nickel affinity chromatography and characterized by Michaelis-Menten kinetics at various temperatures with the chromogenic substrate, o- nitrophenyl-b-galactoside. The three mutated enzymes showed relatively higher KM’s than the wild- type enzyme, consistent with the predicted residues playing key roles in cold-adaptation. Our results confirm the importance of alterations in amino acid size, polarity, and/or charge for improved function of the b-galactosidase enzyme at colder temperatures.

93 ABSTRACT

Characterization of PTP1B function and inhibition James M. Lipchock1, Bonnie Douglas1, Patrick Ginther1, Kelly Bird1, J. Patrick Loria2 1Washington College, Department of Chemistry, 2Yale University, Department of Chemistry and Department of Molecular Biophysics and Biochemistry Protein tyrosine phosphatase 1B (PTP1B) is a known regulator of the insulin and leptin signaling pathways and is an active target for the treatment of type II diabetes and obesity. Recently, cichoric acid (CHA) and chlorogenic acid (CGA) were predicted by molecular docking studies to be allosteric inhibitors that bind distal to the enzyme active site. Given the importance of PTP1B inhibition as a therapeutic target, we have sought to investigate the mechanism by which these molecules allosterically modulate PTP1B activity through a combination of enzyme kinetics, NMR spectroscopy and molecular dynamics simulations. Contrary to expectations, we determined that CHA is a competitive inhibitor of PTP1B. This result was confirmed by titration of 15N-PTP1B with CHA. The resulting CHA-bound structure obtained from ligand docking with the HADDOCK2.2 web portal is consistent with competitive inhibition and suggests the central carboxyl moiety of CHA mimics the phosphoryl group of the natural substrate. Inhibition kinetics confirmed the role of CGA as a non-competitive inhibitor, but analysis of chemical shift perturbations upon titration with CGA does not support binding in the predicted benzbromarone binding pocket. Rather, shifts are localized in a binding pocket adjacent to the active site. This binding site could prove beneficial for the design of new allosteric inhibitors of PTP1B with increased selectively and cell permeability. Given that the rate of conversion for the acid-loop of PTP1B between closed and open conformations was shown previously to match the rate of hydrolysis, we are currently investigating whether CGA inhibits PTP1B by altering protein motions using a combination of CPMG dispersion analysis and molecular dynamics simulations and are exploring which residues play an important role in dictating the rate of catalysis.

Molecular basis of substrate recognition and catalysis in Phosphatidylinositol 3-kinase Sweta Maheshwari, Michelle M. Miller, Robert O’Meally, Robert Cole, Kenneth W. Kinzler, L. Mario Amzel, Bert Vogelstein, Sandra B. Gabelli 1Department of Biophysics and Biophysical Chemistry, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA Phosphatidylinositol 3-kinases (PI3Ks) play a central role in the PI3K/AKT/mTOR pathway involved in receptor mediated signaling in the cell. PI3Ks have dual specificity with a lipid kinase activity that phos- phorylates the 3’ hydroxyl of phosphoinositides, and an intrinsic protein kinase activity. In this study, we have explored the roles of catalytic domain residues in lipid and protein kinase activities of PI3Ka.We have used site-directed mutagenesis and kinetic assays to precisely map the molecular basis of substrate recognition and catalysis in PI3Ka. Kinetic analysis of P-loop residues reveals the involvement of this loop in the recognition of both the lipid substrate and ATP. Mutations of residues in the activation and catalytic loops results in a dramatic change in the kinetics of PI3Ka. While the activation loop mutant inactivates the lipid kinase function without affecting auto-phosphorylation; the catalytic loop mutant abolishes both the lipid and protein kinase activities of PI3Ka. Further, on the basis of structural analysis and modeling, we propose the possible mechanistic role of these critical residues in catalysis.

Biochemical characterization of two haloalkane dehalogenases: DccA from Caulobacter Emily Mundorff1 1Hofstra University, USA Two putative haloalkane dehalogenases (HLDs) of the HLD-I subfamily, DccA from Caulobacter cres- centus and DsaA from Saccharomonospora azurea, have been identified based on sequence compari- sons with functionally characterized HLD enzymes. The two genes were synthesized, functionally

94 ABSTRACT expressed in E. coli and shown to have activity toward a panel of haloalkane substrates. DsaA has a moderate activity level and a preference for long (greater than 3 carbons) brominated substrates, but little activity toward chlorinated alkanes. DccA shows high activity with both long brominated and chlorinated alkanes. The structure of DccA was determined by X-ray crystallography and was refined to 1.5 Å resolution. The enzyme has a large and open binding pocket with two well-defined access tunnels. A structural alignment of HLD-I subfamily members along with mutagenesis results of DccA suggests a possible basis for substrate specificity is due to access tunnel size.

Computational Tool Identifies Catalytically Important Remote Residues of E. coli Ornithine Transcarbamoylase Lisa Ngu, Kevin E. Ramos, Nicholas A. DeLateur, Penny J. Beuning, and Mary Jo Ondrechen 1Department of Chemistry & Chemical Biology, Northeastern University Ornithine transcarbamoylase (OTC) is an important enzyme of the urea cycle that converts excess ammonium to urea and is also important in arginine biosynthesis. OTC catalyzes the reaction of carbamoyl phosphate (CP) and ornithine (ORN) to produce citrulline and inorganic phosphate. In humans, OTC deficiency (OTCD) results in build up of ammonium, glutamine, additional amino acids, and orotic acid. OTCD is the most common urea cycle disorder, occurring in 1 of 800 newborns; patients experience ataxia, Reye’s syndrome, brain damage, and death. In order to better understand the dynamics of OTC, we have applied Par- tial Order Optimum Likelihood (POOL), a machine learning method developed at Northeastern University, to predict catalytically important residues. POOL goes beyond bioinformatics and uses the structure of a protein and computed electrostatic properties to accurately predict residues important for enzyme activity, including those remote to the substrate. POOL predicts a spatially extended active site for E. coli ornithine transcarbamylase, for which an induced-fit conformational change upon binding of CP is believed to play a role in its catalytic mechanism. Conserved mutations of POOL-predicted residues Arg57, Asp140, Tyr160, His272 and Glu299 show significant loss of catalytic efficiency. E. coli and human OTC share 49% homology and POOL-predicted E. coli OTC residues Arg57, His272 and Cys273 align with residues Arg92, His302 and Cys303, respectively, at which are observed human OTCD-associated mutations. Current treatments, including a protein restricted diet, arginine and CIT supplementation, among others, are not fully effective. Our findings give insight to the catalytic mechanism of E. coli OTC that could be useful for understanding the mechanism of human OTC to design an improved OTCD therapy. Acknowledgement: Supported by NSF-MCB-1517290.

Determination of Kinetic Parameters of Trypsin I from Pyloric Caeca of Monterey Sardine (Sardinops sagax caerulea) Using Isothermal Titration Calorimetry Idania Quintero1,3, Enrique Velazquez1,2, Javier Castillo1,2,Rocıo Sugich1,2, Dalia Cruz1,3 1Universidad de Sonora, 2Chemistry Department, 3Health Science Department Pyloric caeca of Monterey sardine (Sardinops sagax caerulea) shows an expression of trypsin I according to a cDNA characterization, is a psychrophilic enzyme according to the catalytic efficiency (kcat/KM) obtained by spectrophotometric essays. This parameters can be obtained by Isothermal titration calorimetry (ITC) using thermal power generated by the enzymatic conversion of substrate to product; were the rate of reaction is directly proportional to thermal power. The objective of this study was to obtain kinetic parameters of trypsin I at different temperatures using ITC. To reach the objective, the enzyme was purified from the pyloric caeca of the sardine using molecular exclusion and affinity chromatography obtaining a yield of 1.0 mg/mL. At 48C kcat and KM of Trypsin I were 4.4 s-1 and 0.3 lM respectively. At 108C were 3.05 s-1 and 0.2 lM (kcat and KM) and at 158C kcat was 4.93 s-1 and KM 0.7 lM. At 208C and 308C no adjust was possible to make, nonetheless at 258C a kcat and KM of 5.30 s-1 0.2 lM were obtained. The higher values of kcat were for 15 and 258C, values of KM were similar at 4, 10

95 ABSTRACT and 258C. The catalytic efficiency (kcat/KM) were 13, 15, 7 and 25 s-1 lM-1 at 4, 10, 15 and 258C respectively, showing a better catalytic efficiency at 48C than 158C; and 3 times fold at 258C. With this results we can reassert the psychrophilic behavior of trypsin as its kcat/KM is higher than mammalian trypsins. The kinetic behavior of enzymes leads to understand biochemical pathways and catalytic mechanisms; were ITC provides a universal approach to determining the kinetic behavior of enzymes.

Cadmium Inhibits MutLa within the Human Mismatch Repair System Shanen M. Sherrer1, Paul L. Modrich1,2 1Department of Biochemistry, 2Howard Hughes Medical Institute, Duke University Medical Center, Durham, NC 27710 DNA mismatch repair (MMR) deficiency increases spontaneous mutability and is the cause of Lynch syndrome, one of the most common hereditary cancers. Yeast studies of others have shown that the mutagenic action of Cd21 is largely due to inhibition of MMR.1,2 We have found that isolated human MutLa (hMutLa, MLH1•PMS2 heterodimer), which is required for MMR, contains 2 bound Zn21 ions and that the bound metal is required for hMutLa function as an endonuclease. We have also found that Cd21, a known inhibitor of Zn-metalloenzymes, is a potent inhibitor of hMutLa endonuclease, an effect that can be reversed by Zn21. In addition, Cd21 selectively inhibits 30-directed MMR in nuclear extracts, a reaction that is absolutely dependent on hMutLa, and that can only be restored with the supplementation of an endonuclease active hMutLa. Based on these findings we suggest that the biological effects of Cd21 on MMR are the result of selective inhibition of hMutLa. References: 1. Jin, Y.H., et al. (2003) Nat. Gen. 34, 326-329. 2. Wieland, M., et al. (2009) Biochem. 48, 9492-9502.

Engineering the reversal of Pseudomonas putida b-hydroxybutyrate dehydrogenase cofactor specificity. Jorna Sojati, Connor Ott, Nadia Galchak, and Jennifer B Palenchar. 1Department of Chemistry, Program in Biochemistry, Villanova University, 800 E. Lancaster Ave., Villanova, PA 19085. Prior work has revealed that trypanosome b-hydroxybutyrate dehydrogenase (b-HBDH) can utilize NADP(H), unlike nearly all other b-HBDHs characterized. Using the trypanosome enzymes as a guide, we sought to alter the cofactor specificity of an NADH-dependent b-HBDH. In the work presented, we describe amino acid changes in the Rossman fold region of an NADH-dependent bacterial b-HBDH from Pseudomonas putida. The histidine-tagged recombinant enzymes were overexpressed, purified from E. coli, and characterized kinetically. In the bacterial b-HBDH, one amino acid change is sufficient to loosen specificity, while a combination of mutations allows for cofactor preference reversal. The kinetic characterization of these mutant enzymes will be presented. Long term, we seek to use this information to alter the cofactor specificity of the parasite enzyme in vivo to better understand the role of this enzyme in the trypanosomes. Acknowledgement: We gratefully acknowledge support for this work from the Villanova University Department of Chem- istry and the Villanova University Research Support Grant program.

The role of the Protein Structural Network (PSN) in the thermostability Valquiria P. Souza1, Cecılia M. Ikegami1, Guilherme M. Arantes1 and Sandro R. Marana1 1Department of Biochemistry, Institute of Chemistry - University of S~ao Paulo Analysis of proteins as networks has been shown a powerful tool to understand their properties and important residues. In this analysis, residues close to each other in, at least 5Å, are called connected.

96 ABSTRACT

Some residues exhibit many connections and can form short contact pathways between distant residues in the protein structure, being called central residues. Central residues have been shown to have important roles in catalysis, thermal stability and allostery. In order to assess the correlation between the residue centrality and its importance in the protein properties, we use two approaches: The first one is to make single mutations at the central residues of a beta-glucosidase, changing those residues to alanine, and measure if those mutations modify the protein thermostability or function. The second one is to perturb a central residue (F251) by changing its environment through single mutations that introduces voids or additional volume. Next we evaluate how those mutations affect the protein as a whole. In general, we have observed that mutations at central residues reduce the Tm in 2 - 58C and increase the unfolding rate in 2 - 9 times, suggesting that damages in the central residues make the protein more unstable. Moreover, replacements of residues close to a central one (F251) decrease the thermostability, but only provided that the mutated residue is pointed toward the central residue. In addition, we have observed no significant effects of the mutations on the beta-glucosidase activity by measuring its specificity for four different substrates and the pKa of the catalytic residues. Shortly, changes in the central residues and in their neighborhoods seem to affect the thermostability of the protein, indicating that they are important to the protein stability. Acknowledgement: Supported by FAPESP, CAPES and CNPq.

Kinetic and mutational studies of the adenosine diphosphate ribose hydrolase from Mycobacterium tuberculosis Puchong Thirawatananond 1Johns Hopkins University School of Medicine, Department of Biophysics and Biophysical Chemistry Background: Mycobacterium tuberculosis represents one of the world’s most devastating infectious agents – with one third of the world’s population infected and 1.5 million people dying each year from this deadly pathogen. We report the kinetic characterization of the ADP-ribose Nudix hydrolase from M. tuberculosis, the product of gene rv1700. The gene was cloned, and the enzyme was expressed, purified, and identified as a pyrophosphohydrolase specific for adenosine diphosphate ribose (ADPR), a compound involved in various pathways including oxidative stress response and tellurite resistance. Results: Optimal catalytic rates were achieved at alkaline pH (7.5-8.5) with either 0.5-1 mM Mg21 or 0.02-1 mM Mn21. Km and kcat values for hydrolysis of ADPR with Mg21 ions are 200 6 19 mM and 14.4 6 0.4 s-1, and with Mn21 ions are 554 6 64 mM and 28.9 6 1.4 s-1. Four residues proposed to be important in the catalytic mechanism of the enzyme were individually mutated and the kinetics of the mutant enzymes were characterized. In the four cases, the Km increased only slightly (2- to 3-fold) but the kcat decreased significantly (300- to 1900-fold), confirming the participation of these residues in catalysis. Conclusions: An analysis of the sequential and structural conservation of Nudix ADPRases defines unambiguously the family and provides insight into residues involved in catalysis. Differences between the M. tuberculosis and other ADPRase protein sequences were identified and contribute to our understanding of the substrate recognition mechanisms particular to Mt-ADPRase.

Enzyme promiscuity and Evolution of Enzyme Kinetic Mechanisms N. Nuray Ulusu 1Koc¸ University, School of Medicine, Rumelifeneri yolu, Sarıyer Istanbul. Enzymes can identify correct substrates over other molecules for their function. Kinetic mechanisms have evolved in response to alterations in ecological and metabolic conditions. The kinetic mechanisms of single-substrate mono-substrate enzyme reactions are easier to understand and much simpler than those of bi-bi substrate enzyme reactions. The increasing complexities of kinetic mechanisms, as well as

97 ABSTRACT the increasing number of enzyme subunits, can be used to shed light on the evolution of kinetic mechanisms. Enzymes with heterogeneous kinetic mechanisms attempt to achieve specific products to sub- sist. In many organisms, kinetic mechanisms have evolved to aid survival in response to changing environmental factors. Enzyme promiscuity is defined as adaptation to changing environmental conditions, such as the introduction of a toxin or a new carbon source. Enzyme promiscuity is defined as adaptation to changing environmental conditions, such as the introduction of a toxin or a new carbon source. Enzymes with broad substrate specificity and promiscuous properties are believed to be more evolved than single-substrate enzymes. This group of enzymes can adapt to changing environmental substrate conditions and adjust catalysing mechanisms according to the substrate’s properties, and their kinetic mechanisms have evolved in response to substrate variability.

Enzyme Mechanism Influences Macromolecular Crowding Effects A.E. Wilcox 1Hobart and William Smith Colleges 300 Pulteney Street Geneva, NY 14456 The intracellular environment is crowded with high concentrations of carbohydrates, proteins, nucleic acids, and other macromolecules. Traditionally, biophysical studies are conducted in dilute conditions, yet experimental and theoretical evidence indicates that enzyme behavior is affected by the presence of these macromolecules, resulting in slowed diffusion, enhanced enzyme-substrate binding, and altered enzyme conformation. To begin to characterize the effects of crowding on enzyme kinetics, Michaelis- Menten parameters were obtained from spectrophotometric assays conducted on the enzyme yeast alco- hol dehydrogenase (YADH) in the absence and presence of dextran. For the standard reaction with ethanol, dextran decreases YADH activity approximately 40%, due to the slowed release of the NADH product, which is rate-limiting. Conversely, when the alternative substrate, isopropanol, is used and the rate-limiting step becomes the chemical hydride transfer, dextran increases YADH activity by 20-40%, possibly due to stabilization of the enzyme. Furthermore, while the effects of crowding in the forward direction of YADH catalysis are independent of dextran size, the reverse direction displays a polymer size dependence as well as potential substrate inhibition and cooperativity. Although further investigation is required, these studies provide insight into how macromolecular crowding can affect enzyme kinetics.

Investigation on the specific recognition between lambda exonuclease and DNA substrates with chemical modification and mismatches Tongbo Wu, Meiping Zhao 1Beijing National Laboratory for Molecular Sciences, MOE Key Laboratory of Bioorganic Chemistry and Molecular Engineering, College of Chemistry and Molecular Engineering, Peking University, Bei- jing 100871, China Lambda exonuclease (k exo) is a well-known and deeply studied enzyme. The 5’ phosphate in the substrates has been proven to be essential for the digestion of k exo, without which the digestion will be significantly slower. In this work, the substrates with a two-nucleotide overhang structure at the 5’ end were found to be digested fast by k exo even without the 5’ phosphate but with 5’ modification by hydrophobic groups (such as C6 spacer). Thus, when the first two nucleotides of the substrates are already unpaired, the hydrophobic groups are likely to play a role similar to the 5’ phosphate. The hydrophobic interaction area in k exo for the 5’ end of the substrates was proposed (Figure 1A). The binding states of k exo and the substrates are either productive or non-productive. An internal labelling with a conjugated structure (such as FAM) on the strand to be digested in substrates could increase the portion of productive binding state if a mismatched base pair is just right on the 5’ side of the labelling. The p-p stacking interaction area in k exo for the internal labelling on the substrate was proposed (Figure 1B). These findings provide a new sight into the digestion mechanism of k exo.

98 ABSTRACT

The intrinsically disordered selenoprotein K has autoproteolytic activity Zhengqi Zhang, Jun Liu, and Sharon Rozovsky 1Department of Chemistry and Biochemistry, University of Delaware, Newark, Delaware 19716, United States Selenoproteins are a diverse family of proteins notable for the presence of the 21st amino acid, selenocysteine. And most of the selenoproteins with known functions are oxidoreductases whereas a selenocysteine is found in the active site. Selenoprotein K (SelK), which contains a selenocysteine (Sec) residue at the C terminus, is an intrinsically disordered membrane protein. SelK plays important role in the regulation of calcium (Ca21) flux in immune cells by complexes with the palmitoyl acyl transferase DHHC6, which is required for palmitoylation of the Ca21 channel protein, inositol-1,4,5-triphosphate receptor (IP3R) in the ER membrane. Here, we report that SelK undergoes autoproteolytic cleavage. We identified the cleavage site by mass spectrometry. These results were further confirmed by preparing SelK by native chemical ligation. This is the first report of a selenocysteine residue being involved in active site of a protease. These results also suggest autoproteolysis may work as a regulatory mechanism for SelK whereby the eventual cellular function could possibly be determined by cleaved or uncleaved SelK.

Experimental evidence for the existence of thermally stable proteins in early life Satoshi Akanuma 1Faculty of Human Sciences, Waseda University, Tokorozawa, Japan A phylogenetic tree can be constructed by comparing the homologous amino acid sequences of modern proteins that were evolved from a common ancestor. Ancestral amino acid sequences of some proteins can be inferred using the topology of the phylogenetic tree and the modern amino acid sequences. Moreover, the ancestral proteins can be reconstructed by synthesizing and expressing the gene encoding the inferred amino acid sequences. Such a reconstructing approach is currently a common technique for understanding the physical characteristics of primitive proteins. Ancestral amino acid sequences of nucleoside diphosphate kinases (NDKs), which might be present more than 3,500 million years ago, were computed using a homogeneous and a non-homogeneous evolution models. Then, the ancestral NDKs were experimentally reconstructed. All of the reconstructed ancestral NDKs were extremely thermally stable. The results were robust to the uncertainties associated with the predictions of the ancestral sequences and to the topologies of the phylogenetic trees (1,2). Ancestral amino acid sequences of NDK were also inferred from a dataset where hyperthermophilic sequences were absent and appeared to be extremely thermally stable (2). Thus, early life most likely contained thermally stable proteins and therefore they were thermophiles that flourished at very high temperatures. Given the thermophilicity of early life, the reconstruction approach will provide an effective method for creating thermally stable proteins. References: 1. Akanuma et al., Proc. Natl. Acad. Sci. USA, 110, 11067–11672 (2013) 2. Akanuma et al., Evolution 69, 2954-2962 (2015)

PI – EVOLUTION The distribution of fitness effects of deletions in TEM-1 beta-lactamase Courtney Gonzalez1, Paul Roberts1, Marc Ostermeier1 1Chemical and Biomolecular Engineering, Johns Hopkins University KEYWORDS: Evolution, TEM-1 beta lactamase, INDELs Insertions and deletions (INDELs) are an important source of genetic diversity in molecular evolution, both in nature and in the laboratory, yet a comprehensive picture of the distribution of fitness effects of INDELs is lacking. Innovations in mutagenesis techniques, selection schemes, and deep-sequencing

99 ABSTRACT technology allow us to make extensive libraries and characterize fitness effects. Here, we study TEM-1 beta-lactamase, a protein that confers resistance to penicillin antibiotics, such as ampicillin (Amp). TEM- 1 is a convenient and well-studied model protein for molecular evolution experiments: when E. coli cells with TEM-1 are challenged to grow in the presence of Amp, organismal fitness can be directly correlated to the resistance conferred by the protein. Past studies include a comprehensive map of the distribution of fitness effects of single codon substitutions in TEM-1. Here, we created a complete library of single codon deletions in TEM-1 beta-lactamase using inverse PCR. Using a bandpass selection scheme, the library was partitioned based on relative fitness, as determined by Amp resistance. Barcoded subli- braries were deep-sequenced to determine the fitness of individual deletions. We find that deletions in secondary structures are generally not tolerated, while deletions in loops are more likely to retain some activity. Two deletions in the signal sequence, which directs export of the protein to the periplasm, were found to increase fitness relative to wildtype. Further studies will characterize the distribution of fitness effects of codon insertions in TEM-1. Together, this will help to further characterize the fitness landscape of TEM-1, as well as bolster molecular evolution studies in general by offering an extensive picture of single codon INDEL effects.

Biophysical mechanisms driving the evolution of androgen specificity Denise Okafor1, Jennifer Colucci1, Eric Ortlund1 1Emory University School of Medicine, Atlanta GA 30322 Understanding the genetic and biophysical mechanisms by which proteins evolve new function is crucial for continued progress in evolutionary biology and biochemistry. Of particular interest is the relationship between shifts in protein function and mechanisms by which amino acid substitutions mediate these changes. Here we seek to understand the evolution of ligand specificity in the androgen receptor (AR), an essential hormone-controlled transcription factor. Through evolutionary time, AR gained the ability to respond to androgens and lost the ability to recognize progestagens, which its immediate phylogenetic ancestor was able to recognize. Three historical amino acid substitutions have been identified that restore progestagen activation to AR. Here, we use molecular dynamics simulations to probe the structural and biophysical effects of these replacements on protein-ligand interactions and their role in ligand specificity. We show that the activation function helix, a key regulatory region of the receptor, is restructured and destabilized in the inactive AR-progestagen complex; additionally, the non- activating complex induces increased conformational dynamics in the ligand. The historical amino acid substitutions are observed to reverse these effects, increasing stabilizing interactions around the ligand binding site and stabilizing the activation function helix. Substitutions also increase the size of the ligand binding pocket, favoring the binding of the larger progestogen molecule. These data suggest that large changes in protein function observed during evolution may be mediated by substitutions in a small number of key residues.

A covariance-based model to estimate antimicrobial drug resistance Frazier N. Baker1, Aleksey Porollo1 1University of Cincinnati and Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA Resistance to antimicrobial drugs through polymorphism in the targeted proteins is a major problem in the treatment of infectious diseases. The identification of mutations and quantification of their effect on susceptibility to a drug would be a major advancement toward personalized therapy. Computational estimation of the effect of mutations is usually conducted on an individual position basis, primarily employing evolutionary information. Such an approach cannot account for context specific changes associated with a given mutation. On the other hand, amino acid covariance analysis enables identification of interacting residues that may be involved in protein folding or function, even when they are

100 ABSTRACT distant in sequence. In this work, we have developed a new model that incorporates both individual conservation properties of amino acids and their covariance information to quantify the effect of mutations on the susceptibility to a drug. Covariance analysis was conducted using the mutual information, chi-squared statistic, and Pearson correlation metrics, all adjusted for phylogeny bias in the multiple sequence alignment. The effect scores were derived for mutations in a number of known antimicrobial drug targets and related to the corresponding experimental values of minimal inhibitory concentrations of a drug. The Pearson correlation coefficient for predicted estimates and experimental values is in the range of 0.5-0.8 for different proteins. The major advantage of the new model is that it can estimate the effect based on haplotypes (multiple simultaneous mutations) and does not require a resolved structure of the targeted protein.

PJ - FOLDING Measuring the effects of vectorial appearance of the Micayla A. Bowman1, Ian M. Walsh1, Patricia L. Clark1 1Department of Chemistry and Biochemistry, University of Notre Dame Proper protein folding is essential for all biological processes. Although the cellular environment can support the folding of an enormously diverse range of protein structures, our ability to refold proteins in vitro is limited to small, simple folds. Molecular chaperones can increase the yield of correctly folded protein for a diverse range of substrates, but a chaperone interaction is required for only 20% of the proteome under normal growth conditions (1). This suggests that other features of the cellular environment fundamentally different from refolding in vitro are responsible for increasing folding yield in vivo. One such difference is vectorial folding of the polypeptide chain during its synthesis from N- to C- terminus by the ribosome. Current in vitro protein refolding techniques cannot recapitulate vectorial protein appearance or folding. We have developed a novel assay to test the effect of vectorial chain appearance on protein refolding in vitro. This assay uses E. coli ClpX, an ATP hydrolysis-driven protein unfoldase/translocase, to drive vectorial appearance of the polypeptide chain and enable folding to begin from either the N- or C-terminus of a substrate. We are currently using this assay to directly test the impact of vectorial appearance of a polypeptide chain on its folding kinetics and refolding yield. Reference: 1. Hartl U, Hayer-Hartl M (2002) Molecular Chaperones in the Cytosol: from Nascent Chain to Folded Protein. Science 295:1852–1858.

Effect of Circular Permutations on Transient Partial Unfolding in Proteins Chen Chen 1Purdue University Under native conditions, proteins can undergo transient partial unfolding, which may cause proteins to misfold or aggregate. A change in sequence connectivity by circular permutation may affect the energetics of transient partial unfolding in proteins without altering the three-dimensional structures. Using Escherichia coli dihydrofolate reductase (DHFR) as a model system, we investigated how circular permutation affects transient partial unfolding in proteins. We constructed three circular permutants, CP18, CP37, and CP87, with the new N-termini at residue 18, 37, and 87, respectively, and probed transient partial unfolding by native-state proteolysis. The new termini in CP18, CP37, and CP87 are within, near, and distal to the Met20 loop, which is known to be dynamic and also part of the region that undergoes transient unfolding in wild-type DHFR. The stabilities of both native and partially unfolded forms of CP18 are similar to those of wild-type DHFR, suggesting that the influence of introducing new termini in a dynamic region to the protein is minimal. CP37 has a significantly more accessible partially unfolded form than wild-type DHFR, demonstrating that introducing new termini near a dynamic region may promote transient partial unfolding. CP87 has significantly destabilized native and partially

101 ABSTRACT unfolded forms, confirming that modification of the folded region in a partially unfolded form destabil- izes the partially unfolded form similar to the native form. Our findings provide valuable guidelines to control transient partial unfolding in designing circular permutants in proteins.

In Silico to In-Cell Folding of Metastable Serpins Anne Gershenson1 1University of Massachusetts, Amherst, USA The serpins a1-antitrypsin (A1AT) and antithrombin III (ATIII) are abundant plasma proteins that help regulate serine proteases involved in inflammation and blood coagulation. Misfolding and polymerization of mutant serpins in the endoplasmic reticulum of hepatocytes can lead to serious genetic diseases due to low serpin circulating levels, and for A1AT mutants polymerization can also lead to serious liver disease. However, it is unclear how serpins fold and misfold in cells. To provide a global view of folding and misfolding, we monitored in-cell folding and secretion using radioactive pulse-chase methods, characterized N- to C-terminal serpin fragments in vitro to mimic co-translational folding and simulated folding in atomistic detail using the Dominant Reaction Pathways transition path sampling approach. Serpins are composed of 400 amino acids organized into two, non-sequential interdigitated domains. Despite this complicated topology, our results indicate that significant amounts of cooperatively folded structure likely form well before these proteins are fully translated. In addition, and in agreement with in vitro folding results, correct incorporation of the C-terminal b hairpin is crucial for proper folding and function. Both cellular experiments and simulations unexpectedly reveal that proper positioning of the C-terminal b hairpin precedes the final structural consolidation of the N-terminal region. Similarly, proper folding of the C- terminal region is disrupted in ATIII disease-associated mutants with the lowest secretion levels. The unexpected effects of mutations on relatively early steps in folding are further supported by simulations of the pathogenic, polymerization prone A1AT Z (Glu342Lys) mutant, which also provide atomistic insight into how misfolding likely occurs. Thus this combination of experimental and computational results provides important, novel insights into how serpins, relatively large and topologically complex proteins, fold and misfold, and provides hints on how disease-associated misfolding might be ameliorated.

Investigating the influence of sequence on folding mechanism using a structure based protein model Elizabeth K. Gichana1, Charles L. Brooks III1,2 1Program in Biophysics, 2Department of Chemistry, University of Michigan A protein’s primary amino acid sequence encodes not only for the equilibrium crystal structure but for its entire energy landscape. Energy landscape theory posits that main-chain topology is the major determinant in the folding mechanism of a protein. What’s less well understood is the role that sequence effects play. Recent studies utilizing ancestral sequence reconstruction have yielded protein sequences along mesophilic and thermophilic lineages of ribonuclease H (RNaseH) leading back to a common ancestor. This family of proteins share function and topology and have high sequence homology but differ in their biophysical properties. We investigate the role that variation in sequence over these evolutionary timescales alters the energy landscape of RNaseH to give rise to these differences by employing coarse-grained molecular dynamics simulations using a structure-based Go-like model to investigate folding mechanisms.

ACPro: A curated database of verified protein folding kinetics Sheila Jaswal1 1Amherst College Chemistry Department, USA The ACPro database (https://www.ats.amherst.edu/protein/) is a freely accessible, curated resource with folding kinetics data for over 120 non-homologous proteins (Wagaman, A.S., Coburn, A., Brand-Thomas

102 ABSTRACT

I., Dash B., and Jaswal, S.S. “A Comprehensive Database of Verified Experimental Data on Protein Folding Kinetics.” Protein Science 23(12), 1808-1812.). While several previous collections of folding data have been published, we generated our database by including only entries for which we could verify the rate constant(s) and the actual structural construct used with an original experimental report. Due to the challenges we encountered in this verification process, we are renewing the call from the 2005 Protein Folding Consortium for standardized reporting of protein folding results (Maxwell KL, et al. “Protein folding: Defining a ’standard’ set of experimental conditions and a preliminary kinetic data set of two- state proteins.” Protein Science 14:602-616.) To facilitate sharing of standardized results within the protein folding community, our database includes a submission form that collects the relevant experimental conditions and results for each protein. We are developing aspects of the database to enhance its utility, including notes about proteins with multiple states and/or isomerization as well as the inclusion of unfolding kinetics data (pending verification). In addition, we plan to develop and share an educational component for incorporation into undergraduate and high school courses. We welcome submissions for new entries, feedback and suggestions for future database improvements. With widespread use, ACPro will support a more robust symbiosis between experiment and theory, facilitating accurate models that deepen our understanding of protein folding, stability and dynamics.

Biophysical studies on 4-Hydroxynonenal modified Human Serum Albumin: A Possible Role in Rheumatoid Farzana Khan 1Department of Biochemistry, JN Medical College, AMU, Aligarh, India Protein carbonylation induced by Lipoxidation-derived reactive aldehydes such as 4-hydroxy-2-nonenal (HNE), acrolein and methylglyoxal plays a significant role in the etiology and/or progression of several human diseases, such as Rheumatoid Arthritis (RA). Among them, HNE is the most abundant cytotoxic chemical entity. It react with proteins to generate stable adducts called advanced lipoxidation end products (ALEs). The aim of the study was to analyze changes in Human Serum Albumin (HSA) upon modification by HNE and evaluation of immunogenicity of native and HNE modified HSA along with possible implications in the pathogenesis of RA. In this study, HSA was incubated with increasing concentrations of HNE to generate ALEs. HNE induced structural changes in the HSA were characterized by UV, fluorescence, CD, polyacrylamide gel electrophoresis, Fourier transform infrared spectroscopic, carbonyl content determination, surface hydrophobicity estimation and HPLC. We thereby report the structural perturbations in HSA upon modification with HNE and the consequential enhanced immunogenicity. The induced antibodies were found to be highly specific for the immunogen and exhibited cross- reactivity with HNE-modified epitopes on proteins, amino acids and nucleic acid. Furthermore, sera from RA patients were screened for the presence of auto-antibodies reactive to native and HNE- modified HSA. A considerable high binding to HNE-modified HSA was observed as compared to native HSA in the serum samples of RA patients. Also inhibition ELISA results points towards appreciable recognition of HNE-modified HSA by auto-antibodies in RA patients. Our results suggest that the HNE induced modification in HSA is capable of modifying and/or cross-linking proteins to the extent of compromising its physiological functions as well as generating neo-epitopes on the molecule, thus making it highly immunogenic.

Validation of Native-state Hydrogen Exchange Mass Spectrometry to map protein folding landscapes Minjee Kim1, Jacob Witten1, Sheila Jaswal1 1Amherst College Biochemistry/Biophysics Program Protein function depends on the proper calibration of stability, dynamics and structure of the native state. Defining a protein’s landscape in terms of the conformations sampled, the differences in their

103 ABSTRACT thermodynamic stabilities, and the rates of transition between them has been an essential approach to investigate the link between function and energetics. Traditionally, landscape parameters are measured using equilibrium denaturation and kinetic chevron analysis monitored by spectroscopic methods. In addition to requiring large numbers of measurements as a function of denaturant and/or temperature to induce bulk unfolding under conditions far from native, such studies are limited to proteins that unfold and refold reversibly on a short timescale without misfolding or aggregating. We have developed an approach using native-state hydrogen exchange coupled with mass spectrometry (N-HXMS) that addresses these challenges. Using numerical simulations to fit N-HXMS time- course data on intact proteins, our method extracts landscape parameters from limited measurements under non-denaturing conditions. We are validating our approach by verifying landscape parameters measured by N-HXMS across HX regimes with those determined by traditional urea denaturation experiments for the model two-state protein, protein L. In addition, we are exploring the potential of N-HXMS under certain conditions to allow simultaneous determination of the unfolding and folding rate constants from a single HXMS time course. This work will help establish N-HXMS as a possible alternative to chevron analysis that requires less time and material for two-state proteins, and a milder method to probe landscapes of proteins in general.

Exploring Protein Stability and Aggregation by nanoDSF Ellen Lee1*, Myriam Badr1, Ana Lazic2, Stefan Duhr3, and Dennis Breitsprecher3 1NanoTemper Technologies, Inc, Cambridge, MA, 2NanoTemper Technologies, Inc, South San Fran- cisco, CA, 3NanoTemper Technologies GmbH, Munich, Germany We will discuss the Prometheus NT.48 instrument and the nanoDSF technology, which allows for in- parallel high-precision characterization of stability and aggregation parameters of biologicals. nanoDSF is an advanced Differential Scanning Fluorimetry technology. It detects the smallest changes in the fluorescence of tryptophan present in virtually all proteins. The fluorescence of tryptophans in a protein is strongly dependent on its close surroundings. By following changes in fluorescence, chemical and thermal stability can be assessed in a truly label-free fashion. The dual-UV technology by NanoTemper allows for rapid fluorescence detection, providing an unmatched scanning speed and data point density. This yields ultra-high resolution unfolding curves which allow for detection of even minute unfolding signals. Furthermore, since no secondary reporter fluorophores are required as in conventional DSF, protein solutions can be analyzed independent of buffer composition, and over a concentration range of 200 mg/ml down to 5 mg/ml. Therefore, nanoDSF is the method of choice for the easy, rapid and accurate analysis of protein folding and stability, with applications in membrane protein research, protein engineering, formulation development and quality control.

Investigation of Confinement Effects on Protein Stability Using Reverse Micelles and Chemical Denaturation Pamela Gallo1, Kayla Schardien1, Hannah Work2, Nathaniel V. Nucci1,2 1Department of Biomedical and Translational Sciences, Rowan University, 2Department of Physics & Astronomy, Rowan University Most studies of protein stability have been performed in bulk aqueous solution. The interior of cells, however, is packed with large molecules that create a highly confined environment, and the effects of confinement on proteins are poorly understood. In order to see how confinement alters the stability of a protein, chemically-induced unfolding of the model proteins hen egg white lysozyme (HEWL) and cytochrome c (cyt c) were examined both in bulk solution and in reverse micelle systems (RMs) composed of either cetyltrimethylammonium bromide (CTAB) and hexanol or decylmonacyl glycerol (10MAG) and lauryldimethylammonium-N-oxide (LDAO). The structural state of the protein was evaluated by native tryptophan fluorescence (HEWL) or visible absorption spectroscopy (cyt c). The extent

104 ABSTRACT of confinement presented by the RM condition was evaluated using dynamic light scattering. In each case, confinement produced by the RM resulted in a destabilization of the native state of the encapsulated protein. This work was supported by startup funds from Rowan University and by summer undergraduate fellowship funds from the New Jersey NASA Space Grant.

Proxy analysis of the structural and functional impact of a pathogenic mutation in a human protein chaperone using an archaeal model Dario Spigolon1,2,3, Travis Gallagher3, Donatella Bulone1, Pier Luigi San Biagio1, Jatin Narang3, Everly Conway de Macario4,5, Francesco Cappello5,6, Alberto J.L. Macario4,5, Frank Robb3,4 1Institute of Biophysics, UOS Palermo, National Research Council, Italy, 2Department of Physics and Chemistry, University of Palermo, Palermo, Italy, 3Institute for Bioscience and Biotechnology Research (IBBR), Rockville, MD, USA, 4Department of Microbiology and Immunology, School of Medicine, Uni- versity of Maryland at Baltimore and Institute of Marine and Environmental Technology (IMET), Columbus Center, Baltimore, MD, USA, 5Euro-Mediterranean Institute of Science and Technology (IEMEST), Palermo, Italy, 6Department of Biomedicine and Clinical Neurosciences, Human Anatomy Section, University of Palermo, Palermo, Italy. Here, we describe analysis of functional impacts of a heritable human mutation in one of the subunits of the complex CCT, a Group II chaperonin. Functional studies on human Group II chaperonins are limited by the complexity of the human CCT particle, which is composed of two octameric rings, each formed from eight similar but non-identical subunits. We used site-directed mutations of an archaeal chaperonin that is composed of one subunit type, effectively amplifying the mutational impact of a point mutation eight-fold. In this study, we analyzed in detail the loss of structural stability of the hexa- decamer formed from monomers with the pathogenic His to Arg mutation, using differential scanning calorimetry (DSC), isothermal titration calorimetry (ITC), High Performance Liquid Chromatography (HPLC), and coupled gel permeation/dynamic light scattering. The disassembly of the complex, which in its wild type version is tightly coupled with subunit denaturation, was decoupled by the mutation without affecting the stability of individual monomers. This mutation in homozygous recessive form causes a crippling neuropathy with high amputation rates. Our work attests to the effectiveness of the homo- hexadecameric archaeal chaperonin as a proxy to amplify the impact of relatively minor deficits in systems with mutations in a single subunit amongst multiple paralogs in the same complex. Acknowledgment: ECdeM, FC, and AJLM were partially supported be IEMEST, FTR was supported by an NSF C-DEBI grant.

Characterization of Truncated Hemoglobins from Chlamydomonas reinhardtii. Miranda Russo1, Eric Johnson, Dillon Nye, Katelyn Jackson, Matthew Preimesberger, Selena Rice, Dagan Marx, Juliette Lecomte 1Johns Hopkins University Truncated hemoglobins (TrHbs) are a group of heme-binding proteins found in many unicellular organisms and plants. TrHbs are capable of diverse chemistries, yet share a common fold. The Chlamydomo- nas reinhardtii genome contains twelve genes, designated THB1-12, with predicted TrHb domains. However little is known about the physiological role or biophysical properties of their encoded proteins. In prior studies, THB1 was shown to possess an unusual heme ligation scheme in the absence of exogenous ligand, with coordination through a histidine on the proximal side of the heme and through a lysine on the distal side (His-Fe-Lys). In this work we present biophysical analysis of two additional TrHbs from C. reinhardtii, THB2 and THB4. NMR and optical spectroscopies support that both THB2 and THB4 utilize the His-Fe-Lys ligation described in THB1, but structural differences between the three proteins suggest each may have unique coordination chemistry under physiologic conditions. We posit that these individual chemistries tailor each protein toward mediation of reactive nitrogen species

105 ABSTRACT

(RNS), and specifically nitric oxide (NO), under different conditions. Present studies focus on enriching our understanding of TrHbs through defining the unique structural and biochemical characteristics of these three related hemoglobins. Acknowledgement: Supported by NSF MCB-1330488 and by NIH T32 GM-008403.

Investigating the relationship between enzyme function, native-state dynamics and kinetic stability in trypsin Abel Samanez1, Cathy Amaya, Sheila Jaswal1 1Amherst College Biochemistry and Biophysics Program Protein misfolding and aggregation are associated with many conditions including type II diabetes, some cancers, and several neurodegenerative diseases. Understanding mechanisms of protein stabilization is crucial for treatment development. Most studied proteins possess thermodynamic stabilization in which the native state is in equilibrium with other conformational states, allowing frequent sampling of vulnerable conformational states that may result in misfolding and aggregation. An investigation of bacterial degradative enzymes shed light on a radically different form of protein stabilization. Unlike thermodynamically stable proteins, alpha-lytic protease (aLP) takes advantage of a large activation free energy barrier between the native and unfolded state to effectively trap the protease in its functional state, prolonging its enzymatic activity in a harsh extracellular environment. Because aLP’s native state is higher in energy than its unfolded states, this represents an extreme form of kinetic stabilization; the high energetic barrier to unfolding is the sole determinant of aLP’s stability and its native-state dynamics are severely dampened. Other proteins that must survive in harsh environments likely display some degree of kinetic stabilization to limit sampling of vulnerable unfolded states as well. Previous work with trypsin, a mammalian homolog of aLP, showed this digestive enzyme has an unfolding barrier similar in height to aLP. We have carried out a comprehensive investigation of trypsin’s conformational dynamics and extent of kinetic stabilization compared to its function by measuring the rates of enzyme inactivation, fluorescence loss, and hydrogen exchange through mass spectrometry. Preliminary results suggest that trypsin’s significant kinetic barrier to unfolding is not coupled to restricted native-state dynamics. However, trypsin remains enzymatically active despite suffering auto-proteolysis in the native state, suggesting an alternative strategy for functional longevity that perhaps relies on its disulfide bonds.

Investigating residue-specific dynamics with native-state hydrogen exchange mass spectrometry Nevon Song1, Jovan Damjanovic1, Jacob Witten1, Sheila Jaswal1 1Amherst College Biochemistry and Biophysics Program A protein’s specific biological function is governed by a precise interplay of structure, stability, and dynamics. It is therefore paramount to investigate the connection between protein function and energetics by measuring the thermodynamics and kinetics that determine how a protein samples its folding conformations. As a result of the large-scale perturbations induced in traditional experimental methods to characterize protein folding, our understanding of landscapes is restricted to a relatively narrow set of proteins that refold reversibly without misfolding or aggregation. In order to expand the investigation of landscapes to a more diverse set of proteins, we have developed a numerical simulations approach to extract landscape parameters from mass spectrometry (MS) measurements of intact proteins undergoing hydrogen exchange (HX) under native conditions. We have implemented the Gillespie algorithm to simulate native-state protein dynamics observed by HXMS as a continuous-time, memory- less Markov process. Because of the unprecedented speed of our algorithm, we can generate simulated data to compare to experimental HXMS data in real time, iteratively optimizing across landscape parameters to determine the best fit values. This has resulted in the discovery of a high-energy intermediate in the native landscape of beta-2-microglubulin (b2M), demonstrating the ability of our approach to

106 ABSTRACT detect, and quantitatively measure rates of transition to, non-native conformations under native conditions. Furthermore, our discovery highlights the importance of sub-global dynamics. We are expanding our modeling to better capture the full range of dynamics accessible through HX. Incorporating residue-level dynamics into our simulations will enable us to expand our approach beyond intact proteins to improve interpretation of complex behavior often observed in peptides generated from the more common method of applying proteolysis following HXMS.

Cooperative folding of a low sequence complexity, PP2 protein lacking a hydrophobic core Zachary Gates1, Michael C Baxa1, Wookyung Yu1, Josh Riback1, Stephen Kent1, Tobin R Sosnick1 1The University of Chicago The burial of hydrophobic residues in a protein core is generally thought to be the basis of the stable, cooperative protein folding. Here, we show that for the snow flea antifreeze protein, both properties can occur without such burial or canonical secondary structure. The protein has low sequence complexity with 46% glycine and an interior filled with backbone H-bonds between six polyproline 2 (PP2) helices. Furthermore, the protein folds in a kinetically two-state manner. Part of the stability arises due high levels of PP2 structure in the unfolded state, which is highly expanded in spite of having a high glycine content. These results challenge our understanding of the origins of cooperativity and stability in protein folding including the balance between solvent and chain entropies.

Cotranslational folding studies of spectrin and Ig-like domains show folding occurs close to the ribosome Annette Steward1, Ola B. Nilsson2, Adrian A. Nickson1, Jeffrey J. Hollins1, Stephan Wickles3, Andrew Marsden1, Carolina Mendonca1, Roland Beckmann3, Gunnar von Heijne2,4, Jane Clarke1 1Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, CB2 1EW, UK, 2Cen- ter for Biomembrane Research, Department of Biochemistry and Biophysics Stockholm University, SE- 106 91 Stockholm, Sweden, 3Gene Center and Center for Integrated Protein Science Munich, CiPS-M, Feodor-Lynen-Strasse 25, University of Munich, 81377 Munich, Germany, 4Science for Life Laboratory Stockholm University, Box 1031, SE-171 21 Solna, Sweden The vectorial nature of protein synthesis means that N-terminal regions of a protein are potentially able to fold before C-terminal regions have emerged from the ribosome. We therefore ask: how do the key

107 ABSTRACT features of protein folding, elucidated from in vitro studies, manifest in cotranslational folding on the ribosome? To address this question, we have employed an arrest peptide-based force measurement assay to investigate the cotranslational folding of a family of all a-helical spectrin domains and also two b-sheet Greek key domains, I27 and MATH. Our results show that for all these domains, cotranslational folding occurs in the vestibule or close to the ribosome. In the case of spectrin domains, force measurement and cryo EM data show cotranslational folding can occur vectorially, via a pathway not observed in vitro; however the ability to do so depends on the stability of partly folded structures that can form within the ribosome. A circular permu- tant of MATH also shows that a change in connectivity has little effect on when this protein can fold on the ribosome, suggesting that cotranslational folding of b-sheet Greek key domains is similar to the cooperative folding observed in vitro.

Effects of the pro-region on the Vibrio Cholerae cytolysin folding landscape Han Vu1, Kathryn Sundheim1, Sheila Jaswal1 1Amherst College, Biochemistry & Biophysics Program Protein stability can be modeled on a spectrum ranging from purely thermodynamic to purely kinetic stabilization. Proteins stabilized through thermodynamics maintain their native form due to the free energy difference between native and unfolded states, while in kinetically stabilized proteins, a high free energy barrier prevents sampling unfolded states. While thermodynamic stability has been observed for most proteins studied as models for protein folding, evidence is increasing that some degree of kinetic stabilization is important in the function or pathology of many proteins. Here we investigated the balance of thermodynamic and kinetic stability in Vibrio cholerae cytolysin (VCC), a b- pore forming toxin. VCC is synthesized with a pro-region which is cleaved from a precursor (pro-VCC) to form mature VCC. This is similar to the canonical examples of kinetically stabilized proteases, which depend on the folding assistance of their pro-regions to overcome high energy barriers to folding in reaching native states that lack thermodynamic stability. By monitoring urea-induced unfolding by fluorescence, both pro-VCC and VCC were found to be thermodynamically stable by 4 kcal/mol and to possess significant kinetic barriers to unfolding (22 kcal/mol). Thus VCC is an example of combined thermodynamic and kinetic stability. In addition, the remarkable similarity of the VCC and pro-VCC landscapes indicates that the VCC pro-region likely does not assist in folding and holds some other role.

Folding of Metastable Serpin at Atomic Resolution Fang Wang1, Silvio a Beccara3,4, Pietro Faccioli3,6,Anne Gershenson2,5 & Patrick L. Wintrode1 1Department of Pharmaceutical Sciences, School of Pharmacy, University of Maryland, Baltimore, MD 21201, 2Department of Biochemistry & Molecular Biology, University of Massachusetts Amherst, Amherst, MA 01003, 3Trento Institute for Fundamental Physics and Applications, 38123 Povo (Trento), Italy, 4Interdisciplinary Laboratory for Computational Science, Fondazione Bruno Kessler, 38123 Povo (Trento), Italy, 5Program in Molecular and Cellular Biology, University of Massachusetts Amherst, Amherst, MA 01003;, 6Dipartimento di Fisica, Universita degli Studi di Trento, 38100 Povo (Trento), Italy Inhibitory serpins are conformationally labile proteins that fold to a kinetically trapped, metastable state. Disease linked naturally occurring mutations can result in misfolding and polymerization and the folding and misfolding mechanisms of serpins are therefore of intense medical interest. However, at nearly 400 amino acids, serpins are much too large for their folding to be simulated using conventional molecular dynamics. Recently, the Dominant Reaction Pathways (DRP) method, a path integral based variational framework, has proven capable of simulating protein folding with extremely high computational efficiency. Utilizing this method we have simulated the folding of the serpin alpha-1 antitrypsin (A1AT) in

108 ABSTRACT all atom detail using a physics based force field. We find that folding begins with the independent formation of local structure, and these structural units then dock to each other in a defined order to successfully reach the native state. Introducing the most common disease associated mutation (the “Z” mutation E342K) leads to a state in which two of the three beta sheets, A and B, fail to fully form. Intro- ducing Ser at position 290 in the Z mutant rescues folding in our simulations, and preliminary experimental data indicate that this mutation may suppress Z polymerization in cells. These results show that the DRP method can reveal the folding mechanisms of large complex proteins in unprecedented detail without resort to coarse graining or structure based force fields.

PK – INTRINSICALLY DISORDERED PROTEINS Functional protein disorder in coupled binding-folding and liquid-liquid demixing Priya R. Banerjee1, Diana M. Mitrea1, Richard W. Kriwacki1, Ashok A. Deniz1 1The Scripps Research Institute Intrinsically disordered proteins (IDPs) are wide-spread in eukaryotic proteome and recognized as key elements of a variety of cellular functions. At the molecular level, post-translational modifications (PTMs) such as phosphorylation and binding to physiological targets modulate order-disorder transitions in IDPs, thereby regulating their cellular functions. Furthermore, liquid- liquid phase separation (LLPS), primarily driven by IDPs and/or low complexity intrinsically disordered regions (IDRs) in signaling proteins, is an emerging theme underlying the dynamic assembly of membrane-less organelles in cells. Here we discuss biophysical studies on nucleophosmin (NPM1), a nucleolar phospho-protein responsible for the liquid-like structure of the nucleolus, in both molecular and LLPS states. Using single- molecule and ensemble experiments, we demonstrated that The N-terminal oligomerization domain of NPM1 (Npm-N, residues 1-130), a necessary and sufficient element to trigger phase separation via binding to nucleolar partners, can interconvert between a globally disordered monomer and a b-sheet enriched pentamer state via a tunable mechanism. Our data are consistent with a folding-induced- assembly pathway at physiological ionic strength, while an assembly-induced-folding pathway is favored at low salt condition, showcasing a remarkable mechanistic plasticity in this protein system. Phosphoryl- ation in Npm-N results in pathway-specific effects in temporally decoupling the folding and assembly, thereby increasing the lifetimes of the intermediate states. Conversely, binding to nucleolar partners locked Npm-N in the ordered pentameric state promoting phase separation, and counter the effects of phosphorylation. The liquid demixing of Npm-N is mediated by the formation of binary protein-partner complexes via weak multi-valent interactions. Finally, single molecule experiments directly reveal subtle conformational rearrangements in Npm-N during LLPS, providing structural insights into the liquid demixed state. Together, our results demonstrate a rich conformational complexity at play in the biophysics of IDPs, driving their functional diversity.

A bacterial antitoxin’s conformational ensemble provides insight into distinct functional roles of its distinct disordered states Virginia M. Burger1, Alexandra Vandervelde, Albert Konijnenberg, Jelle Hendrix, Frank Sobott, Remy Loris, Collin M. Stultz 1Massachusetts Institute of Technology Intrinsic disorder plays a key role in the regulation of cell death by bacterial toxin-antitoxin (TA) modules. In TA modules, an unstable antitoxin normally inhibits a protein toxin. Cellular stress triggers increased degradation of the labile antitoxin, thereby releasing the toxin. The activated toxin then disrupts essential cellular processes, causing cell death or quiescence. The CcdA-CcdB TA module in E. coli consists of the antitoxin CcdA and the toxin CcdB. CcdA is comprised of a folded DNA-binding domain and two intrinsically disordered regions (IDRs), which regulate binding to CcdB. NMR studies suggest that CcdA, in the absence of CcdB, predominantly samples conformations belonging to either a closed

109 ABSTRACT or an open state, distinguished by the distance from the IDR termini to the folded domain. We use all- atom explicit-water molecular dynamics simulations, native ion mobility-mass spectrometry, and single- pair FRET to determine the conformational ensemble of unbound CcdA, with the goal of inferring functional roles from structural details. Our results indicate that CcdA samples metastable states of varying compactness, in which one IDR opens up away from the folded domain. Contrary to previously published CcdA structures, our data suggest that CcdA preferentially adopts compact structures in its unbound form. Further analysis of the proclivity of the IDRs for CcdB-bound conformations and the solvent exposure of predicted Lon-recognition sites within each state provides insight into the functional role of these states. As intrinsically disordered antitoxin proteins like CcdA are plentiful in bacteria, understanding how disorder facilitates their functions could lead to novel antibiotics that harness TA modules to kill bacterial cells.

Localized assembly of RNA granules by intrinsically-disordered proteins that bind RNA Deepika Calidas1, Jarrett Smith, Tu Lu, Dominique Rasoloson, Helen Schmidt and Geraldine Seydoux 1Johns Hopkins School of Medicine and Howard Hughes Medical Institute P granules are dynamic RNA granules that assemble in the posterior of the C. elegans 1-cell embryo. How P granule assembly is restricted to a specific region of the zygote cytoplasm is not known. Previ- ously, our lab identified a novel class of serine-rich, intrinsically-disordered proteins (MEG proteins) that localize dynamically to the periphery of each P granule (Wang et al, e-Life, 2014). We now show, using genome editing to generate truncation and deletion alleles in the meg genes, that the MEG proteins are essential assembly scaffolds for P granules that utilize distinct domains to recruit RNA and RNA- binding proteins into the granules. In vitro, recombinant MEG-3 binds RNA and the RNA-binding protein PGL-1, and undergoes liquid-liquid phase separation in a salt and concentration-dependent manner. In vivo, MEG-3 forms an anterior-low/posterior-high cytoplasmic gradient, and high levels of MEG- 3 correlate with regions of P granule assembly. The N-terminal region of MEG-3 (1-544), which has greater predicted intrinsic disorder, drives phase separation in vitro, and forms a gradient in vivo. How- ever, this region cannot recruit PGL-1 to granules in vivo. In contrast, the C-terminal region of MEG- 3(545-862) binds PGL-1 in vitro and recruits it to granules in vivo, but cannot phase separate or form a gradient. These findings suggest a multi-layered assembly process that involves both spatially-patterned phase separation and the formation of protein complexes.

110 ABSTRACT

Structural details of RNA-binding protein disordered domain phase separation in ALS and cancers Nicolas L. Fawzi1, Veronica H. Ryan, Alexander E. Conicella, Kathleen A. Burke, Abigail M. Janke 1Department of Molecular Pharmacology, Physiology, and Biotechnology, Brown University Phase separation of RNA-binding proteins via multivalent interactions between aromatic/polar-rich disordered domains is the basis for formation of functional cytoplasmic granules and nuclear puncta. These domains have also been identified as the nucleators of neuronal inclusions in amyotrophic lateral sclerosis (ALS) and frontotemporal dementia. Here, we describe our atomic resolution nuclear magnetic resonance spectroscopy approaches for visualizing low complexity domain structure and interactions along the pathway from monomer, to liquid-liquid phase separated state, to static aggregates and hydrogels. Our recent results on RNA-binding protein Fused in Sarcoma1 suggest that the low-complexity domain (FUS LC) remains disordered even within liquid phase separated states and recruits unphosphorylated RNA-polymerase II C-terminal domain into the liquid phase separated state, adding a potential explanation for FUS LC transcriptional activation in cancer. Importantly, phase separation is reversible and is modulated by ionic strength and interaction with RNA, distinguishing these assemblies from static inclusions. We also present our latest results expanding atomic resolution characterization to TDP-43 self-assembly associated with ALS and to the ability of post-translational modification to act as switch controlling disordered domain assembly and conversion to disease-associated aggregates. Reference: 1. Burke, K.A., Janke, A.M., Rhine, C.L., Fawzi, N.L. (2015). Residue-by-Residue View of in Vitro FUS Granules that Bind the C-Terminal Domain of RNA Polymerase II. Molecular Cell. 60, 231-41.

111 ABSTRACT

Autoregulation of the p53 Binding Activity of Mdm2 by Intrinsically Disordered Regions Ferrolino, M.C.1, Phillips, A.H.1, Kriwacki, R.W.1,2 1Department of Structural Biology, St. Jude Children’s Research Hospital, Memphis, TN 38103, USA, 2Department of Microbiology, Immunology and Biochemistry, University of Tennessee Health Sciences Center, Memphis, TN, 38163, USA Autoinhibition is a critical protein regulatory mechanism that often involves cis-interacting motifs. Auto- inhibitory modules that flank functional domains commonly contain intrinsically disordered regions (IDRs). Intrinsic disorder in inhibitory motifs provides an advantage of structural heterogeneity as well as larger interaction potential compared to ordered domains. Mdm2, the major negative regulator of tumor suppressor p53, contains long disordered regions. Mdm2 inhibits p53 activity by interacting with the transactivation domain (p53-TAD) and also functions as an E3 ubiquitin ligase that promotes p53 proteosomal degradation. The p53 binding domain (PBD) of Mdm2 is flanked by two IDRs, the autoinhi- bitory Lid at the N-terminus and the acidic domain (AcD) at the C-terminus. We have shown that, aside from the Lid, the AcD of Mdm2 modulates the interaction of Mdm2 with p53-TAD. This involves disordered regions within the AcD that weakly interact with the p53 binding pocket. The AcD is also known to directly interact with the RING domain to stimulate its ubiquitination activity. The ability of the AcD to interact with both the p53 binding domain and the RING domain provides a molecular basis for the allosteric link between substrate recognition and ubiquitination activity of Mdm2 toward p53. Our study illustrates the crucial role of IDRs in communication between functional domains in proteins.

Disorder within cysteine-rich protein and its implications for multifunctional roles: the curious case of Granulin-B. Gaurav Ghag1, Lauren M. Wolf2, Randi G. Reed1, Nicholas P. van der Munnik3, Claudius Mundoma4, Melissa A. Moss2,3 and Vijay Rangachari1 1Department of Chemistry and Biochemistry, University of Southern Mississippi, Hattiesburg, MS, 2Biomedical Engineering Program and, 3Department of Chemical Engineering, University of South Car- olina, Columbia, SC, 4Institute of Molecular Biophysics, Florida State University, Tallahassee, FL Granulins (Grns) are a family of small, cysteine-rich proteins that are generated upon proteolytic cleavage of their precursor, progranulin during inflammatory events. All seven Grns (A-G) contain twelve conserved cysteines ( 20%) that are believed to form six intramolecular disulfide bonds, rendering this family of proteins unique. Grns are involved in multi-functional roles, including wound healing, embryonic growth, and signal transduction, and are also implicated in neurodegenerative diseases such as frontotemporal dementia and Alzheimer disease. Despite their manifold functions, there exists a dearth of information regarding their structure-function relationship. Here, we sought to establish the role of disulfide bonds in structure and function by studying the native, oxidized (intramolecular disulfide bonds) form of GrnB and the completely reduced GrnB (rGrnB). We establish that both rGrnB and native GrnB are intrinsically disordered (IDP) at low concentrations. We also observed that the rGrnB forms a fuzzy homodimer without a net gain in the structure. Interestingly, rGrnB activates NF-kB in human neuroblastoma cells in a concentration-dependent manner, which correlates with the observed monomer-dimer dynamics. We observed that native GrnB constitute less than 10% of recombinantly expressed protein in E. coli, the rest being multimeric with intermolecular disulfide bonds. Our findings question the role of disulfide bonds in GrnB structure and functions. Efforts are underway to ascertain this by investigating the behavior of various forms of GrnB on inflammatory responses in glial cells.

Stabilization of a globular protein by a disordered tardigrade protein Aakash Mehta1, Samantha Piszkiewicz1, Thomas Boothby1, Bob Goldstein1, and Gary Pielak1 1University of North Carolina at Chapel Hill – Department of Chemistry Tardigrades, commonly known as “water bears,” are a phylum of microscopic animals that have evolved mechanisms to survive radiation levels 1,000-times greater than the average animal, the vacuum of

112 ABSTRACT outer space for 10 days, pressures up to six times that of the deepest ocean trenches, temperatures from 22728Cto1518C, and over 10 years of desiccation. Assays that examine the survival of tardigrades at different drying rates suggest that their macromolecules are not more stable than those of other microorganisms. Therefore, it is likely that other molecules aid in stabilization during dehydration. We have found that a novel class of disordered proteins in tardigrades, Cytoplasmic Abundant Heat Soluble (CAHS) proteins, plays a critical role in surviving desiccation-induced stress. CAHS proteins form heat-labile gels. We examined the effect of these gels on globular proteins stability using a model protein, the N-terminus SH3 domain of Drosophila drkN protein, which unfolds by a simple, two-state process at equilibrium. 19F NMR spectroscopy was used to quantify the population of the folded and unfolded forms of fluorine- labeled SH3. CAHS proteins were found to stabilize SH3 in a folded state. This observation suggests that CAHS proteins may be useful as pharmaceutical excipients for formulating protein-based drugs.

Nucleophosmin-mediated molecular networks reveal insights into the structural organization of the granular component of the nucleolus Diana M. Mitrea1, Christopher S. Stanley, Jaclyn A. Cika, Clifford S. Guy, David Ban, Priya R. Bannerjee, Amanda Nourse, Ashok A. Deniz, Richard R. Kriwacki 1St Jude Children’s Research Hospital The nucleolus is the site where ribosome are synthesized and assembled. It is a membrane-less organelle formed through liquid-liquid phase separation of its components (> 700 proteins, and various forms of nucleic acids) from the surrounding nucleoplasm. The nucleolus is further partitioned into three compartments: Fibrillar centers (FC), dense fibrillar component (DFC) and granular component (GC). Within these compartments, the pre-ribosomal RNA (rRNA) is synthesized, spliced and modified, and assembled with ribosomal proteins, respectively. Here, we show that nucleophosmin (NPM1), a major constituent of the GC, integrates within the nucleolus via a multi-modal mechanism, involving multivalent interactions with proteins containing arginine-rich linear motifs (R-motifs) and ribosomal RNA (rRNA). Importantly, these R- motifs are found in canonical nucleolar localization signals and are often present in multiple copies, which enable these proteins to phase separate with NPM1 and/or rRNA. Here, we identify multivalency of acidic tracts and folded nucleic acid binding domains, mediated by N-terminal domain oligomerization, as structural features required for phase separation of NPM1 with other nucleolar components in vitro and for localization within mammalian nucleoli. We propose that one mechanism of nucleolar localization involves collective phase separation of proteins and RNA within the nucleolus. Based on a novel combination of biophysical approaches, we propose a model for the molecular organization within liquid-like droplets formed by the N-terminal domain of NPM1 and R-motif peptides, thus providing the first of their kind insights into the structural organization of the nucleolus.

PL – MEMBRANE PROTEINS Solution NMR studies on membrane proteins in lipid bilayers Philipp Ansorge1, Oliver Zerbe1 1Department of Chemistry, University of Zurich, Winterthurerstrasse 190, CH-8057 Zurich G protein-coupled receptors (GPCRs) are membrane proteins that contain 7 trans-membrane alpha helices (TMs) and are of high pharmaceutical relevance. Unfortunately, structures of this class of receptors by NMR are rare because of their size and intrinsic conformational instability, their low expression yields, poor stability and tendency to aggregate. Bacteriorhodopsin (bR), a light-driven proton pump from Halobacteria, serves as a model protein to develop biochemical and spectroscopic methods for GPCRs. bR forms a stable 7 TM bundle, and can be expressed in high yields. The crystal structure of bR is published, but only flexible parts of the protein in detergent micelles have been assigned by solution NMR so far [1].

113 ABSTRACT

To avoid problems associated with artificial environments such as detergent micelles, we study bR inside a real lipid bilayer. Therefore the protein is incorporated into nanodiscs (NDs) that are optimized for solution NMR studies [2]. We have used bR for the optimization of biochemical protocols for expression and reconstitution of 7-TM proteins into NDs, to study back-exchange of amide protons in TM helix locations, and to optimize spectroscopic tools. In addition, we have developed labeling schemes and assignment methods that aim at obtaining a sufficient number of restraints for structure calculations of real GPCRs. This includes various methyl-labeling strategies. We could demonstrate that bR in NDs presents a sample that is stable at 478C for weeks, and provides TROSY and NOESY spectra of high quality that display approximately the correct number of peaks, including TM residues. References: 1. M. Schubert, M. Kolbe, B. Kessler, D. Oesterheldt, P. Schmieder, ChemBioChem, 2002, 10, 1019-1023 2. F. Hagn, M. Etzkorn, T. Raschle, G. Wagner, JACS, 2013, 103, 1919-1925

Molecular explanations for metal selectivity and the conformational change process in Nramp- family divalent metal transporters Aaron Bozzi1, Lukas Bane, Wilhelm Weihofen, Brandon Lee, Rachelle Gaudet 1Harvard University, USA Nramp (Slc11) transporters secure the transition metal ions that all organisms require as enzymatic cofactors. We have determined the crystal structure of a bacterial Nramp homolog showing an inward-

114 ABSTRACT facing LeuT-fold, and used cysteine accessibility measurements to identify the outward metal permeation pathway. From investigating the transport impairment mechanisms of two anemia-causing glycine- to-arginine disease mutants, we explain the Nramp conformational change process. The G45R mutant acts as a steric wedge to block the critical movement of TM1a, thus locking the transporter in the inward-facing state. The G153R mutation on TM4 interferes with the extracellular gating process and perturbs the substrate specificity of the metal binding site. We further explore Nramp metal selectivity by demonstrating that a conserved metal-binding methionine is not essential for transport of Nramp’s physiological substrates like iron and manganese, as mutations to alanine and other small residues retain function. Instead, we provide a wealth of functional data to illustrate an apparent evolutionary tradeoff in which the methionine’s unique sulfur ligand serves to reject abundant alkaline earth metal competitor ions while promoting uptake of the toxic metal cadmium.

Heterologous expression of teleost trace amine-associated receptor fused with N-terminal sequence of rodopsin Mi-Jin Choi1, Jong-Myoung Kim1 1PuKyong National University, Busan, Republic of Korea Olfaction is a chemosensory perception important for feeding, reproduction, migration and predator avoidance in teleost. Among four types of olfactory receptors, trace amine-associated receptor (TAAR) is responsible for detecting the trace amine-related molecules. Gene encoding TAARs were obtained from giant grouper (Epinephelus lanceolatus), a commercially important fish species in Asia. Teleost TAAR consists of 324 amino acids with a predicted seven transmembrane domains and conserved motifs (NPXXY and DRY) common to GPCR. Full-length TAAR gene was amplified from cDNA template and modified to contain sequences encoding HA tag (YPYDVPDYA) and Hist6 sequences inserted to its N-terminus and C-terminus, respectively, to identify the orientation of the heterologously expressed TAAR protein. To further facilitate the expression of the TAAR, chimeric receptors were constructed by replacing the N-termini of the TAARs with 20 (RhoI) or 70 (RhoII) amino acids of bovine opsin. Genes encoding the modified TAARs were clone into an expression vector pMT4 to analyze its functional properties and the ligands. Upon tranfection into HEK293T cell, expression levels and the topological orientation of TAARs were analyzed by western blotting and immunohistochemical staining. Expression level of the TAAR chimeric receptor containing RhoII sequence (N-terminal 70 a.a. of bovine opsin) was higher than that those of the chimeric receptor containing Rho I sequence (N-terminal 20 a.a. of bovine opsin), and wild type TAAR. The results suggest that the level of functional expression of TAARs can be enhanced by an addition of the N-terminal sequence of rod opsin.

Distinct structural elements govern folding, stability and catalysis in the outer membrane enzyme PagP Bharat Ramasubramanian Iyer1 Radhakrishnan Mahalakshmi1 1Molecular Biophysics Laboratory, Department of Biological Sciences, Indian Institute of Science Edu- cation and Research, Bhopal, India. The outer membrane enzyme PagP of Gram-negative bacteria palmitoylates the lipid A component of lipopolysaccharides, and is therefore implicated in resistance to host immune defenses. PagP homologs from Escherichia coli (PagP-Ec) and Salmonella typhimurium (PagP-St) share >80% sequence identity, and are anticipated to display similar functional and thermodynamic properties. Surprisingly, our equilibrium unfolding analyses reveal that the PagP homologs exhibit differences in folding and stability, specifically suggesting that PagP-Ec is more stable than PagP-St, by two-fold. This stability is offset by the catalytic efficiency in both proteins, with PagP-St showing 20 fold higher enzymatic activity than PagP-Ec (Iyer, BR and Mahalakshmi, R. (2015) Biochemistry 54:5712-5722). We present the implications of subtle sequence variations between both proteins on their biophysical properties and catalytic

115 ABSTRACT function. Our findings reveal that the effect of residue substitutions on the barrel stability and catalytic efficiency is non-additive in both PagP-Ec and PagP-St. Thermodynamic analyses suggest that even single residue mutations can considerably affect the cooperativity of folding, and resistance to denaturation of the polypeptide. The outcome of the mutation is influenced by the residue position in the folded barrel, with residues distal from the catalytic site having an impact on the catalytic mechanism. Our study also reveals that the N-terminal helix plays an important role in the switch in the thermodynamic stabilities of both PagP-Ec and PagP-St, but has no effect on the catalytic efficiency of the barrel. On the contrary, select point mutations can considerably alter the activity of the protein without influencing scaffold stability. While the basic characteristics of outer membrane b-barrels are retained in both proteins, a tradeoff between activity and stability emerges from the biophysical and functional studies of these homologous enzymes. These results provide insight on the distinct factors that govern the folding and stability of membrane proteins and the role of key residues in PagP function and recy- cling in the bacterial outer membrane.

Characterization of the interaction between cyt c and cardiolipin-incorporated bicelles by solution NMR. Hisashi Kobayashi1, Satoshi Nagao1, Shun Hirota 1 1Graduate School of Materials Science, Nara Institute of Science and Technology, 8916-5 Takayama- cho, Ikoma, Nara, 630-0192, Japan Cytochrome c (cyt c) is a heme protein located in the mitochondrial intermembrane space, where it transfers electrons in the respiratory chain. Cyt c also plays an important role in apoptosis by its release to the cytosol. Cyt c associates with the inner mitochondrial membrane by interaction with cardiolipin (CL), which may increase the peroxidase activity of cyt c. Although the cyt c2CL interaction may alter the cyt c function between electron transfer and peroxidase activity, the detailed mechanism of the function conversion is unclear. Solution NMR is one of the most powerful methods for analyzing molecular interactions in proteins. However, it is difficult to study membrane-associated proteins by solution NMR due to their high molecular weight. In this study, we investigated the CL-interaction site of cyt c by solution NMR using small isotropic bicelles containing CL. 1H–15N Heteronuclear Single Quantum Correlation spectra of cyt c were measured in the absence and presence of CL-containing bicelles. The backbone NH signals of the N- and C-terminal helices and the region around K72-K73 exhibited substantial chemical shift perturbations by an addition of CL- containing bicelles, whereas the signals did not shift by an addition of bicelles not containing CL. These results indicate that a specific surface area containing the N- and C-terminal helices and the region around K72-K73 of cyt c interacts with CL in bicelles.

Cell-Free Protein Expression for Soluble Expression of Functional Class A G Protein Coupled Receptors Lisa Maria Ko€gler1, Max Steinhagen2, Anette Kaiser1, Jan Stichel1, Annette G. Beck-Sickinger1 1Institute of Biochemistry, Leipzig University, Bruderstraße€ 34, 04103 Leipzig, Germany, 2Institute of Microbiology, TU Dresden, Helmholtzstr. 10, 01069 Dresden, Germany To investigate structure, binding modes and molecular dynamics in proteins spectroscopic methods like nuclear magnetic resonance (NMR) spectroscopy or electron paramagnetic resonance (EPR) spectroscopy are useful techniques. However, these methods remain challenging for G protein coupled receptors (GPCR). Their hydrophobic and flexible character leads to difficulties obtaining GPCR in a soluble, homogenous form in concentrations suitable for structural investigation. Furthermore, site-specific labeling is required to insert isotopic or spin label. In contrast to other expression systems, the cell-free protein expression offers some advantages. It is a fast and directed expression and it is possible to modulate the reaction environments with additives like detergents or lipids for a soluble expression of

116 ABSTRACT membrane proteins. Furthermore, the reaction conditions, like the amino acid pool, can be fully controlled, which enables an efficient incorporation of unnatural amino acids. For GPCR expression we used a continuous exchange cell free (CECF) system suitable for coupled in vitro transcription/translation. As source for the required cellular components, a S30 cell extract from Escherichia coli BL21(DE3) was prepared. Five different class A GPCR, the human neuropeptide Y receptor type 1 (hY1R), type 2 (hY2R) and type 4 (hY4R), as well as the C-X-C chemokine receptor type 4 (CXCR4) and the Chemokine like receptor 1b (CMKLR1b) were expressed by using this system. The expression was successful for all five tested class A GPCR. By addition of mild detergents during expression soluble GPCR could be obtained directly. Furthermore, ligand binding properties could be verified for about 25-30% of individually expressed receptor. In addition, functional receptors could be separated from unfolded ones. This enables cell-free protein expression for the production of GPCR for spectroscopic measurements.

A new Rosetta refinement algorithm to improve membrane protein structures Julia Koehler Leman1 1John Hopkins University, USA Despite their importance, membrane proteins are extremely challenging to study and restraints for structure determination are typically sparse or of low resolution. Further, their structures are influenced by the inhomogeneous membrane environment. When membrane protein structures are determined by different techniques in different environments, the naturally raised question becomes ‘which structure is correct’? Towards answering this question, we compiled a database of membrane proteins that have known structures determined by both NMR and crystallography. We then investigated their differences in backbone RMSDs, convergence within the NMR ensemble, straightness within the transmembrane region, stereo-chemical correctness, and packing. After quantifying these differences, we used a newly developed Rosetta high-resolution refinement protocol on the NMR structures in an attempt to reduce them. While many questions concerning the influence of membrane mimetics, specific lipids and inherent features of structure determination remain unanswered, our results pave the way for improving the structural quality of membrane proteins.

Application of the dual split protein reporter to a high throughput screening of potential fusion inhibitors of HIV-1 envelope protein Mizuki Yamamoto1,2, Yasushi Kawaguchi2,3, Jun-ichiro Inoue1,2, Zene Matsuda2,4 1Division of Cellular and Molecular Biology, 2Research Center for Asian Infectious Diseases, 3Division of Molecular Virology, Institute of Medical Science, the University of Tokyo, Tokyo, Japan, 4Laboratory of Structural Virology and Immunology, Institute of Biophysics, Chinese Academy of Sciences, Beijing, China HIV-1 is a causative agent of AIDS affecting more than 35 million people worldwide. Some inhibitors targeted at membrane fusion are already in clinical use, and they block the first essential step of HIV-1 replication, membrane fusion. In this study, to identify more new fusion inhibitors, we set up a cell- based screening system using the split protein reporter, called dual split protein (DSP). The DSP assay relies on a pair of chimeric split reporters expressed separately in effector and target cells. Each DSP (DSP1-7 or DSP8-11) is a split chimeric protein between Renilla luciferase (RL) and GFP. Membrane fusion can be detected by monitoring the recovery of either GFP or RL activity during the merge of the effector and target cells. Here we established the effector cells constitutively expressing the HIV-1 envelope proteins with a different coreceptor tropism together with DSP1-7. Using target cells stably expressing CD4, the respective coreceptors (CXCR4 or CXCR4/CCR5), and DSP8-11, we screened 1017 drugs that are already in clinical use. The degree of inhibition of membrane fusion was measured from the RL activity in live cells using membrane-permeant substrate for RL. The specificity of our assay was validated, because Maraviroc, a known CCR5 antagonist, was identified as a potent inhibitor of the

117 ABSTRACT

CCR5-tropic envelope in our DSP assay. We found several hits (including Maraviroc) and are currently analyzing them.

Essential Oil Antifungals: the Search for Novel Targets Brittney L. Murray1, Robert L. McFeeters1 1The University of Alabama in Huntsville, USA Fungi are ubiquitous microorganisms capable of living in soil, food, water, and colonizing the human body. Pathogenic fungi lead to the death of millions of individuals globally each year. In addition to having significant deleterious side effects, current antifungals are becoming less effective due to the development of resistance. Consequently, the need for new antifungals and new antifungal targets is urgent and growing. Essential oils are a rich source and underdeveloped source of antifungal compounds capable of limiting the growth of or completely killing fungi. This study presents the antifungal activity of essential oils against the pathogenic fungi Aspergillus niger, Cryptococcus neoformans, and Candida albicans. Following the discovery of inhibition and isolation of active compounds, identification of targets will follow with an emphasis on discovering novel extracellular glycoproteins susceptible to small molecule inhibition.

The intrinsically disordered membrane enzymes selenoprotein S and Selenoprotein K Jun Liu, Zhenqui Zhang, Sharon Rozovsky1 1University of Delaware, UJSA Selenoproteins constitute a family of enzymes involved in the management and regulation of reactive oxygen species, signaling molecules that are also affiliated with molecular damage and disease. The family contains two intrinsically disordered membrane proteins: selenoprotein S (SelS) and selenoprotein K (SelK). Both are members of the Endoplasmic Reticulum Associated Protein Degradation (ERAD) pathway, which is responsible for dislocation of misfolded proteins from the ER for degradation in the cytoplasm. We have shown that SelK and SelS belong to the very small group of intrinsically disordered proteins that exhibit enzymatic functions. We demonstrate that SelS is an efficient disulfide reductase. It also influences the activity and conformation of its protein partner, the AAA ATPase valosin-containing protein (VCP) p97, that provides the energy to translocates misfolded proteins. We recently discovered that SelK is able to cleave its own peptide bond, releasing a selenocysteine–containing peptide, and thus terminating enzymatic activity. We propose that this autoproteolysis is a regulatory mechanism responsible for SelK associations with different membrane complexes. We discuss the unprecedented contribution of selenocysteine to the peptide bond cleavage.

118 ABSTRACT

Role of Transmembrane Domain of Arabidopsis CRINKLY4 Receptor-like Kinase (ACR4) in a Membrane-like Environment Shweta Shah1, A. Gururaj Rao1 1Roy J. Carver Dept. of Biochemistry, Biophysics and Molecular Biology, Iowa State University ACR4 is a Ser/Thr receptor like kinase whose architecture consists of an extracellular ligand binding domain, a transmembrane (TM) helix and an intracellular domain (ICD). Most recent studies on the ICD of this protein have established the importance of the TM and its molecular association with plasma membrane that in turn affects binding of downstream signaling partners. The main objective of this study is to characterize and compare the function of the ICD of ACR4 with/without TM domain. To this end, we have recombinantly expressed active and soluble TM domain fused ICD protein (SUMO-TM-ICD) in pE-SUMO vector. While the far UV-CD spectra suggest that proteins are properly folded, differences in thermal stability of SUMO-TM-ICD and SUMO-ICD indicate that the TM domain may provide structural stability to the kinase domain. Stability is also reflected in limited proteolysis studies. The time course of autophosphorylation, monitored by Phos-tagTM, indicates clear differences in the phosphoryation pattern when the ICD is separated or fused with the TM domain. Autophosphorylation site mapping of TM-ICD by LC/MS/MS identified 20 different phosphorylation site, among which are two tyrosine phosphorylation sites that suggests that in vitro kinase activity of ACR4 may not be restricted to Ser/Thr residues. The alanine mutants and phosphomimetic glutamic acid mutants of these Tyr residues are inactive, while phenylalanine mutants are able to restore kinase activity, indicating the structurally important role of the Tyr residues. Furthermore, to mimic the natural membrane bound environment, we have prepared functionally active SUMO-TM-ICD Nanodiscs. These nanodiscs are being utilized to characterize SUMO- TM-ICD in its near-native environment and to effectively elucidate protein-protein interaction that are mediated through TM domains.

Using thermodynamics to cure disease – nucleotide analogs provide significant correction of the temperature-dependent defect in F508del-CFTR Chi Wang1, Andrei A. Aleksandrov2, Zhengrong Yang3, Elizabeth Proctor2, Pradeep Kota2, Jianli An3, Farhad Forouhar4, Gregory Boel1, Nikolay V. Dokholyan2, John R. Riordan2, Christie G. Brouillette3, John F. Hunt1,4 1Department of Biological Sciences, Columbia University, New York, NY 10027, USA, 2Department of Biochemistry and Biophysics, School of Medicine, University of North Carolina, Chapel Hill, NC 27599, USA, 3Department of Chemistry, University of Alabama, Birmingham, AL 35294, USA, 4Northeast Structural Genomics Consortium, Columbia University, New York, NY 10027 Cystic fibrosis (CF) is a chronic lethal genetic disease, which is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) protein. Deletion of a single phenylalanine residue at position 508, F508del is the most common lethal mutation found in CF patients1-3. Fully synthesized F508del-CFTR is rapidly degraded in the endoplasmic reticulum due to F508del-induced misfolding, preventing proper transport of the protein to the epithelial cell membrane4-6. Previously, we have demonstrated that F508del facilitates partial unfolding and consequent aggregation of the first nucleotide-binding domain in human CFTR (hNBD1)7-9. This F508del-faciliated aggregation is likely a key contributor in F508del-CF pathogenesis. These findings established a foundation to develop a new high-throughput biophysical screen to identify compounds that directly bind to the hNBD1 domain to prevent the F508del-hNBD1 aggregation in vitro and thereby stabilize F508del- CFTR in vivo. We used computational protein engineering10,11 to construct an equivalently stable hNBD1 variant that covalently binds to a visible fluorophore at a single labeling site. Fluorescence self-quenching of this labeled domain upon unfolding/aggregation provides a robust and accurate assay for corrector compounds. We demonstrated hNBD1 interaction with nucleotide analogs, which

119 ABSTRACT bind 20 Å away from F508. Unexpectedly, several common nucleotide analogs bind 3-fold tighter than ATP, the physiological ligand of hNBD1. These analogs offset half of the defect in thermal stability in F508del-hNBD1 and preserves the gating function of F508del-hCFTR channels even at 37 ˚C where ATP cannot. The results strongly suggested these analogs merit further exploration as cystic fibrosis drugs, and future coordinated crystallographic and isothermal titration calorimetry studies of the analogs will provide valuable guidance for development of compounds with enhanced affinity to the channel. References: 1. Riordan, JR et al. Science 245, 1066, 1989 2. Riordan, JR Annu. Rev. Biochem. 77, 701, 2008 3. Hunt, JF et al. Cold Spring Harb. Perspect. Med. 3, 1, 2013 4. Denning, G M et al. Nature 358, 761, 1992 5. Yang, Y et al. PNAS 90, 9480, 1993 6. Cheng, SH et al. Cell 63, 827, 1990 7. Lewis, HA et al. JMB 396, 406, 2010 8. Protasevich, I et al. Protein Sci. 19, 1917, 2010 9. Wang, C. et al. Protein Sci. 19, 1932, 2010 10. Yin, S. et al.Nat. Methods 4, 466, 2007 11. Yin, S. et al. Structure 15, 1567, 2007

Enhanced resolution in size exclusion chromatography for separation of proteins with Mr 3000- 70 000 Berit Zade1, Lars Andersson1, Karin Torstenson1, Staffan Lindqvist1 and Jon Lundqvist1 1GE Healthcare Bio-Sciences AB, Bjorkgatan€ 30, SE 751 84 Uppsala, Sweden The elucidation of protein function and the role of protein-protein interactions is an area of intense research efforts, where chromatographic methods are used for analytical and preparative separations. A common challenge is to separate proteins with similar charge and size properties. High resolution size exclusion chromatography (SEC) is the method of choice for detection and quantitation of protein aggregation or truncation. SEC offers separation by size under mild, non-dissociative conditions and can be used for both isolation and analysis of proteins. Here, data from applications using a new SEC resin, SuperdexTM 75 Increase are shown. The design is made for rapid separation and analysis of proteins with molecular weights ranging from Mr 3000 to 70 000. Results obtained with Superdex 75 Increase, showed that separations are (i) up to three times quicker with preserved resolution and (ii) give up to 50% higher resolution with the same flow rate compared to the predecessor Superdex 75. For protein preparations, it is essential that the protein does not aggregate, oligomerize, or degrade. With Superdex 75 Increase 5/150 GL (aimed for rapid SEC analysis with run times of 6-10 min), aggregation tendencies and other size changes can be studied under different conditions, for example, for storage stability analyses of biopharmaceuticals. We used the column to monitor small changes in size homogeneity in a large number of storage samples under different conditions and over different time periods (weeks). Separation between oligomeric and monomeric forms of a histidine tagged protein was successfully performed using the Superdex 75 Increase 10/300 GL column. The purification gave 5 mg of essentially size-homogeneous target protein that was obtained in a single SEC run. In that purification work, the 5/ 150 GL column format was used to rapidly determine which of the collected fractions could be pooled (monomeric protein) and used for further characterization.

120 ABSTRACT

PM – MOTORS AND MACHINES Mechanisms and Applications of R bodies, Membrane-Breaking Protein Needles Jessica Polka1, Pamela Silver1 1Department of Systems Biology, Harvard Medical School, 2Wyss Institute for Biologically Inspired Engineering Background: R bodies are polymeric protein structures produced in the cytoplasm of bacterial endo- symbionts of Paramecium, where they function as toxin delivery devices. At cytoplasmic pH, they resemble a coils of ribbon 500nm in diameter and approximately 400nm deep. At low pH, however, they violently and reversibly extend to form hollow needles with pointed ends that can be up to 20um long. Despite their scale, they are assembled from just two small (100aa) proteins. Methods: To understand R body assembly, we used Tandem Mass Tag Mass Spec (TMT-MS) and mutagenesis to identify regions involved in a covalent assembly mechanism. We employed Structured Illumination Microscopy (SIM) to watch super-resolution dynamics of the structural and catalytic R body components. To understand pH-driven extension, we employed a high-throughput visual screen to identify mutants defective in pH response. We then characterized these mutants with circular dichroism, phase contrast timelapse microscopy, and negative stain EM to identify the role of specific amino acids in pH- mediated conformational changes. Results: The R body assembly process requires specific lysine residues, suggesting that the covalent assembly process relies on isopeptide bond formation. We also find evidence for a molecular counting mechanism and present data suggesting that a protein required for R body assembly – RebC – forms a long-range assembly template. Our mutagenesis implicates a small unstructured region in a helical transition that drives R body extension. Video microscopy reveals that this conformational change propagates like a wave through the R body polymer. Conclusions: Taken together with structural data, our results lead us to propose a curvature-driven model for R body extension. We also use R bodies as tools to rupture foreign membranes, triggering the release of encapsulated cargo. Further uses of R bodies as intracellular delivery agents are also explored.

PN – PEPTIDES Uptake Mechanism of the Cell-penetrating Peptide pVEC: How Does the Hydrophobic N-terminus Contribute? Begum Alaybeyoglu1, Alper Duranel2, Elif Ozkirimli1,2 1Department of Chemical Engineering, Bogazici University, 34342, Istanbul – Turkey, 2Computational Science and Engineering Program, Bogazici University, 34342, Istanbul – Turkey Peptide-based drugs are promising alternatives to small molecule drugs due to their high specificity, target affinity, and low toxicity. Although these drugs have been favored, their use as therapeutics

121 ABSTRACT against intracellular targets is limited because of their larger size and hydrophobicity restricting their uptake. The discovery of cell penetrating peptides (CPPs), which can translocate through the cell membrane without the need for a receptor, has accelerated the studies on peptide-based drugs and novel peptide drugs that target intracellular proteins have been suggested. pVEC (LLIILRRRIRKQAHAHSK), a cell-penetrating peptide derived from murine VE-cadherin [1], was shown to carry proteins and oligomers as cargo [1, 2]. The significance of the N-terminal residues (LLIIL) was discovered by mutational analysis that resulted in 50-75% reduction in uptake into human Bowes melanoma cells [2]. In our previous studies, we have shown that a novel chimeric peptide encompassing the N-terminal residues of pVEC and a beta-lactamase inhibitory peptide, acts as a potential beta-lactamase inhibitor that can transport across the cell wall to inhibit its target protein [3]. In an effort to characterize the translocation of pVEC and the contribution of its hydrophobic N-terminus to its uptake, we performed steered molecular dynamics (SMD), and replica-exchange umbrella sampling simulations to compare the free energy profiles for the membrane translocation of pVEC and del5 pVEC (RRRIRKQAHAHSK). We repeated the MD simulations with different peptide/lipid ratios to examine the concentration dependence of spontaneous adsorption and translocation of pVEC. We believe expanding the knowledge base of membrane active peptides will accelerate and advance peptide based drug discovery studies. Acknowledgement: This project was funded by TUBITAK 114M179. References: 1. A. Elmquist, M. Lindgren, T. Bartfai, U. Langel, VE-cadherin-derived cell-penetrating peptide, pVEC, with carrier functions, Exp Cell Res. 269 (2001) 237–244. 2. A. Elmquist, M. Hansen, U. Langel, Structure-activity relationship study of the cell-penetrating peptide pVEC, Biochim Biophys Acta. 1758 (2006) 721–729. 3. B. Alaybeyoglu, B. Sariyar Akbulut, E. Ozkirimli, A novel chimeric peptide with antimicrobial activity, Journal of Peptide Science. 21 (2015) 294–301.

Developing Novel Peptide Scaffold for Protein-Protein Interactions (PPIs) Inhibition Bobo Dang1, Vikram K. Mulligan2, Marco Mravic1, Yibing Wu1, Thomas Lemmin1, David Baker2, William F. DeGrado1 1Dept. of Pharmaceutical Chemistry, UCSF, 2Dept. of Biochemistry, University of Washington Protein protein interactions (PPIs) function in many aspects of biological processes, [1,2] and often underlie serious human diseases. [3] Thus, developing strategies to inhibit protein-protein interactions is of clinical significance. Peptides, despite their high selectivity, good efficacy and tolerability, have been difficult to develop as therapeutics due to their instability, lack of defined structure and poor cell permeability. [4] Here, we propose to develop a template-constrained tricyclic peptide as a new class of organic-protein hybrid scaffold with considerable potential to inhibit complex protein-protein interfaces. In this scaffold, we use macrocyclization and crosslinking to constrain the peptide to adopt an exquisitely well-defined tertiary structure which has not been seen before. Cyclization and crosslinking renders these scaffold based peptides much more resistant to protease digestion. [4] The exquisitely well-defined secondary and tertiary structures provide many different surfaces for potential binding with proteins to inhibit protein-protein interactions. The size of this scaffold peptide (50 amino acids) is smaller than almost all natural proteins, thus it gives us the power to produce the molecule relatively easily either through chemical synthesis or protein expression, and we also have total control to modify the molecules beyond natural occurring amino acids. Through computational design, [5,6] phage display [7,8] and chemical protein synthesis [9], we plan to develop this intrinsically stable and versatile peptide scaffold as a robust platform for the discovery of various PPIs inhibitors. To showcase the effectiveness and power of our designed peptide scaffold, we aim to develop an

122 ABSTRACT inhibitor for PD-1, and ultimately disrupt the interaction of PD-1 and PD-L1, a clinically significant PPI in cancer References: 1. Lu, H. C., Fornili, A., Fraternali, F. Expert Rev. Proteomics, 10, 511–520 (2013). 2. Pawson, T., Nash, P. Genes Dev., 14, 1027–1047 (2000). 3. Ryan, D. P., Matthews, J. M. Curr. Opin. Struct. Biol., 15, 441–446 (2005). 4. Fosgerau, K., Hoffmann, T. Drug Discov. Today., 20, 122-128 (2015). 5. Rohl, C. A., Strauss, C. E., Misura, K. M., Baker, D. Methods Enzymol., 383, 66-93 (2004). 6. Das, R., Baker, D. Annu. Rev. Biochem., 77, 363–382 (2008). 7. Smith, G. P. Science, 228, 1315-1317 (1985). 8. Heinis, C., Rutherford, T., Freund, S., Winter, G. Nat. Chem. Biol., 5, 502-507 (2009). 9. Kent, S. B. H. Chem. Soc. Rev., 38, 338-351 (2009). 10. Pardoll, D. M. Nat. Rev. Cancer, 12, 252-264 (2012).

Leucine and isoleucine contribute differently to the binding of amphipathic peptides to membranes Antje Pokorny1, Mia A. Rosenfeld1, Paulo F. Almeida1 1University of North Carolina Wilmington, USA Lysette is a 22 amino acid peptide derived from staphylococcal d-lysin that forms an amphipathic a- helix when bound at membrane-water interfaces. We previously found that the experimentally determined DG8 of binding for lysette is more favorable than that predicted by the Wimley-White interfacial hydrophobicity scale by about 4 kcal/mol. Closer investigation of the amino acid composition of other peptides that are well described by the Wimley-White interfacial scale, such as melittin, led us to hypothesize that a preponderance of isoleucine (Ile) over leucine residues (Leu), as found in lysette, may be responsible for the deviation from the Wimley-White prediction. To test our hypothesis, we designed lysette variants lysette-I and lysette-L. In lysette-I, all Leu residues were replaced by Ile, and in lysette-L, all Ile were replaced by Leu. We also designed a melittin variant, iso-melittin, in which all Leu residues were replace by Ile. Peptide-lipid equilibrium dissociation constants and helicities of peptides bound to zwitterionic phosphatidylcholine (POPC) vesicles were determined by stopped-flow fluorescence and circular dichroism. If the hypothesis were correct, Lysette-I and iso-melittin should bind significantly better to zwitterionic bilayers than their Leurich counterparts, if corrected for the helicities of the bound states. We found that both lysette-I and iso-melittin bound significantly better to POPC bilayers than predicted by the Wimley-White interfacial scale, whereas the Leu-rich variants bound as predicted. Acknowledgement: This work was supported in part by NIH grants AI088567 and GM072507

Comparison of Solid-Phase Extraction and Size Exclusion Chromatography applied for preparation of urine samples from pregnant women for LC-MSMS V.A. Shirokova1, A.E. Bugrova1,4, N.L. Starodubtseva1,2, A.S. Kononikhin1,2, Z.S. Khodzhaeva1,K. Muminova1, I.A. Popov1,2,3, V.E. Frankevich1, E.N. Nikolaev2,3,4, G.T. Sukhikh1 1Moscow Institute of Physics and Technology, Moscow, Russia;, 2V. I. Kulakov Research Center for Obstetrics, Gynecology and Perinatology, Ministry of Healthcare of the Russian Federation, Moscow, Russia;, 3V.L. Talrose Institute for Energy Problems of Chemical Physics, Russian Academy of Sciences, Moscow, Russia;, 4Emanuel Institute for Biochemical Physics, Russian Academy of Sciences, Moscow, Russia. Purpose: Nowadays urine peptidome profiling is widely used for studying various diseases which affect metabolic ways and/or kidney reabsorption. Our group has been researching a complex pregnancy disorder known as preeclampsia (PE). This pathology manifests with high blood pressure and proteinuria.

123 ABSTRACT

Looking for potential biomarkers we have tried two different protocols of peptides extraction from pregnant women’s urine: solid-phase extraction and size exclusion chromatography in order to choose the suitable one. Whereas the first one gave poor results, the second one showed significant advantage. Experimental description: Urine samples from women with mild and severe preeclampsia and the control group of healthy pregnant women were collected at the V. I. Kulakov Research Center of Obstetrics, Gynecology and Perinatology. All patients included in the study provided written informed consent. All procedures and study methods were approved by the Commission of biomedical ethics at V.I. Kulakov Research Center for Obstetrics, Gynecology and Perinatology. They were united in three pools of 10 samples, mixed and stored at 2808C. Beforehand we used a urine sample from a healthy person to establish the protocol. In the first stage of preparation each aliquot of sample was centri- fuged in Centrisart ultracentrifugation filter devices (Sartorius, Gottingen, Germany) to remove proteins with a molecular weight cut-off 20kDa. Then the filtrate was processed using Strata-X cartridges by Phe- nomenex or PD-10 desalting columns by GE Bioscience. Finally, it was lyophilized, diluted and analyzed on MALDI-TOF mass spectrometer and HPLC-MS/MS. Results: The method based on size exclusion chromatography showed a significant increase in the number of peptide identifications, than the one based on solid-phase extraction. Notably, it allowed finding several specific PE peptides, for example, fragments of collagens [1]. Conclusion: In case of urinary biomarker search, it is necessary to remove contaminants (e.g. salts) and extract the maximum of peptides. We eventually choose preparation via size exclusion chromatography because it fits these requirements better than other extraction techniques. Acknowledgement: Russian Science Foundation (grant No. 16-14-00181). Reference: 1. Kononikhin AS, Starodubtseva NL, Bugrova AE, Shirokova VA, Chagovets VV, Indeykina MI, Popov IA, Kostyukevich YI, Vavina OV, Muminova KT, Khodzhaeva ZS, Kan NE, Frankevich VE, Nikolaev EN, Sukhikh GT. An untargeted approach for the analysis of the urine peptidome of women with preeclampsia. J Proteomics. 2016 Apr 21. pii: S1874-3919(16)30145-2. doi: 10.1016/j.jprot.2016.04.024.

Developing potent and specific inhibitors of the Grb7 breast cancer target using phosphotyrosine mimetics and bicyclic peptides Gabrielle Watson, Menachem Gunzburg1, Ketav Kulkarni1,3, Katie Cergol2, Richard Payne2, Patrick Perlmutter3, Matthew Wilce1, Jackie Wilce1. 1Department of Biochemistry and Molecular Biology, Monash University VIC, 2Department of Chemis- try, The University of Sydney, NSW, 3Department of Chemistry, Monash University VIC Signalling pathways governing proliferation and migration are frequently deregulated in cancerous cells and are an attractive target for the development of novel therapeutics. Growth factor receptor bound protein 7 (Grb7) is an intracellular adaptor protein with an established role in these signalling processes. Due to the role of Grb7 and its overexpression in an abundance of cancers, Grb7 has been identified as an exciting and opportune therapeutic target. The non-phosphorylated cyclic peptide, G7-18NATE, binds to the Grb7 SH2 domain, blocking Grb7 interactions with its upstream binding partners. Cell permeable G7-18NATE has been shown to inhibit pancreatic cell migration as well as reducing cellular growth and migration in breast cancer cell lines. The peptide binds to Grb7 with moderate affinity (18 lM); therefore, to improve on this, derivatives have been developed and tested using a structure-based drug design approach. G7-18NATE derivatives with either a carboxymethylphenylalanine (M1) or caboxyphenylalanine (M2) substitution have 3-fold and 9-fold increased affinity for the Grb7 SH2 domain compared with G7-18NATE, respectively. The X- ray crystal structure of the Grb7 SH2: M1 complex clearly reveals additional protein- peptide

124 ABSTRACT interactions occurring in the Grb7 phosphotyrosine binding pocket4. Bicyclic derivatives with varying additional linkages produce nM affinity with the structural analysis identifying unexpected modes of binding occurring at an alternate interface to the phosphotyrosine mimetic. Combining both these strategies, and hence covering a broad Grb7 interface, has produced a tight binding inhibitor of Grb7 (KD 5 380 nM) and marks considerable progress in Grb7 based anti-cancer drug development. Developing these G7-18NATE derivatives will establish fundamental methods that can be readily applied to other intracellular therapeutic targets. References: 1. Stein, D., Wu, J., Fuqua, S.A., Roonprapunt, C., Yajnik, V., D’Eustachio, P., Moskow, J., Buchberg, A.M., Osborne, C.K. and Margolis, B. (1994) The SH2 domain protein GRB-7 is co-amplified, overexpressed and in a tight complex with HER2 in breast cancer. EMBO J. 13, 1331 2. Tanaka, S., Pero, S.C., Taguchi, K., Shimada, M., Mori, M., Krag, D.N., and Arii, S. (2006) Specific peptide ligand for Grb7 signal transduction protein and pancreatic cancer metastasis. J. Natl. Cancer Inst. 98, 491 3. Pero, S.C., Shukla, G.S., Cookson, M. M., Flemer, S., and Krag, D.N. (2007) Combination treatment with Grb7 peptide and Doxorubicin or Trastuzumab (Herceptin) results in cooperative cell growth inhibition in breast cancer cells. Br. J. Cancer 96, 1520 4. Watson, G.M., Gunzburg, M.J., Kulkarni, K., Ambaye, N., Cergol, K., Payne, R., Pero, S., Perlmutter, P., Wilce, M.J. and Wilce, J.A. (2015) Cyclic peptides incorporating phosphotyrosine mimetics as potent and specific inhibitors of the Grb7 breast cancer target. J. Med. Chem, 58, 7707

Inhibitor peptide design – improving affinity without losing specificity Menachem J. Gunzburg1, Gabrielle Watson1, Ketav Kulkarni1,2, Nigus Ambaye1, Katie Cergol3, Richard Payne3, Mark P. Del Borgo1, Patrick Perlmutter2, Matthew C.J. Wilce1, Jacqueline A. Wilce1 1Department of Biochemistry and Molecular Biology, Monash University, VIC, Australia, 2School of Chemistry, Monash University, VIC, Australia, 3School of Chemistry, The University of Sydney, NSW, Australia. The design of potent and specific peptide inhibitors to therapeutic targets is of enormous utility for both proof-of-concept studies and for the development of potential new therapeutics. Here we describe the development of a specific inhibitor of the Grb7-SH2 domain involved in cancer progression. Grb7 is an adapter protein, aberrantly co-overexpressed with erbB-2 and identified as an independent prognostic marker in breast cancer. Grb7 signals the activation of erbB-2 which plays a key role in disregulated cell growth in cancer. Grb7 also mediates signalling from focal adhesion kinase (FAK) exacerbating cell migration and the metastatic potential of cells. It is thus a prime target for the development of novel anti-cancer therapies. We have structurally characterised a cyclic peptide (G7-18NATE) that is a specific inhibitor of Grb7 and inhibits cellular growth and migration in cancer cell lines1. Based on this we have developed a series of second generation bicyclic peptides, constrained via O-ally-serine ring-closure metathesis, that show enhanced affinity and maintained specificity for the Grb7-SH2 domain2. Interestingly, structural studies reveal an unexpected involvement of the O-ally-serine linker in target binding. We have also developed cyclic peptides that incorporate carboxymethylphenylalanine and carboxyphenylalanine as phosphotyrosine mimetics, and shown using X-ray crystallography the way in which this also contributes to improved binding3. Finally, we have shown that by combining these two strategies we are able to achieve peptides with affinities in the nM range that still maintain target specificity. References: 1. Ambaye et al., (2011) J. Mol. Biol. 412, 397-411. 2. Gunzburg et al., (2013) Biopolymers. 100, 543-549. 3. Watson (2015) J. Med. Chem. 58, 7707-7718.

125 ABSTRACT

PO - PROTEIN IN CELLS Breaking the fourth wall: Quaternary organizations forge a link to the novel non-enzymatic function of RNR-a Yimon Aye1, Yuan Fu2, and Marcus J C Long2 1Cornell University and Weill Cornell Medicine, 2Cornell University The classical textbook description of the enzyme ribonucleotide reductase (RNR) has been the rate- limiting catalytic role in the de novo biosynthesis of dNTPs required for DNA replication and repair. The nucleotide-regulated a subunit of RNR in unison with the metalloprotein parter subunit b, or its p53- regulated isoform, p53b, constitutes the active enzyme that catalyzes NDP reduction to dNDPs in humans. RNR activity is positively correlated with cell proliferation and is a known target of nucleoside anticancer agents. However, seemingly at odds with the major role of RNR in cancer proliferation, selective overexpression of a alone confers anti-metastasis activity in transformed cells and in transgenic mice, and is correlated with longer lifespan in cancer patients. Indeed, a-specific gene therapy approaches had been in clinical trials. The basic biology underlying this enigmatic a-specific tumor sup- pression is unknown and is long postulated to involve non-reductase functions. Our most recent seren- dipitous discovery of a new nucleotide-driven a-specific signaling role has shed the first light on the long proposed noncanonical function of a. We present the detailed mechanistic deconvolution of the newly discovered function of a underpinning the clinically relevant tumor suppressor phenotype. Research support: The authors acknowledge an NIH Director’s New Innovator award (1DP2GM114850), an NSF CAREER award (CHE-1351400), a Beckman Young Investigator award, a Bur- roughs Wellcome Fund CRTG, a Sloan Research Fellowship, and an intercampus seed grant from Cornell University and Weill Cornell Medicine (to Y.A.)

Creative destruction of DNA builds CRISPR immunity: Identification of E. coli host helicase, nuclease and polymerase enzymes that target replication forks to promote Cas1-Cas2 CRISPR adaptation. Tom Killelea1, Ivana Ivancic-Bace2, Simon D Cass1 and Edward L. Bolt1 1University of Nottingham, UK, 2University of Zagreb, Croatia. Prokaryotic CRISPR-Cas systems give adaptive immunity against mobile genetic elements (MGEs). Immunity is built from capture of MGE DNA fragments and their integration into CRISPR loci, termed “Adaptation”. This creates a DNA memory of encounters between host and MGE. Transcription and processing of CRISPR RNA into ribonucleoprotein complexes leads to targeted destruction of a compli- mentary MGE by nucleases, termed “Interference”. Adaptation requires the CRISPR-associated proteins Cas1 and Cas2, and Cascade or Cas9 complexes deliver interference in many prokaryotes. Cas1-Cas2 can catalyze adaptation alone, but interference stimulates adaptation. There is much interest in adaptation reaction mechanisms, with and without interference, to understand the genesis of CRISPR immunity and to develop new DNA editing tools. Recent work in E. coli1, 2 identified that CRISPR adaptation relies on enzymes more generally known for processing DNA at broken or blocked replication forks. Here we describe roles for three host cell proteins that are distinct from CRISPR-Cas but are required for adaptation in E. coli: RecG helicase, RecB nuclease and PolI polymerase. We report biochemical analyses, using physiologically relevant Cas1-Cas2 concentrations, identifying substrate preference for Cas1- catalysed DNA nicking and transesterification, which accords with Cas1-Cas2-DNA structures 3, 4, and with a bias for adaptation to target replication termination (ter) sites. New data is presented indicating that Cascade interference complexes are effective roadblocks to replication, which we propose are sensed by RecG helicase provoking adaptation. A model is proposed adaptation targeting MGEs at blocked, terminating or collapsed replication forks, allowing capture of DNA fragments into CRISPR loci1. This may have implications for how Cascade or Cas9 may provoke genome instability when deployed tactically to block replication or transcription.

126 ABSTRACT

1 Ivancic-Bace et al Nucleic Acids Res. (2015) 2 Levy et al Nature (2015) 3 Wang et al Cell (2015) 4 Nunez et al Nature (2015)

Effects of macromolecular crowding on protein folding kinetics Annelise H. Gorensek1, Austin E. Smith1, Gerardo M. Perez Goncalves1 and Gary J. Pielak1,2,3 1Department of Chemistry, 2Department of Biochemistry and Biophysics, 3Lineberger Comprehensive Cancer Center The native environment of fully-folded, globular proteins—the inside of cells—has a macromolecule concentration exceeding 300 g/L and is predicted to affect the rate at which proteins fold. Traditional crowding theory suggests all crowding effects are entropic; crowding agents limit the space available to the unfolded ensemble, enabling the protein to fold faster. However, in vitro folding studies with different crowders indicate enthalpic contributions affect folding, while the entropic effects are more complex than previously understood. In this work, 19F NMR and the 7 kDa N-terminal SH3 domain of the Dro- sophila signaling protein drk (SH3) are used to quantify the activation entropy and enthalpy of folding and unfolding under crowded conditions. Our results indicate the effects of both synthetic polymer and biologically-relevant protein crowders are not simply entropic. The interpretation of these data informs our understanding of how proteins fold in living cells.

Crowding and protein dimerization Alex J Guseman1,2, Gary J. Pielak2 1Molecular and Cellular Biophysics Training Program, and, 2Department of Chemistry, University of North Carolina-Chapel Hill In cells, proteins are surround by macromolecules at concentrations of greater than 100 g/L, yet the majority of our knowledge comes from experiments conducted in dilute buffer solutions. The structure and stability of proteins in cells is influenced by two interactions, hard core repulsions, which arise from excluded volume effects, and transient chemical interactions with surrounding molecules. High concentrations of inert polymers, small molecule cosolutes, and proteins are often used to mimic cellular conditions. Here, we describe the effects of crowding on a protein-protein interaction by using 19F NMR spectroscopy and a variant of the 6 kDa globular B1 domain of protein G. The A34F variant was previously shown to form a side by side dimer in buffer. Using the 3-flourotyorsine labeled variant we measured a dissociation constant of 59 6 5 lM, consistent with previous the study. We then proceed to show the influences of co-solutes on dimer dissociation. Crowding agents such as sucrose, Ficoll-70, glycine betaine, trimethylamineoxide, and bovine serum albumin stabilize the dimer, whereas urea, eth- ylene glycol, 8 kDa polyethylene glycol and lysozyme destabilize the dimer. Measuring the temperature dependence of dimerization allows for us to measure the van’t Hoff enthalpy of dimer formation. We are now extending this methodology to in-cell NMR studies of the dimer.

Acetyl-coa carboxylases in dinoflagellates: fueling the polyketide synthase pathways Saddef Haq1, Tsvetan R. Bachvaroff2, David R. Goodlett3 and Allen R. Place2 1Graduate Program in Life Sciences, University of Maryland, Baltimore, Baltimore, MD, 2Institute of Marine and Environmental Technology, University of Maryland Center for Environmental Science, Bal- timore, MD, 3School of Pharmacy, University of Maryland, Baltimore, Baltimore MD Dinoflagellates are known to make a diverse array of fatty acids and polyketides. A necessary precursor for their synthesis is malonyl-CoA which is formed by carboxylating acetyl CoA using the enzyme acetyl-CoA carboxylase (ACC). In plastid-containing organisms, ACCs are present in the cytosol and the

127 ABSTRACT plastid (chloroplast). Two different forms of these naturally-biotinylated enzymes exist, the heteromeric (prokaryotic) and homomeric (eukaryotic) form. Through transcriptome analysis in Amphidinium car- terae (CCMP 1314) we were able to find two full-length homomeric type ACC sequences; no heteromeric type ACCs were found. Based on phylogenetic analysis we were able to assign the putative cellular location for these two ACCs. These assignments were validated using mass spectrometry proteomics on isolated gel bands, along with streptavidin western blotting which shows two bands corresponding to the calculated sizes of these ACCs. Additional bands showing other naturally biotinylated proteins were also observed. Transcript abundance for these ACCs follow the established global pattern of expression for dinoflagellate mRNA messages over a diel cycle. This is the first description at the transcriptomic and protein level of ACCs in dinoflagellates. Future work will involve subcellular fractio- nation and kinetic properties of these two ACCs as well as identification of the remaining biotinylated proteins in dinoflagellates.

Importance of protein kinetic stability in extremophiles: A study of thermoacidophilic archaea Sulfolobus acidocaldarius Jayeeta Sen1, Ke Xia1, Wilfredo Colon 1 1Rensselaer Polytechnic Institute, Center for Biotechnology and Interdisciplinary Studies, Department of Chemistry and Chemical Biology Kinetically stable proteins (KSPs) represent a group of slow unfolding proteins due to a high energy barrier for unfolding. They are resistant to aggregation or premature degradation since they are virtually conformationally trapped in their native states. From previous studies, we have observed that thermophilic and mesophilic prokaryotes showed higher abundance of KSPs than eukaryotes suggesting that kinetic stability might be crucial for the survival and adaptation of less complex organisms like prokaryotes. These prokaryotes have simpler cellular organization and thus rely heavily on resilient proteins to endure harsh conditions. Hence the study of such hyperstable proteins in these organisms gives us an insight about the role of kinetic stability in their survival, adaptation and proliferation in extreme environments. My current research focuses on identifying the KSPs from Sulfolobus acidocaldarius which is a thermo-acidophilic archaea that grows under high temperature (708C) and low pH (pH53) conditions. The KSPs are screened from the non-KSPs using diagonal-two-dimensional (D2D) SDS-PAGE technique. The protein spots are then excised from the gel, trypsin digested and identified using MALDI-MS and peptide mass fingerprinting. Analysis of these proteins and comparing them with other thermophilic and mesophilic counterparts could help us decipher a common factor for their hyperstability and increase our understanding of the structural basis, biological and pathological significance of protein kinetic stability in nature.

PP – PROTEIN INTERACTIONS AND ASSEMBLIES Deciphering the molecular and functional basis of the RhoGAP family proteins: A systematic approach towards selective inactivation of the Rho family proteins Ehsan Amin‡, Mamta Jaiswal‡, Urszula Derewenda§, Katarina Reis¶, Kazem Nouri‡, Katja T. Kossemeier‡, Pontus Aspenstrom€ ¶, Avril Somlyo§, Radovan Dvorsky‡, and Mohammad R. Ahmadian‡ ‡Institute of Biochemistry and Molecular Biology II, Medical Faculty, Heinrich-Heine-University, 40225 Dusseldorf,€ Germany, §Department of Molecular Physiology and Biological Physics, University of Virginia, Charlottesville, VA 22908, USA; ¶Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, SE-171 77 Stockholm, Sweden The Rho GTPase-activating proteins (RhoGAPs) are one of the major classes of regulators of the Rho- related protein family that are crucial in cell adhesion, motility, contractility, growth, differentiation, and development. Using database searches, we extracted 66 distinct human RhoGAPs with a common catalytic GAP domain that terminate signal transduction of the Rho proteins by stimulating the slow intrinsic

128 ABSTRACT

GTP hydrolysis (GTPase) reaction. The specificity of the majority of the members of RhoGAP family is largely uncharacterized. Thus, in this study we comprehensively investigated the structure-function relationships of RhoGAPs by combining our in vitro data with in silico data. The activity of fourteen repre- sentatives of the RhoGAP family was measured in real-time towards 12 Rho family proteins. We identified and structurally verified hotspots in the interface between RhoGAPs and Rho proteins as critical determinants for binding and catalysis. We have found that the RhoGAP domain itself is nonselec- tive and in some cases rather inefficient under cell-free conditions. Thus, we propose that other domains of RhoGAPs confer substrate specificity and fine-tune their catalytic efficiency in cells.

Biomolecular interaction determination and quantification by MicroScale Thermophoresis Myriam Badr1, Ellen Lee1, Ana Lazic2, Stefan Duhr3, and Dennis Breitsprecher3 1NanoTemper Technologies, Inc, Cambridge, MA, 2NanoTemper Technologies, Inc, South San Fran- cisco, CA, 3NanoTemper Technologies GmbH, Munich, Germany MicroScale Thermophoresis (MST), an immobilization-free technology, is used to quantify biomolecular interactions (pM-mM), ranging from protein-protein interactions to small molecule-target binding. MST, the movement of molecules in an optically generated microscopic temperature gradient, is monitored by fluorescence. This thermophoretic movement is affected by the entropy of the hydration shell around molecules and is highly sensitive to binding reactions, which affect the size, charge, conformation, and/or hydration shell. We show how MST can be used to identify and quantify interactions between biomolecules of interest: proteins, nucleic acids, ions, etc. In one study, the binding site of a protein-protein complex is investigated with protein engineering. We also demonstrate how interactions with proteins such as GPCRs can be analyzed in a Label-Free manner using tryptophan fluorescence. With MST, affinities of membrane proteins can be probed in detergents or liposomes.

129 ABSTRACT

Watch the Clock is Ticking: Real time study of biological clock at atomic resolution by NMR Archana G. Chavan1, Joel Heisler1, Yong-Gang Chang1, Roger Tseng1 and Andy LiWang1,2 1School of Natural Sciences, University of California, 2Center for Chronobiology, Division of Biological Sciences, University of California, San Diego All living organisms have evolved to respond the day and night cycle through the system called circadian clocks. Animals use these circadian clocks to regulate healthy cycles of activity and rest in anticipa- tion of sunrise and sunset. Like gears of a mechanical clock, the clock proteins keep time by interacting with one another. Despite of their importance to life on Earth, molecular mechanisms of biological clocks remain mysterious, and face two major challenges: (1) gears of biological clock are tiny proteins that needs to be observed at atomic resolution and (2) as these protein gears are constantly moving in synchronized interplay, real time measurements of their states is crucial for elucidation of molecular mechanism. While there are several atomic-resolution structures of circadian clock proteins are available, they represent only few frozen states within crystals. Alternatively, previously known real-time measurements are at resolutions that are too low to see how clock proteins move to tell the time. Cur- rent research has finally developed an approach using nuclear magnetic resonance (NMR) spectroscopy that can observe in real time and at atomic resolution a fully reassembled and functional biological circadian clock as it ticks over many days in a test tube. Measuring the movements of individual atoms inside clock proteins and in real time will be one of the most substantial advances in the field.

The Determining Factors of the Self-assembly of a Collagen Mimetic Triple Helix Fangfang Chen1, Sam Wong1, Parminder Jeet Kaur1 1Department of Biochemistry, Hunter College, The Graduate Center of The City University of New York, NY We recently reported that a designed triple helix Col108 self assembled to form collagen-like mini- fibrils having a d-periodicity of 35nm. Our Study further tested the role of periodicity in the amino acid sequence on the assembly of triple helix. Recombinant peptide Col108 consists 378-residue triple helix domain organized into three repeating sequence units, and a C-terminal fodon domain. The 35nm d- period of the mini-fibrils is consistent with the one unit staggered arrangement of the associating helices. To further investigate this hypothesis, we developed several new peptides. The 2U108 contains two repeating sequences, and the 1U108 contains only one. We also made another peptide Col877 which consists of three repeating units but the amino acid sequence of each unit is very different from that of Col108. As predicted, 2U108 and Col877, formed mini-fibrils having the same periodicity as that observed in Col108. But no fibril-like assemblies were found in 1U108. Since both 2U108 and Col877 have repeating sequence units while 1U108 does not, these findings support the essential roles of the repeating sequence units on the self-assembly of collagen-like, staggered fibrils.

The Role of Dimerization in the Methylation Activity of EcoP15I DNA Methyltransferase Madhusoodanan Urulangodi1 and Desirazu N Rao1 1Department of Biochemistry, Indian Institute of Science, Bangalore, INDIA EcoP15I DNA methyltransferase (M.EcoP15I) recognizes a short asymmetric sequence, 5’-CAGCAG-3’, and methylates the second adenine only on one strand of the double-stranded DNA (dsDNA) in the presence of magnesium ions. Biochemical and structural characterization support the notion that purified M.EcoP15I exists and functions as dimer. However, the exact role of dimerization in M.EcoP15I reaction mechanism remains elusive. We generated a three-dimensional model of the M.EcoP15I dimer based on the structure of the M.MboIIa dimer. The side-chains of three amino acid residues (D223,V225 and V392) from one Mod subunit make extensive contacts with the atoms of the other subunit. These residues were indeed present at the vicinity of the extensive dimeric interface (4000 A2) in the structure

130 ABSTRACT of EcoP15I. Next, we engineered M.EcoP15I into a stable monomeric form and studied the role of dimerization in enzyme catalyzed methylation reaction. While the monomeric form binds single- stranded DNA (ssDNA) containing the recognition sequence it is unable to methylate it. Furthermore we show that while the monomeric form has AdoMet binding and Mg21 binding motifs intact optimal dsDNA binding required for methylation is dependent on dimerization. Deletion of the N-terminal domain (NTD) of M.EcoP15I resulted in complete disruption of the dimer to form catalytically inactive but structurally folded monomers. The proximity of NTD helices at the dimeric surfaces support the importance of this domain in the dimerization of M.EcoP15I. Together, our biochemical data supports a unique subunit organization for M.EcoP15I to catalyze Mg21 dependent methylation reaction.

Expression, Purification, and Characterization of Human M6PR Extra Cellular Domain Brian Dwyer1, Andrea Iskenderian1, Dianna Lundberg1, Muthu Meiyappan1, and Bohong Zhang1 1University of Massachusetts The cation-independent mannose-6-phosphate receptor (M6PR) is a multifunctional protein that binds ligands like mannose-6-phosphate (M6P) and IGFII. Its major function is to transport enzymes to lyso- somes. In order to use M6PR to characterize M6PR binding proteins, a human M6PR (hM6PR) overexpression cell line was developed. It was discovered that a significant fraction of hM6PR was released by the cell into the culture media (soluble hM6PR or shM6PR). An affinity purification process was developed to purify the soluble form of hM6PR. The purified hM6PR was characterized through biochemical and biophysical assays. It was determined the entire extracellular domain of hM6PR was present in the soluble hM6PR and that it was fully functional in regards to binding M6P containing protein ligands.

Structure of serum amyloid a suggests a mechanism for high-density lipoprotein binding and function as a protein hub Nicholas Frame1 1Boston University School of Medicine Serum Amyloid A (SAA) is a major acute-phase plasma protein conserved from sea cucumbers to humans. SAA functions in the immune response and cholesterol homeostasis via unclear mechanisms; elevated SAA causes inflammatory amyloidosis and contributes to atherosclerosis. During the acute- phase response, SAA secretion into plasma increases up to 1000-fold. Most circulating SAA is bound to high-density lipoprotein (HDL, or Good Cholesterol). Although such binding is central to lipid mobiliza- tion for cell repair, the mechanism by which SAA selectively binds HDL is unclear. We combine the recently solved x-ray crystal structures of lipid-free SAA with our amino acid sequence analysis to propose such a mechanism. We identify two amphipathic alpha-helices in the N-terminal domain of SAA (residues 1-69) which bind lipids and form a concave solvent-exposed hydrophobic surface. The curvature of this surface (radius r4.2 nm) matches that of HDL, explaining the binding preference of SAA for HDL over larger lipoproteins. This curvature is maintained through a novel structural motif including GPGG residues. Our proposed mode of protein-lipid surface binding is unique to SAA and differs from any other apolipoproteins. In our model, a flexible linker connects the HDL-binding N-domain to the C- domain that binds various polar/charged ligands including cell receptors (CD36, LOX-1) and heparan sulfate proteoglycans. Ligand binding via both domains of SAA brings HDL and cellular receptors into contact facilitating their functional interactions. Therefore, we propose that SAA acts as a protein hub mediating interactions among diverse proteins, lipids, and proteoglycans. Our model is supported by the observation that SAA residues 1-76, termed amyloid-A, form the major protein constituent in inflammation-linked amyloidosis.

131 ABSTRACT

An Autoinhibited Dimeric Form Of Pro-Apoptotic Bax Regulates Apoptosis Thomas P. Garner1, Denis E. Reyna1, Amit Priyadarshi1, Hui-Chen Chen3, Vladimir N. Malashkevich2, Steve S. Almo2, Emily H. Cheng3 and Evripidis Gavathiotis1 1Department of Biochemistry, Department of Medicine, Albert Einstein Cancer Center, Wilf Family Cardiovascular Research Institute, Albert Einstein College of Medicine, 2Department of Biochemistry, Department of Biophysics and Physiology, Albert Einstein Cancer Center, Albert Einstein College of Medicine, 3Human Oncology and Pathogenesis Program, Memorial Sloan-Kettering Cancer Center, New York Pro-apoptotic BAX is a cell fate regulator playing an important role in cellular homeostasis, immune cell differentiation and maturation and several diseases including cancer and heart disease. In nonapoptotic cells, BAX is predominantly localized in the cytosol where it has a quiescent conformation. Following a pro-apoptotic stimulus, cytosolic BAX is activated and translocates to the mitochondria to initiate mitochondrial dysfunction and apoptosis. Here, cellular and biochemical data unexpectedly demonstrate that BAX has a newly identified inactive dimer conformation that prevents BAX activation compared to the inactive monomer conformation. The crystal structure of the full-length BAX dimer conformation was determined and revealed an asymmetric dimer structure of two inactive BAX protomers. Dimeriza- tion is formed between the N-terminal trigger site of one protomer and a novel C-terminal binding surface of the second protomer. This dimeric conformation inhibits either the N-terminal conformational change of one protomer or the displacement of the C-terminal helix a9 of the second protomer; both conformational changes are required for the activation of a single BAX molecule. Functional analysis shows that preventing the formation of the cytosolic dimer is sufficient to increase the levels of BAX activation and apoptosis induction, and promoting dimerization of cytosolic BAX causes resistance to BAX activation and apoptosis induction. Our data support a model that apoptosis induction can be regulated by the equilibrium between the cytosolic BAX monomer and the autoinhibited dimer. Our study identified an unprecedented mechanism of cytosolic BAX regulation providing a novel target for therapeutic modulation of pathological apoptosis.

Domain Swapping in c-Type Cytochromes and Myoglobin Shun Hirota1, Satoshi Nagao1, Partha Pratim Parui1, Megha Subhash Deshpande1, Chunguang Ren1, Yugo Hayashi1, Takaaki Miyamoto1, Yin-Wu Lin1, Yoshiki Higuchi2,3 1Graduate School of Materials Science, Nara Institute of Science and Technology, 2Graduate School of Life Science, University of Hyogo, 3RIKEN SPring-8 Center We have previously shown that horse cytochrome (cyt) c forms polymers from monomers by 3D domain swapping its C-terminal a-helix successively (runaway domain swapping) [1]. The C-terminal a- helix of dimeric horse cyt c was displaced from its original position in the monomer, and the Met–heme coordination was perturbed significantly in the dimer, causing higher cyanide ion binding affinity and peroxidase activity compared to those in the monomer. Horse cyt c formed domain-swapped oligomers by the interaction between the N- and C-terminal a-helices at the early stage of folding from its unfolded state [2], and the interaction important for formation of domain-swapped oligomers existed in the molten globule state [3]. Psudomonas aeruginosa cyt c551 and Hydrogenobacter thermophilus (HT) cyt c552 formed oligomers by domain swapping the N-terminal region containing the heme [4]. By elongation of the hinge loop in HT cyt c552, we also observed domain swapping of the C-terminal region [5]. High-order domain-swapped oligomers of HT cyt c552 were produced during E. coli culture, whereas the domain-swapped protein amount decreased when the protein stability was decreased by mutation [6]. Cyt cb562 also domain swapped and formed a dimer, where the two helices in the N- terminal region of one protomer interacted with the other two helices in the C-terminal region of the other protomer. In the crystal, three domain-swapped cyt cb562 dimers formed a unique cage structure

132 ABSTRACT with a Zn-SO4 cluster inside the cavity [7]. Myoglobin (Mb) also formed a domain-swapped dimer with two extended a-helices [8]. Each new long a-helix was formed by the E and F helices and the EF-loop of the original monomer, and as a result the proximal and distal histidines of the heme originated from different protomers. We succeeded in construction of an artificial heterodimeric protein with two different heme active sites (a bis-histidine-coordinated heme and a H2O/histidine-coordinated heme) using domain swapping for Mb [9]. 1. S. Hirota, et al., Proc. Natl. Acad. Sci. USA, 12854 (2010) 2. P. P. Parui, et al., Biochemistry, 52, 8732 (2013) 3. M. S. Deshpande, et al., 53, 4696 (2014) 4. S. Nagao, et al., PLoS One, 10, e0123653 (2015) 5. C. Ren, et al., Mol. BioSyst., 11, 3218 (2015) 6. Y. Hayashi, et al., Sci. Rep., 6, 19334 (2016) 7. T. Miyamoto, et al., Chem. Sci, 6, 7336 (2015) 8. S. Nagao, et al., Dalton Trans., 41, 11378 (2012) 9. Y.-W. Lin et al., Angew. Chem. Int. Ed., 54, 511 (2015).

Subunit interaction and asymmetry of cold-active alkaline phosphatase studied using bimane fluorescence label at dimer interface. Jens G. Hjo€rleifsson1 and Bjarni Asgeirsson 1 1Science Institute, Department of Biochemistry, University of Iceland One of the most interesting challenges in biology at present is to define protein-protein interactions within the proteome and subunit interactions within oligomeric proteins (cooperativity). Alkaline phosphatase (AP) is a homodimeric enzyme. Negative cooperativity based on asymmetry of subunits has been suggested using both experimental methods (1-3) and computer simulations (4). However, detailed knowledge as to how and where the cooperativity takes place is lacking, e.q. which specific interfacial residues mediate information transfer between subunits. Here, we employ a recent method to study dimer dissociation/association of a cold-active alkaline phosphatase (Vibrio sp.) by incorporating a bimane fluorescent probe at the dimer interface. Bimane fluorescence is quenched intrinsically by Trp and Tyr within a certain well-defined distance (5,6). A Trp residue was introduced at the interface to statically quench a bimane probe located on a short loop (a.a. 55-63) in juxtaposition on the opposite subunit. Two lifetimes with similar amplitudes where observed for bimane fluorescence that may indicate asymmetry. Using bimane fluorescence, urea denaturation curves indicated a four-state path involving a seemingly inactive dimer intermediate. Introducing a Cys residues at the dimer interface for bimane attachment greatly reduced the activity of the enzyme, indicating a role for loop 55-63 in maintaining high catalytic efficiency. (1) Sun et al., Eur. J. Biochem. 245 (1997) 32–39. (2) Orhanovic & Pavela-Vrancic, Eur. J. Biochem. 270 (2003) 4356–4364 (3) Hoylaerts et al., J. Biol. Chem. 272 (1997) 22781–22787. (4) Asgeirsson et al., FEBS J. 280 (2013) 157. (5) Mansoor et al. 49 (2010) 9722-9731. (6) Jones et al., Biochemistry, 53 (2014) 6290-6301.

Novel antimicrobial lectin Myxovirin Tyler H. Jones1 and Robert L. McFeeters1 1Department of Chemistry, University of Alabama in Huntsville Several high mannose binding lectins have been extensively studied for their potent antimicrobial properties. One such lectin is Scytovirin, a standout in terms of safety profile and lack of toxicity. Scytovirin is a small, 9 kDa protein with two nearly identical domains, differing in only 3 of 39 amino acids. Each

133 ABSTRACT

domain of Scytovirin is known to specifically bind Mana(1!2)Mana(1!6)-Mana(1!6)Man, the D3 arm of Man9. Carbohydrate binding and resultant antimicrobial activity is influenced by the constituencies of the binding pocket, in particular 3 aromatic residues. From the genome of the myxobacteria Myxo- coccus fulvus, a novel protein domain with high sequence homology to the carbohydrate binding domains of Scytovirin was discovered. This protein domain, termed Myxovirin, demonstrates several key differences compared to Scytovirin, including a substitution of one of the key aromatic binding pocket residues. Whereas, alteration or extension of the N-terminus of Scytovirin disrupts folding, native Myxo- virin is expressed following an N-terminal peptide, opening the possibility to N-terminal modifications or engineering. After successful cloning, recombinant expression, and purification, solution NMR backbone chemical shifts indicate Myxovirin, like Scytovirin, has no regular secondary structure elements. Moreover, Myxovirin retains the ability to bind Man9 and demonstrates novel antimicrobial activity. Understanding the structural similarities and differences between Myxovirin and Scytovirin will help aid in developing high mannose-binding lectins as tools to combat emerging pathogens.

Canonical Mapping of bZip Protein-DNA Recognition Hyun Joo1 and Jerry Tsai1 1University of The Pacific Understanding protein-nucleic acid interactions is a fundamental step in deciphering how proteins recognize DNA sequences. A knob-socket analysis of the bZIP domain protein-DNA interfaces characterizes the set of canonical packing contributions that determine a protein’s binding specificity for DNA. Because the knob-socket model decomposes the complex packing into identifiable units of interaction, a specific residue/structural code provides a descriptive model about how the a-helix recognizes DNA

134 ABSTRACT bases. The dominant types of residue packing in the protein-DNA interface were found to be the 304,000 four-body and 71,000 five-body packing cliques. Among the four-body cliques, 27% consist of protein sockets packing against a DNA base knob and 9% involve DNA sockets formed by 2 adjacent bases and a complementary base packed with a protein residue knob. A canonical pattern can be derived from these knob-socket packing patterns for the a-helix binding to DNA. Along the a- helical i64 positions, the conserved Asn, Ile/Val, and Arg residues from a ridge of knobs that pack into very regular sets of sockets on the DNA double helix. This i64 ridge of knobs divides the interaction surface of the a-helix into 2 regions of sockets that recognize blocks of DNA bases. The N-terminal side of this ridge specifically binds 2 consecutive bases of the DNA coding (1) strand and the C- terminal side specifically binds 2 consecutive bases on non-coding (-) strand. Because of DNA based complementarity, the a-helix is able to recognize a sequence of 4 consecutive bases. The pattern of sockets and their amino acid composition on the a-helix is shown to determine the specificity for particular DNA bases. This canonical model of protein-DNA recognition demonstrates that the knobsocket model can rationally investigate and produce clear insight into protein-DNA binding interfaces.

Improving the affinity of the phosphoepitope-binding FHA1 domain Sehar Khosla1 and Brian K. Kay1 1Department of Biological Sciences, University of Illinois at Chicago JunB is a component of activator protein transcription factor, AP-1, and can act as a tumor suppressor and as an oncogene. While in the past polyclonal and monoclonal have been used to follow its phosphorylation, we have engineered the Forkhead Associated (FHA) domain to recognize specific phosphothreonine-containing epitopes. FHA is an attractive scaffold for generating affinity reagents owing to its small size (12 kDa), ability to be displayed on the surface of M13 bacteriophage particles, and specificity in binding phosphothreonine-containing peptides. From a library of over one billion variants, we have isolated several that bind specifically to a JunB phosphopeptide (Biotin- EAR- SRDA(pT)PPVSPYKK-NH2). To improve their affinity, we have constructed dimeric forms of the reagent by fusing the FHA sequence to a leucine zipper sequence (from the yeast GCN4 protein). Both monomeric and dimeric forms of the engineered FHA domain were expressed in Escherichia coli, purified, and demonstrated to be thermostable (Tm of 70-75oC). In an Enzyme Linked Immunosorbent Assay (ELISA), the dimer showed an increase in affinity as determined by EC50 and IC50 values. These results indicate that dimerization of FHA domain increases its apparent affinity to its target, without affecting its specificity. Experiments are currently being performed to display the FHA domain as a dimer on virions, by fusing a leucine zipper sequence followed by an amber codon in frame with the pIII coat protein.

Gene Ontology in Comparative Protein Docking Anna Hadarovich1,2, Ivan Anishchenko1, Alexander V. Tuzikov2, Petras J. Kundrotas1, Ilya A. Vakser1 1Center for Computational Biology and Department of Molecular Biosciences, The University of Kan- sas, 2United Institute of Informatics Problems, National Academy of Sciences Structural characterization of protein-protein interactions (PPI) is important for understanding life processes at the molecular level. The experimental techniques, due to inherent limitations, have resolved only a fraction of known PPI. Thus the rest of the PPI structures should be modeled by computational docking techniques. Comparative protein docking generates the structure of a protein-protein complex based on known structures from PDB (templates). It has proven to be the method of choice in large- scale structural studies of PPI. Typically, the detection of templates is based on structural similarity between the target and the templates, quantified by the TM-score of the TM-align algorithm. However, the performance of the template-based docking deteriorates significantly when targets are only weakly

135 ABSTRACT similar to the templates. We present a non-structural supplement to the structure-based scoring, based on the similarity of the Gene Ontology (GO) annotations (GO-scores) for the target and the template complexes. A combined scoring function consisting of the TM-score and three GO-scores (one per each of the three ontology domains, “molecular function,” “biological process,” and “cellular component”) was tested on a non-redundant set of 587 protein-protein complexes, utilizing 4,950 template structures from the DOCKGROUND resource (http://dockground.compbio.ku.edu). The results show that this function discriminates incorrect docking models significantly better than the scoring by the TM-score only. The results of this work could be applied to modeling of protein-protein complexes, including structural reconstruction of protein interaction networks, by improving the quality of the docking and providing a reliability score for the generated models.

Silk Fibroin/Sericin Interaction at the Biomimetic Interface Hyo Won Kwak1 and Ki Hoon Lee1 1Department of Biosystems & Biomaterials Science and Engineering, Seoul National University, Republic of Korea Silk is an outstanding natural fiber having high toughness. People tried to mimic the spinning process of silkworm or spider, but there are still unknown secrets to be solved. The biggest question is how the concentrated protein is not precipitated until spinning. In the silkgland of silkworm, fibroin and sericin are stored until spinning. The inner core is filled with fibroin, and sericin envelops the core. These proteins are in direct contact to each other without mixing. We mimic this by loading sericin solution on the top of fibroin solution, and observed the gelation behavior of fibroin layer. The gelation of fibroin occurs due to the irreversible conformational transition of fibroin from random coil to b-sheet structure, which resembles the conformational transition during the native spinning process. The gelation time of fibroin was retarded in the presence of sericin due to the delayed conformational transition from random coil to b-sheet structure. We furthermore confirmed that the fibroin gel has Silk I structure which is the native structure in the silkgland. This Silk I structure fibroin gel is converted to Silk II structure (irreversible b-sheet structure) after removing the upper sericin layer. Our results clearly shows that the presence of sericin layer contributes to the stabilization of fibroin even they were phase separated.

Towards quantitative mapping of protein interactions in vivo O. Matalon1, ED. Levy1 1Weizmann Institute of science, department of structural biology Keywords: PCA, Network, Affinity, in vivo, protein interactions Protein Interaction Networks reflect the organization of physical matter in cells. Several methods have thus been developed to measure protein interactions at the scale of whole organisms to elucidate how proteomes are physically organized and make up a living cell. Current such methods, however, only provide binary data and are not able to quantify the strength or affinity of protein-protein interactions. A quantitative description of protein interactions is yet fundamental to our understanding of cellular processes that involve competitive binding, dosage sensitivity or dynamics of complex assembly. Here we present a strategy for the quantitative measurement of protein interactions that is both applicable in vivo and at a proteome-wide scale. Our strategy is based on the DHFR Protein fragment Comple- mentation Assay (DHFR-PCA) in which growth was shown to be proportional to protein complex concentration. Similarly to classic quantitative methods, our strategy uses titration by increasing the concentration of one protein while maintaining the concentration of the partner constant. Calibration is done by expressing mutants of E9-colicin and its inhibitors, Immunity proteins 9 and 2, which have known affinities. The results show that our strategy can reliably discriminate between interactions that differ in their dissociation constants by less than an order of magnitude, and this across a wide dynamic

136 ABSTRACT range spanning 10-6 to 10-11M. The novel quantitative understanding brought about by this analysis enables us to measure interactions of highly abundant proteins that were excluded from previous DHFR-PCA analyses, and probe the network of interactions between ribosomal proteins and the translocation machinery.

Versatile Nucleotide Recognition with the Peptidyl-tRNA Hydrolase Fold Robert McFeeters1 1University of Alabama, Huntsville Peptidyl-tRNA hydrolase (Pth) activity is ubiquitous and essential in all living cells. The single Pth enzyme found in bacteria is very different from the multi-component Pth system in eukaryotes. Bacterial Pths were originally thought to be very similar based on a high degree of sequence homology. How- ever, phylogenetic characterization reveals two distinct clades and small molecule inhibitor screening has uncovered possibilities for broad, clade, and species specific inhibition. From high resolution crystal structures and SANS characterization of the Pth:peptidyl-tRNA complex in conjunction with chemical shift perturbation mapping and computational docking, details of the active site are emerging. Addi- tionally, the hydrolase fold with nucleotide recognition is found in a variety of proteins. For example, human PTRHD1 was originally classified as a Pth based on sequence similarity and structure prediction. We show PTRHD1 lacks hydrolase activity, but binds DNA in a Zn21 dependent fashion. Thus, Pths share similarities with nucleotide binding proteins but have a unique capability to distinguish between aminoacyl- and peptidyl-tRNA. Increased understanding of Pths and their ability to recognize nucleotides provides insight into a vital cellular process and supports development of this emerging antibiotic target.

Architecture of the yeast exocyst complex Margaret Heider1, Mingyu Gu3, Caroline Duffy1, Anne Mirza1, Laura L. Marcotte1, Alexandra Walls1, Zhanna Hakhverdyan2, Raghav Kalia3, Nicholas Farrall3, Jeff Gelles4, Larry Friedman4, Michael Rout2, Adam Frost3 & Mary Munson1 1University of Massachusetts Medical School, Worcester, MA, 2The Rockefeller University, New York, NY, 3University of Utah, Salt Lake City, UT, and UCSF, San Francisco, CA, 4Brandeis University, MA The exocyst is a hetero-octameric complex that has been proposed to serve as a tethering complex for both exocytosis and endocytosis, and may also play a role in autophagy. Its overall structure and

137 ABSTRACT function remain poorly understood at the molecular level. Here, we show that we can purify endogenous exocyst complexes from Saccharomyces cerevisiae and that they are stable and consist of all eight subunits with equal stoichiometry. Using a combination of biochemical and auxin induced-degradation experiments in yeast, we mapped the subunit connectivity, identified two stable four-subunit modules within the octamer. We also demonstrate that several known exocyst-binding partners, including the Sec4 Rab GTPase, Cdc42 and the myosin motor Myo2, are not necessary for exocyst assembly and stability. Furthermore, we visualized the intact structure of the yeast complex by using negative-stain electron microscopy; our results indicate that the exocyst exists predominantly as a stable, octameric complex with an elongated architecture that suggests that the subunits are contiguous helical bundles packed together into a bundle of long rods.

Characterisation of Non-Histone Lysine Acetyltransferases and Deacetylases in Probiotics Sita Vaag Olesen1,2, Per H€agglund2, and Birte Svensson1 1Enzyme and Protein Chemistry, Department of Systems Biology, Technical University of Denmark, DK-2800 Lyngby, 2Protein and Immune Systems Biology, Department of Systems Biology, Technical University of Denmark, DK-2800 Lyngby N-lysine acetylation of non-histone proteins is a common posttranslational modification [1]. Acetylation and deacetylation of histones have, in humans, been shown to be involved in cancer [3,4]. It has also been shown that acetylation of non-histone proteins are involved in cell metabolism [2], and they may therefore have the potential to function as biomarker for human diseases. Furthermore, studies indicate that deacetylation may be related to the response to oxidative stress in cells [5,6]. Lysine acetyltransferases (KATs) and lysine deacetylases (KDACs), as well as acetylation and deacetylation in general, have been investigated in various organisms such as Escherichia coli, Bacillus subtilis, Mycobacterium tuberculosis, and Salmonella enterica [7], including some 3D-structures. However, to our knowledge it has not been investigated in probiotics. In the beginning of this project, 11 putative non-histone KATs and 3 KDACs in Lactobacillus acidophilus NCFM were identified by bioinformatic analyses, and some of these putative enzymes have been successfully cloned and expressed. Activity assays will be established based on work done by Toro et al. [8] and Berndsen et al. [9], and the enzymes will be characterized. Another goal of the project is to establish a mass spectrometry method in order to investigate the acetylome of L. acidophilus NCFM. Additionally, a method to measure the organism’s response to oxidative stress will be established. The different methods will be applied on clinical samples and, if relevant, on other organisms such as fungi. Acknowledgements: This project is supported by a collaborative DTU-Nordic Bioscience Ph.D. scholarship [1] Ouidir et al. (2015) Proteomics, [2] Yu et al. (2008) J. Microbiol. Biotechnol., [3] Lin et al. (1998) Nature, [4] Grignani et al. (1998) Nature, [5] Erjavec et al. (2007) Proc. Natl. Acad. Sci. U. S. A., [6] Aguila- niu et al. (2003) Science, [7] Bernal et al. (2014) New Biotech, [8] Toro et al. (2015) Prot. Sci., [9] Bernd- sen et al. (2005) Methods

Site Specific 1H-13C Phenylalanine in the Study of the Pth1:Peptidyl tRNA Complex Bhargavi Ramaraju1, Hana McFeeters1, Robert McFeeters1 1Department of Chemistry, University of Alabama in Huntsville, AL Structural insight into biological macromolecules furthers understanding of function and provides a variety of information about cellular processes. Though NMR spectroscopy can be utilized to provide such insight, limitations, including size, are significant barriers. Site specific isotope labeling can be used to provide information on a variety of systems not amenable to study by traditional means including large, slow tumbling systems and membrane proteins. In effect, site specific isotope labeling increases the size and

138 ABSTRACT

thereby number of macromolecular systems able to be studied by NMR spectroscopy. Methyl labeling has been established as a leading method for site specific studies. However, more options are needed to extend utility and supplement systems where methyls are sparse or inadequately located. Aromatic amino acids are well positioned to be additional site specific probes, complementing existing methyl labeling. Aromatics are often found at important interaction interfaces and play significant roles in terms of structure and interactions. They have the added properties of having distinct 13C sidechain chemical shifts, multiple magnetically equivalent 1H positions in the side chains, and reduced effective correlation times due to ring flipping. Thus, they are excellent probes for studying large systems. Our group has developed site specifically 13C labeled aromatic amino acids. Presented here is the use of 1H-13Ca,E phenylalanine and tyrosine for study of the enzyme Pth1 with its natural substrate, peptidyl-tRNA.

Expression and purification of Arabidopsis phosphatase PP2A-3 catalytic subunit in E.coli and approaches to its activation in trans. Priyanka Sandal1, Shweta Shah1 and A. Gururaj Rao1 1Roy J. Carver Dept. of Biochemistry, Biophysics & Molecular Biology, Iowa State University We have previously demonstrated the interaction between the intracellular kinase domain of Arabidop- sis CRINKLY4 (ACR4) receptor-like kinase and PP2A-3c, the catalytic subunit of the oligomeric PP2A phosphatase that regulates formative cell division in roots [Yue et al, PNAS (2016)113:1447]. Importantly, PP2A-3c is also the first described substrate for ACR4 phosphorylation. The hetero-trimeric PP2A holoenzyme is a Serine/Threonine phosphatase that is conserved in plants and animals and is composed of Scaffold ‘A’ subunit, catalytic ‘C’ subunit and substrate specific ‘B’ subunit. The formation of an active PP2A holoenzyme, and its regulation, is mediated through specific protein-protein interactions that include the activator protein, Phosphotyrosyl phosphatase activator (PTPA) that activates the catalytic subunit by stabilizing the active site conformation [Guo et al, Cell Research (2014) 24:190], and the enzyme, Leucine carboxyl methyltransferase 1 (LCMT-1) that methylates the conserved leucine at the C- terminus of the catalytic subunit and facilitates binding of the B subunit. In light of the requirement of these “accessory” proteins, eukaryotic systems such as insect cells & mammalian cells are the preferred routes for producing active PP2A. In our laboratory, we are exploring the feasibility of producing active PP2A-3c phosphatase, in trans, after it has been purified from E. coli. Thus, the N-terminally MBP- tagged PP2A3-c protein has been purified from E. coli but found to be initially inactive in a colorimetric assay against a synthetic phospho-peptide substrate. Interestingly, however, overnight incubation at

139 ABSTRACT room temperature results in a slow but measurable “activation” when similarly assayed. Furthermore, differences in the intrinsic fluorescence spectrum of the non-incubated and incubated proteins suggest that the observed mild activity of the incubated protein may be attributed to a slow conformational change to a more active conformation. These results suggest that greater activation of the MBP-fused PP2A-3c may be achieved in trans using the “accessory” proteins PTPA and LCMT-1.

A New Dimension of Detection in Analytical Ultracentrifugation with Fluorescence Detection System Using Photoswitchable GFPs as Time Domain Probes Huaying Zhao1, Yan Fu2, Carla Glasser3, Eric Andrade2, Mark L. Mayer3, George Patterson2, Peter Schuck1 1Dynamics of Macromolecular Assembly Section, Laboratory of Cellular Imaging and Macromolecular Biophysics, 2Section on Biophotonics, National Institute of Biomedical Imaging and Bioengineering; and, 3Laboratory of Cellular and Molecular Neurophysiology, Porter Neuroscience Research Center, National Institute of Child Health and Human Development National Institutes of Health Multi-component protein complex formation is of great interest in physiological and biochemical studies as it is ubiquitous in numerous biological systems. However, such systems are generally challenging to investigate due to the complexity of the assembly mechanism and limited number of signals offered by the current experimental resources for identifying each species. Analytical ultracentrifugation (AUC) provides powerful methods for studying such systems. It offers information on size, shape and binding energies for reversible systems from analysis of sedimentation profiles of molecular mixtures in free solution. By virtue of the superb hydrodynamic resolution achieved in sedimentation velocity, multiple co-existing complexes can be identified, even in the presence of impurities and aggregates. A recently introduced fluorescence optical detection system (FDS) for AUC offers specific advantages for studying high-affinity protein interactions because of the high sensitivity and selectivity of fluorescence. In the current study, we employed photoswitchable GFP molecules as fluorescent probes in FDS-AUC and took the advantage of their time-dependent fluorescent signal change due to photoswitching as a new dimension of detection. We have developed computational approaches to account for such unique time-dependent signal so the size and shape of the species carrying the photoswitchable molecule can be resolved in the time domain spectrum. The photoswitchable molecules with different switching kinetics were examined with FDS and the different time domain spectra were demonstrated to be dis- tinguishable for the model system under the current study. This time domain detection adds a new dimension of detection in FDS-AUC and will expand AUC application to more complicated protein system with multiple components and allow for quantitative and reliable analysis of the binding mechanism and stoichiometry under very low macromolecular concentrations. Using this technique we studied the high-affinity interactions between the amino-terminal domains of GluA2 and GluA3 AMPA receptors.

Signaling by assembly: how the host innate immune system responds to pathogen dsDNA Seamus R. Morrone1, Sarah Stratmann2, Mariusz Matyszewski1, Xiong Yu3, Edward H. Egelman3, Antoine van Oijen2,4, and Jungsan Sohn1 1Johns Hopkins School of Medicine, 2University of Groningen, 3University of Virginia, 4University of Wollongong In the human innate immune system, Absent-in-melanoma-2-like receptors (ALRs) assemble signaling filaments on double-stranded (ds)DNA arising from invading pathogens (Fig. 1). Unlike conventional enzymatic relays, ALRs then transduce signals by inducing the assembly of downstream filaments, eventually generating inflammatory responses (Fig. 1). Although essential for defense against a number of pathogens (e.g. HIV and L. monocytogenes), persistent ALR complexes promote autoimmune disorders (e.g. Sjogren’s€ syndrome and lupus).

140 ABSTRACT

Here, we aim to elucidate the underlying mechanisms by which ALRs carry out pathogen-dsDNA sensing pathways. We focus on two prototypical family members, namely IFI16 and AIM2; our approach includes solution assays, electron microscopy, and single-molecule techniques. IFI16 recognizes pathogen-dsDNA both in the host nucleus and cytoplasm, and thus raising a question as to how it distinguishes self from nonself dsDNA. We found that IFI16 uses dsDNA as a one- dimensional diffusion-scaffold to assemble into filaments, consequently allowing the size of dsDNA to regulate its assembly activity. Importantly, nucleosomes represent barriers that prevent IFI16 from targeting host dsDNA by interfering with its assembly. This unique scanning-assisted assembly mechanism would allow IFI16 to distinguish self- from nonself-dsDNA in the nucleus. AIM2 assembles into a signaling filament on pathogen-dsDNA in the host cytoplasm. It was previously proposed that the pyrin domain of AIM2 (PYD, AIM2PYD) inhibits dsDNA binding; however, our new assays revealed that AIM2PYD plays a critical positive role both in filament formation and dsDNA binding. As observed for IFI16, the size of exposed dsDNA acts a key regulator for the filament assembly by AIM2. The helical symmetry of the upstream AIM2PYD filament is consistent with the filament assembled by the PYD of the downstream ASC adaptor, suggesting that upstream ALR filaments act as a structural template for assembling downstream filaments. Relevant publications: Morrone et al., 2014, PNAS Morrone et al., 2015, Nature Communications Stratmann et al., 2015, eLife

Analysis of Eukaryotic Heme a Synthase Cox15 and Associated Dysfunctions Samantha Swenson1, Andrew Cannon1, Nicholas Harris2, Nicholas Taylor2, Jennifer L. Fox2, and Oleh Khalimonchuk1 1University of Nebraska-Lincoln, Department of Biochemistry, Redox Biology Center, 2College of Char- leston, Department of Chemistry and Biochemistry Heme is an essential but highly reactive cofactor in the cell. Tight regulation of its trafficking process is required as heme can be cytotoxic in its free form. Cytochrome c oxidase (CcO), the key enzyme of the mitochondrial electron transport chain, requires modified heme cofactors for enzymatic activity. How- ever, the exact mechanism of heme trafficking to CcO is unclear. Two conserved proteins, heme o synthase (HOS) Cox10 and heme a synthase (HAS) Cox15, mediate the process along with the assembly factors Coa2 and Shy1/SURF1. Mutations in Cox10 and Cox15 are linked to several congenital diseases including fatal infantile hypertrophic cardiomyopathy. Therefore, better understanding of how the heme a biosynthetic pathway functions as well as how heme is delivered to CcO is required to design treatments for this presently non-curable disease. It is known that Cox10 oligomerization is important for coupling heme a synthesis and its insertion into CcO. However, the nature of interactions between

141 ABSTRACT

Cox10 and Cox15, as well as the physiological role of Cox15 oligomerization remain to be determined. Here, we identify important residues for Cox15 catalysis, determine the oligomerization state of HAS and how those oligomers are stabilized. We also interrogate the role of the unstructured linker between the two domains of Cox15, and elucidate the importance of conserved residues that lead to fatale infantile hypertrophic cardiomyopathy.

On the role of C-Terminal tail helical domain of anabaena sensory rhodopsin transducer in unusual high stability, ligand and receptor interaction. Vishwa Trivedi1, Tashmay Jones1 and Rene Walker1, Rory Henderson2 and TK S Kumar2 1Department of Natural Science, College of Science, Engineering and Mathematics, Bethune Cookman University, 2Department of Chemistry and Biochemistry, University of Arkansas The Anabaena sensory rhodopsin transducer [ASRT] is a 125 amino acid protein that has been indicated to function as a signaling molecule downstream of the cyanobacterial sensory rhodopsin photoreceptor. The crystal structure along with solution state NMR has reveled that this beta stranded protein exists in tetrameric state. Interestingly a recent study has demonstrated the eukaryotic-like interaction of ASRT with DNA. Besides interaction with photoreceptor, ASR and DNA binding ability, the ASRT display a common structural fold that may be transform it as an unique carbohydrate binding module. However, the signaling state/mechanism of ASRT is obscure. Our initial data supports the hypothesis that carboxyl terminus of ASRT is involved in highly stable oligomeric assembly and may be linked to signaling state of this novel transducer molecule. We observed that position of hexa-histidine tag, used for ease of single step purification at carboxyl terminus destabilize the SDS resistant oligomeric state. We explored further the unique stability of ASRT by selectively removing carboxyl end domain [aa sequence 100-125 deletion]. The unfolding profile suggests that this region of 100-125 is critical in unusual higher stability. The truncated form of ASRT does not exhibit any significant influence due to presence of the hexahistidine tag on oligomeric state. In addition to unique stability, we have observed xylan binding to ASRT. Our preliminary modeling of receptor’s cytoplasmic, flexible region and ASRT using CABS-dock is supported by pull down assay and isothermal titration calorimetry. The unfolding of ASRT and its deletion forms using chemical denaturant, xylan binding and docking model will be presented to outline the putative signaling state of this transducer protein. [Supported by National Institute of Health, NIGMS SCORE award to VT [SC3 GM113803].

Creation of an artificial protein fiber by an easy-to-use designing of a protein-protein interaction Sota Yagi1, Satoshi Akanuma2, Tatsuya Uchida3, Akihiko Yamagishi1 1Department of Applied Sciences, Tokyo University of Pharmacy and Life Sciences, 2Faculty of Human Sciences, Waseda University, 3Department of Molecular Life Sciences, Tokyo University of Pharmacy and Life Sciences Although there are a limited number of successful de novo designs of protein-protein interaction (PPI) with computational methods, it is still difficult to design PPIs between arbitrary proteins. In contrast, recent works of de novo design of coiled-coil proteins have shown that the creation of interaction between a-helical peptides is relatively easy. In this study, we applied the design principle of helix-helix interaction to create the de novo PPI between two proteins, sulerythrin and LARFH. We first engineered a mutated sulerythrin that has six additional leucine and six aspartate residues on its two a-helices (1). We also engineered a mutated LARFH so that three leucines and three arginines were aligned on an a- helix. The resulting mutated proteins, sulerythrin-6L6D and LARFH-IV-3L3R were expected to interact to each other through inter-molecular coiled-coil formation that was stabilized by hydrophobic and ionic interactions. A pull-down assay demonstrated that sulerythrin-6L6D and LARFH-IV-3L3R indeed interact to each other specifically. The dissociation constant of the interaction determined by a FRET experiment was 0.42 mM. Thus, we successfully generated the de novo PPI with the medium binding affinity by

142 ABSTRACT modifying the exposed residues on the helices of the targeted proteins. We also produced mutated sulerythrin and LARFH so that they had two identical interfaces at both ends of their teriary structures. After the resulting proteins were mixed at a 1:1 molar ratio, fibrous structures were observed by atomic force microscopy (2). Therefore, our PPI design method also contributed to generating the artificial protein fiber. (1) Yagi et al. (2014) BBA Proteins and Proteomics 1844, 553–560 (2) Yagi et al. (2016) BBA Proteins and Proteomics 1864, 479–487

Peptide binders based on complementary Armadillo Repeat Protein Fragments Erich Michel1, Randall Watson1, Andreas Pluckthun€ 2, Oliver Zerbe1 1Department of Chemistry and 2Department of Biochemistry, University of Zurich We have recently discovered that it is possible to reconstitute a designed Armadillo repeat protein (ArmRP) from complementary fragments (Watson et al., Structure 2014). The two fragments, when mixed, will form a complex with nanomolar affinity. The structure of the complex is almost identical to the structure of the entire (single-chain) protein. ArmRP can be split at several places, within repeats as well as between repeats. The affinity of the fragments various between nanomolar and micromolar Kd’s, depending on the exact location of the split site. We have also seen that a repeat protein, that can bind a peptide, can be split into two fragments, and that the reconstituted complex again recognizes the peptide. We have then created a small library of C-terminal fragments, that all form complexes with a particular N-terminal fragment. Using NMR in combination with differentially labeled fragments we could demonstrate that upon addition of a peptide, for which only one of the present protein complexes presents a high-affinity binder, the equilibrium is shifted towards the complex that represents the best binder for that peptide. We present spectroscopic (NMR) and other biophysical data to characterize the underlying proteins and protein complexes. We feel we have discovered a novel interesting property of ArmRP, in that they can be reconstituted from a large number of different complementary fragments, and that interesting biochemical applications may emerge from these systems. Reference: R. Watson, M. Christen, C. Ewald, F. Bumback, C. Reichen, M. Mihajlovic, E. Schmidt, P. Guntert,€ A. Pluckthun,€ A. Caflisch, O. Zerbe (2014): Spontaneous Self-Assembly of Engineered Armadillo Repeat Pro- tein Fragments into a Folded Structure, Structure, 22, 985-995.

Global Analysis of Data from Multiple Biophysical Methods for Studying Protein Complex Stoichiometry and Affinity Huaying Zhao1 and Peter Schuck1 1Dynamics of Macromolecular Assembly Section, Laboratory of Cellular Imaging and Macromolecular Biophysics, National Institute of Biomedical Imaging and Bioengineering, National Institutes of Health Multi-protein interactions are ubiquitous and play a key role in cellular regulation and signal transduction mechanisms. It is essential to unravel their binding schemes and elucidate the role of each

143 ABSTRACT component in order to reach a mechanistic understanding of the relevant physiological functions. How- ever, the biophysical characterization of stoichiometry and energetic states of multi-protein complexes in solution is far from trivial, and generally beyond the capability of classical methods of studying protein interactions, such as analytical ultracentrifugation, isothermal titration microcalorimetry, biosensor, and fluorescence methods. To address this problem, we have implemented in the software SEDPHAT a global multi-method analysis (GMMA) approach combining the information content from different complementary biophysical disciplines into a comprehensive joint analysis that can provide a more detailed view of protein interactions. Even for simple systems, fundamentally such an analysis would constitute a convincing proof that we understand in full detail both the experimental techniques and the macromolecules under study. Indeed, the observed discrepancies might enable us to identify systematic errors in techniques, which are otherwise very difficult or even impossible to assess and could be vastly underestimated. By exploiting synergy between different techniques, GMMA improves the accuracy of thermodynamic parameters, and increases the complexity of interacting systems suitable for solution binding studies.

Mapping Integrin I Domain Binding Sites on Type I Collagen Monomers and Fibrils by microscopy Jie Zhu1, Ana Monica Nunes1 and Jean Baum1 1Rutgers University Collagen interactions with integrin cell surface receptors are essential for platelet aggregation, cell development and hemostasis. Fibrillar collagen is the most abundant protein in the body and provides the structural basis of connective tissues, bones and extracellular matrix (ECM). Integrin a1b1 and a2b1 are the principle cell surface proteins that interact with collagen in the ECM, binding to the GXX’GEX’’ motifs of collagen. The mechanism of binding of collagen to integrin is not understood. Here we present and compare the binding of integrin a2b1 I domains to type I collagen in both monomeric and fibrillar conformations. Using atomic force microscopy (AFM), transmission electron microscopy (TEM), and biolayer interferometry (BLI) we visualize integrin I domain - collagen monomer and integrin I domain - collagen fibril complexes and find that integrin binds to collagen monomers tighter than collagen fibrils. We propose that this difference originates from the accessibility of binding motifs on free collagen monomers versus fibrils. This work sheds light on the orientation of type I collagen fibril surfaces and helps elucidate the mechanism of collagen interactions with its receptors.

Biochemical Determination of APOBEC3A Interactions with ssDNA Samantha Ziegler1 and Yong Xiong1 1Yale University The APOBEC3 (A3) family of proteins are zinc-coordinating cytidine deaminases that mutate cytidine to uridine in single-stranded DNA (ssDNA). Each A3 protein recognizes a specific DNA hotspot sequence. The A3 proteins function in the innate immune response to a variety of viruses, resulting in the hypermutation of the viral single-stranded DNA. This hypermutation often makes the virus nonviable. In addition, the A3 family has the potential to mutate genomic DNA, as both A3A and A3B have been implicated in somatic mutations found in different cancers, highlighting the importance of deciphering the interaction between DNA and the A3 proteins. The purpose of our work is to determine how A3A recognizes its ssDNA substrate, providing mechanistic insight into both specificity and mode of A3A- ssDNA recognition. We have investigated how the length of ssDNA and localization of the A3A-specific hotspot within the oligomer affect the A3A-ssDNA interaction. We found that the ssDNA must be at least 9 nucleotides long in order for binding to occur. These results shed light onto how A3A targets viral DNA and host ssDNA generated during DNA replication.

144 ABSTRACT

PQ – PROTEOMICS Mass spectrometry analysis of NXS/T glycosylation sites in recombinant glycoproteins Izabela Sokolowska1, Armand G. Ngounou Wetie1, Alisa G. Woods1 and Costel C. Darie1 1Biochemistry & Proteomics Group Department of Chemistry & Biomolecular Science, Clarkson University Full structural characterization of chimeric recombinant proteins that are used as therapeutics includes analysis of post-translational modifications (PTMs) such as disulfide bridge assignment or investigation of glycosylation sites. These PTMs are essential to understanding the complete three-dimensional structure or the conformation of these proteins, as well as their solubility or stability. An important feature of IgG-based chimeric proteins is the NXS/T N-linked glycosylation site from the Fc part of the IgG, a site that may be critical in determining the solubility or stability of the chimeric protein. Here, we employed a targeted proteomics approach to investigate whether introduction of new N-linked glycosylation sites into a chimeric recombinant protein influences the glycosylation of existing glycosylation sites. Specifically, we over-expressed the chimeric construct containing the Fc region of the IgG fused to the exons 7-8 of mouse ZP3 (IgG-Fc-ZP3E7) and purified the protein product. We used an IgG-heavy chain (IgG-HC) control. We analyzed glycosylation of chimeric and control protein by trypsin-AspN double digest by nanoliquid chromatography-tandem mass spectrometry (nanoLC-MS/MS) using data dependent analysis (DDA) and information dependent analysis (IDA or DDA with inclusion list). Our findings suggest that manipulation of N-glycosylation of recombinant proteins may aid development of new approaches for controlling solubility and stability of recombinant proteins. Preliminary data: Both IgG-HC and IgG-Fc-ZP3E7 proteins were glycosylated and contained N-linked oligosaccharides that are removed upon PNGaseF treatment. Our experimental approach allowed us to identify the N-linked glycosylation sites in both IgG-HC and IgG-Fc-ZP3E7 proteins. In control experiments, the glycosylation site was occupied as expected. After PNGaseF treatment we identified a glycosylation site sequence with Asn residue (96NST98) converted into Asp (96DST98) showing that the 96NST98 glycosylation site within IgG-HC was occupied by an oligosaccharide. However, in the IgG-Fc- ZP3E7 protein only one out of three NXS/T glycosylation sites was occupied by N-linked oligosaccharides. Only the 291NSS293 glycosylation site was occupied by an oligosaccharide residue as expected, as it was identified as a 291DSS293-containing peptide upon PNGaseF treatment. The 288NCS290 site was not occupied by any oligosaccharide, as determined Western blotting shift assay. However, the 96NST98 site, which is part of the Fc part of the IgG was not glycosylated, as determined by nanoLC- MS/MS. Therefore, we concluded that in the IgG-Fc-ZP3E7 protein, upon introduction of additional potential NXS/T glycosylation sites within its sequence, the original 96NST98 glycosylation site from the Fc region of the IgG-Fc-ZP3E7 protein is no longer glycosylated. Novelty: Manipulation of N-glycosylation of recombinant proteins may aid development of new approaches for controlling solubility and stability of recombinant proteins. Preliminary data: Both IgG-HC and IgG-Fc-ZP3E7 proteins were glycosylated and contained N-linked oligosaccharides that are removed upon PNGaseF treatment. Our experimental approach allowed us to identify the N-linked glycosylation sites in both IgG-HC and IgG-Fc-ZP3E7 proteins. In control experiments, the glycosylation site was occupied as expected. After PNGaseF treatment we identified a glycosylation site sequence with Asn residue (96NST98) converted into Asp (96DST98) showing that the 96NST98 glycosylation site within IgG-HC was occupied by an oligosaccharide. However, in the IgG-Fc- ZP3E7 protein only one out of three NXS/T glycosylation sites was occupied by N-linked oligosaccharides. Only the 291NSS293 glycosylation site was occupied by an oligosaccharide residue as expected, as it was identified as a 291DSS293-containing peptide upon PNGaseF treatment. The 288NCS290 site was not occupied by any oligosaccharide, as determined Western blotting shift assay. However, the 96NST98 site, which is part of the Fc part of the IgG was not glycosylated, as determined by nanoLC- MS/MS. Therefore, we concluded that in the IgG-Fc-ZP3E7 protein, upon introduction of additional

145 ABSTRACT potential NXS/T glycosylation sites within its sequence, the original 96NST98 glycosylation site from the Fc region of the IgG-Fc-ZP3E7 protein is no longer glycosylated.

Spatial and Temporal control of Lysine Acetyl transferases (KATs): Ligand gated split KATs deSilva,C. S.1, Restituyo, E.1, Ghosh, I1 1Department of Chemistry and Biochemistry, University of Arizona The post-translational modification of proteins is central to genome regulation and cell signaling. Lysine acetyl transferase (KATs) were formerly known as histone acetylases (HATs) as they acetylated histones and thereby control transcription. However, recent studies strongly suggest that the acetylation mark is far more common protein modification with possibly over 7,000 acetylated proteins in the human proteome. There are numerous KATs and currently available small molecule based inhibition methods are not uniquely specific while RNAi based gene knock down studies can fail to provide details related to the true role of any enzyme as compensatory acetylation may occur. To address this problem, we have developed a method to selectively turn-on the activity of individual enzymes utilizing small molecules. We have successfully created the first generation of ligand inducible split-KATs, GCN5 and PCAF. Here inactive N- and C fragments are attached to the FKBP/FRB proteins, and heterodimerization is achieved by the small molecule ligand, rapamycin. We are currently using this approach for designing several new KATs, Myst2 and HAT1, which belong to different families, with the intention of exploring the possibility to establish temporal and spatial control over multiple KATs simultaneously in relevant cell lines.

Oncogenic epithelial cell-derived exosomes containing Rac1 and PAK2 induce angiogenesis in recipient endothelial cells David Greening1 1La Trobe Institute for Molecular Biology The metastatic cascade describes the process by which tumour cells escape their primary site and colo- nize secondary locations. Tumour angiogenesis facilitates passage, and cells at the leading edge of the primary tumour are thought to undergo epithelial-mesenchymal transition (EMT) to acquire increased motility and invasiveness. Whether leading edge EMT cells directly promote endothelial cell recruitment and angiogenesis remains largely unknown, and the role of exosomes (30-150 nm diameter extracellular vesicles) in this process has not yet been explored. Using Ras-transformed MDCK cells (21D1 cells) that exhibit a complete EMT phenotype, and MDCK cells stably expressing the master transcriptional regulator YBX1 (MDCKYBX1) that display an intermediate EMT phenotype, we examined the effects of exosomes from these cells on recipient endothelial cells. Firstly, we monitored in vitro cell motility and tube formation of 2F-2B cells supplemented with exosomes. MDCK exosomes had no effect, while MDCKYBX1 and 21D1 exosomes significantly enhanced 2F-2B cell motility, and increased the tube length and tube branch points. Next, exosome-supplemented 2F-2B cells were subcutaneously injected as Matrigel plugs into NOD/SCID mice. After 21 days, tail vein injection of FITC-dextran revealed that only 2F-2B cell plugs treated with MDCKYBX1 and 21D1 exosomes were perfused systemically. Compar- ative proteomic analysis highlighted that 21D1 exosomes contained VEGF-associated proteins (NRP1, NRP2, and TNFRSF12A), while MDCKYBX1 exosomes were enriched with activated Rac1 and PAK2. To validate this, 2F-2B cells and HUVECs were pre-treated with PAK inhibitors (FRAX597 and PF-3758309) then supplemented with exosomes. While PAK inhibitor treatment did not significantly impede tube formation promoted by 21D1 exosomes, tube length and branching was reduced to baseline levels despite treatment with MDCKYBX1 exosomes. Our results demonstrate for the first time that oncogenic cells undergoing EMT can communicate with endothelial cells via exosomes, and establish exosomal Rac1/ PAK2 as angiogenic promoters that may function from very early stages of the metastatic cascade.

146 ABSTRACT

Spatially targeted optical microproteomics (STOMP): Isolation and proteomic analysis of micron- scale features in pathological specimens Hadley, K.C.1,2, Rakhit R.3, Guo, H.4, Sun, Y.1,2, Jonkman, J.E.5, McLaurin, J.6, Hazrati, L.N.7, Emili, A.4, Chakrabartty, A.1,2,8 1Department of Medical Biophysics, University of Toronto, Toronto, Canada, 2Princess Margaret Can- cer Centre (formerly Ontario Cancer Institute), University Health Network, Toronto, Canada, 3Depart- ment of Chemical and Systems Biology, Stanford University, Stanford, United States, 4Terrence Donnelly Centre for Cellular & Biomolecular Research, Department of Molecular Genetics, University of Toronto, Toronto, Canada, 5Advanced Optical Microscopy Facility, Princess Margaret Cancer Centre, University Health Network, Toronto, Canada, 6Biological Sciences, Sunnybrook Research Institute, Toronto, Canada, 7Department of Paediatric Laboratory Medicine, The Hospital for Sick Children, Toronto, Canada, 8Department of Biochemistry, University of Toronto, Toronto, Canada. Spatially targeted optical microproteomics (STOMP) is a novel proteomics technique for interrogating micron-scale regions of interest (ROIs) in tissue, with no requirement for genetic manipulation (1). Methanol or formalin-fixed specimens are stained with fluorescent dyes or antibodies to identify ROIs, then soaked in solutions containing the photo-tag 4-benzoylbenzyl-glycyl-hexahistidine. Confocal imaging along with two photon excitation are used to covalently couple photo-tags to all proteins within each ROI. Spatial resolution of photo-tagging was measured at 0.67 mm in the xy-plane and 1.48 mm axially: approximately a ten-fold resolution improvement over laser capture microdissection. The tissue is then solubilized, and photo-tagged proteins are isolated by affinity chromatography and identified by mass spectrometry. As a test case, we examined amyloid plaques in the Alzheimer’s disease (AD) TgCRND8 mouse model (methanol post-fixed frozen sections) and a post-mortem human AD case (formalin fixed tissue), confirming known plaque constituents and discovering new proteins not previously identified in senile plaques. STOMP can be applied to various biological samples including cell lines, primary cell cultures, ex vivo specimens, biopsy samples, and fixed post-mortem tissue. References: 1. Hadley, K.C. et al. (2015) “Determining composition of micron-scale protein deposits in neurodegenerative disease by spatially targeted optical microproteomics.” eLife 4:e09579 PMID: 26418743.

Structural and functional effects of SNPs in the hydrophobic core of hGSTP1-1: Y80C mutation Mathura, S1; Makeleni, N1 and Dirr, HW1 1Protein Structure-Function Research Unit, School of Molecular and Cell Biology, University of the Witwatersrand, Johannesburg, South Africa Keywords: hGSTP1-1, SNP, Y80C, chemotherapeutics The structure-function relationship of human cytosolic glutathione-s-transferase P1-1 (hGSTP1-1) and its implication for anticancer drug-resistance remains intriguing. Here, the classical Phase II detoxification enzyme of the GST superfamily was found to overexpress in chemotherapeutic-resistant tumor cell lines. This leads to accelerated detoxification of drug substrates and an acquired resistance.1 Detoxification inhibition may therefore steer drug development towards a new class of hGSTP1-1 inhibitor-based chemotherapeutics. Structurally, homodimeric hGSTP1-1 has two ligand-binding sites per subunit: a GSH-specific G-site (contained within the N-terminal thioredoxin-like domain), and nonspecific hydrophobic substrate-binding H-site (contained within the all-helical C-terminal domain).2 Interestingly, the conserved Y80 residue sits between both subunit interface and domain interface. The prevailing view is that Y80 facilitates the binding of other ligands (e.g. tocopherol) to the Hsite, which was found to inhibit detoxification at the G-site.3 Our strategy is therefore to examine the structure- function significance of the Y80 residue using an interesting SNP mutation, Y80C. We compare the structure-function features relative to those of the wild-type (WT) to elucidate how the identity of the residue at this position influences the catalytic, structural and stability properties at the binding sites.

147 ABSTRACT

Circular dichroism and fluorescence spectra reveal minor secondary and tertiary alterations of Y80C structure relative to WT. Enzymatic activity was maintained suggesting that structural alterations have minimal effect on the G-site enzymatic function due to the induced mutation (specific activity/ lmol.min-1.mg-1: WT 5 43.9 6 1.6; Y80C 5 36.5 6 1.3). Interestingly, thermal denaturation studies, reveal some destabilisation in the enzyme as compared to WT (TmWT 5 52 6 3oC; TmY80C 5 47 6 3oC). This trend was further examined by pulse proteolysis unfolding (CmWT 5 4.2 6 0.02 M; CmY80C 5 3.2 6 0.01 M). 1. McIlwain, CC; Townsend, DM and Tew, KD; Oncogene, (2006) 25, 1639-1648.; 2. Dirr, HW; Reinemer, P and Huber, R; J. Mol. Biol., (1994) 243, 72-92.; 3. Ralat, LA and Colman, RF; Biochemistry, (2006) 45, 12491-12499.

PR – PROTEOSTASIS AND QUALITY CONTROL Dynamic Periplasmic Chaperone Reservoir Facilitates Biogenesis of Outer Membrane Proteins Shawn M. Costello1, Ashlee M. Plummer1, Patrick J. Fleming1, and Karen G. Fleming1 1Thomas C. Jenkins Department of Biophysics, Johns Hopkins University Outer membrane protein (OMP) biogenesis is critical to bacterial physiology including pathogenesis and antibiotic resistance. The process of OMP biogenesis has been thoroughly studied in vivo and each of its components have been studied in isolation in vitro. This work integrates parameters and observations from both in vivo and in vitro experiments into an integrated model. We utilize this model to computationally assess OMP biogenesis in a global manner. Using deterministic and stochastic methods we are able to simulate OMP biogenesis under several genetic conditions, each of which successfully replicates experimental observations. We observe that unfolded OMP has a prolonged lifetime in the periplasm where it makes, on average, hundreds of short-lived interactions with chaperones before folding into its native state. We find that some periplasmic chaperones function primarily as quality control factors, which complements the folding catalysis function of other chaperones. Mechanistic hypotheses proposed in the literature include either a physical bridge between the inner and outer membranes or parallel folding pathways catalyzed by different chaperones. The implementation of these mechanisms is not required to reproduce experimental results. Finally, we find a finely tuned balance between thermodynamic and kinetic parameters maximizes OMP folding flux and minimizes aggregation and unnecessary degradation. This work provides a holistic interpretation of OMP biogenesis that provides unique insight into this essential pathway.

Substrate Ubiquitination Controls the Unfolding Ability of the Proteasome Eden L. Reichard1, Giavanna G. Chirico1, William J. Dewey1, Nicholas D. Nassif1, Katelyn E. Bard1, Nickolas E. Millas1 & Daniel A. Kraut1 1Department of Chemistry, Villanova University In eukaryotic cells, proteins are targeted to the proteasome for degradation by polyubiquitination. These proteins bind to ubiquitin receptors, are engaged and unfolded by proteasomal ATPases, and are processively degraded. The factors determining to what extent the proteasome can successfully unfold and degrade a substrate are still poorly understood. We find that the architecture of polyubiquitin chains attached to a substrate affects the proteasome’s ability to unfold and degrade the substrate, with K48- or mixed-linkage chains leading to greater processivity than K63-linked chains. Ubiquitin- independent targeting of substrates to the proteasome gave substantially lower processivity of degradation than ubiquitin-dependent targeting. Thus, even though ubiquitin chains are removed early in degradation, during substrate engagement, remarkably they dramatically affect the later unfolding of a protein domain. Our work supports a model in which a polyubiquitin chain associated with a substrate switches the proteasome into an activated state that persists throughout the degradation process.

148 ABSTRACT

Assembly and dysregulation of the M. tuberculosis Clp protease Karl R Schmitz1, Alvaro J Amor1, Jason K Sello2, Robert T Sauer1 1Department of Biology, Massachusetts Institute of Technology, 2Department of Chemistry, Brown University Clp proteases carry out regulated ATP-fueled degradation of protein substrates through the collaboration of a AAA1 unfoldase and a ClpP peptidase. In Mycobacterium tuberculosis, a major global pathogen, ClpP is composed of discrete ClpP1 and ClpP2 heptameric rings. Assembly of inactive ClpP1 and ClpP2 rings into an active ClpP1P2 complex is stimulated by binding of unfoldases, by translocation of substrates into the peptidase barrel, and by peptides that mimic unfoldase or substrate interactions. To better understand how the active protease assembles, we used bio-layer interferometry to monitor the kinetics of assembly and disassembly, and the effects of individual binding partners on this process. We further show that peptide antibiotics capable of substituting for unfoldase binding can have opposing effects on proteolysis activity. These findings add to our understanding of how Clp proteases function in M. tuberculosis, and may influence the development of improved peptide antibiotics.

PS – SINGLE MOLECULE STUDIES Identification of the Molecular Origin of Disease with Single Molecular Optical Tweezers Jeneffer England1, Yuxin Hao1, Susan S. Taylor2, Rodrigo A Maillard1 1Department of Chemistry, Georgetown University, Washington, DC, 2Department of Pharmacology, University of California, San Diego, La Jolla, CA Protein Kinase A (PKA) plays a major role in the cell by acting as a signaling hub in all eukaryotic cells. PKA is composed of a catalytic (C) and regulatory (R) subunit that combine to form an inactive holoenzyme complex. The binding of two molecules of cAMP molecules to each cyclic nucleotide-binding (CNB) domain, CNB-A and CNB-B, enables the release of the active C-subunit. Mutations that affect the interaction between the R and C subunit result in the misregulation of PKA and consequently, in tumors that are associated with Cushing’s syndrome and Fibrolamellar Hepatocellular Carcinoma. The crucial involvement of PKA in cellular process and disease emphasize the importance of understanding the basic mechanisms of activation and regulation of this protein. In this study, a single R-subunit molecule has been unfolded in the absence and presence of the C-subunit with optical tweezers. Interestingly, two distributions of unfolding forces were observed with a force similar to and greater than the unfolding forces observed for apo R. This study provides direct insight into the thermodynamicforces between

149 ABSTRACT

R:C surface interactions. Future work includes the incorporation of disease related mutations of the R- subunit to probe the effect on the energies between the R:C interface and the unfolding of a PKA chimera in which the R and C subunits are connected by a peptide linker. These studies may provide insight into the identification of the molecular origin of diseases due to mutations of PKA and investigate the effect of allosteric inhibitors on PKA function.

Single-protein free-solution optical trap for studying their dynamics, structure, and behavior Ryan M Gelfand1 1CREOL The College of Optics and Photonics, UCF Nano-aperture optical trapping (NAOT) is a method that can be used to trap single protein molecules in free-solution; no tethers, labels, or tags are required so that the dynamics and the structure of the protein in full three dimensions can be studied with little to no steric hindrance. Light transmission is monitored through a nano-aperture milled into an opaque gold film, once a single protein is trapped the transmission increases due to the dielectric loading of the aperture; furthermore, the signal fluctua- tion increases due to both the micro and macro motion of the trapped protein. These contain valuable information about the system. By analyzing them we can study protein-protein and protein-small molecule interactions, protein behavior in different solution environments, and protein structure using two- photon stimulated anti-stokes Raman spectroscopy. This new tool has successfully been used to study the binding kinetics of a variety of protein-small molecule complexes. For example, the equilibrium constant as well as the association and dissociation rates have been determined for human serum albumin with small ligands such as phenytoin and tolbu- tamide, and the results agreed with those previously published. And since this method is only limited by the speed of the photodetector it can characterize kinetics on the order of single nanoseconds and theoretically down to femtoseconds. Another advantage of NAOT is that since it is a single molecule technique it can detect processes, for example, rare events or heterogeneity that the ensemble studies may miss during the observation of complex protein binding dynamics. In the future this method which has already been successful in studying a variety of protein-small molecule interactions will be used to expand our knowledge base and provide a simple and inexpensive tool for biophysics researchers.

PT – STRUCTURE (X-RAY/NMR/EM) NMR and X-Ray Crystallography-Based Structural Investigations of Peptide Catalysts Nadia C. Abascal, Anthony J. Metrano, Phillip A. Lichtor, Michael W. Giuliano, & Scott J. Miller* *Yale University While investigations of protein structure have steadily become more and more refined over the years, similar investigations in the realm of peptide structure have been less numerous. This poster will present studies of peptide-based catalyst structures in the context of oxidation-competent tetra-, penta-, and hexamers. Employing both X-ray crystallography and two-dimensional NMR spectroscopy, we have improved our understanding of the conformational preferences of the catalysts, and their possible connections to the selectivity the catalysts induce in bromination, Baeyer-Villiger, and epoxidation reactions. This poster will discuss structure-function trends in two classes of catalytic peptides, which contain N- terminal dimethylaminoalanine (Dmaa) or aspartic acid as essential catalytic residues, respectively.

The Unique Myristoylation Signal of the Feline Immunodeficiency Virus Matrix Protein Janae L. Baptiste1 and Michael F. Summers1 1Howard Hughes Medical Institute, Department of Chemistry and Biochemistry, University of Maryland The human immunodeficiency virus (HIV) remains a global health concern. Drug resistance and non- compliance to current treatment demonstrate the necessity for novel therapeutics, but development of

150 ABSTRACT new therapies has been limited by the absence of an appropriate animal model. Because of strong resemblance between the feline immunodeficiency virus (FIV) and HIV, felines are a tenable animal model for HIV in humans. The matrix protein (MA) mediates viral assembly, a process common to both HIV and FIV for which no therapies are commercially available. Both HIV and FIV MA require the presence of the myristate moiety, a saturated fatty acid that is covalently linked to the N-terminus of MA by N-myristyltransferase (NMT) via its recognition of the myristoylation signal of the substrate MA. Myris- toylated mammalian proteins commonly have the myristoylation signal M-G-X-X-X-S/T, where X is any amino acid residue, however FIV MA exhibits the unique myristoylation signal M-G-X-X-X-G. Because myristoylated feline proteins follow the common mammalian sequence and this unique myristoylation signal is conserved in several FIV isolates, it is hypothesized that the adaptation of FIV MA to bear a non-consensus myristoylation sequence may be an evolutionary advantage. This work aims to determine the three-dimensional structure of FIV MA and characterize its function by means of nuclear magnetic resonance spectroscopy in order to provide a framework to study the structural and functional impacts of mutations to FIV MA to bear the common mammalian myristoylation signal.

Investigating zinc finger recognition of epigenetically modified DNA Bethany Buck-Koehntop1 1University of Utah The Cys2His2 zinc finger domain is a well-defined compact structure stabilized by a coordinated zinc atom. Tandem arrays of these concerted motifs are often utilized to elicit specific recognition of nucleic acid sequences. While structurally conserved, this general fold can be utilized to distinguish a diverse range of biological targets, requiring a deeper understanding for the molecular basis of recognition exhibited by these motifs. Our laboratory investigates a family of Cys2His2 zinc finger containing methyl-CpG binding proteins (MBPs), termed the ZBTB MBPs, that mediate transcription by specifically targeting both methylated and sequence-specific non-methylated DNA sites. DNA methylation is an essential epigenetic modification in eukaryotes required for genomic stability and control of gene expression. Consequently, aberrant alterations in genomic methyl-CpG patterns are associated with a multitude of diseases, including cancer. Thus, understanding the mechanisms by which MBPs mediate disease relevant transcription is an essential component required for discerning disease onset and progression. Here we utilize an interdisciplinary approach to begin delineating the mechanisms by which this family of MBPs recognize DNA and regulate transcription. Specifically, we have used biophysical characterization of the binding interactions for the ZF domains of this MBP family with their cognate DNA targets to begin discerning the molecular basis for their differential mode of DNA recognition. Further, our lab couples in vitro findings with in cell methodologies to better understand the transcriptional effects of these proteins in cells. Together, these findings provide initial insights into the molecular basis by which these proteins preferentially recognize, interpret and translate DNA signals into cellular transcriptional responses.

Uncovering Stimulator Binding Mechanism to Soluble Guanylyl Cyclase by Solution NMR Cheng-Yu Chen†, Jessica A. Wales†, Andrzej Weichsel†, James E. Sheppeck‡, Joon Jung‡, Paul A. Renhowe‡, and William R. Montfort† †Department of Chemistry and Biochemistry, University of Arizona, ‡Ironwood Pharmaceuticals Soluble guanylate cyclase (sGC) is the primary nitric oxide (NO) receptor and the key enzyme in regulating NO-signaling pathways in mammals. Binding of NO stimulates the production of 3’,5’-cyclic guano- sine monophosphate (cGMP), which acts as a secondary messenger in regulating physiological functions including vasodilation, platelet aggregation, neurotransmission, and myocardial function. Dysfunction of sGC is associated to various forms of cardiovascular diseases. sGC stimulators such as the FDA approved

151 ABSTRACT compound riociguat for treating pulmonary hypertension stimulate the cyclase activity by stabilizing the NO-activated conformation. Nonetheless, the molecular mechanism of stimulation is lacking due to a poor understanding of sGC structure/function, and an unclear drug binding mechanism. In this study, we focused on sGC stimulators YC-1 and its derivatives, Bay 41-2272 and IWP-051, a newly developed stimulator compound, for their binding to truncated Manduca sexta sGC (Ms sGC) as well as the heme domain (H-NOX) homologs from bacteria including Closteridium botulinum (Cb), She- wanella woodyi (Sw), Fischerella sp (Fs). and Nostoc sp (Ns). We showed significant transferred NOESY enhancement for IWP-051 upon binding to Ms sGC, Cb H-NOX, and Sw H-NOX. These findings suggest specific binding of stimulators to the H-NOX domains. To further identify residues involved in stimulator binding, we used chemical shift perturbation induced by compounds IWP-051 and Bay 41-2272 upon binding to Sw H-NOX, and Fs H-NOX. The assigned resonances of Sw H-NOX allowed us to locate the compound binding site. Together, our studies suggest a common binding mode conserved throughout evolution and provide fundamental insight for developing improved compounds targeting cardiovascular diseases.

Uncovering the Membrane Interacting Regions of the HSV-1 Fusogen gB Rebecca S. Cooper1, Elka R. Georgieva2, Henry B. Rogalin1, Jack H. Freed2, and Ekaterina E. Heldwein1 1Department of Molecular Biology and Microbiology, Tufts University School of Medicine, 2National Biomedical Center for Advanced ESR Technology (ACERT), Cornell University Herpesviruses cause lifelong, latent infections and a wide range of ailments. Remarkably, four viral glycoproteins, gD, gH/gL and gB, are required to achieve fusion of the virion and its target cell. This process is ultimately mediated and powered by a large rearrangement of the fusogen gB. A major barrier to elucidating this refolding process and its regulation is that, while the structure of the inactive, postfusion form of the gB ectodomain is available, the structure of active, prefusion gB remains elusive. Furthermore, important membrane-interacting regulatory regions of gB, including its membrane proximal region (MPR), transmembrane domain (TMD), and cytoplasmic domain (cytodomain), have not yet been resolved. We hypothesize that, collectively, these domains both facilitate fusion and stabilize the metastable prefusion conformation of gB through their interactions with the membrane. To determine the structure of the MPR, TMD and cytodomain, we have crystallized full-length gB in a detergent environment. In each protomer of the gB trimer, the MPR appears to be a lipid embedded helix. It surfaces just prior to taking a sharp turn into the TMD helix, which forms a funnel with the other TMD helices as it traverses the membrane. Since the cytodomain remains unstructured in these crystals, we also used electron spin resonance (ESR) spectroscopy to characterize its secondary structure and association with liposomes. We found that the cytodomain C-terminus adopts a helical conformation upon embedding in anionic membranes, potentially enabling this domain to restrain fusion by acting as a clamp that braces against the bilayer. These experiments represent a major departure from the typical practice of analyzing soluble fusogen ectodomains and are expected to reveal long-sought details of the fusogenic conformational changes in gB, clarifying the complex herpesvirus entry process and facilitating vaccine development.

Recognition of diverse NES peptides by the Exportin CRM1 Ho Yee Joyce Fung1, Yuh Min Chook1 1University of Texas Southwestern Medical Center CRM1 (Exportin-1 or XPO1) is the most prominent nuclear exporter in the cell, which recognizes hundreds of protein cargoes by binding their nuclear export signals (NESs). NESs are 8-15 residues-long sequences with regularly spaced hydrophobic residues that bind to five hydrophobic pockets (P0-4) in the CRM1 NES-binding groove. We recently discovered that the binding groove can bind NESs in a

152 ABSTRACT bi-directional manner and this knowledge led to a large expansion of NES consensus patterns. However, structures of only 5 NESs bound to CRM1 are available at this time, representing only 4 of the 10 NES consensus patterns. The extent of structural diversity of CRM1-bound NESs remains underexplored. Here, we report 8 new CRM1-NES structures, which revealed both new NES conformations and unexpected features of the CRM1 groove. Structural comparison of all CRM1-bound NESs revealed that the peptides are made up of two structural elements. One element is often a 3-turn a-helix that occupies the wide and shallow P0-P2 pockets in the CRM1 groove, and the other element is a 3-residue strand that binds the narrow but deep P3 and P4 pockets. A common feature of all NESs is the loss of helical character at the transition between the two structural elements, which coincides with narrowing of the NES-binding groove by CRM1 residue Lys568. We found that CRM1 Lys568 is critical for binding NESs and for steric filtering to generate selectivity for true NESs. In summary, we have obtained structures for all known NES consensus patterns and identified a key specificity-determining feature of the CRM1 groove. Currently available NES prediction algorithms that use only sequence-derived information suffer from low precision. Our findings provide new structure-informed criteria for development of structure- based computational approaches to improve NES prediction accuracy.

Isolating the inner domain of the HIV-1 envelope as an independent molecule exposing the C1- C2 gp120 region involved in potent Fc mediated effector function to HIV-1 Neelakshi Gohain1, William D. Tolbert1, Maxime Veillette1, Jean-Philippe Chapleau1, Chiara Orlandi1, Andres Finzi1, George Lewis1, Marzena Pazgier1 1University of Maryland, USA Antiretroviral therapy (ART) is an effective and lifesaving approach to control HIV-1 infection but it does not eradicate the virus. A preventive vaccine still remains the ultimate hope for controlling and/or stop- ping the global HIV pandemic. A growing consensus, supported by the Rv144 clinical trial, suggests a relevant role of Fc mediated effector functions of antibodies such as antibody-dependent cellular cytotoxicity (ADCC) in vaccine induced protective responses. The conformational and non-neutralizing epitopes in the first and second constant (C1/C2) region of the HIV-1 gp120 envelope (Env) (Cluster A or A32-like epitopes) were found to be major targets for antibodies acting via ADCC induced by the Rv144 vaccine regiment. We describe the first successful attempt at isolating the inner domain of HIV-1 gp120 (ID) as an independent molecule that encapsulates the conformational A32-like epitopes within a minimal structural unit without the complication of other known epitope specificities. The ID design was guided by our atomic level description of the A32 epitope region gained from crystal structures of several Cluster A antibodies in complex with CD4-triggered gp120. Through two steps of structure-based design, we developed ID-2, consisting of the inner domain of gp120 stabilized in the CD4-bound conformation by an inter-layer disulfide bond expressed independently of the outer domain. ID2 stably expresses the transitional C1-C2 epitopes involved in potent FcR-effector function to HIV-1 as indicated by physiochemical, antigenicity and functional testing. Our inner domain design recapitulates the specificities of the potent Fc mediated effector functions in the humoral response of RV144 vaccinees without the complication of other known epitope regions and thereby represents a novel probe/immunogen candidate for inducing antibody responses to the A32-like epitope region.

Structural characterization of FHA-mediated regulation of a virulent ABC transporter in Mycobac- terium tuberculosis Florian Heinkel1, Mark Okon1,Jorg€ Gsponer1, Lawrence McIntosh1 1Department of Biochemistry and Molecular Biology, University of British Columbia Auto-inhibition, the ability of one part of a protein to inhibit the function of another, is a common “on- site” strategy for biological regulation. Intrinsically disordered regions of proteins and their post- translational modifications often contribute to this inhibitory function. One such example for a

153 ABSTRACT disordered region performing phosphorylation-dependent auto-inhibition is seen with proteins containing a Fork-Head Associated (FHA) domain. These domains bind phospho-threonine (pThr) containing peptides to help regulate phospho-dependent protein-protein interaction networks. These intermolecular interactions are auto-inhibited by the competitive intramolecular binding of a phosphorylated linker region adjacent to the FHA domain. Here we characterize structural and dynamic properties underlying the regulation of the unique ABC transporter Rv1747 from Mycobacterium tuberculosis (Mtb) by its tandem FHA domains and the phospho-acceptor sites in the intrinsically disordered linker connecting the two. Using various NMR experiments probing both structure and dynamics, we were able to show that while structurally very similar, the two FHA domains have a strikingly different stability and affinity to the target phospho-peptides, pointing to a correlation of conformational adaptability and binding affinity. Furthermore, we show evidence for the tandem FHA domain acting as a multi-interface dimerization domain of the transporter which, together with the affinity difference between the two domains, points to a sophisticated and tunable regulatory mechanism of this important transporter in Mtb.

Structural and functional studies of Mycobacterium tuberculosis MazF-mt6 toxin provide insight into RNA substrate specificity. Eric Hoffer1,2, Stacey J. Miles, Samantha Schwartz2, and Christine M. Dunhama2 1Department of Biochemistry, Emory University School of Medicine, 2Biochemistry, Cell and Develop- mental Biology program Toxin-antitoxin systems are found ubiquitously in prokaryotes and play key roles in bacterial survival by regulating essential processes during stress. E. coli has 19 putative toxin-antitoxin gene pairs, while the family is expanded to 90 in Mycobacterium tuberculosis (Mtb) suggesting that TA systems play fundamental survival roles in bacteria that undergo prolonged latency. At least nine of those 90 putative toxin-antitoxin gene pairs in Mtb belong to the MazEF family. In E. coli, the homologous MazF toxin cleaves free mRNA during stress, while the antitoxin MazE inhibits MazF during non-stress growth. However, recent studies have shown that certain Mtb MazF toxins are capable of cleaving RNA molecules that are very different from the canonical E. coli mRNA substrates. Mtb MazF-mt6 cleaves Helix 70 (H70) of the 23S rRNA in the 50S ribosomal subunit, suggesting that MazF-mt6 has the potential to regulate protein synthesis on a global level by disrupting ribosome function. Here we report the 2.7 Å x-ray crystal structure of MazF-mt6 as well as biochemical analyses to address function. Our work demonstrates that MazF-mt6 cleaves not only an intact 50S ribosomal subunit, but also a short 35 mer RNA containing the H70 sequence. We identify MazF residues R13 and T36 as being the most critical for mRNA cleavage and furthermore show D10 may contribute to the positioning of critical catalytic residues. Given that toxin proteins of toxin-antitoxin systems are proposed novel antimicrobial targets, understanding the molecular mechanism of RNA target recognition is critical in effectively designing potential inhibitors for toxins.

Structure and carbohydrate-binding properties of a C-type lectin SPL with novel carbohydrate- recognition motifs Shuhei Itakura1, Hideaki Unno1, Shuichiro Goda1, and Tomomitsu Hatakeyama1 1Graduate School of Engineering, Nagasaki University Carbohydrate-binding proteins (lectins) play important roles in self-defense systems in marine invertebrates. One of the well-known animal lectins is the C-type lectin family, which shares C-type carbohydrate-recognition domains with similar basic folds. They are also commonly characterized by the tripeptide sequences QPD or EPN suggesting galactose- or mannose-specificities, respectively. In this study, we have purified a C-type lectin SPL from the bivalve Saxidomus purpuratus, and found that this lectin has an interesting carbohydrate-recognition mode using a novel carbohydrate-recognition motifs.

154 ABSTRACT

SPL was isolated from S. purpuratus homogenate by the affinity chromatography using a GlcNAc- immobilized column. Its cDNA was cloned by PCR using the primers designed based on the N-terminal amino acid sequence. The deduced amino acid sequence as well as X-ray crystallographic analysis indicated that SPL belongs to the C-type lectin family, composed of a dimer with either identical or similar polypeptide chains. In its carbohydrate-binding sites, two novel recognition motifs RPD and KPD were identified. SPL exhibited preferential binding to GlcNAc and GalNAc, but not to Glc and Gal, suggesting that acetamido group is important for recognition by SPL with these motifs. This was further confirmed by the X-ray crystallographic analysis of SPL/GalNAc complex, which revealed that the acetamido group of GalNAc made intimate interactions with surrounding residues, but without an involvement of Ca21 ion, unlike other canonical C-type lectins.

Crystal Structure of PKG I:cGMP Complex Reveals a cGMP-Mediated Dimeric Interface that Facili- tates cGMP-induced Activation Jeong Joo Kim1,2, Robin Lorenz2, Stefan T. Arold3, Albert S. Reger1,7, Banumathi Sankaran4, Darren E. Casteel5, Friedrich W. Herberg2, and Choel Kim1,6* 1Department of Pharmacology, Baylor College of Medicine, 2Department of Biochemistry, University of Kassel, 3Division of Biological and Environmental Sciences and Engineering, Computational Bio- science Research Center, King Abdullah University of Science and Technology (KAUST), 4Berkeley Cen- ter for Structural Biology, Lawrence Berkeley National Laboratory, 5Department of Medicine, University of California, San Diego, 6Verna and Marrs McLean Department of Biochemistry and Molecular Biology, Baylor College of Medicine, 7Current address:Patheon Biologics-STL, St. Louis, Mis- souri 63134, United States of America Cyclic GMP-dependent protein kinase (PKG) is a key regulator of smooth muscle and vascular tone and represents an important drug target for treating hypertensive diseases and erectile dysfunction. Despite its importance, its activation mechanism is not fully understood due to a lack of structural information. To understand molecular details of its activation mechanism, we determined a 2.5 Å crystal structure of the PKG I regulatory (R)-domain bound with cGMP, which represents the activated state. Though we used a monomeric domain for crystallization, the structure reveals that two regulatory domains form a symmetric dimer where the cGMP bound high affinity pockets provide critical dimeric contacts. Using small angle X-ray scattering and mutagenesis, we demonstrate that this dimer exists in solution when it tethered by the N-terminal leucine zipper domain and that the dimer interface is crucial for activation. Finally, the structural comparison with the homologous cAMP-dependent protein kinase (PKA) reveals that PKG is drastically different from PKA in its R-domain arrangement upon cGMP binding, suggesting a novel activation mechanism for PKG.

Structure, inhibition, and regulation of a two-pore channel TPC1 Alexander Kintzer1 1University of California at San Francisco Two-pore channels (TPCs) comprise a subfamily of eukaryotic voltage- and ligand-gated cation channels that contain two non-equivalent tandem pore-forming subunits that then dimerize to form quasi- tetramers. Found in vacuolar or endolysosomal membranes, they regulate the conductance of Na1 and Ca21 ions, intravesicular pH, trafficking of filoviruses, excitability, and cellular amino acid homeostasis. Membrane potential and cytosolic Ca21-ions activate TPCs, whereas luminal low pH and Ca21, phosphorylation, and binding of pharmacophores are inhibitory. We report the crystal structure of TPC1 from Arabidopsis thaliana at 2.87Å resolution as a basis for understanding ion permeation, channel activation, the location of voltage-sensing domains, and regulatory ion-binding sites. We determined sites of phosphorylation in the N-terminal and C-terminal domains that are positioned to allosterically modulate cytoplasmic Ca21-activation. One of the two voltage sensing domains (VSD2) encodes voltage

155 ABSTRACT

sensitivity and inhibition by luminal Ca21 locks VSD2 in a resting conformation, distinct from the activated VSDs observed in structures of other voltage-gated ion channels. The structure shows how potent pharmacophore trans-Ned-19 allosterically acts to inhibit channel opening. In animals, trans-Ned-19 prevents infection by Ebola virus and Filoviruses by blocking fusion of the viral and endolysosomal membranes, thereby preventing delivery of their RNA into the host cytoplasm. The structure of TPC1 paves the way for understanding the complex function of these channels and may aid the development of antiviral compounds.

Structural and functional studies of the Mycobacterium tuberculosis vapbc30 toxin-antitoxin system In-Gyun Lee1, Sang Jae Lee1, Susanna Chae1, Ki-Young Lee1, Ji-Hun Kim1, and Bong-Jin Lee1* 1College of Pharmacy, Seoul National University, Seoul Korea Toxin-antitoxin (TA) systems play important roles in bacterial physiology, such as multidrug tolerance, biofilm formation, and arrest of cellular growth under stress conditions. To develop novel antimicrobial agents against tuberculosis, we focused on VapBC systems, which encompass more than half of TA systems in Mycobacterium tuberculosis. Here, we report that theMycobacterium tuberculosis VapC30 toxin regulates cellular growth through both magnesium and manganese ion- dependent ribonuclease activity and is inhibited by the cognate VapB30 antitoxin. We also determined the 2.7-Å resolution crystal structure of the M. tuberculosis VapBC30 complex, which revealed a novel process of inactivation of the VapC30 toxin via swapped blocking by the VapB30 antitoxin. Our study on M. tuberculosis VapBC30 leads us to design two kinds of VapB30 and VapC30-based novel peptides which successfully disrupt the toxin-antitoxin complex and thus activate the ribonuclease activity of the VapC30 toxin. Our discovery herein possibly paves the way to treat tuberculosis for next generation.

156 ABSTRACT

References [1] Lee KY, Lee KY, Kim JH, Lee YG, Lee SH, Sim DW, Won HS, Lee BJ. Nucl. Acids. Res., 43(10), 5194- 207 (2015) [2] Lee IG, Lee SJ, Chae S, Lee KY, Kim JH, Lee BJ Nucl. Acids. Res., 43(10), 7624-7637 (2015)

Anatomy of the b-branching enzyme of polyketide biosynthesis and its interaction with an acyl- ACP substrate Finn P. Maloney1,2, Lena Gerwick3, William H. Gerwick3, David H. Sherman1,4, and Janet L. Smith1,5 1Life Sciences Institute, University of Michigan, 2Chemical Biology Doctoral Program, University of Michigan, 3Center for Marine Biotechnology and Biomedicine, Scripps Institution of Oceanography, 4Department of Medicinal Chemistry, University of Michigan, 5Department of Biological Chemistry, University of Michigan Engineering of modular polyketide synthase (PKS) biosynthetic pathways is a potential route to the discovery of new pharmaceuticals not accesible via chemical synthesis. Some PKS contain b-branching enzymes that enhance polyketide diversity by performing an alkyl-substitution to a specific intermedia- tein the pathway. b-branching proceeds by a conserved branch-formation step and variable branch- processing steps that tailor the final alkyl substituent. A 3-hydroxy-3-methylglutaryl synthase (HMGS) catalyzes an aldol-addition between an acetyl donor and a polyketide acceptor, forming an initial car- boxyalkyl branch. Each acyl substrate is tethered by a specialized donor or acceptor acyl carrier protein (ACPD, ACPA) via a phosphopantetheine (Ppant) cofactor. ACP selectivity by HMGS is essential to catalytic fidelity, and to preventing aberrant b-branching. High-resolution crystal structures of the HMGS- ACPD complex reveal a unique structural motif in ACPD that contributes to a striking shape and electrostatic complementarity with HMGS. The affinity ACPD-HMGS is unusually strong for a PKS enzyme-ACP pair, and is dependent on hydrophobic and ionic interactions of conserved residues. The Ppant adopts different conformations in the HMGS active site pre- and post-acetyl transfer, and a pocket in the active site is responsible for this substrate-dependent positioning. These observations provide a basis for analyzing ACP selectivity by HMGS, which may be used for future efforts to engineer PKS by b-branch incorporation. This work was supported by NIH grants R01-DK042303 to JLS and R01-CA108874 to WHG, DHS & JLS.

Hydration Dynamics of Hen Egg-White Lysozyme Bryan S. Marques1, Nathaniel V. Nucci1, Matthew A. Stetz1, and A. Joshua Wand1 1Johnson Research Foundation and Department of Biochemistry and Biophysics, University of Penn- sylvania Perelman School of Medicine Water, the universal solvent, is of utmost importance for virtually all biological processes, especially in its interaction with proteins and other macromolecules. By encapsulating proteins in reverse micelles, it is possible to use NMR nuclear Overhauser effect (NOE) spectroscopy in order to quantify interactions between protein surfaces and water molecules in a site-specific manner. This approach was used to investigate the surface hydration of hen egg-white lysozyme, particularly within the hydrophilic peptidoglycan binding cleft. While it is clear there is a wide distribution of hydration dynamics across the surface of the protein, within the binding cleft of the apo protein the waters are relatively dynamic. Similarly dynamic waters are also observed within a partially hydrophobic core more than 6 Å away from the surface of the protein. Binding of a peptidoglycan inhibitor traps slow interfacial waters. Finally, remote locations on the surface of the protein show an increase in hydration dynamics upon inhibitor binding.

157 ABSTRACT

The unique structure of HSV-1 UL21, a multifunctional tegument protein. Claire M. Metrick1 and Ekaterina E. Heldwein1 1Department of Molecular Biology and Microbiology and Graduate Program in Biochemistry, Sackler School of Graduate Biomedical Sciences, Tufts University School of Medicine The tegument layer of herpesviruses contains over a dozen proteins that play important structural roles in virion morphogenesis. Many of these proteins also facilitate early stages of viral replication. A detailed understanding of how these often multifunctional tegument proteins operate is lacking. HSV-1 UL21 is a 535-amino acid tegument protein that is conserved among alphaherpesviruses and is found in the cytoplasm and nucleus in transfected and infected cells. UL21 has been implicated in cytoplasmic budding, intracellular trafficking, and nuclear egress, yet remains largely uncharacterized. To provide a blueprint for exploration of the multiple roles of UL21, we determined the crystal structure of UL21 from HSV-1, which revealed a novel protein fold. Through structure analysis, we identified several regions of potential functional importance for future targeting by mutagenesis. We have also observed a novel ability of UL21 to bind nucleic acids. Current work focuses on carefully characterizing this binding interaction and determining the relevance of these in vitro observations during infection.

Crystal structure of a novel Ca21-dependent lectin GJL-I from the sea anemone Gyractis japonica and its carbohydrate-recognition mechanism Azusa Nakamura1, Hideaki Unno1, Shuichiro Goda1, and Tomomitsu Hatakeyama1 1Graduate School of Engineering, Nagasaki University In some marine invertebrates, lectins are known to be involved in innate immunity by recognizing specific carbohydrate chains on foreign microorganisms, leading to following responses, such as phagocytosis. However, there have been limited numbers of the reports on the structure and function of lectins from Cnidaria. In this study, we investigated the structure and function of a novel lectin GJL-I from the sea anemone Gyractis japonica. Amino acid sequence of GJL-I was determined by a combination of peptide sequencing and cDNA cloning. The resulting amino acid sequence is composed of 186 amino acid residues, in which N-terminal 22 residues correspond to the signal peptide. Since assumed N- terminal glutamine residue could not be detected by the Edman degradation of native GJL-I, its tryptic digests were analyzed by MALDI-TOF-MS. As a result, N-terminal glutamine residue was found to be converted to pyroglutamic acid. GJL-I forms a homodimer linked by disulfide bonds. Homology search on the database resulted in no homologous protein except for a putative lectin from the sea anemone Nematostella vectensis. Glycoconjugate microarray analysis demonstrated that GJL-I preferably binds several oligosaccharides with galactose moieties at their termini. Three-dimensional structure of GJL-I/ lactose complex was determined by the X-ray crystallographic analysis. In this structure, galactose moiety of lactose is recognized through coordinate bonds with Ca21 ion and hydrogen bonds with nearby residues in the binding site, which resembles that of C-type lectins. However, overall fold of GJL-I is totally different from that of C-type lectins, suggesting that they acquired similar carbohydrate recognition modes using Ca21 in spite of their independent evolutionary origins.

Structural characterization of protein Sis1 by Nuclear Magnetic Resonance PINHEIRO, GLAUCIA M.S.1, ALMEIDA, FABIO C.L.2, AMORIM, GISELE C. AND RAMOS, C.H.I.1 1Institute of Chemistry, University of Campinas UNICAMP, Campinas SP. Brazil, 2Institute of Medical Biochemistry and Nucleus for Structural Biology and Bioimaging (CENABIO) - Federal University of Rio de Janeiro, Rio de Janeiro RJ. Although the three-dimensional structure of a protein is coded in its amino acid sequence, during the cellular folding process, assistance of other proteins named as molecular chaperones are necessary for the achievement of the proper structure. Thus, chaperones are involved in protein homeostasis,

158 ABSTRACT preventing aggregation and refolding aggregates. Hsp70 (70 kDa heat shock protein) is among the most important chaperone family by helping folding of both newly synthesized and misfolded proteins. To perform its function Hsp70 requires the help of co-chaperones, such as that of the Hsp40 family that are associated with partially unfolded proteins, preventing their aggregation and delivering them to Hsp70. Structural studies in high resolution of these proteins are difficult due to the large molecular mass of these macromolecules. However, our research group has made important contributions on the structure and function of a yeast Hsp40, SIS1, solving the quaternary structure of wild-type and mutants by SAXS (Ramos e col., 2008) and identifying by NMR randomic regions that are involved in the interaction with client proteins (Borges e col., 2012). Recently, we have been executing experiments with SIS1, in order to obtain data on the structure and dynamics of this co-chaperone by high resolution NMR. Initial data obtained are very promising and suggest that we are close of a great achievement in Brazil, i.e. the determination, at high resolution, of the structure of a dimeric protein, wherein each monomer has close to 40 kDa.

Structures of cGMP-dependent Protein Kinase (PKG) I? Leucine Zippers Reveal an Interchain Disulfide Bond Important for Dimer Stability Liying Qin,#,§ Albert S. Reger,#,k,† Elaine Guo,D,† Matthew P. Yang,‡,† Peter Zwart,? Darren E. Casteel, & and Choel Kim*,§k §Verna and Marrs McLean Department of Biochemistry and Molecular Biology, kDepartment of Phar- macology, Baylor College of Medicine Department of Chemistry, ‡Department of Biochemistry, Rice University ?Lawrence Berkeley National Laboratory & Department of Medicine, University of Califor- nia at San Diego PKG Ia is a central regulator of smooth muscle tone and vasorelaxation. The N-terminal leucine zipper (LZ) domain dimerizes and targets PKG Ia by interacting with G-kinase anchoring proteins. The PKG Ia LZ contains C42 that is known to form a disulfide bond upon oxidation and activate PKG Ia. To understand the molecular details of the PKG Ia LZ and C42-C420 disulfide bond, we solved crystal structures of PKG Ia WT LZ and C42L LZ. Our data demonstrate that the disulfide bond C42-C420 dramatically stabilizes PKG Ia and that the C42L mimics the oxidized WT LZ structurally.

Crystallization by controlled evaporation with acoustic monitoring Limone Rosa1, Anna Tsygelnytska1, Andrea Kocmarek1, Olivia Wiggins1, Denise Laspina1, Alexei Soares1 1Brookhaven National Laboratory Finding conditions that induce protein crystallization is a slow process because conventional hanging drop methods screen a narrow band of precipitant concentrations. Furthermore, when a crystal inducing condition is identified, the exact precipitant concentration that yields nucleation and growth remains unknown (requiring additional experiments to optimize the crystallization conditions). Here, we describe a technique for screening crystallization conditions by gradually increasing the protein and precipitant concentrations through controlled evaporation of the solvent. The rate of evaporation (and hence the increase in protein and precipitant concentration) is regulated by adjusting the size of the apertures through which evaporation takes place. A phase diagram is generated for each trial using an acoustic Echo 550 instrument that continuously monitors the protein and precipitant concentrations using acoustic pulses. This sonar-like monitoring [1] allows a wider range of concentrations to be tested and continuously tracks the concentration of protein and precipitant (including the concentration where nucleation occurs). Evaporation is halted when crystallization is complete, when precipitate forms, or when the fluid level falls below a threshold. This strategy expedites the design of an optimized crystallization protocol without the need for additional fine tuning once an initial crystallization condition is detected. We have also observed that widening the range of tested precipitant concentrations

159 ABSTRACT can change the crystal packing (for example, exposing new areas of the protein surface to fragment screening assays). Conveniently, the same acoustic instrument that is used to monitor crystallization can generate a strong sound pulse to harvest the crystals [2] [3]. By coupling a wide concentration screen, acoustic monitoring, and acoustic harvesting, we describe a robust system for discovering crystallization conditions. [1] D. L. Ericson. (2016). J. Lab. Automation 21: 107-114. [2] A. Soares et al. (2011). Biochemistry 50: 4399-4401. [3] X. Yin et al. (2014). Acta Cryst D70: 1177-1189.

Molecular Understanding of USP7 Substrate Recognition and C-Terminal Activation Lionel Rouge1, Travis W. Bainbridge1, Paola Di Lello1, Ingrid E. Wertz1, Till Maurer1, James A. Ernst1, Jeremy Murray1 1Genentech The deubiquitinating enzyme USP7 has a pivotal role in regulating the stability of proteins involved in fundamental cellular processes of normal biology and disease. Despite the importance of USP7, the mechanisms underlying substrate recognition and catalytic activation are poorly understood. Through structural, biochemical and biophysical analyses, we have elucidated the molecular mechanism by which the C-terminal 19 amino acids of USP7 (residues 1084-1102) enhance the ubiquitin cleavage activity of the deubiquitinase (DUB) domain. Our data demonstrate the C-terminal peptide binds the activation cleft in the catalytic domain and stabilizes the catalytically competent conformation of USP7. Additional structures of longer fragments of USP7, as well as solution studies, provide insight into full-length USP7, the role of the UBL domains, and demonstrate that both substrate recognition and deubiquitinase activity are highly regulated by the catalytic and noncatalytic domains of USP7; a feature that could be essential for the proper function of other multi-domain DUBs.

NSLS-II macromolecular crystallography beamlines: opportunities for advanced data collection Alexei S. Soares1, Jean Jakoncic1, Martin R. Fuchs1, Dieter K. Schneider1, Lonny E. Berman1, Stuart Myers1, Edwin Lazo1, Dileep K. Bhogadi1, Herbert J. Bernstein1, Wuxian Shi1, John Skinner1, Bruno Martins1, Vivian Stojanoff1, Robert M. Sweet1, and Sean McSweeney1 1Brookhaven National Laboratory The two new beamlines at the National Synchrotron Light Source-II, for Frontier Macromolecular Crys- tallography (FMX) and for highly Automated Macromolecular Crystallography (AMX), will begin user operation in fall 2016. The low emittance of the NSLS-II storage ring is the basis for providing previously unattainable brilliance to address current and upcoming challenges in crystallography. With a target flux of 1013 ph/s at 1 Å, and beam sizes of 1 – 20 mm (FMX) and 4 – 100 mm (AMX), the new

160 ABSTRACT beamlines will provide up to two orders of magnitude higher flux densities than the brightest MX beamlines currently in user operation. One mission for these unprecedented flux densities is to leverage ultra-fast data rates to enhance experiments in serial crystallography. In addition, combinatorial crystallography will be advanced by the high level of automation. New approaches to specimen preparation, high density specimen holders, and beamline automation are required to keep up with the anticipated specimen throughput. We will address this challenge using a state of the art sample preparation facility for high throughput crystal deposition on high density formats such as MiniSpine bases, coupled to highly automated (/autono- mous) specimen handling in the beamlines.

Tet3 CXXC domain is an epigenetic reader for 5-carboxylcytosine Jikui Song1, Seung-Gi Jin2,3, Zhi-Min Zhang1, Thomas L. Dunwell3, Matthew R. Harter1, Xiwei Wu4, Jennifer Johnson2, Zheng Li5, Jiancheng Liu6, Piroska E. Szabo2,4, Qiang Lu6, Guo-liang Xu5, and Gerd P. Pfeifer2,3 1Department of Biochemistry, University of California Riverside, 2Center for Epigenetics, Van Andel Research Institute, 3Department of Cancer Biology, Beckman Research Institute of the City of Hope, 4Department of Molecular and Cellular Biology, Beckman Research Institute of the City of Hope, 5The State Key Laboratory of Molecular Biology, Institute of Biochemistry and Cell Biology, Chinese Acad- emy of Sciences, 6Department of Neurosciences, Beckman Research Institute of the City of Hope During development, DNA methylation is dynamically regulated by ten-eleven translocation (Tet) proteins, which modulate the DNA methylation patterns via oxidization of 5-methylcytosine (5mC) to 5- hydroxymethylcytosine (5hmC), and further to 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC). However, the functional implication of these 5mC derivatives in gene regulation remains unclear. We identified that the CXXC domain of the full-length Tet3 (Tet3FL) isoform is a specific reader for 5caC- containing DNA, and determined the crystal structure of the Tet3-CXXC domain in complex with 5caC DNA at 1.3 Å resolution. The structure reveals that the Tet3-CXXC domain binds to the 5caC DNA in a fashion distinct from its binding to unmodified cytosine. Our cellular assays further suggest that binding of the CXXC domain of Tet3FL to 5caC or C restricts its genome-wide 5mC oxidation capacity. Mapping of Tet3FL in neuronal cells shows that Tet3 is localized precisely at the transcription start sites (TSS) of genes involved in lysosome function, mRNA processing and key genes of the base excision repair pathway. These data suggest an anchoring mechanism for Tet3 in maintaining the unmethylated state of CpG islands at TSS regions, in which the epigenetic readout of 5caC by the Tet3 CXXC domain promotes its catalytic domain to rapidly oxidize additional 5mC bases in the immediate vicinity.

Structural characterization of non-active site, TrkA selective kinase inhibitors Hua Su1, Keith Rickert1, Kartik Narayan1, Christine Burlein1, Marina Bukhtiyarova1, Danielle Hurzy1, Craig Stump1, Xufang Zhang1, John Reid1, Srivanya Tummala1, Jennifer M. Shipman1, Maria Kornienko1, Abdelghani Achab1, Chad Chamberlin1, Peter Saradjian1, Berengere Sauvagnat1, Xianshu Yang1, Michael Ziebell1, Nickbarg Elliott1, John Sanders1, Steve Carroll1, Darrell Henze1, Andy Cooke1 1Merck and Co. Current therapies for chronic pain can have insufficient efficacy and lead to side effects, necessitating research on novel targets against pain. Although originally identified as an oncogene, TrkA is linked to pain and elevated levels of NGF, the ligand for TrkA, is associated with chronic pain. Antibodies that block TrkA interaction with its ligand, NGF, are in clinical trials for pain relief. Here, we describe the identification of TrkA-specific inhibitors and the structural basis for their selectivity over other Trk family kinases. The x-ray structures reveal a novel binding site outside the kinase active site that utilizes residues from the kinase domain and the juxtamembrane region. Three modes of binding with the

161 ABSTRACT juxtamembrane region are characterized through a series of ligand-bound complexes. The structures indicate a critical pharmacophore on the compounds that leads to the distinct binding modes. The mode of interaction can allow TrkA selectivity over TrkB and TrkC or promiscuous, pan-Trk inhibition. This highlights the difficulty in characterizing the structure-activity relationship of a chemical series in the absence of structural information due to large shifts in the interacting residues. These structures illustrate the flexibility of binding to sequences outside but adjacent to the kinase domain of TrkA. This knowledge allows development of compounds with specificity for TrkA or the family of Trk proteins.

Crystal structure of the D444V disease-causing mutant of human dihydrolipoamide dehydrogenase Eszter Szabo1*, Reka Mizsei1*, Zsofia Zambo1, Beata Torocsik1, Manfred S. Weiss2, Vera Adam-Vizi1 and Attila Ambrus1 1Department of Medical Biochemistry, Semmelweis University, Budapest, Hungary, 2Helmholtz- Zentrum Berlin fur€ Materialien und Energie, Berlin, Germany *these authors contributed equally to this work Dihydrolipoamide dehydrogenase is a common subunit (E3) of the alpha-ketoglutarate dehydrogenase (KGDHc), the pyruvate dehydrogenase and the branched-chain ketoacid dehydrogenase complexes and the glycine cleavage system. While the KGDHc is a key enzyme in the Krebs cycle, it is also the major producer of reactive oxygen species (ROS) in the mitochondrion as well as a very sensitive target of ROS. ROS generation by and dysfunctions of the human (h) KGDHc are implicated in senescence/ aging, neurodegenerative diseases, ischemia-reperfusion, hypoxia- and glutamate-induced cerebral damage and E3-deficiency, among other pathologies. KGDHc generates ROS via the E3 subunit, the pathogenic mutations of which cause the often lethal human disease, the E3-deficiency. The D444V pathogenic amino acid substitution of human (h) E3 leads to hypotonia, microcephaly, lactic acidosis, developmental delay, and hypoglycemia. Isolated hE3-D444V generates ROS at a significantly higher rate than does the isolated hE3, which may contribute to the severity of the pathologies. We determined the crystal structure of the isolated hE3-D444V at 1.8 Å resolution. In light of the structure we provide a putative explanation for the greatly compromised physiological activity as well as for the enhanced ROS-generation.

New labware for high throughput screening of crystallization conditions and chemical libraries Ella Teplitsky1, Karan Joshi2, Lauren Zipper3, Anthony Borelli4, Alexander Scalia5, Daniel Ericson6, Denise Laspina7, Alexei Soares8 1Department of Biochemistry and Cell Biology, Stony Brook University, 2Department of Electrical and computer Engineering, Stony Brook University, 3Department of Mechanical Engineering, Binghamton University, 4Department of Biology, Stony Brook University, 5Department of Biological Sciences, 6Department of Biomedical Engineering, University at Buffalo, SUNY, Buffalo, NY, USA, 7Department of Biology, Stony Brook University, 8Energy Sciences Directorate, NSLS-II, Brookhaven National Laboratory We describe “Flex Plate”, a 96 well crystallization plate that disassembles into individual pieces (called plate fragments) such that crystals in each of the 96 chambers can be individually subjected to diffraction experiments using a conventional magnetic goniometer head (figure 1). The Flex Plate is designed to operate in connection with standard liquid handling robots for high throughput screening of crystallization conditions or fragment libraries. This function allows a library of chemical fragments to be screened to map the binding properties of a region of interest on a protein. Two different designs are optimized for screening (a) crystallization conditions and (b) fragment libraries. When screening crystallization conditions with design (a), each crystallization chamber is paired with a single precipitant reservoir, so that many kinds of precipitants can be tested. When screening chemical libraries using design (b), crystallization chambers share a single reservoir. The crystallization area is transparent to X-rays so that data can be

162 ABSTRACT

obtained in situ from crystals in each plate fragment. A snap-on adapter couples each plate fragment to a magnetic base so that diffraction data can be obtained using a conventional goniometer either at room temperature [1] or cryogenically [2]. When coupled with advanced liquid handling robots such as the Echo 550, each Flex Plate can accommodate 1728 screened chemicals or precipitant conditions.

An Allosteric Model for Control of Pore Opening by Substrate Binding in the EutL Microcompartment Shell Protein Michael C. Thompson1,†, Duilio Cascio2, David J. Leibly1, and Todd O. Yeates1,2 1Department of Chemistry and Biochemistry, University of California, Los Angeles, 2UCLA-DOE Insti- tute for Genomics and Proteomics, University of California, Los Angeles The ethanolamine utilization (Eut) microcompartment is a protein-based metabolic organelle that is strongly associated with pathogenesis in bacteria that inhabit the human gut. The exterior shell of this elaborate protein complex is composed from a few thousand copies of BMC-domain shell proteins, which form a semi-permeable diffusion barrier that provides the interior enzymes with substrates and cofactors while simultaneously retaining metabolic intermediates. The ability of this protein shell to regulate passage of substrate and cofactor molecules is critical for microcompartment function, but the details of how this diffusion barrier can allow the passage of large cofactors while still retaining small intermediates remain unclear. Here we report structural and biophysical evidence to show that ethanolamine, the substrate of the Eut microcompartment, acts as a negative allosteric regulator of pore opening in a particular shell protein, EutL. Specifically, a series of X-ray crystal structures of EutL, along with equilibrium binding studies, reveal that ethanolamine binds to EutL at a site that exists in the closed- pore conformation and which is incompatible with opening of the large pore for cofactor transport. The allosteric mechanism we propose is consistent with the cofactor requirements of the Eut microcompartment, leading to a new model for EutL function. †Currently in the Department of Bioengineering & Therapeutic Sciences, University of California, San Francisco

Structural basis for effective nitric oxide decomposition in microbial denitrification Tosha, Takehiko1; Terasaka, Erina1, 2; Matsumoto, Kimi1, 2; Sugimoto, Hiroshi1; Shiro, Yoshitsugu1, 2 1RIKEN SPring-8, Sayo, Hyogo, Japan. 2.University of Hyogo, Ako, Hyogo, Japan. Nitric oxide (NO) plays diverse roles in biological processes including blood pressure regulation, neurotransmission, and immune response in mammal. On the other hand, NO showed cytotoxicity through tissue damaging due to its high reactivity with biomolecules. It is, therefore, essential to understand how biological systems control the action of NO to minimize its cytotoxic effect in cells. Microbial denitrification is a good model for elucidating in the regulating systems for the action of NO, since NO is produced as an intermediate product in denitrification, but NO is not accumulated in cells. Given that NO is produced by periplasmic nitrite reductase (NiR) and is decomposed by membrane-integrated NO

163 ABSTRACT reductase (NOR), NiR likely interacts with NOR to avoid diffusing NO into the cellular environment. We, then, tried to crystalize the cytochrome cd1-type NiR (cd1NiR)- cytochrome c-dependent NOR (cNOR) complex, and solved the structure at a resolution of 3.2 Å. The structure showed that cytochrome c domain of cd1NiR binds to a periplasmic hydrophilic domain of cNOR through van der Waals interactions and one salt bridge formation. The structure implies that positively charged patch of cd1NiR interacts with the phosphate groups of lipid bilayer, which could also contribute to the formation of the complex with cNOR. However, no obvious pathways for NO channeling from the heme d1 active site of cd1NiR to the heme/non-heme iron active center of cNOR was observed in the structure of the complex. This observation suggests that cd1NiR produces NO at the close proximity to cNOR by the complex formation, which facilitates NO binding to cNOR and rapid NO decomposition.

Structure and Physical Studies of the AlgH protein from Pseudomonas aeruginosa and the AlgH Protein Family Jeffrey L. Urbauer1, Aaron B. Cowley1, Henry T. Niedermaier1, Hayley E. Broussard1 and Ramona J. Bieber Urbauer1 1Department of Chemistry, and Department of Biochemistry and Molecular Biology, The University of Georgia The worsening problem of antibiotic resistance of pathogenic bacteria encourages not only efforts towards new antibiotics and antibiotic targets, but alternate strategies to control bacterial infections. One such strategy, complementary to antibiotic therapies, is to target virulence regulation, reasoning that targeting virulence avoids the strong evolutionary pressure exerted by antibiotics for selecting resistance. Along with other antibiotic resistant bacteria considered serious threats by the Centers for Disease Con- trol and Prevention (such as methicillin-resistant Staphyloccoccus aureus (MRSA), vancomycin-resistant Enterococcus (VRE) and others), is Multidrug-Resistant Pseudomonas aeruginosa (MDRPA). Some of these MDRPA strains are resistant to all known antibiotics, and comprise a high percentage (13%) of all hospital-acquired Pseudomonas aeruginosa infections. We are studying the AlgH protein from Pseudo- monas aeruginosa, the namesake for a family of proteins of unknown function. Although the precise mechanism by which AlgH and its orthologues function is unknown, studies have shown that the algH gene regulates production of many virulence factors, and presumably AlgH then serves a global regulatory function. So, controlling AlgH function may help mitigate virulence and consequently assist in controlling infections. Using NMR spectroscopy we determined the high-resolution solution structure of AlgH from Pseudomonas aeruginosa and are studying the structural and dynamics properties of AlgH and its orthologs from other bacteria. Employing evolutionary trace analysis, hydrogen-deuterium exchange and other methods we have identified important residues in AlgH, regions of stability and instability of the protein, and how these are correlated. Using molecular dynamics simulations we are studying conformational and structural transitions of these proteins and how they might participate in AlgH function. The results should assist in understanding the underlying physical bases for AlgH function.

Detection of Small Molecule Effectors of Pyruvate Carboxylase via the Development of a Novel High-throughput Screen Brittney N. Wyatt1 Martin St. Maurice1 1Marquette University Pyruvate carboxylase (PC) is a metabolic enzyme that catalyzes the conversion of pyruvate to oxaloacetate (OAA) in a wide range of organisms1-4. Small molecules directed against PC would greatly benefit studies on the cellular role of this enzyme and have the potential to be developed into antimicrobial agents. However, there are currently no known specific and potent small molecule regulators of PC. We aim to find, develop, and characterize novel small molecule effectors of PC that will aid in understanding the structure, mechanism, and regulation of this enzyme and will ultimately lead to new chemical probes to study PC in its cellular context. A novel fixed-time assay has been developed based on the reaction of

164 ABSTRACT

OAA with the diazonium salt, Fast Violet B (FVB), which produces a colored complex with an absorbance maximum at 530 nm5. Current results indicate that the fixed time assay is reproducible, sensitive and responsive to known effectors of Rhizobium etli PC, Staphylococcus aureus PC, and L. monocytogenes PC, and has a strong potential to be amenable to high-throughput screening with excellent Z-factors of 0.85 for the activation assay and 0.87 for the inhibition assay. The assay has been validated with a plate uniformity assessment test and was used in an initial pilot screen to test the assay’s response to 1,280 compounds. The pilot screen results indicate that the assay has the strong potential to find novel small molecule effectors of PC that can be developed into valuable research tools. [1 (2008) Biochem J. 3, 367-387. [2] (2001) PNAS 108, 8674-8679 [3] (2008) J. Biol. Chem. 283, 28048-28059. [4] (2010) J Bacteriol 192, 1774-1784. [5] (1990) Anal. Biochem. 189, 95-98.

The structure basis of ceramide transfer protein functional regulation Xiaolan Yao1 and Jennifer Prashek1 1School of Biological sciences, University of Missouri Kansas City Sphingolipids are not only important components of biological membranes but also regulators of a variety of cellular processes. The ceramide transfer (CERT) protein is a critical regulator of sphingolipids homeostasis. It carries out nonvesicular trafficking of ceramide from ER to Golgi, where ceramide is converted to sphingomyelin. CERT contains several domains and motifs. The amino terminal PH domain targets CERT to the Golgi by recognizing phosphatidylinositol-4-phosphate lipids (PI4P) that are enriched in the Golgi. The carboxyl terminal START domain is responsible for the ceramide transfer activity of CERT. Following the PH domain is a short stretch rich in serine and threonine residues called the SR motif. Phosphorylation of the SR motif inhibits both the PH domain binding to PI4P and START domain transfer of ceramide. Previous studies show this inhibition requires both the PH and START domains of CERT, suggesting an auto-inhibitory interaction between the two. In this study, we use a combination of structural and biochemical methods to investigate the structure basis of CERT functional regulation. Our data show PH and START domains indeed interact with each other. Importantly, the interaction interface overlaps with PH domain interface for PI4P binding, consistent with earlier finding that phosphorylation inhibition of PH binding to PI4P requires START domain. However, to our surprise, mutations that disrupt PH and START interaction increase the affinity for PI4P and ceramide transfer activity in both nonphosphory- lated CERT and a phosphorylation mimic of CERT, suggesting additional elements in CERT are required for phosphorylation inhibition. This finding also leads to the interesting implication that the unphosphorylated CERT is only active upon targeting to the Golgi membrane by PI4P binding, thus preventing any potential deliver of ceramide to unintended organelles inside the cell.

PU – SYNTHETIC BIOLOGY Producing Membrane Bound Proteins as Countermeasures to Infectious Diseases Wei He1, Angela C. Evans1, Martina Felderman2, Delia F. Tifrea3, Amy Rasley1, David Homan1, Kurt Kamrud2, Nathaniel Wang2, Sukumar Pal3, Luis M. de la Maza3, Bolyn Hubby2, Todd Peterson2, Nicholas O. Fischer1 and Matthew A. Coleman1,4 1Lawrence Livermore National Laboratory, 2Synthetic Genomics Vaccine Inc. La Jolla, CA, 3University of California Irvine, Pathology and Laboratory Medicine, 4University of California Davis, School of Medicine, Radiation Oncology Infectious diseases constitute a growing public health threat. Due to rapid population growth and frequent traveling, infectious pathogens can quickly spread to thousands or millions of individuals. Despite

165 ABSTRACT medical advances, infectious diseases continue to be a problem to our society. Here we developed nanolipoprotein (NLP) technology for incorporating membrane bound proteins as a medical counter- measure to infectious disease threats. NLPs are 8-20 nm disc-shaped particles formed by the spontaneous self-assembly of phospholipids and scaffold apolipoproteins. Our lab has developed NLP technology coupled with in vitro protein production (cell-free) to synthesize native-like pathogen proteins. Here, we show that NLPs can be used to incorporate functional membrane bound proteins, such as YopB and MOMP. NLPs have multiple biodefense and human health applications, and these NLP: membrane protein complexes can provide insight on the mechanism of infectious diseases as well as a platform for screening mitigation agents. NLP:membrane protein complexes can also be used to elicit protective immune responses.

Mutagenesis of Ribonuclease U2: Rational, Semi-rational and Random Approaches Beulah Solivio1 1University of Toronto, Canada Ribonuclease (RNase) U2 is a purine- specific enzyme that cuts the 3’- end of a stretch of ribonucleoti- des. It is a widely used enzyme in RNA sequencing by mass spectrometry, where post-transcriptionally modified nucleotide information can be determined. Our goal is to produce a variant of RNase U2 with an enhanced specificity in cleaving after adenosines. With the availability of the RNase U2 crystal structure, we were able to attempt a mutagenesis method based on a guided approach using bioinformatics to determine binding energies and in silico alanine scanning. Due to unforeseen factors, the RNase U2 variants generated from these methods yielded more non-specific cleavage than the wild-type enzyme. Site-saturation mutagenesis was also employed to produce a variety of enzymes carrying a mutation at key residues in the enzyme’s binding site. Finally, phage display method is used to randomize the entire binding site of RNase U2. The advantage and disadvantage of each method is discussed and the results of each are compared.

PV – THERAPEUTICS AND ANTIBODIES Effects of rearranging the domain order of a diabody-based IgG-like bispecific antibody on antitumor activity, degradation resistance, and pharmacokinetics Ryutaro Asano1,2, Ippei Shimomura1, Shozo Furumoto3, Mitsuo Umetsu1, and Izumi Kumagai1 1Department of Biomolecular Engineering, Graduate School of Engineering, Tohoku University, 2Pres- ent address: Department of Biotechnology and Life Science, Graduate School of Engineering, Tokyo University of Agriculture and Technology, 3Department of Radiopharmaceutical Chemistry, Graduate School of Pharmaceutical Sciences, Tohoku University One approach to creating more effective therapeutic antibodies is to develop IgG-like bispecific antibodies with tetravalency, which may provide several advantages such as bivalent binding to each target antigen. Although the effects of configuration and antibody-fragment type on the function of IgG-like bispecific antibodies have been studied, there have been only a few detailed studies of the influence of the variable fragment domain order. Here, we prepared four types of hEx3-scDb-Fc, IgG-like bispecific antibodies, built from a single-chain hEx3-Db [humanized bispecific diabody (bsDb) that targets EGFR and CD3], to investigate the influence of domain order and fusion manner on the function of a bsDb with an Fc fusion format. Higher cytotoxicities were observed with hEx3-scDb-Fcs with a VL–VH order (hEx3-scDb-Fc-LHs) compared with a VH–VL order, indicating that differences in the Fc fusion manner do not affect bsDb activity. In addition, the results of flow cytometry suggested that the higher cytotoxicities of hEx3-scDb-Fc-LH may be attributable to structural superiority in cross-linking between target cells and antigens. Interestingly, enhanced degradation resistance and prolonged in vivo half-life were also observed with LH type. Our results show that merely rearranging the domain order of diabody- based IgG-like bispecific antibodies can enhance not only their antitumor activity, but also their

166 ABSTRACT degradation resistance and in vivo half-life, and that hEx3-scDb-Fc-LHs are potent candidates for next- generation therapeutic antibodies.

Monoclonal antibodies specifically targeting amyloidogenic forms of transthyretin (TTR) with potential to treat TTR-related cardiomyopathy and polyneuropathy Natalie J. Galant1, Jeffrey N. Higaki2, Kevin C. Hadley1,AmyWon3, Stephen J. Tam2,Ken Flanagan2,Tarlochan Nijjar2, Ronald Torres2, Jose R. Tapia2, Joshua Salmans2, Robin Barbour2, Wagner Zago2, Gene G. Kinney2, Christopher M. Yip3, Avi Chakrabartty2 1Princess Margaret Cancer Centre, University Health Network, University of Toronto, Toronto, Ontario M5G 2C1, Canada, 2Prothena Biosciences Inc, South San Francisco, California 94080, USA, 3Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Ontario M5S 3E1, Canada Transthyretin (TTR, or prealbumin) is an abundant serum protein which normally forms soluble, stable homotetrameric complexes. Point mutations and unknown pathological conditions can favour the dissociation of the TTR tetramer into non-native monomers. These monomers aggregate and accumulate as amyloid throughout the body, particularly in the heart and peripheral nerves. This deposition of TTR amyloid in cardiac tissue and nerves results in the development of cardiomyopathy and polyneuropathy, respectively. We have recently developed conformation-specific polyclonal and monoclonal antibodies (mAbs) which can potentially treat both of these diseases via their ability to specifically recognize and bind to the disease-associated forms of TTR via a cryptotope (an epitope normally buried and inaccessi- ble in the native protein, but exposed in its altered conformation). These mAbs were demonstrated in vitro to specifically binding to misfolded TTR, inhibit fibril formation, induce phagocytic clearance of non-native and aggregated TTR, and immunoreact with TTR amyloid in diseased heart tissue (Galant et al., 2016; Higaki et al., Amyloid, 2016). We further investigated the mechanism of mAb-mediated inhibition of fibrillogenesis using immunogold transmission electron microscopy (TEM) and atomic force microscopy (AFM). These high resolution imaging techniques have confirmed the cryptotope as an effective mAb target due to its exposure within both pathological TTR misfolding intermediates and end-point insoluble TTR fibrils. These results further support the use of monoclonal antibodies to target pathological protein conformations as potentially effective immunotherapies for TTR amyloidosis. Galant, N.J., et al. (2016). Substoichiometric inhibition of transthyretin misfolding by immune- targeting sparsely populated misfolding intermediates: a potential diagnostic and therapeutic for TTR amyloidosis. In revision. Higaki, J.N., et al. (2016). Novel conformation-specific monoclonal antibodies against amyloidogenic forms of transthyretin. Amyloid. Mar 16:1-12.

An antibody discovery platform of KBIO and its application to developing antibody therapeutics and diagnostics Kyungjae Kang1, Hyung-Kyung Choi1, Keunwan Park1, Sungjin Kim1, Kiweon Cha1, and Daeyoung Kim1 1Division of Drug Development and Optimization, KBIO-New Drug Development Center, Osong-eup, Cheongju-si, Chungcheongbuk-do, Republic of Korea. Phage display technology has been one of the most effective, successful molecular evolutionary technologies since it was introduced by G. Smith in 1985. The technology, of particular, has been pivotal in selecting synthetic antibodies against numerous of diseases-specific targets through rationally-designed CDR structures built on VH, VL, or both scaffolds. KBIO-New Drug Development Center (KBIO-NDDC) in Korea has been equipped with a number of phage display antibody libraries ranging from single domain to Fab libraries. Recently, we have constructed synthetic human Fab phage display libraries on a VH-VL pair (VH3-23/Vk1-39) derived from a Fab clone with desirable properties including high bacterial expression, high non-aggregation, and high thermal stability. The CDRs presenting on the VH-VL pair were randomized in a manner to mimic

167 ABSTRACT natural diversity of antibodies, to avoid non-functional antibodies such as truncated forms of antibodies, and to exclude potential post-translational modification sites. We are currently validating those libraries using several targets in terms of selectability and developability. We as a R&D supporting public organization of Korea aim to contribute to developing antibodies for therapeutics and diagnostics so domes- tic or global antibody-related industries will get benefits through the efforts. In the presentation, some case studies performed with our antibody library line-ups will be introduced and our ongoing efforts to fulfill our mission will also be shared.

Contributions of loop histidine residues for Zinc ion binding and its stability of an anti-ZnO VHH antibody Ryosuke Sasaki, Hikaru Nakazawa, Izumi Kumagai, Mitsuo Umetsu, Koki Makabe1 1Yamagata University, Japan The interaction between biomolecules and inorganic materials has gained much interest for the development of sensors and highly controlled materials. For this purpose, 4F2, a VHH antibody with high affinity and specificity toward zinc oxide (ZnO) surfaces has been constructed. We have reported that 4F2 has an affinity not only toward ZnO but toward zinc ion (Zn21). The binding affinity between 4F2 and Zn21 determined by isothermal calorimetry showed a sub-milli molar dissociation constant. The binding event mainly occurred near histidine residues of the CDR3 loop. Here, we report that the histidine residues also contribute its stability. We measured several loop variants of 4F2 and determined their stability using differential scanning calorimetry. We found that the replacement of CDR1 induces the destabilization of 4F2 and that of CDR3 induces stability recover. Our results showing here provide the basis for the recognition mechanism of ZnO by 4F2.

Energetic Basis for Optimization of Cysteine Protease Inhibitors Conceic¸~ao A. Minetti1, Igor M. Prokopczyk2, Erika V.M. Orozco2, Fabiana Rosini2, David P. Remeta1, and Carlos A. Montanari2 1Department of Chemistry & Chemical Biology, Rutgers University, 2Institute of Chemistry, University of S~ao Paulo, S~ao Carlos, SP, Brazil Cysteine proteases of the papain superfamily are present in nearly all groups of eukaryotes and participate in a wide range of biological processes and associated pathologies. Considering their critical role in the life-cycle progression of numerous pathogenic protozoa, these proteases represent potential targets for selective inhibitors. Cruzain is the major cysteine protease of Trypanosoma Cruzi (T. cruzi) and has been the subject of extensive structure-based drug design strategies. While the latter is generally regarded as a viable approach, the optimization of lead candidates benefits from evaluation of physicochemical properties including characterization of the thermodynamic forces governing cruzain-inhibitor interactions via isothermal titration calorimetry (ITC). Stepwise modifications of a prototype inhibitor yield a subset of structural analogs that are evaluated on the basis of their cruzain inhibition activity. The compounds include analogs with a number of functional group substitutions relative to the parent compound. We find remarkable correlation between the binding affinities (Ka) and inhibitor potencies (Ki) for the structural analogs evaluated herein. Resolving the ITC-derived Gibbs energies into their enthalpic and entropic contributions furnishes significant insight regarding differential binding modes. Our combined structural/energetic approach extends the comparative analysis beyond simple correlations of inhibitor potency and binding affinity. Dissec- tion of the enthalpic-entropic balance in terms of stepwise chemical substitutions facilitates optimization of lead candidates amongst ligand sub-populations. Collectively, X-ray orthogonal validation, molecular dynamics simulations, and calorimetric measurements provide the requisite structural and energetic parameters to identify cruzain inhibitors that operate with the highest efficiency at subna- nomolar concentrations.

168 ABSTRACT

Recombinant single chain variable fragment antibody production in Escherichia coli and purification with affinity chromatography using Arista Slice Purification system Adam Vandergriff1,2, Vikas Padhye3, Ke Cheng1,2 1UNC/NCSU Joint Department of Biomedical Engineering, 2Molecular Biomedical Sciences, College of Veterinary Medicine, North Carolina State University, 3Practichem Recent work in both life science research and pharmaceutical industry have led to an increasing number of antibody based drugs. Of particular interest is the use of recombinant antibodies for targeting stem cells to injuries such as myocardial infarction. We have previously used an antibody based targeting system known to target hematopoietic stem cells to the heart. Many experiments involving the use of immunotherapy in animal models requires large amounts of antibody. In order to obtain the large amount of antibodies required for these experiments, we have begun to use E. coli to produce recombinant single chain antibody fragments (scFv). E. coli expression has many advantages over mammalian expression including easier transfection, lower cost media, and overall simpler use. Following expression, it is necessary to purify the antibody to remove any bacterial proteins. Using histidine affinity chromatography on the Practichem’s Arista Slice, we have been able to isolate sufficient amounts (4-16mg) of high purity scFv antibodies.

How insulin binds: structure of a micro-receptor complex and implications for analog design Nischay Rege1, Michael Glidden1, Nelson Phillips1,YanwuYang1, Faramarz Ismail-Beigi1, & Michael A. Weiss1 1Departments of Biochemistry and Medicine, Case Western Reserve University School of Medicine The discovery of insulin in 1921 represented a landmark in molecular medicine and led to extensive investigation of the structure and function of this globular protein hormone with application to therapeutic analog design. This presentation provides a summary of the current structural understanding of the active conformation of insulin in relation to its mechanism of receptor binding. Implications of recent crystal structures and NMR studies will be discussed as a foundation for the engineering of novel ultra-stable single-chain analogs, intended as basal (long-acting) insulin formulations for use in regions of the developing world lacking access to refrigeration yet challenged by a global pandemic of Type 2 diabetes mellitus (“diabesity”). A key constraint in the design of therapeutic insulin analogs is posed by their physical degradation to form amyloid. Spectroscopic studies of insulin fibrils have provided structural constraints regarding the molecular structure of a protofilament and distorted conformations of insulin proposed as intermediates in the process of fibrillation. Consideration of such constraints has highlighted the potential utility of (a) single-chain insulin analogs containing foreshortened connection domains and (b) engineered additional disulfide bridges as pioneered in elegant studies of two-chain analogs by Vinther, T. N. et al. Protein Science 24, 779-88 (2013). These modifications must accommodate induced fit of the hormone on receptor binding, a binding mechanism that entails splaying of the C-terminal segment of the B chain and potential changes in its N-terminal segment. Insight into the hormone-binding surfaces of the ectodomain of the insulin receptor and “micro-receptor” models of the hormone-receptor complex have enabled visualization at low resolution of how the splayed B chain inserts between domains of the receptor. These recent structures provide a new and promising foundation for analysis of structure-activity relationships with direct application to the design of novel insulin analogs. Efforts are underway toward the optimization of insulin analogs to address unmet needs of patients with diabetes mellitus in affluent societies and in the developing world. The speaker is grateful to Drs. M. C. Lawrence, J. G. Menting and C. Ward (WEHI, Melbourne) for their generous collaboration and encouragement in the course of these long-term studies. Additional collaborative ties with V. Chauhan, S.B. Kent, N. Phillips, D. F. Steiner (deceased), R. Tycko, J. Whittaker, and N. Wickramasinghe are warmly acknowledged. References: 1. Menting, J.G., Whittaker, J., Margetts, M.B., Whittaker, L.J., Kong, G.K.-W., Smith, B.J., Watson, C.J., Zakov a, L., Kletvikova, E., Jiracˇek, J., Steiner, D.F., Chan, S.J., Dodson, G.G., Brzozowski, A.M., Weiss, M.A., Ward, C.W., & Lawrence, M.C. (2013) Nature 493, 241–245.

169 ABSTRACT

2. Menting, J.G., Yang, Y., Chan, S.-J., Phillips, N.B., Smith, B.J., Whittaker, J., Wickramasinghe, N.P., Whittaker, L., Pandyarajan, V., Wan, Z.-l., Yadav, S.P., Carroll, J.M., Strokes, N., Roberts, Jr., C. T., Ismail- Beigi. F., Milewski, M., Donald F. Steiner, D.F., Chauhan, V.C., Ward, C.W., Weiss, M.A., & Michael C. Law- rence, M.C. (2014) Proc. Natl. Acad. Sci. USA. (PNAS-Plus) 111, E3395-404.

Finding novel therapeutic compounds against infectious diseases Ana C. Romero1 1University of Texas at Austin, College of Natural Sciences, Texas Institute for Discovery Education in Science (TIDES), Freshman Research Initiative (FRI), Department of Molecular Biosciences, Virtual Drug Screening Stream Enzymatic proteins can serve as an Achilles’s heel for many infectious diseases. The ability to effectively target and inhibit their molecular function provides an opportunity to mitigate the deleterious outcomes of disease states in humans. However, identifying new drug leads using traditional methods is an expensive and time consuming process. This research stream uses computers to sift through libraries of chemical structures and predict which ones may bind most effectively to a protein that is a potential drug target. One of the many diseases targeted in the stream is the Bubonic plague, caused by the bacterium Yersinia pestis. The plague is an emergent threat in many parts of the world as a public health concern and a potential bioterrorist weapon. The focus for the stream this semester is to search for chemical compounds that can bind specifically to a protein called YopH, also known as Tyrosine protein specific phosphatase, may lead to an effective therapeutic drug since this enzyme is essential for the virulence and infectivity of the bacterium. Virtual drug screening software is used to make predictions about which compounds may bind the best. The results are then visualized and interpreted with molecular graphics programs. Several of the best candidate molecules can then be tested in the wet lab to determine their efficacy in comparison to the computational predictions.

Phosphopeptide mapping and mass spectrometry analysis of human tristetraprolin Heping Cao1 and Kandan Sethumadhavan1 1U.S. Department of Agriculture, Agricultural Research Service, Southern Regional Research Center, New Orleans, LA, USA Cottonseeds are classified as either glanded or glandless seeds depending on the presence or absence of pigment glands which contain toxic gossypol. Glanded seeds compose of approximately 10% linters, 40% hulls and 50% kernels. The kernels contain about 35% of oil and 40% of protein. After oil extraction, com- mercial cottonseed meal contains up to 2% residual level of lipid. The presence of gossypol limits its use of meals primarily to feed ruminants, which have a relative high tolerance for the compound. The presence of gossypol, which can bind to protein, also makes it more difficult to recover concentrated protein fractions from the meal. Glandless seeds do not contain pigment glands and have only trace levels of gossypol which may be useful for potential utilization of the protein as a food ingredient or as a feed for non-ruminant animals. This development generated considerable interest within the cotton industry. Developing improved extraction techniques and new industrial uses for cottonseed proteins would provide additional markets for the cottonseed meal fraction and would provide a diversification of uses that would improve overall profit- ability of the seed. The objective of this study was to compare protein compositions between glanded and glandless cottonseeds. The seed were grounded into fine powder and extracted with aqueous buffer. The results showed that glandless seeds contained more proteins than glanded seeds. SDS-PAGE showed that several novel soluble proteins were present in glandless seeds but missing in glanded seeds. These proteins were characterized chromatographically and identified by mass spectrometry.

170 ABSTRACT

PW – TRANSCRIPTION/TRANSLATION/POST-TRANSLATIONAL MODIFICATIONS Nonredundant function for repeats within the RNA polymerase II C-terminal domain Stephen M. Fuchs1, Michael Babokhov1, Nihaarika Sharma1, Alexandra E. Exner1, Mohammad M. Mosaheb1, Summer A. Morrill1 1Tufts University, Department of Biology The C-terminal domain (CTD) of Rpb1, the largest subunit of eukaryotic RNA polymerase II, consists of repeating units of a seven amino acid sequence, YSPTSPS. The CTD coordinates the temporal association of dozens of protein factors with RNAPII as it transcribes a gene in a manner dependent on phosphorylation of CTD repeat residues. The work described here explores whether all repeats have redundant function or whether there is a regiospecific nature to the CTD. Using a library of CTD mutant plasmids containing serine to alanine mutations in specific regions of the CTD and a novel, tet-off gene regulation system in yeast, we have identified specific phenotypes for some repeats but not others. We go on to demonstrate that many of these phenotypes are dependent on the presence of Ser5 within repeats and believe they are consistent with impaired recruitment of Mediator. To our knowledge these data demonstrate, for the first time, a role for individual CTD repeats.

Epigenetic Regulation and the O-GlcNAcase: OGA as a Reader and Eraser of the Histone Code Katryn R Harwood1, John A. Hanover1 1NIDDK at the NIH One often-overlooked posttranslational modification (PTM) is protein O-GlcNAcylation. O-GlcNAc is added/removed by a single pair of proteins, OGT and OGA, respectively. O-GlcNAc transfer is the final step of the hexosamine biosynthetic pathway and its addition/removal likely informs upon organisms’ nutritive states. Of interest is not only the ability to translate the nutritive state into the up/down-regulation of genes, but to transmit this information to future generations. Many modifications including O- GlcNAc have been implicated in transgenerational inheritance. PTMs of histones that change chromatin structure and subsequent gene expression are also of note: these are largely conserved between humans and C. elegans. Potentially, this “code” of histone PTMs facilitates intergenerational communication on a transcriptional level. We hypothesize that interplay between O-GlcNAc cycling and histone PTM may allow for fine-tuning of transcriptional activity. In C. elegans, the early embryo reflects the germline gene expression profile and overall maternal contribution to transcriptional activity. Known histone PTMs maintain an important gene expression profile: methylation by the protein MES-4 is associated with active transcription primarily on autosomes while methylation by MES-2,3,6 is associated with repressive chromatin restricted largely to the X chromosome. In vitro work done in our lab identified a domain of OGA that is responsible for its specific interaction with MES-4 methyl marks. Current work has focused on “decoding” the interplay between O-GlcNAc cycling and these histone PTMs both in vitro and in C. elegans mutants to further interrogate the roles each play in transcriptional regulation as associated with metabolism, particularly as they relate to the transmission to future progeny. These experiments will provide insight into the mechanism by which the nutritive state is translated into changes in transcriptional profiles on the level of an entire organism across generations.

Determining histone deacetylase 8 substrates using non-natural amino acids Jeffrey Lopez1 1University of Michigan The histone deacetylase family is comprised of 18 enzymes that catalyze the removal of an acetyl group from the e-position of lysine residues. Lysine deacetylation is a dynamic post-translational modification that affects over 4000 different sites in macromolecules and has been implicated in various types of cancers (references) and neurodegenerative diseases. However, much of the specificity of each isozyme

171 ABSTRACT remains largely unknown, and conventional co-immunoprecipitation methods have proven ineffective in unmasking substrates. In this study, we have incorporated a photo reactive, non-natural amino acid, p- benzonyl-4-phenylalanine (BPA), into HDAC8, a structurally well characterized isozyme of this family with a still incomplete in vivo substrate pool and used photo crosslinking to trap both substrates and binding partners of HDAC8 in HEK293 lysates. Using photo crosslinking with the addition of both co- immunoprecipitation and proteomic analysis provides better insight into HDAC8’s in vivo substrates and its role in cell regulation. Potential substrates were evaluated using short acetylated peptides and an enzyme-coupled assay and their kinetic parameters were calculated. Overall, this approach provides a better approach to elucidating HDAC substrates and binding partners and could potentially be adapted to other isozymes of the deacetylase family.

Caspase-9 is regulated by phosphorylation through diverse mechanisms Banyuhay P. Serrano1, Kristen L. Huber1 and Jeanne A. Hardy1 1University of Massachusetts-Amherst A dynamic interplay between caspases and their cognate kinases exists. The battle between them can tip the scale towards cell death or survival; however the molecular basis of regulation by phosphorylation of caspases, particularly of caspase-9, is vastly understudied. Phosphoregulation of caspase-9 is the most complex of all caspases with 11 reported sites of phosphorylation by multiple kinases. At present we have uncovered the molecular basis for the action of two of important kinases, PKA and c-Abl. We identified Ser-183 as the critical site for caspase-9 inactivation by PKA by blocking substrate binding. In addition to direct inhibition, our studies revealed that phosphorylation of this site disrupts the interaction between the core of caspase and the regulatory CARD domain. This result uncovers a whole new level of caspase-9 regulation that targets interdomain interactions, the changes in which may influence the formation of the apoptosome, the most upstream of all apoptotic machinery. We also unambiguously identified Tyr-397 as the major and functionally relevant site of phosphorylation by c-Abl. Moreover, phosphorylation of Tyr-397 directly correlates with caspases-9 inhibition. Our work is the first to define the molecular mechanism by which c-Abl confers protection from apoptosis via inactivation of caspase-9. Given that caspase-9 is more extensively phosphorylated than any other caspase, understanding the wide variety of mechanisms of phosphorylation-mediated control of caspase-9 also provides critical insights across the caspase family. These molecular details of phosphoregulation will certainly contribute to the next generation of caspases-directed therapeutics and also has important implication for caspases-kinase co-therapies. Funding: National Institutes of Health (NIH) GM 080532 and in part by the Chemistry-Biology Interface Training Program Fellowship (National Research Service Award T32 GM08515) from UMass Amherst

172 ABSTRACT

Refolding and Purification of Unmodified Human Elongation Factor 2 Nirja B. Patel1, Joshua R. Ostovitz2, Nathaniel Donahue1, and John E.Weldon1,2 1Molecular Biology, Biochemistry and Bioinformatics Program &, 2Department of Biological Sciences, Towson University Translation elongation factor 2 (EF2) is an essential component of protein synthesis that advances ribosomes along mRNA in eukaryotes and archaea. EF2 contains a single unique amino acid called diphthamide that is a post-translational modification of a histidine. Previous work and circumstantial evidence suggest that diphthamide is important, but its precise function remains elusive. Diphthamide has been found only in EF2, and its bacterial homolog (elongation factor G) lacks the modification. The diphthamide synthesis pathway is also complex, with a mechanism divided into three stages that requires seven different proteins in yeast (DPH 1-7). The synthesis of diphthamide, however, has yet to be reconstituted in vitro. In order to study both the function of diphthamide and its synthesis, we are developing a protocol to purify human EF2 (hEF2) without the diphthamide modification. We chose to express hEF2 in E. coli due to the lack of diphthamide synthesis in bacteria. When overexpressed, hEF2 partitioned into the insoluble fraction of E. coli lysates. After washing and denaturing the insoluble pellet, we were able to obtain a highly pure hEF2 fraction. We subsequently evaluated refolding conditions designed to maximize hEF2 solubility. Our results suggest a strategy for the preparation of hEF2 unmodified with diphthamide. Future directions include evaluating unmodified hEF2 for structural and functional differences from diphthamide-modified hEF2.

Protein targets of tyrosine nitration in human astrocytomas Dr. Xianquan Zhan1 1Key Laboratory of Cancer Proteomics of Chinese Ministry of Health, Xiangya Hospital, Central South University, People’s Republic of China Tyrosine nitration in a protein is associated with the pathogenesis of highly fatal astrocytomas. To explore the molecular mechanisms of astrocytomas and to discover new biomarkers/therapeutic targets, nitroproteins from a high-grade astrocytoma tissue were detected with two-dimensional gel electrophoresis (2DE)-based nitrotyrosine immunoblots, characterized with liquid chromatography-tandem mass spectrometry (LC-MS/MS), followed by bioinformatics analysis, and verified with Western blot and immunoprecipitation. Fifty-seven nitrotyrosine immunopositive protein spots were detected. A total of 870 proteins (nitrated and non-nitrated) in nitrotyrosine- immunopositive 2D gel spots were identified, and 18 nitroproteins and their 20 nitrotyrosine sites were identified with MS/MS analysis. Those nitroproteins are involved in multiple biological processes, such as signal transduction, transcription and translation, drug-resistance, cytoskeleton, apoptosis, cell proliferation and immune response, phenotypic dedifferentiation, cell migration, and metastasis. Among those nitroproteins that might play roles in astrocytoma biological system was nitrosorcin, which is associated with drug-resistance and metastasis and might play a role in the spread of and treatment of an astrocytoma. Semi-quantitative immunoreactivity of sorcin revealed increased expressions among different grades of astrocytomas relative to controls, and a semi- quantitative increased level of sorcin nitration in high-grade astrocytoma relative to control. Nitro- b–tubulin is involved in cytoskeleton and cell migration. Semi-quantitative immunoreactivity of b– tubulin revealed increased expressions among different grades of astrocytomas relative to controls and a semi-quantitatively increased level of b–tubulin nitration in high-grade astrocytoma relative to control. Each nitroprotein was rationalized and related to the corresponding functional system to provide new insights into tyrosine nitration and its potential role in the pathogenesis of astrocytoma formation.

173 ABSTRACT

PX - METABOLIC ENGINEERING/ENERGY APPLICATIONS Characterization of protein compositions of glanded and glandless cotton seeds Heping Cao1, Leesa J. Deterding2, Perry J. Blackshear3 1U.S. Department of Agriculture, Agricultural Research Service, Southern Regional Research Center, 1100 Robert E. Lee Blvd, New Orleans, LA 70124, United States, 2Laboratory of Structural Biology, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, NC 27709, United States, 3Laboratory of Signal Transduction, National Institute of Environmental Health Sci- ences, National Institutes of Health, Research Triangle Park, NC 27709, and Departments of Biochemistry and Medicine, Duke University Medical Center, Durham, North Carolina 27710, United States Cytokines including tumor necrosis factor alpha (TNF) are important in the development and progression of arthritis, obesity, diabetes and other diseases. However, anti-TNF therapies targeting TNF and its receptors failed in clinical trials. An alternative strategy to reduce pro-inflammatory cytokines is to promote degradation of cytokine mRNAs because these mRNAs possess AU-rich elements (AREs) and that their stabilities are largely controlled by ARE binding proteins. Tristetraprolin (TTP), an anti-inflammatory protein, binds to the AREs of clinically important mRNAs such as TNF mRNA and promotes the destruction of those transcripts. TTP-deficient mice develop a profound inflammatory syndrome with cachexia, dermati- tis, erosive arthritis, autoimmunity and myeloid hyperplasia, due to excessive production of TNF and other cytokines. TTP expression is induced by various factors including insulin and plant polyphenol extracts. TTP is highly phosphorylated by several protein kinases. Multiple phosphorylation sites are identified in human TTP, but it is difficult to assign major vs. minor phosphorylation sites. The objective of this study was to generate information on TTP phosphorylation using phosphopeptide mapping and mass spectrometry (MS). Wild-type and site-directed mutant TTP proteins were expressed in transfected human cells followed by in vivo radiolabeling with [32P]-orthophosphate. Histidine-tagged TTP proteins were purified with Ni- NTA affinity beads and digested with trypsin and lysyl endopeptidase. The digested peptides were separated by C18 column. Wild-type and all mutant TTP proteins were localized in the cytosol, phosphorylated extensively in vivo and capable of binding to ARE-containing RNA probes. Mutant TTP with S90 and S93 mutations resulted in the disappearance of a major phosphopeptide peak. Mutant TTP with an S197 mutation resulted in another major phosphopeptide peak being eluted earlier than the wild-type. Additional mutations at S186, S296 and T271 exhibited little effect on phosphopeptide profiles. MS analysis identified the peptide that was missing in the S90 and S93 mutant protein as LGPELSPSPTSPTATSTTPSR. MS also identified a major phosphopeptide associated with the first zinc-finger region. These analyses suggest that the tryptic peptide containing S90 and S93 is a major phosphopeptide in human TTP.

Abstracts of the 2016 Award Winners

EMIL THOMAS KAISER AWARD Global Analysis of Post Translational Modification in Disease Charles S. Craik1, Sam Ivry1, Nicole Olson1, Michael Winter1 and Anthony O’Donoghue1 1Department of Pharmaceutical Chemistry, University of California San Francisco Uncovering the substrate specificity of post-translational modifying (PTM) enzymes is central to understanding their physiological role in homeostasis and disease. A chemically defined and rapid multiplex substrate profiling method has been developed to reveal the substrate specificity of various post-translational modifying enzymes using LC-MS/MS sequencing. Design of the sequences is based on the hypothesis that recognition by many PTM enzymes requires no more than two amino acids suitably positioned within a peptide substrate for initial detection. A physicochemically diverse library of peptides was generated by incorporating all combinations of neighbor and near-neighbor amino acid pairs into decapeptide sequences that are flanked by unique dipeptides at each terminus. Use of the tetradecapeptide library on a panel of evolutionarily diverse peptidases in both purified form and in complex biological mixtures generated

174 ABSTRACT prime and non-prime site information and substrate specificity that either matched or expanded upon previous substrate motifs. We refer to this method as Multiplex Substrate Profiling by Mass Spectrometry (MSP-MS). In addition to peptidases we developed quantitative motifs for a selection of kinases from each branch of the kinome and were able to increase the information content of the quantitative motifs by generating sub-libraries to expand the testable sequence space. This method biochemically confirmed the activity of numerous highly selective as well as broadly promiscuous peptidases and kinases and showed unparalleled sensitivity, specificity and quantitation of PTM enzyme activity. The motifs are proving useful in the development of diagnostics, prognostics and therapeutics for disease intervention.

DOROTHY CROWFOOT HODGKIN AWARD From BRCA1 to Parkin and Back Again Rachel E. Klevit1 1Dept. of Biochemistry, Univ. of Washington, Seattle WA The attachment of the small protein “Ubiquitin” to other cellular proteins serves as a signal that can lead to myriad outcomes that regulate virtually every cellular process. Given its central role in protein homeostasis, DNA damage response, and inflammation, it is not surprising that dysfunction of protein ubiquitylation is at the root of many human diseases including cancer, Parkinson’s disease, and neurodegenerative diseases. I will present how our quest to understand the function of the breast cancer protein BRCA1, a ubiquitin ligating enzyme, led us to make an unexpected fundamental discovery about another ligating enzyme, Parkin.

CARL BRAND€ EN AWARD A New Role for Protein Surfaces Gary J. Pielak1 1Department of Chemistry, University of North Carolina at Chapel Hill Quinary structure, which is responsible for organizing the cellular interior, is brought about by weak protein- protein interactions that occur under crowded conditions. These interactions also appear at the high protein concentrations used to deliver biologicals. Unfortunately, we know almost nothing about quinary interactions, because they are completely absent in the dilute buffered solutions used for most experiments. The fact that proteins are almost always studied in dilute solutions rather than in cells means there is much more information about interiors than there is about exteriors. It is well known that interior side chains must be hydrophobic, the buried atoms must be exquisitely well packed and nearly all hydrogen bond donors and acceptors must be satisfied. This knowledge has yielded tremendous assets, including the ability to predict, manipulate and design protein structure and stability. The same knowledge-to- benefits claim cannot yet be made for exteriors. Our recent work sheds new light on protein surfaces. We find that exteriors harbor as much information as interiors, but this information is only available when proteins are studied in cells or under crowded conditions in vitro. Specifically, changing the protein exterior affects the equilibrium thermodynamics of globular protein unfolding in cells in a manner that emphasizes a key role of surface charge-charge interactions. Our findings are important both for understanding fundamental aspects of biology and formulating biologics.

STEIN AND MOORE AWARD Protein folding – on and off the ribosome: Using Physics and Chemistry to understand Biology Jane Clarke1 1Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, CB2 1EW, UK Most eukaryotic proteins contain more than one independently folding domain. Since the local concentration of domains is high, how do natural proteins avoid misfolding? Our single molecule studies of the folding and misfolding of multidomain proteins suggest that the early folding events of

175 ABSTRACT multidomain proteins are much more intricate than previously thought. The single molecule data suggest that 50% of the total population form transient non-native species. Some of the misfolded species are short-lived amyloid-like and likely to aggregate. Low sequence identity between neighboring domains does not prevent this misfolding, but it prevents the formation of stable non-native species. Since co-translational folding is likely to be important to prevent misfolding particularly for multidomain proteins, we are now investigating the folding of multidomain proteins on stalled ribosomes. We conclude that individual domains can fold close to the ribosome, certainly before the following domain has been translated and that this will prevent inter-domain misfolding.

PROTEIN SCIENCE BEST PAPER AWARD An Allosteric Model for Control of Pore Opening by Substrate Binding in the EutL Microcompartment Shell Protein Michael C. Thompson1,†, Duilio Cascio2, David J. Leibly1, and Todd O. Yeates1,2 1Department of Chemistry and Biochemistry, University of California, Los Angeles, 2UCLA-DOE Insti- tute for Genomics and Proteomics, University of California, Los Angeles The ethanolamine utilization (Eut) microcompartment is a protein-based metabolic organelle that is strongly associated with pathogenesis in bacteria that inhabit the human gut. The exterior shell of this elaborate protein complex is composed from a few thousand copies of BMC-domain shell proteins, which form a semi-permeable diffusion barrier that provides the interior enzymes with substrates and cofactors while simultaneously retaining metabolic intermediates. The ability of this protein shell to regulate passage of substrate and cofactor molecules is critical for microcompartment function, but the details of how this diffusion barrier can allow the passage of large cofactors while still retaining small intermediates remain unclear. Here we report structural and biophysical evidence to show that ethanolamine, the substrate of the Eut microcompartment, acts as a negative allosteric regulator of pore opening in a particular shell protein, EutL. Specifically, a series of X-ray crystal structures of EutL, along with equilibrium binding studies, reveal that ethanolamine binds to EutL at a site that exists in the closed- pore conformation and which is incompatible with opening of the large pore for cofactor transport. The allosteric mechanism we propose is consistent with the cofactor requirements of the Eut microcompartment, leading to a new model for EutL function. †Currently in the Department of Bioengineering & Therapeutic Sciences, University of California, San Francisco

176