Role of Acetyl-Coa Synthetase and Sirtuin Proteins Naazneen Sofeo Iowa State University

Iowa State University Capstones, Theses and Graduate Theses and Dissertations Dissertations

2019 Regulation of acetyl-CoA metabolism in plants: Role of acetyl-CoA synthetase and sirtuin proteins Naazneen Sofeo Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/etd Part of the Biochemistry Commons

Recommended Citation Sofeo, Naazneen, "Regulation of acetyl-CoA metabolism in plants: Role of acetyl-CoA synthetase and sirtuin proteins" (2019). Graduate Theses and Dissertations. 17322. https://lib.dr.iastate.edu/etd/17322

This Dissertation is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected].

Regulation of acetyl-CoA metabolism in plants: Role of acetyl-CoA synthetase and sirtuin proteins

Naazneen Sofeo

A dissertation submitted to the graduate faculty

in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

Major: Biochemistry

Program of Study Committee: Basil J Nikolau, Co-major Professor Marna D. Yandeau-Nelson, Co-major Professor Richard B Honzatko Scott W Nelson Zengyi Shao

The student author, whose presentation of the scholarship herein was approved by the program of study committee, is solely responsible for the content of this dissertation. The Graduate College will ensure this dissertation is globally accessible and will not permit alterations after a degree is conferred.

Iowa State University

Ames, Iowa

2019

TABLE OF CONTENTS

ACKNOWLEDGMENTS ...... iv

ABSTRACT ...... v

CHAPTER 1. INTRODUCTION AND DISSERTATION ORGANIZATION ...... 1 General Introduction ...... 1 Significance ...... 4 Dissertation Organization ...... 6 References...... 8

CHAPTER 2. ALTERING THE SUBSTRATE SPECIFICITY OF ACETYL-COA SYNTHETASE BY RATIONAL MUTAGENESIS OF THE CARBOXYLATE BINDING POCKET ...... 12 Abstract ...... 13 Introduction...... 14 Results & Discussion ...... 16 Conclusion ...... 26 Methods ...... 27 Abbreviations ...... 31 Author Contributions ...... 31 Acknowledment ...... 31 References...... 32 Figures ...... 36 Tables ...... 43 Supporting Information ...... 45

CHAPTER 3. UNDERSTANDING EFFECTS OF ACETYLATION IN REGULATING ARABIDOPSIS ACETYL-COA SYNTHETASE ...... 47 Abstract ...... 47 Introduction...... 48 Material & Methods ...... 49 Results...... 54 Discussion ...... 57 References...... 58 Figures ...... 62

CHAPTER 4. SIRTUIN-MEDIATED REGULATION OF PLANT ACETYLOME AND METABOLOME ...... 70 Abstract ...... 70 Introduction...... 70 Results...... 72 Discussion ...... 78 Methods ...... 80 iii

References...... 91 Figures ...... 96 Tables ...... 104 Supplemental Information ...... 104

CHAPTER 5. GENERAL CONCLUSION ...... 105 Conclusions...... 105 References...... 108

APPENDIX A. USING A STRATEGY THAT EXPANDS THE GENETIC CODE TO GENERATE IN VIVO ACETYLATED LYS-622 MODIFICATION OF THE ARABIDOPSIS ACETYL-COA SYNTHETASE ...... 110 Introduction...... 110 Methods ...... 111 Results and Discussion ...... 114 References...... 116 Figures ...... 118 Tables ...... 124

APPENDIX B. SUBCELLULAR LOCALIZATION OF THE SRT1 PROTEIN IN ARABIDOPSIS THALIANA ...... 125 Introduction...... 125 Methods ...... 125 Results and Conclusion...... 126 References...... 127 Figures ...... 128 Tables ...... 133

ACKNOWLEDGEMENTS

I would like to first thank my major professor, Dr. Basil J Nikolau for giving me an opportunity to be a part of his research group. It has been a true privilege and joy to work in this group. I also would like to thank my co-major professor, Dr. Marna D. Yandeau-Nelson for her valuable support, guidance and encouragement throughout. I would like to thank Dr.

Richard B Honzatko, Dr. Scott W. Nelson, and Dr. Zengyi Shao, for serving as my valuable

Program of Study Committee members. I would also like to thank all the former and present

Nikolau lab members for their guidance, friendship and support throughout.

I thank Jason Hart for his work, which set a foundation to the Chapter 2 of this dissertation. I want to thank Dr. Justin W. Walley and Maxwell R. McReynolds for the collaborative work done on Arabidopsis sirtuins. I also thank ISU Protein Facility (especially

Joel Nott) for providing assistance with protein mass spectrometry analysis and the ISU WM

Keck Metabolomics Research Laboratory (especially Dr. Lucas Showman and Dr. Kirthi

Narayanaswamy) for providing assistance with GC-MS/MS. I thank Dr. Gustavo MacIntosh for providing access to his FPLC, the ISU DNA facility and ISU Roy J. Carver High Resolution

Microscopy Facility (especially Margie Carter) for providing assistance with Confocal

Microscopy. I would also like to thank the department faculty and staff for making my time at

Iowa State University a wonderful experience.

Finally, I would like to thank my parents, husband and brother for all their love and support. Without them, I would not have been able to pursue and fulfill my dreams. I dedicate this dissertation to them. v

ABSTRACT

As an activated form of acetate, acetyl-CoA is a key metabolic intermediate that links many metabolic processes, including the TCA cycle, amino acid metabolism, fatty acid metabolism and biosynthetic processes that generate many polyketides and some terpenes. This molecule is also involved in regulation of proteins, through the post translational modification of reversible acetylation. It is the most prevalent thioester-containing intermediate of metabolism, and biological systems utilize multiple enzymatic mechanisms to generate this central metabolite. Acetyl-CoA synthetase (ACS) is a member of a large super-family of enzymes that display diverse substrate specificities, with a common mechanism of catalyzing the formation of a thioester bond between Coenzyme A and a carboxylic acid, while hydrolyzing ATP to

AMP and pyrophosphate. We explored the structural basis of the substrate specificity of ACS which specifically makes acetyl-CoA. This enzyme was redesigned successfully, which switched a highly specific enzyme from using only acetate, to be equally specific for using longer linear (up to hexanoate) or branched chain (methylvalerate) carboxylate substrates.

ACS is known to be regulated through the post translational modification of reversible acetylation. In organisms like bacteria and humans, it has been shown that this modification leads to reduction of activity of the enzyme. In this study we have analyzed the effect of acetylation on Arabidopsis ACS activity and the conserved lysine involved in this modification. This study shows that effect of acetylation on this plant enzyme is very different from what is observed in homologous ACS. According to our knowledge, this is the first report of how acetylation affects a plant ACS, using in vitro studies.

Sirtuins are histone deacetylases involved in deacetylating other proteins and thus involved in epigenetic and metabolic regulation. We characterized and compared Arabidopsis sirtuins, vi

SRT1 and SRT2 both in vitro and in vivo, through total untargeted metabolomics, acetylome analysis and through heterologous protein expression and characterization. A comparative integrated biochemical-genetic study of the acetylome and metabolome of srt1 or srt2 single mutants and srt1srt2 double mutant plants indicate that although there are undetectable effects on the growth phenotype of the plants, there are considerable changes in the molecular features that were assayed. The changes induced by the double mutant indicate that these proteins deacetylate both histone and non-histone proteins. In addition, the mutant-induced changes in the metabolome indicate that the two SRT homologs are involved in regulating different metabolic processes, generating distinct metabolomes. This study also explores the effect of these two deacetylases on Arabidopsis ACS acetylation. In summary the study of the proteins, acetyl-CoA synthetase and sirtuins in Arabidopsis, will help to further the understanding of acetyl-CoA metabolism in plants. 1

CHAPTER 1. INTRODUCTION AND DISSERTATION ORGANIZATION

General Introduction

Acetyl-CoenzymeA (CoA) synthetase is a ~72 kDa protein which plays an important role in fatty acid biosynthesis and polyketide biosynthesis pathways. In both these pathways, the priming or precursor molecule is a two to four carbon short chain acyl-CoA molecule. More appropriately, it is an activated acyl molecule. In fatty acid biosynthesis, this activated molecule provides the starting carbons for the pathway in which the fatty acid molecule to be synthesized is elongated through reiterative addition of two carbons. The elongation precursors

(e.g. Malonyl-CoA) for the pathway can also be made from the priming molecule by its carboxylation.

Acetyl-CoA synthetase (ACS) catalyzes this activation of the acetate molecule by adding it to the thiol group of CoA, thus forming the priming molecule for the pathway. In this process it hydrolyzes ATP to AMP and pyrophosphate. This pathway generates fatty acids which are straight chain and have an even number of carbon atoms. ACS functions via a two-step reaction. The first step is the formation of an acetyl-AMP intermediate and second step is the formation of the thioester of acetate and CoA [1].

The most common substrate for acyl-CoA synthetase family is acetate, however certain enzymes in this family can utilize longer chain or branched chain carboxylate substrates such as propionate or isobutyrate [2,3]. One such example is isobutyryl-CoA synthetase from

Pseudomonas chlororaphis [3].

In plant cells, acetyl-CoA cannot be transported across membranes [4], hence there are multiple acetyl-CoA pools [5–7]. These pools are localized in cytosol, mitochondria, plastids and 2 peroxisomes. The de-novo fatty acid biosynthesis pathway is located in plastids and therefore uses the plastidic acetyl-CoA pool.

Two separate enzymes lead to the formation of acetyl-CoA in plastids: Pyruvate dehydrogenase complex (PDHC) and Acetyl-CoA synthetase (ACS). Among these two enzymes, PDHC is the preferred enzyme in plants during normal growth [8,9]. The conditions under which ACS act is still unclear.

The binding pocket chemistry that defines the substrate specificity of Arabidopsis thaliana

ACS is not yet known, however studies on prokaryotic acyl-CoA synthetases have provided insight to its carboxylate binding site. It was found in Methanothermobacter thermoautotrophicus acetyl CoA synthetase that the hydrophobic cavity between the N and C terminal domains is the probable binding pocket for carboxylate and the residues in it are in close proximity to the substrate [10]. This study was based on the crystal structure of acetyl-

CoA synthetase from Salmonella enterica [11]. Mutations of these probable binding site residues affected enzyme function and also enabled the enzyme to bind not only acetate but also butyrate [10]. Rational mutagenesis based upon known crystal structures of homologous

ACS from other species can be used to find the binding pocket residues of A. thaliana ACS.

The regulation of ACS in Arabidopsis thaliana is also not completely understood. It is most probably modulated by the post translational modification of reversible acetylation like other homologous acetyl-CoA synthetases from prokaryotes. Reversible acetylation is a crucial process for not only transcriptional regulation but also other cellular pathways. This is an abundant and evolutionarily conserved post translational modification [12]. 3

Structures of both acetylated and non-acetylated ACS from Salmonella enterica suggested that the conserved lysine play an important role in the first step of the reaction presumably by aligning the acetate close to the active site [11] and does not affect the second step [13].

Structural studies in S. enterica have been done to understand what changes occur in ACS when it is acetylated. A rotation of 140° of the C-terminal domain is required for the binding of CoA and the catalysis of the second step, which is the thioester formation step [14].

However, there is no evidence of any conformational changes induced specifically by the acetylation of the protein [11]. It has been shown that these acetylation reactions mask the positive charge of lysine side chains, and thus potentially affect the nature of the interactions within the tertiary structure of the modified protein [15,16].

The enzymatic mechanism of ACS regulation through lysine acetylation has been extensively studied in bacterial systems [17–19]. It has been shown in Salmonella enterica that acetylation regulates the activity of the AMP-forming family of enzymes. The non-acetylated ACS from

S. enterica is about 480-fold more active than acetylated ACS [13]. This deacetylation is catalyzed by CobB enzyme which is a homolog of yeast Sir2 deacetylase [20]. The acetylation of ACS occurs at a conserved lysine residue in the A10 motif. This residue is conserved across species.

In yeast, ACS is involved in both carbon metabolism and regulation. ACS generates acetyl-

CoA for histone acetylation in the nucleus, in global gene transcription [21,22] and in various pathways outside the nucleus [23].

In plants, hundreds of Nε-amine lysine sites have been identified [24], and these post translational modification affects both histone and non-histone proteins; the latter are involved 4 in a variety of metabolic processes including photorespiration, redox regulation, TCA cycle, and amino acid and protein metabolism [25]. Two types of proteins play a role in this process, lysine acetyl transferases (KATs) that acetylate the protein and lysine deacetylases (KDACs) that remove the acetyl group. There are various classes of KATs and KDACs that have been extensively reviewed [12,17]. There are 14 KDACs identified in Arabidopsis that deacetylate other proteins [26]. Among these, there are two sirtuin family (NAD+ dependent deacetylases) proteins, SRT1 and SRT2 [27–29].

Activity of sirtuins is shown to be inhibited by nicotinamide [30]. Both primary sequence and protein structure determines the substrate specificity of sirtuins [31]. In Arabidopsis, sirtuin

SRT2 has been shown to be localized in the nucleus and mitochondria and interact with proteins involved in energy metabolism and transport of metabolites [28]. SRT2 knockout leads to higher resistance against pathogens possibly due to its effect on salicylic acid biosynthesis [32]. SRT1 has been shown to be upregulated during leaf to callus transition in

Arabidopsis [33] and is involved in stress response [27,29]. Most studies focus on effect of these sirtuins on regulation of one specific pathway and a more integrated analysis of acetylation would aid in understanding how it affects different proteins in a pathway [15].

Hence, this study also aims to analyze and understand the broader picture of how sirtuins affect various cellular pathways in Arabidopsis.

Significance

Manipulation or bioengineering of enzymes using rational biochemical or genetic approaches is done to produce novel altered end product or alter activity to increase enzymatic yields. The

National Science Foundation Engineering Research Center, Center of Biorenewable 5

Chemicals (CBiRC) aims to shift the basis of production of industrially useful chemical from a petroleum- to a renewable resource (http://www.cbirc.iastate.edu/overview/mission/).

This goal can be made possible by engineering two common pathways in organisms, the fatty acid synthesis and the polyketide synthesis pathways. Modified end product molecules that can be formed from these pathways can be used to generate economically useful chemicals which have similar chemistries when compared to petrochemicals [34]. The modified end products from these pathways may contain ω-end chemical functionality (such as hydroxy or amine group) that can be used as precursors to generate industrially significant products like lubricants or cosmetics.

Bioengineered acetyl-CoA synthetase can produce acyl-CoA primers that are odd numbered

(i.e., propionyl-CoA) or branched chain (i.e., isobutyryl-CoA, isovaleryl-CoA, 2- methylvaleryl-CoA or 2-methylbutyryl-CoA) or are chemically functionalized at the omega- end of the acyl-CoA (i.e., hydroxypropionyl-CoA, hydroxybutyryl-CoA or 4-aminobutyryl-

CoA). The incorporation of such alternative primers into the Fatty Acid Synthesis/PolyKetide

Synthesis biosynthetic machinery would diversify the products and create renewable chemical products with various industrial applications.

The majority of acyl-CoA synthetases that are biochemically characterized are from prokaryotic sources with only a few from mammalian sources[10,35,36]. Acetyl-CoA synthetases from plants have not been studied extensively except the enzyme from Pisum sativa (pea) [37]. In plant systems, such as Arabidopsis, the metabolic function of ACS is not completely clear. Unlike other biological systems, ACS is not the physiological source of acetyl-CoA that is the precursor of de novo fatty acid biosynthesis in plastids [8]; this is in 6 spite of the fact that ACS is plastid localized [9]. In this study, we explored if the acetylation- based regulation of ACS is extrapolatable from bacteria, yeast and algae to affect activity of this enzyme in plant systems.

The discovery and characterization of the acetylases and/or deacetylases that act on ACS will prove vital to understanding ACS regulation and function in A. thaliana. The study of the regulation of ACS in Arabidopsis thaliana might provide insights into the conditions in which the enzyme is active. Understanding the regulation of ACS would also help in understanding its function in A. thaliana, a fact which still remains unclear. Understanding how sirtuins affect metabolism and regulation in a plant cell would shed light on the regulatory aspect of acetyl-

CoA. It will help to understand targets of sirtuins, and also demonstrate how the post translational modification of reversible acetylation controls and regulates various pathways in a plant cell. Hence this study helps unravel understanding of both acetyl-CoA metabolism and regulation in plants, using various biochemical, metabolomic, proteomic and synthetic biology approaches.

Dissertation Organization

This dissertation was under the guidance, support and supervision of my major professor Dr.

Basil J. Nikolau. This dissertation is organized into five chapters and two appendices.

Chapter 1 describes general introduction and significance of the work in this dissertation. It contains general background on the Arabidopsis protein Acetyl-CoA Synthetase, its enzymatic role in the fatty acid biosynthesis pathway and the regulation of its homologous proteins. This chapter also describes the various deacetylases involved in the post translational modification 7 of proteins, especially sirtuins, and a general introduction to it. This chapter sets the context and significance for the research described in Chapter 2, 3 and 4.

Chapter 2 has been modified from an already submitted manuscript to the journal ACS

Synthetic Biology. This manuscript is currently under the review process. This chapter describes the enzymatic characterization of the Arabidopsis acetyl-CoA synthetase protein and how the substrate specificity of this protein was successfully modulated using targeted rational mutagenesis.

Chapter 3 has been drafted to be submitted to the journal Archives of Biochemistry and

Biophysics. This chapter describes studies performed to understand how the modification of acetylation affects the protein acetyl-CoA synthetase and its regulation. Western blot and mass spectrometry analysis were done to determine acetylation status. The mass spectrometry studies were done at the ISU Protein Facility with help and expertise of Joel Nott. Mutational studies, protein characterization and enzymatic analysis were also performed for the protein.

Chapter 4 describes a collaborative project with Dr. Justin W. Walley’s lab at Iowa State

University. This chapter describes the study to understand how sirtuins affect the total metabolome and total acetylome of Arabidopsis. The homozygous single and double knockout sirtuin lines were generated for this study. This chapter conatins total metabolite analysis of different sirtuin knockout mutants and wild-type Arabidopsis lines. The tissue growth and collection, total metabolite extraction, and metabolomics mass spectrometry analysis was done by Naazneen Sofeo. GC-MS/MS experiment was performed at ISU WM Keck Metabolomics

Research Laboratory with the help and guidance of Dr. Lucas Showman and Dr. Kirthi

Narayanaswamy. The chapter also contains total acetylome analysis of wild-type and sirtuin double knockout Arabidopsis lines. The protein extraction, trypsin digestion, di-methyl 8 labelling, acetyl enrichment and mass spectrometry analysis was done by graduate student

Maxwell R. McReynolds, of Dr. Justin W. Walley’s lab.

Chapter 5 describes general conclusions from the work described in Chapter 2, 3 and 4.

Appendix A describes the generation and characterization of an acetyl-CoA synthetase with an incorporated non-canonical amino acid (acetylated lysine) using genetic code expansion strategy. This was done as part of the study to understand regulation of acetyl-CoA synthetase but was not included in Chapter 3. The clone to express the orthogonal tRNA synthetase machinery was generously provided by Dr. Chenguang Fan (Department of Chemistry &

Biochemistry, University of Arkansas).

Appendix B describes the study to determine the sub-cellular location of the Arabidopsis sirtuin SRT1. This was part of the study to characterize Arabidopsis sirtuins as described in

Chapter 4. This appendix describes the confocal microscopy studies done on GFP-tagged

SRT1 proteins. The microscopy studies were done under the guidance of Margie Carter at ISU

Roy J. Carver High Resolution Microscopy Facility.

References

[1] P. Berg, Acyl adenylates; an enzymatic mechanism of acetate activation., J. Biol. Chem. 222 (1956) 991–1013. [2] A.R. Horswill, J.C. Escalante-Semerena, The prpE gene of Salmonella typhimurium LT2 encodes propionyl-CoA synthetase, Microbiology. 145 (1999) 1381–1388. doi:10.1099/13500872-145-6-1381. [3] Y. Hashimoto, H. Hosaka, K.-I. Oinuma, M. Goda, H. Higashibata, M. Kobayashi, Nitrile pathway involving acyl-CoA synthetase: overall metabolic gene organization and purification and characterization of the enzyme., J. Biol. Chem. 280 (2005) 8660– 7. doi:10.1074/jbc.M405686200. [4] J.L. Brooks, P.K. Stumpf, Fat metabolism in higher plants. XXXIX. Properties of a soluble fatty acid synthesizing system from lettuce chloroplasts., Arch. Biochem. Biophys. 116 (1966) 108–16. 9

[5] B. Liedvogel, Acetyl Coenzyme A and Isopentenylpyrophosphate as Lipid Precursors in Plant Cells - Biosynthesis and Compartmentation, J. Plant Physiol. 124 (1986) 211– 222. doi:10.1016/S0176-1617(86)80035-9. [6] B.L. Fatland, B.J. Nikolau, E.S. Wurtele, Reverse genetic characterization of cytosolic acetyl-CoA generation by ATP-citrate lyase in Arabidopsis., Plant Cell. 17 (2005) 182– 203. doi:10.1105/tpc.104.026211. [7] J. Schwender, J.B. Ohlrogge, Probing in vivo metabolism by stable isotope labeling of storage lipids and proteins in developing Brassica napus embryos., Plant Physiol. 130 (2002) 347–61. doi:10.1104/pp.004275. [8] J. Ke, R.H. Behal, S.L. Back, B.J. Nikolau, E.S. Wurtele, D.J. Oliver, The Role of Pyruvate Dehydrogenase and Acetyl-Coenzyme A Synthetase in Fatty Acid Synthesis in Developing Arabidopsis Seeds, Plant Physiol. 123 (2000) 497–508. doi:10.1104/pp.123.2.497. [9] M. Lin, D.J. Oliver, The Role of Acetyl-Coenzyme A Synthetase in Arabidopsis, PLANT Physiol. 147 (2008) 1822–1829. doi:10.1104/pp.108.121269. [10] C. Ingram-Smith, B.I. Woods, K.S. Smith, Characterization of the acyl substrate binding pocket of acetyl-CoA synthetase., Biochemistry. 45 (2006) 11482–90. doi:10.1021/bi061023e. [11] A.M. Gulick, V.J. Starai, A.R. Horswill, K.M. Homick, J.C. Escalante-Semerena, The 1.75 Å crystal structure of acetyl-CoA synthetase bound to adenosine-5???- propylphosphate and coenzyme A, Biochemistry. 42 (2003) 2866–2873. doi:10.1021/bi0271603. [12] A. Drazic, L.M. Myklebust, R. Ree, T. Arnesen, The world of protein acetylation, Biochim. Biophys. Acta - Proteins Proteomics. 1864 (2016) 1372–1401. doi:10.1016/j.bbapap.2016.06.007. [13] V.J. Starai, I. Celic, R.N. Cole, J.D. Boeke, J.C. Escalante-Semerena, Sir2-dependent activation of acetyl-CoA synthetase by deacetylation of active lysine, Science (80-. ). 298 (2002) 2390–2392. doi:10.1126/science.1077650. [14] G. Jogl, L. Tong, Crystal structure of yeast acetyl-coenzyme A synthetase in complex with AMP., Biochemistry. 43 (2004) 1425–31. doi:10.1021/bi035911a. [15] T. Narita, B.T. Weinert, C. Choudhary, Functions and mechanisms of non-histone protein acetylation, Nat. Rev. Mol. Cell Biol. (2018) 1. doi:10.1038/s41580-018-0081- 3. [16] A.M. James, K. Hoogewijs, A. Logan, A.R. Hall, S. Ding, I.M. Fearnley, M.P. Murphy, Non-enzymatic N-acetylation of Lysine Residues by AcetylCoA Often Occurs via a Proximal S-acetylated Thiol Intermediate Sensitive to Glyoxalase II., Cell Rep. 18 (2017) 2105–2112. doi:10.1016/j.celrep.2017.02.018. [17] V. Bernal, S. Castaño-Cerezo, J. Gallego-Jara, A. Écija-Conesa, T. de Diego, J.L. Iborra, M. Cánovas, Regulation of bacterial physiology by lysine acetylation of proteins, N. Biotechnol. 31 (2014) 586–595. doi:10.1016/j.nbt.2014.03.002. 10

[18] T. Noy, H. Xu, J.S. Blanchard, Acetylation of acetyl-CoA synthetase from Mycobacterium tuberculosis leads to specific inactivation of the adenylation reaction, Arch. Biochem. Biophys. 550–551 (2014) 42–49. doi:10.1016/j.abb.2014.04.004. [19] J.G. Gardner, F.J. Grundy, T.M. Henkin, J.C. Escalante-Semerena, Control of acetyl- coenzyme A synthetase (AcsA) activity by acetylation/deacetylation without NAD+involvement in Bacillus subtilis, J. Bacteriol. 188 (2006) 6715. doi:10.1128/JB.01181-06. [20] A.W. Tsang, J.C. Escalante-Semerenat, CobB, a new member of the SIR2 family of eucaryotic regulatory proteins, is required to compensate for the lack of nicotinate mononucleotide:5,6- dimethylbenzimidazole phosphoribosyltransferase activity in cobT mutants during cobalamin biosynthesis in Sal, J. Biol. Chem. 273 (1999) 31788– 31794. doi:10.1074/jbc.273.48.31788. [21] L. Galdieri, T. Zhang, D. Rogerson, R. Lleshi, A. Vancura, Protein acetylation and acetyl coenzyme a metabolism in budding yeast., Eukaryot. Cell. 13 (2014) 1472–83. doi:10.1128/EC.00189-14. [22] H. Takahashi, J.M. McCaffery, R.A. Irizarry, J.D. Boeke, Nucleocytosolic acetyl- coenzyme a synthetase is required for histone acetylation and global transcription., Mol. Cell. 23 (2006) 207–17. doi:10.1016/j.molcel.2006.05.040. [23] S.C. Kim, R. Sprung, Y. Chen, Y. Xu, H. Ball, J. Pei, T. Cheng, Y. Kho, H. Xiao, L. Xiao, N. V Grishin, M. White, X.-J. Yang, Y. Zhao, Substrate and functional diversity of lysine acetylation revealed by a proteomics survey., Mol. Cell. 23 (2006) 607–18. doi:10.1016/j.molcel.2006.06.026. [24] R.S.P. Rao, J.J. Thelen, J.A. Miernyk, In silico analysis of protein Lys-N(?)-acetylation in plants, Front. Plant Sci. 5 (2014) 381. doi:10.3389/fpls.2014.00381. [25] A.C. König, M. Hartl, P.J. Boersema, M. Mann, I. Finkemeier, The mitochondrial lysine acetylome of Arabidopsis, Mitochondrion. 19 (2014) 252–260. doi:10.1016/j.mito.2014.03.004. [26] R. Pandey, A. Müller, C.A. Napoli, D.A. Selinger, C.S. Pikaard, E.J. Richards, J. Bender, D.W. Mount, R.A. Jorgensen, Analysis of histone acetyltransferase and histone deacetylase families of Arabidopsis thaliana suggests functional diversification of chromatin modification among multicellular eukaryotes., Nucleic Acids Res. 30 (2002) 5036–55. [27] F. Zhang, L. Wang, E.E. Ko, K. Shao, H. Qiao, Histone Deacetylases SRT1 and SRT2 Interact with ENAP1 to Mediate Ethylene-Induced Transcriptional Repression, Plant Cell. 30 (2018) 153–166. doi:10.1105/tpc.17.00671. [28] A.-C. Konig, M. Hartl, P.A. Pham, M. Laxa, P.J. Boersema, A. Orwat, I. Kalitventseva, M. Plochinger, H.-P. Braun, D. Leister, M. Mann, A. Wachter, A.R. Fernie, I. Finkemeier, The Arabidopsis Class II Sirtuin Is a Lysine Deacetylase and Interacts with Mitochondrial Energy Metabolism, PLANT Physiol. 164 (2014) 1401–1414. doi:10.1104/pp.113.232496. 11

[29] X. Liu, W. Wei, W. Zhu, L. Su, Z. Xiong, M. Zhou, Y. Zheng, D.-X. Zhou, Histone Deacetylase AtSRT1 Links Metabolic Flux and Stress Response in Arabidopsis, Mol. Plant. 10 (2017) 1510–1522. doi:10.1016/j.molp.2017.10.010. [30] J. Gallego-Jara, A. Écija Conesa, T. de Diego Puente, G. Lozano Terol, M. Cánovas Díaz, Characterization of CobB kinetics and inhibition by nicotinamide., PLoS One. 12 (2017) e0189689. doi:10.1371/journal.pone.0189689. [31] P. Knyphausen, S. de Boor, N. Kuhlmann, L. Scislowski, A. Extra, L. Baldus, M. Schacherl, U. Baumann, I. Neundorf, M. Lammers, Insights into Lysine Deacetylation of Natively Folded Substrate Proteins by Sirtuins., J. Biol. Chem. 291 (2016) 14677– 94. doi:10.1074/jbc.M116.726307. [32] C. Wang, F. Gao, J. Wu, J. Dai, C. Wei, Y. Li, Arabidopsis putative deacetylase AtSRT2 regulates basal defense by suppressing PAD4, EDS5 and SID2 expression., Plant Cell Physiol. 51 (2010) 1291–9. doi:10.1093/pcp/pcq087. [33] K. Lee, O.S. Park, S.J. Jung, P.J. Seo, Histone deacetylation-mediated cellular dedifferentiation in Arabidopsis, J. Plant Physiol. 191 (2016) 95–100. doi:10.1016/j.jplph.2015.12.006. [34] B.J. Nikolau, M.A.D.N. Perera, L. Brachova, B. Shanks, Platform biochemicals for a biorenewable chemical industry., Plant J. 54 (2008) 536–45. doi:10.1111/j.1365- 313X.2008.03484.x. [35] A.S. Reger, J.M. Carney, A.M. Gulick, Biochemical and crystallographic analysis of substrate binding and conformational changes in acetyl-CoA synthetase., Biochemistry. 46 (2007) 6536–46. doi:10.1021/bi6026506. [36] T. Fujino, J. Kondo, M. Ishikawa, K. Morikawa, T.T. Yamamoto, Acetyl-CoA synthetase 2, a mitochondrial matrix enzyme involved in the oxidation of acetate., J. Biol. Chem. 276 (2001) 11420–6. doi:10.1074/jbc.M008782200. [37] R.H. Behal, M. Lin, S. Back, D.J. Oliver, Role of acetyl-coenzyme A synthetase in leaves of Arabidopsis thaliana., Arch. Biochem. Biophys. 402 (2002) 259–67. doi:10.1016/S0003-9861(02)00086-3.

CHAPTER 2. ALTERING THE SUBSTRATE SPECIFICITY OF ACETYL-COA

SYNTHETASE BY RATIONAL MUTAGENESIS OF THE CARBOXYLATE

BINDING POCKET

This chapter has been modified from the manuscript submitted to journal ACS Synthetic

Biology

Naazneen Sofeo1,3,4, Jason H. Hart1,3, Brandon Butler3, David J. Oliver2,3, Marna D.

Yandeau-Nelson1,2,3,4, Basil J. Nikolau1,3,4,5

1 Roy J. Carver Department of Biochemistry, Biophysics and Molecular Biology, Iowa State

University, Ames, IA 50011, USA

2 Department of Genetics, Development and Cell Biology, Iowa State University, Ames, IA

50011, US

3 Engineering Research Center for Biorenewable Chemicals Iowa State University, Ames, IA

50011, USA

4 Center for Metabolic Biology, Iowa State University, Ames, IA 50011, USA

5 Corresponding Author

Abstract

Acetyl-CoA synthetase (ACS) is a member of a large super-family of enzymes that display diverse substrate specificities, with a common mechanism of catalyzing the formation of a thioester bond between Coenzyme A and a carboxylic acid, while hydrolyzing ATP to AMP and pyrophosphate. As an activated form of acetate, acetyl-CoA is a key metabolic intermediate that links many metabolic processes, including the TCA cycle, amino acid metabolism, fatty acid metabolism and biosynthetic processes that generate many polyketides and some terpenes. We explored the structural basis of the specificity of ACS for only activating acetate, whereas other members of this superfamily utilize a broad range of other carboxylate substrates. By computationally modeling the structure of the Arabidopsis ACS and the Pseudomonas chlororaphis isobutyryl-CoA synthetase using the experimentally determined tertiary structures of homologous ACS enzymes as templates, we identified residues that potentially comprise the carboxylate binding pocket. These predictions were systematically tested by mutagenesis of four specific residues. The resulting rationally redesigned carboxylate binding pocket modified the size and chemo-physical properties of the carboxylate binding pocket. This redesign successfully switched a highly specific enzyme from using only acetate, to be equally specific for using longer linear (up to hexanoate) or branched chain (methylvalerate) carboxylate substrates. The significance of this achievement is that it sets a precedent for understanding the structure-function relationship of an enzyme without the need for an experimentally determined tertiary structure of that target enzyme, and rationally generates new biocatalysts for metabolic engineering of a broad range of metabolic processes.

Keywords: Acetyl CoA synthetase, Targeted mutagenesis, Enzyme engineering, Substrate specificity, Homologous enzymes, Arabidopsis 14

Introduction

The promise of synthetic biology is based on the premise that the systematic application of engineering principles to redesign individual molecular entities can outcompete natural evolutionary-based optimized design of complex biological processes. Furthermore, these individual genetically encoded redesigned entities can be fabricated into assemblages of novel systems (chassis), which do not already exist in the natural world 1, 2. Enzyme catalysts that provide specificity and enhanced rates of chemical conversions are key biological entities that are targets of such human-driven optimization. A limiting factor in such optimization redesigns is the fact that enzymatic functionality is based on tertiary structural design principles that cannot a priori be deduced from primary structural data, which are the most readily available datasets concerning structures of enzymes. This study illustrates a research path for redesigning a specific enzyme whose tertiary structure is unknown, but could be computationally modelled based on tertiary structures of functionally related enzymes 3, 4. The systematic targeted engineering of an AMP-generating acetyl-CoA synthetase from Arabidopsis thaliana (atACS) led to the redesign of this enzyme to generate new synthetic enzymes that display distinct catalytic capabilities not present in any of the enzymes that served as templates for the redesign.

The atACS enzyme is a member of the AMP-forming acyl-CoA synthetase family, which is part of the adenylate forming superfamily of enzymes 5-7. The initial half reaction catalyzed by this superfamily of enzymes activates a carboxylate substrate by converting the hydroxyl leaving group of the substrate to an adenosine monophosphate (adenylate intermediate) using the

7-9 hydrolysis of PPi to drive the reaction in the forward direction (Figure 1). In the second half reaction, the adenylate intermediate reacts with the thiol group of the phosphopantothenate group of Coenzyme A, generating an acyl-CoA product, which can be considered as the “activated” form 15 of the carboxylate substrate 7, 10. Different acyl-CoA synthetases in this family act upon different carboxylate substrates, and these include for example, acetate, propionate, malonate, chlorobenzoate 4, 11-15, which vary in the nature of the alkyl group of the carboxylate substrate.

Thioester bonds, as occurs in acyl-CoA, are crucial to a large number of anabolic and catabolic processes. These include fatty acid and polyketide metabolism, assembly of complex lipids, amino acid metabolism, assembly of mevalonate for the biosynthesis of isoprenoids, and porphyrin biosynthesis 10, 16. Chemically the formation of the thioester bond facilitates the subsequent reactions of the acyl moiety because of the high energy content of the bond, and because the thiol group is a better leaving group than the hydroxyl group of the initial carboxylate substrate for the nucleophilic reactions, that lead to the formation of different carbonyl-containing products (e.g., esters, amides and ketones). Indeed, because of the high energy content of the thioester bond, and its chemical flexibility, it has been theorized that it could have been a precursor to life, preceding the role of ATP as the common energy carrier in a so called "thioester world" 17.

The most prevalent thioester-containing intermediate of metabolism is acetyl-CoA, and biological systems utilize multiple enzymatic mechanisms to generate this central metabolite.

Most prevalent of these reactions is likely the oxidative decarboxylation of pyruvate, catalyzed by the pyruvate dehydrogenase complex 18, and the retro-Claisen cleavage of citrate, catalyzed by ATP-citrate lyase 19. However, ACS is part of a large family of enzymes that is widely prevalent in biological systems, and in some cases, such as in plants, the physiological relevance of this enzyme to metabolism is not clear 18, 20-23. Due to this potential metabolic redundancy, engineering ACS to utilize different carboxylate substrates has considerable potential in impacting the end-products of metabolism, particularly in plant systems. Acetyl- 16

CoA is a key precursor of fatty acid biosynthesis and many polyketide biosynthetic processes, and it both primes these complex processes, and is the precursor of the malonyl-CoA substrate that is used to elongate the fatty acids or polyketides (e.g., chalcone synthase)24-26. Hence, engineering ACS to utilize novel carboxylate substrates and hence generate alternative acyl-

CoA substrates for fatty acid or polyketide biosynthesis would provide the means to explore the enzymological flexibilities of these biosynthetic machineries and build a technology platform to produce a diversity of chemical products with wide ranging applications 27. As a prerequisite to such redesign of biological chasses, in this study we have been successful in not only broadening the substrate specificity of a eukaryotic ACS, but shifting its substrate specificity, using a systematic targeted mutagenesis strategy.

Results & Discussion

Identification of the Residues Forming the Putative Carboxylate Binding Pocket of atACS

Alignment of sequences of multiple acyl-CoA synthetases, which utilize different carboxylate substrates, reveal that atACS shares higher sequence identity (> 40%) with other acyl-CoA synthetases that utilize acetate as the preferred substrate, rather than those that utilize propionate, isobutyrate or malonate (Table 1). This conservation in sequence is focused in ten motifs (A1 to A10) (Figure 2), three of which (motifs A3, A4 and A5) have been associated with: a) the carboxylate substrate binding pocket of acyl-CoA synthetases; b) the amino acid binding pocket of non-ribosomal peptide synthases, and c) in interacting with the pyrophosphate product of these enzymes 7, 10, 28. We explored the basis for the high substrate specificity of atACS for acetate, by comparing and contrasting the structure of this enzyme with that of the Pseudomonas chlororaphis isobutyryl-CoA synthetase (pcICS), which is equally specific, but for isobutyrate. Higher order structural insights of atACS and 17 pcICS were obtained by using 3D structural models generated with I-TASSER 29-31, using as a guide the experimentally determined structures of the acetate-utilizing homologous ACS from Salmonella enterica (PDB: 1PG4) 32 and Saccharomyces cerevisiae (PDB: 1RY2) 33.

Despite the fact that atACS and pcICS differ in primary structure (sharing only 30% sequence identity), both computationally generated 3D models of atACS and pcICS share a high degree of tertiary structural alignment with each other and with both S. enterica ACS and S. cerevisiae ACS. The 3D structural alignment of atACS and pcICS structures has an all-atom Root Mean Square Deviation (RMSD) of 14 Å for 3159 atoms. The RMSD provides a numerical insight on the structural alignment between two tertiary structures; a smaller

RMSD value indicates a higher degree of structural alignment. Comparison of the computationally generated structures of atACS and pcICS with the experimentally determined structures of S. enterica ACS (1PG4) and S. cerevisiae ACS (1RY2) indicates a higher degree of 3D structural alignment among the enzymes that share the common substrate, acetate. Specifically, the all-atom RMSD comparison between atACS and S. enterica ACS or S. cerevisiae ACS is 1.8 Å and 10.5 Å, respectively, whereas the homologous comparisons between pcICS and S. enterica ACS or S. cerevisiae ACS is 13.7

Å and 8.8 Å, respectively (Table 1 & Figure 3A & 3B). Therefore, the atACS model has smaller RMSD values when compared to enzymes that act on smaller carboxylate substrates

(e.g., acetate, malonate and propionate) as compared to enzymes that act on larger carboxylate substrates (e.g., isobutyrate or benzoate) (Table 1). These insights provide confidence that the

3D structural comparisons provide valuable and accurate insights into design principles for the substrate binding pocket that can be experimentally validated by directed mutagenesis.

These computational comparisons identified four residues that may comprise the carboxylate binding pocket of atACS (i.e., residues Ile323, Thr324, Val399 and Trp427), and of pcICS (i.e., residues Ala278, Tyr279, Ala313 and Gly338) (Figure 3C & 3D). Confidence that these identified residues are important for enzymatic function is further enhanced by the fact that the former set are conserved in all acyl-CoA synthetases that utilize acetate as the preferred carboxylate substrate (Figure 2). The accuracy of these predictions was evaluated by singularly and in combination mutating the atACS residues (i.e., Ile323, Thr324, Val399 and Trp427) to the corresponding residues that comprise the binding pocket of pcICS (Figure 2). The expectation is that these mutations would affect catalysis and/or alter the carboxylate substrate specificity of the atACS enzyme. The kinetic parameters (Vmax, Km and kcat) of the resulting mutant proteins were determined with a range of different carboxylate substrates, including a series of straight chain carboxylates (acetate to octanoate) and branched chain carboxylates (isobutyrate,

3-methylvalerate, 4-methylvalerate and 4-methylhexanoate), and these data were compared to the parallel data generated with the wild-type atACS.

Mutations that identify an important residue for catalysis

The wild-type atACS follows traditional Michaelis-Menten kinetics, with a strong preference for acetate as the carboxylate substrate. Specifically, the catalytic efficiency

(kcat/Km) is approximately 55-fold higher with acetate than propionate, and there is no detectable activity with longer chain carboxylates (Supporting information, Table S1). The singular replacement of the Thr324 residue with the Tyr residue that occurs at this position in the pcICS structure inactivates atACS with all tested substrates. Furthermore, each multi-site mutant that contained this Thr324Tyr mutation (i.e., the double mutant Ile323Ala, Thr324Tyr and the quadruple mutant Ile323Ala, Thr324Tyr, Val399Ala, Trp427Gly) also fails to exhibit enzymatic 19 activity with any of the tested carboxylate substrates (Supporting information, Table S1). The

CD spectra of these mutants indicate near normal folding of the mutant proteins (Figure 4).

However, homology modelling of the Thr324Tyr mutant protein revealed a change in the surface topology of the predicted substrate binding pocket resulting in a decreased volume for binding of the carboxylate substrate (Figure 5). This is consistent therefore with steric hindrance hampering the ability of the carboxylate substrates to access the catalytic site and thereby inactivating the enzyme.

Mutations that switch substrate specificity to longer, linear chain carboxylate substrates

Two single site mutants, Trp427Gly and Val399Ala, resulted in altered substrate specificity of atACS. The former mutation causes the most dramatic alteration, switching the substrate preference from acetate to longer, straight-chain carboxylate substrates. Namely, this mutant enzyme is able to utilize “new” carboxylate substrates that range between 3- to 8- carbons in length (Supporting information, Table S1). Furthermore, this switch is accompanied by a 6.5-fold decrease in catalytic efficiency (kcat/Km) with acetate, the native substrate of the wild-type enzyme (Figure 6A).

The other single-site mutant that affected catalytic capabilities of atACS is the

Val399Ala switch, which broadened the substrate specificity of the enzyme so that it could equally utilize both acetate and propionate. Thus, the catalytic efficiency (kcat/Km) with either of these carboxylates is similar to the catalytic efficiency of the wild-type enzyme with its native acetate-substrate. However, unlike the Trp427Gly mutant, the Val399Ala mutant enzyme could not utilize longer chain carboxylate substrates of 4-8 carbon atoms (Supporting information, Table S1). 20

Based on the fact that these two single site mutations at Val399 and Trp427 individually generate somewhat distinct changes in catalytic activity of atACS, we evaluated the effect of combining the two mutations in the double mutant, Val399Ala,

Trp427Gly. This double mutant further enhanced the switch in substrate specificity that was observed with the Trp427Gly single mutant (Figure 6A). Specifically, as compared to the wild-type enzyme, the Val399Ala, Trp427Gly double mutant shows an 80-fold decrease in catalytic efficiency with the acetate-substrate, and a concomitant increase in catalytic efficiency with the longer, straight chain carboxylate substrates (i.e., propionate to octanoate). With propionate as the substrate, this double mutant enzyme shows a 2.5-fold increase in catalytic efficiency as compared to either the wild-type enzyme or the Trp427Gly single mutant (Figure 6A). The catalytic efficiency with butyrate, the best substrate for this mutant enzyme, matched the catalytic efficiency expressed by the wild-type enzyme with acetate. With these “new” longer chain-length substrates, the double mutant enzyme

427 displayed a lower Km value (ranging from 1.5 to 7.5-fold lower values) than the Trp Gly single mutant (Supporting information, Table S1). Analysis of the structural homology models indicate that both the Trp427Gly and Val399Ala single mutations enlarge the carboxylate binding pocket significantly, and further enlargement is apparent in the

Val399Ala, Trp427Gly double mutant. Thus, these models are consistent with the ability of these variant enzymes to use the larger substrates, and in general it is likely that steric hinderance is a major factor in narrowing the substrate specificity for the native atACS enzyme

(Figure 5).

Mutations that switch substrate specificity to branched chain carboxylate substrates

The Thr324 residue of atACS, which we showed to be critical for catalytic competence because its replacement with the equivalent residue that occurs in pcICS (i.e.,

Tyr) resulted in an inactive enzyme, was subsequently replaced with residues that have smaller side chains. These mutations were systematically implemented to explore whether maintaining either the hydroxyl-group functionality (i.e., a serine switch) or the branched alkyl chain functionality (i.e., a valine switch) or completely eliminating the side chain (i.e., a glycine switch) of the existing Thr residue is important for catalysis.

The Thr324Ser variant did not show any activity with acetate or longer straight chain substrates (Supporting information, Table S1), however the Thr324Val and Thr324Gly variants maintain the ability to utilize acetate as a substrate, but with quantitative differences between them (Figure 6B; Supporting information, Table S1). Namely, the catalytic efficiency of the Thr324Val mutant is about 2-fold higher than either the wild-type or

324 Thr Gly variant with acetate, and this was primarily due to a lower Km value for this substrate. Moreover, the Thr324Val and Thr324Gly single mutants also showed qualitative differences in substrate specificity, gaining the ability to utilize branched chain carboxylate substrates, such as isobutyrate and 4-methylvalerate, a capability that is undetectable with the wild-type atACS (Figure 6B; Supporting information, Table S1).

The 3D homology models of these Thr324 single mutants reveal the potential explanations of these effects on catalytic ability of the variant enzymes. In contrast to the

Thr324Tyr single mutant, which inactivates the enzyme due to an apparent reduction in the size of the carboxylate binding pocket (Figure 5), the Thr324Gly single mutant is predicted to have a larger carboxylate binding pocket as compared to the wild-type enzyme, thus 22 explaining the ability of this mutant to utilize the larger, branched chain carboxylate substrates (Figure 5). Similarly, the Thr324Val single mutant is also predicted to have a larger carboxylate binding pocket, although smaller than the Thr324Gly single mutant, which explains the ability of this single mutant to utilize straight-chain substrates, up to three carbon atoms in length (i.e. propionate) and also branched chain substrates, albeit with lower efficiency as compared to the Thr324Gly single mutant. The size of the binding pocket of the

Thr324Ser mutant is similar to the wild-type enzyme (Figure 5), thus the inactivity of this single site mutant cannot be explained solely based on steric interference between the substrate and pocket-size, but instead may be attributable to the difference in the polarity of the pocket affected by the orientation of the hydroxyl group of the Ser side chain as compared to Thr in wild-type (Figure 5). These findings therefore indicate that both steric effects and the chemo-physical properties (polarity versus hydrophobicity) of the binding pocket affect the carboxylate specificity of the enzyme.

Based on these observations, we systematically combined the singular Thr324Gly or

Thr324Ser mutations with the double mutant, Val399Ala, Trp427Gly, to evaluate the interrelationships among these residues to affect catalytic capabilities of atACS. The enzymatic capabilities of the resulting Thr324Gly, Val399Ala, Trp427Gly and Thr324Ser, Val399Ala,

Trp427Gly triple mutants were compared to the wild-type and the parental double mutant enzymes.

These triple mutants express enhanced catalytic capabilities in utilizing longer chain linear carboxylate substrates than the Val399Ala, Trp427Gly double mutant enzyme (Figure 6C).

Thus, whereas the double mutant can utilize butyrate with the highest efficiency (Figure 6A), the Thr324Gly, Val399Ala, Trp427Gly triple mutant can utilize hexanoate with a catalytic 23 efficiency that is comparable to that of the wild-type enzyme with its native acetate substrate

(Figure 6C). Moreover, these triple mutants also utilize branched chain substrates, such as 3- methylvalerate, 4-methylvalerate and 4-methylhexanoate. Compared to the Val399Ala,

Trp427Gly double mutant, this additional switch in the preferred carboxylate substrate by the triple mutants is associated with decreased Km values for the longer chain length carboxylate substrates of more than 5-carbon atoms, and increased Km values for the shorter chain substrates, such as propionate, butyrate and isobutyrate (Supporting information, Table S1).

Thus, comparing the modeled structure of the carboxylate binding pockets of the

Val399Ala, Trp427Gly double mutant, with the Thr324Gly, Val399Ala, Trp427Gly triple mutant indicates that the replacement of Thr324 with a smaller side chain residue, i.e., glycine, affects both the shape and size of the binding pocket, and its polarity (Figure 5), resulting in a shift in substrate specificity from a 2-carbon substrate, to more than 4-carbon long, straight chain and branched substrates, which are also less polar. These findings further strengthen the conclusion that both steric interference and chemo-physical properties of the binding pocket plays in determining the carboxylate substrate specificity of atACS.

Even more dramatic changes in substrate specificity were generated when the

Ile323Ala mutation was combined with the Val399Ala, Trp427Gly or Val399Gly, Trp427Gly double mutations. Thus, Figure 6C shows that both the Ile323Ala, Val399Gly, Trp427Gly and

Ile323Ala, Val399Ala, Trp427Gly triple mutants are inactive with acetate and only shows barely detectable activity with propionate, both of these triple mutants can utilize 4- to 8- carbon straight chain carboxylates as substrates, as well as the branched chain carboxylates,

3-methylvalerate and 4-methylvalerate. The catalytic efficiency of these triple mutant enzymes with branched chain substrates is comparable to the activity expressed by the wild- 24 type enzyme with acetate as the substrate. Detailed examination of the kinetic properties of all the triple mutants indicate that they prefer 4-methyl branched chain carboxylates (i.e., 4- methylvalerate and 4-methylhexanoate) over the 3-methyl branched carboxylate (i.e., 3- methylvalerate), manifested by lower Km values and higher catalytic efficiencies.

The structural homology models of the triple mutants indicate that these mutations enlarge the carboxylate binding pocket, thus explaining the capability of these mutants to utilize longer chain substrates (Figure 5). The triple mutant Thr324Gly, Val399Ala, Trp427Gly which shows the largest shift in substrate specificity (Figure 6C), shows the most non-polar binding pocket cavity (Figure 5), which provides the explanation of why this variant is more active with the most hydrophobic carboxylate substrates.

Because the Trp427Gly mutation has a large effect on substrate specificity and promiscuity (Supporting information, Table S1), we explored the consequence of changing the Trp427 residue to all other possible amino acids (the consequence of changing Trp427 to

Gly had already been evaluated; Fig. 6). These variants thereby evaluated how the change in the chemo-physical property of this amino acid side chain affects substrate specificity

(Supporting information, Table S1). Only five of the 18 additional variants were catalytically active (Val, Ala, Phe, Ser or His), and the other 13 were inactive with all tested carboxylates. These five single site mutants utilize acetate as a substrate, though with lower catalytic efficiency (Supporting information, Table S1). The mutations that incorporate small hydrophobic side chain residues (i.e., Trp427Val or Trp427Ala) enable the enzyme to better utilize straight chain carboxylate substrates, of 3- and 4-carbon atoms. In addition, the mutations that conserve the ring structure of the native Trp427 (i.e. Trp427Phe or

Trp427His) maintain enzymatic activity with acetate. These findings indicate that Trp427 is 25 an important residue in the carboxylate binding pocket, and in addition to its aromaticity and/or polarity associated with the indole ring, the size of the side chain has an effect on restricting the specificity of the wild-type enzyme to acetate.

Kinetics of ATP and CoA dependence

The wild-type atACS and selected mutants showing altered preference for the carboxylate substrate were also characterized to evaluate the effects of these mutations on the kinetic parameters associated with the other two substrates of the reaction, ATP and CoA

(Table 2). The assays used to determine the Km and Vmax values for these substrates were conducted with saturating concentrations of the most preferred carboxylate substrate for the individual variant enzymes. These characterizations demonstrate that the mutations that were designed to affect the specificity of the carboxylate substrate also affected the enzyme’s dependence on these two additional substrates, although the magnitude of these changes are considerably smaller than the changes in the kinetic parameters for the carboxylate substrates.

The most dramatic of these changes is associated with the kinetic parameters for ATP, with up to 10-fold decrease in Km in mutants that enable the enzyme to better utilize longer

427 chain linear carboxylate substrates (e.g., Trp Gly), and similar 10-fold decrease in Vmax in the triple-site mutant (i.e., Ile323Ala Val399Ala Trp427Gly) that switches the carboxylate substrate preference from acetate to the branched chain substrates. In contrast, there are relatively minor effects on the kinetic parameters associated with CoA. The alterations in these kinetic parameters are consistent with the mechanism of the ACS-catalyzed reaction; namely ATP and the carboxylate substrates first react to form the adenylate-intermediate, which subsequently

26 reacts with CoA. Thus, changing the carboxylate binding-site would be expected to have greater effects on ATP kinetics as these two substrates have to come together at an active site that is juxtaposed next to each substrate binding pocket.

Conclusion

Despite the conservation of ACS enzymes among a wide range of evolutionary phyla, this enzyme has maintained a restricted substrate specificity 32. For example, the plant ACS that was the focus of this study only utilizes acetate efficiently as a substrate, and propionate to a much lesser extent. In this study we extrapolated from experimentally determined 3-D structural data of two phylogenetically distantly related ACS enzymes, and modelled the potential carboxylate substrate binding pockets of two enzymes that also assemble thioester bonds but display different specificities toward the carboxylate substrate (acetate or isobutyrate). Rational mutagenesis experiments guided by this modeling were successful in initially introducing substrate promiscuity to the enzyme, and ultimately altering its substrate specificity. We have been particularly successful in shifting the substrate specificity of this eukaryotic ACS away from its native substrate (acetate), to be equally efficient in utilizing novel branched chain carboxylates that are not substrates of any of the guiding model enzymes.

Similar studies have targeted microbial ACS enzymes34-36, which expanded the carboxylate substrates that these enzymes can utilize, thus converting a very specific enzyme to a promiscuous enzyme. The variant enzymes that we have generated are distinct in that they are not promiscuous, but they show shifted substrate specificities, being equally efficient with the new substrates, as the wild-type enzyme is with its native substrate.

Hence, this is an exemplary study of a research path to generate novel enzymes.

Moreover, the novel acyl-CoA synthetase enzymes that were generated herein can serve as 27 genetically encoded catalytic elements for bioengineering multiple metabolic pathways and potentially producing useful novel chemicals 27, 37-38. Specifically, the integration of these new enzymes into fatty acid or polyketide biosynthetic processes can lead to the generation of bioproducts with distinct chemo-physical properties.

Methods

Generation of atACS variants

The atACS ORF was cloned into the pET24b vector (Merck KGaA, Darmstadt,

Germany) between the HindIII and XhoI sites. atACS was also subcloned into the pET30f vector at the BamHI restriction site. pET30f vector is a modified pET30 vector (Takara, Clontech,

Mountain View, CA) encoding an N-terminal poly-histidine coding sequence, followed by a

TEV cleavage site. Mutagenesis of specific residues was performed either using the

QuikChange Lightning site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA) via manufacturer instructions or by recombination PCR in a two-step fashion 13. The atACS

ORF cloned into the plasmid pET24b, was used as template for QuikChange or recombination

PCR. The full-length atACS ORF PCR product containing the mutation was purified and cloned into a pGEM T-Easy vector, and then transformed into Top-10 competent E. coli cells.

Sequence-confirmed atACS variants in pGEM T-Easy were digested with HindIII and XhoI to release the atACS ORF and were subcloned into the corresponding restriction sites in the pET24b vector. Sequence confirmed pET30f and pET24b constructs were subsequently transformed into E. coli ArcticExpress (DE3) cells (Agilent Technologies, Santa Clara, CA) for protein expression.

Overexpression and purification of wild-type and variant atACS proteins

Arctic Express strains harboring atACS variants were grown overnight at 37 °C with shaking at 250 rpm, in 5-10 ml LB medium containing the antibiotics kanamycin (50 μg/ml) and gentamycin (20 μg/ml). The overnight culture was used to inoculate 0.5-1.0 L LB media containing no antibiotics, and these cultures were grown at 30 °C with shaking at 250 rpm, until an OD600 of ~0.6. Expression was then induced by the addition of IPTG to a final concentration of 0.1 mM, and the culture was grown at 13 °C with shaking at 250 rpm for 24-

48 hours.

Cells were harvested by centrifugation at 10,000g for 10 minutes and resuspended in a buffer containing 0.5 M NaCl, 5 mM imidazole, 0.1% (v/v) Triton X-100, 0.1 mg/ml phenyl methyl sulfonyl fluoride (PMSF), and 10 μl/ml Protease Inhibitor Cocktail (Sigma-Aldrich

Co., St. Louis, MO) and disrupted by sonication on ice. Two different buffer systems were used in this cell disruption step, either 10 mM HEPES-KOH, pH 7.5, or 20 mM Tris-HCl, pH

7.5. The resulting extract was centrifuged at 20,000xg for 30 minutes, and the supernatant was retained, filtered through a 0.45 μm filter disc (Corning Inc., Corning, NY) and applied to a column containing 2-5 ml PerfectPro Ni-NTA agarose (5 Prime, Inc., Gaithersburg, MD) at 4 °C. The column was washed with increasing concentrations of imidazole, and the atACS protein was eluted with 0.2 M imidazole. Samples were immediately dialyzed at 4 °C into either 10 mM HEPES-KOH, pH 7.5, 10 mM KCl or 50mM Tris-HCl pH 7.5; these two methods yielded enzyme preparations that expressed different specific activities, and they are

specifically identified in the Results. In all comparative studies of mutants, both the wild-type and all variant enzymes were prepared identically, so as to avoid this buffer variable, whose origins are as of yet unknown.

Spectrophotometric atACS Activity Assay

ACS activity was measured by coupling the carboxylate-dependent formation of the

AMP product to the oxidation of NADH, using the enzymes myokinase, pyruvate kinase and lactate dehydrogenase 4 (Sigma-Aldrich Co., St. Louis, MO). The oxidation of NADH was monitored by the rate of decrease in absorbance at 340 nm. Each enzyme variant was assayed with 11 different carboxylate substrates that differed in the alkyl chain-length (2-8 carbons) or whether the alkyl chain was linear or branched, and their concentrations in assays ranged between 0-10 mM. The specific substrates were: acetate, propionate, butyrate, valerate, hexanoate, heptanoate, octanoate, isobutyrate, 3-methyl valerate, 4-methyl valerate and 4- methyl hexanoate. Assays were performed in 96-well micro-titer dishes at 37 °C with a final well-volume of 100 μl, and A340 of each well was monitored with a BioTek ELx808TM

Absorbance Microplate Reader, using Gen5TM Data Analysis software (BioTek Instruments,

Winooski, VT). The standard reaction contained 50 mM Tris-HCl pH 7.5, and 5mM Tris(2- carboxylethyl) phosphine hydrochloride (Thermo Fisher Scientific Inc., Waltham, MA); 5%

(v/v) ethanol; 5 mM ATP; 6mM MgCl2; 5 mM phospho(enol)pyruvate; 0.4 mM NADH; 2-20

U Pyruvate Kinase/Lactate Dehydrogenase mix (Sigma-Aldrich Co., St. Louis, MO); 10U

Myokinase from rabbit muscle or 1U Myokinase from chicken muscle (Sigma-Aldrich Co., St.

Louis, MO); and between 2.5-5 μg of purified recombinant atACS protein.

Computational Experiments

The chloroplast targeting sequence was identified in the atACS sequence using TargetP

1.1 (http://www.cbs.dtu.dk/services/TargetP). Amino acid similarity and identity values of homologous proteins were obtained using BLASTp (http://blast.ncbi.nlm.nih.gov). Structural predictions were performed using the I-TASSER server 28, 37, 40. Predicted structures and known crystal structures were visualized and analyzed using PyMOL (The PyMOL Molecular

Graphics System, Version 1.3.0 Schrödinger, LLC). Multiple sequence alignments were generated using Clustal Omega 39.

Circular Dichroism

CD spectra were obtained at the Iowa State University Protein Facility

(http://www.protein.iastate.edu/) using a J700 spectropolarimeter (JASCO Inc., Easton, MD).

Spectra were taken from 190-250 nm with readings every 0.2 nm. Samples were measured in a 1 mm dichroically neutral quartz cuvette and the average of 3-5 scans was used for analysis.

Protein samples in 50 mM Tris-HCl were diluted into ddH2O to a final concentration of 0.15-

0.25 mg/ml for spectra acquisition. A baseline spectrum for 5 mM Tris-HCl was obtained in the same manner and subtracted from the experimental protein spectra. After data collection,

2 -1 spectra were converted from millidegrees to molar ellipticity ([θ], degrees cm dmol ). Spectra were analyzed using a suite of algorithms collectively called CDPro 40, utilizing algorithms:

SELCON3, CDSSTR, and CONTIN/LL.

Abbreviations atACS: Arabidopsis thaliana acetyl CoA synthetase; ACS: Acetyl CoA synthetase; pcICS:

Pseudomonas chlororaphis isobutyryl CoA synthetase; ATP: Adenosine triphosphate; CoA:

Coenzyme A

Author contributions

B.J.N, M.D.Y-N and D.J.O conceived the study. N.S, J.H.H, M.D.Y-N, and B.J.N designed the experiments. N.S, J.H.H and B.B conducted the experiments. N.S, J.H.H, M.D.Y-N, and

B.J.N wrote the manuscript.

Notes: The authors declare no competing financial interest

Acknowledgement

This work was partially supported by the State of Iowa, through the Center of Metabolic

Biology, and by the National Science Foundation, which supported the Engineering Research

Center for Biorenewable Chemicals (Award No. EEC-0813570). Brandon Butler acknowledges support from the Research Experience for Undergraduates program of the

National Science Foundation, provided through the Engineering Research Center for

Biorenewable Chemicals. The authors thank Dr. Fuyuan Jing of Iowa State University for gifting the pET30f vector used for protein expression, and Dr. Alan T. Culbertson, who evaluated the phylogenetic-based interrelationships among ACS enzymes.

References

1. McCarty, N. S.; Ledesma-Amaro, R., Synthetic Biology Tools to Engineer Microbial Communities for Biotechnology. Trends in Biotechnology 2018, 0 (0).

2. Liu, Y.; Liu, L.; Du, G.; Chen, J.; Synthetic Biology Toolbox and Chassis Development in Bacillus subtilis. Trends in Biotechnology 2018, 0 (0).

3. Hashimoto, Y.; Hosaka, H.; Oinuma, K.-I.; Goda, M.; Higashibata, H.; Kobayashi, M., Nitrile Pathway Involving Acyl-CoA Synthetase. 2005.

4. Horswill, A. R.; Escalante-Semerena, J. C., The prpE gene of Salmonella typhimurium LT2 encodes propionyl-CoA synthetase. Microbiology 1999, 145 (Pt 6), 1381-8.

5. Conti, E.; Stachelhaus, T.; Marahiel, M. A.; Brick, P., Structural basis for the activation of phenylalanine in the non-ribosomal biosynthesis of gramicidin S. Embo j 1997, 16 (14), 4174- 83.

6. Gulick, A. M., Conformational dynamics in the Acyl-CoA synthetases, adenylation domains of non-ribosomal peptide synthetases, and firefly luciferase. ACS Chem Biol 2009, 4 (10), 811-27.

7. Marahiel, M. A.; Stachelhaus, T.; Mootz, H. D., Modular Peptide Synthetases Involved in Nonribosomal Peptide Synthesis. Chem Rev 1997, 97 (7), 2651-2674.

8. Berg, P.;With the technical assistance of Georgia, N., ACYL ADENYLATES: AN ENZYMATIC MECHANISM OF ACETATE ACTIVATION. Journal of Biological Chemistry 1956, 222 (2), 991-1013.

9. Schmelz, S.; Naismith, J. H., Adenylate-forming enzymes. Curr Opin Struct Biol 2009, 19 (6), 666-71.

10. Starai, V. J.; Escalante-Semerena, J. C., Acetyl-coenzyme A synthetase (AMP forming). Cell Mol Life Sci 2004, 61 (16), 2020-30.

11. Andrew M. Gulick,; Xuefeng Lu, a.; Dunaway-Mariano , D., Crystal Structure of 4- Chlorobenzoate:CoA Ligase/Synthetase in the Unliganded and Aryl Substrate-Bound States. 2004. ⊥

12. Guan, X.; Nikolau, B. J., AAE13 encodes a dual-localized malonyl-CoA synthetase that is crucial for mitochondrial fatty acid biosynthesis. Plant J 2016, 85 (5), 581-93.

13. Guo, Y.; Oliver, D. J., E. coli propionyl-CoA synthetase is regulated in vitro by an intramolecular disulfide bond. Prikl Biokhim Mikrobiol 2012, 48 (3), 289-93. 33

14. Hughes, A. J.; Keatinge-Clay, A., Enzymatic extender unit generation for in vitro polyketide synthase reactions: structural and functional showcasing of Streptomyces coelicolor MatB. Chem Biol 2011, 18 (2), 165-76.

15. Kai-Hsuan Chang; Hong Xiang, a.; Dunaway-Mariano, D., Acyl-Adenylate Motif of the Acyl-Adenylate/Thioester-Forming Enzyme Superfamily: A Site-Directed Mutagenesis Study with the Pseudomonas sp. Strain CBS3 4-Chlorobenzoate: Coenzyme A Ligase. 1997.

16. Pandey, S. K.; Yadav, S.; Temre, M. K.; Singh, S. M., Tracking acetate through a journey of living world: Evolution as alternative cellular fuel with potential for application in cancer therapeutics. Life Sci 2018, 215, 86-95.

17. de Duve, C., The Beginnings of Life on Earth. American Scientist 1995, 83 (5), 428-437.

18. Ke, J.; Behal, R. H.; Back, S. L.; Nikolau, B. J.; Wurtele, E. S.; Oliver, D. J., The Role of Pyruvate Dehydrogenase and Acetyl-Coenzyme A Synthetase in Fatty Acid Synthesis in Developing Arabidopsis Seeds1. In Plant Physiol, 2000; Vol. 123, pp 497-508.

19. Fatland, B. L.; Nikolau, B. J.; Wurtele, E. S., Reverse genetic characterization of cytosolic acetyl-CoA generation by ATP-citrate lyase in Arabidopsis. Plant Cell 2005, 17 (1), 182-203.

20. Bao, X.; Pollard, M.; Ohlrogge, J., The biosynthesis of erucic acid in developing embryos of brassica rapa. Plant Physiol 1998, 118 (1), 183-90.

21. Behal, R. H.; Lin, M.; Back, S.; Oliver, D. J., Role of acetyl-coenzyme A synthetase in leaves of Arabidopsis thaliana. Arch Biochem Biophys 2002, 402 (2), 259-67.

22. Lin, M.; Oliver, D. J.,The Role of Acetyl-Coenzyme A Synthetase in Arabidopsis. In Plant Physiol, 2008; Vol. 147, pp 1822-9.

23. Zeiher, C. A.; Randall, D. D., Identification and Characterization of Mitochondrial Acetyl- Coenzyme A Hydrolase from Pisum sativum L. Seedlings. Plant Physiol 1990, 94 (1), 20-7.

24. Austin, M. B.; Noel, J. P., The chalcone synthase superfamily of type III polyketide synthases. Nat Prod Rep 2003, 20 (1), 79-110.

25. Liedvogel, B., Acetyl Coenzyme A and Isopentenylpyrophosphate as Lipid Precursors in Plant Cells - Biosynthesis and Compartmentation. Journal of Plant Physiology 2015, 124 (3- 4), 211-222.

26. Brooks, J. L.; Stumpf, P. K., Fat metabolism in higher plants. XXXIX. Properties of a soluble fatty acid synthesizing system from lettuce chloroplasts. Arch Biochem Biophys 1966, 116 (1), 108-16.

27. Nikolau, B. J.; Perera, M. A.; Brachova, L.; Shanks, B., Platform biochemicals for a biorenewable chemical industry. Plant J 2008, 54 (4), 536-45. 34

28. Gulick, A. M.; Starai, V. J.; Horswill, A. R.; Homick, K. M.; Escalante-Semerena, J. C.,The 1.75 A crystal structure of acetyl-CoA synthetase bound to adenosine-5'-propylphosphate and coenzyme A. Biochemistry 2003, 42 (10), 2866-73.

29. Roy, A.; Kucukural, A.; Zhang, Y., I-TASSER: a unified platform for automated protein structure and function prediction. Nat Protoc 2010, 5 (4), 725-38.

30. Yang, J.; Yan, R.; Roy, A.; Xu, D.; Poisson, J.; Zhang, Y., The I-TASSER Suite: protein structure and function prediction. In Nat Methods, United States, 2015; Vol. 12, pp 7-8.

31. Zhang, Y., I-TASSER server for protein 3D structure prediction. BMC Bioinformatics 2008, 9, 40.

32. Reger, A. S.; Carney, J. M.; Gulick, A. M., Biochemical and Crystallographic Analysis of Substrate Binding and Conformational Changes in Acetyl-CoA Synthetase. Biochemistry 2007, 46 (22), 6536-46.

33. Jogl, G.; Tong, L., Crystal structure of yeast acetyl-coenzyme A synthetase in complex with AMP. Biochemistry 2004, 43 (6), 1425-31.

34. Ingram-Smith, C.; Woods, B. I.; Smith, K. S., Characterization of the acyl substrate binding pocket of acetyl-CoA synthetase. Biochemistry 2006, 45 (38), 11482-90.

35. Khersonsky, O.; Lipsh, R.; Avizemer, Z.; Ashani, Y.; Goldsmith, M.; Leader, H.; Dym, O.; Rogotner, S.; Trudeau, D. L.; Prilusky, J.; Amengual-Rigo, P.; Guallar, V.; Tawfik, D. S.; Fleishman, S. J., Automated Design of Efficient and Functionally Diverse Enzyme Repertoires. Mol Cell 2018, 72 (1), 178-186.e5.

36. Trudeau, D. L.; Edlich-Muth, C.; Zarzycki, J.; Scheffen, M.; Goldsmith, M.; Khersonsky, O.; Avizemer, Z.; Fleishman, S. J.; Cotton, C. A. R.; Erb, T. J.; Tawfik, D. S.; Bar-Even, A., Design and in vitro realization of carbon-conserving photorespiration. Proc Natl Acad Sci USA 2018.

37. Steen, E. J.; Kang, Y.; Bokinsky, G.; Hu, Z.; Schirmer, A.; McClure, A.; Del Cardayre, S. B.; Keasling, J. D., Microbial production of fatty-acid-derived fuels and chemicals from plant biomass. Nature 2010, 463 (7280), 559-62.

38. Schwartz, T. J.; O’Neill, B. J.; Shanks, B. H.; Dumesic, J. A., Bridging the Chemical and Biological Catalysis Gap: Challenges and Outlooks for Producing Sustainable Chemicals. 2014.

39. Thompson, J. D.; Higgins, D. G.; Gibson, T. J., CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 1994, 22 (22), 4673-80.

40. Sreerama, N.; Woody, R. W., Estimation of protein secondary structure from circular dichroism spectra: comparison of CONTIN, SELCON, and CDSSTR methods with an expanded reference set. Anal Biochem 2000, 287 (2), 252-60. 36

Figures

Figure 1: Enzymatic mechanism of AMP-forming acetyl CoA synthetase. The reaction occurs in two steps, where an acetyl-adenylate intermediate is formed in the first half-reaction. This intermediate then reacts with Coenzyme A to form acetyl-CoA and release AMP.

Figure 2: Multiple sequence alignment of the amino acid sequences of acyl CoA synthetases that utilize acetate (ACS), propionate (PCS), isobutyrate (ICS), malonate (MCS or MCL) and

2,3-dihyroxybenzoate (DHBe) as the preferred substrate. Alignment was performed using

Clustal Omega. Asterisks indicate the residues proposed to be forming the carboxylate binding pocket. White letters on a black background indicate identity, white letters on a gray background indicate similarity, and black letters on a white background indicate no conservation. Ten conserved motifs of the AMP-forming family of acyl-CoA synthetases are indicated (A1-A10).

Figure 3: A. Comparisons of the 3-D structural homology models of atACS (cyan ribbon diagram) and pcICS (green ribbon diagram) predicted using I-TASSER, and the experimentally determined structures of scACS (1RY2) (pink ribbon diagram) and seACS

(1PG4) (yellow ribbon diagram). These structures were aligned using PyMOL. B.

Comparisons of the 3-D structural homology models of atACS (purple ribbon diagram) and pcICS (green ribbon diagram) computationally predicted using I-TASSER and aligned using

PyMOL. C. Predicted position of the residues that constitute the putative carboxylate binding pocket of atACS (purple), ecACS (blue), ecPCS (pink), and pcICS (green). These alignments were guided with the experimentally determined tertiary structure of the seACS (beige).

Numbers represent the position of the homologous residue in the atACS protein. D. Putative binding pocket residues of atACS from the homology structure predicted using I-TASSER, aligned with the derived structures of Coenzyme A, AMP and the acetyl group from the experimentally determined structure of these substrates in 1RY2 and 1PG4. 39

Figure 4: Comparison of CD spectra of wild-type atACS and the mutated variants.

Measurements are an average of three scans with a reading every 0.2 nm. Units are in molar ellipticity [θ]. 40

Figure 5: Comparison of the carboxylate binding pocket surface computationally generated from 3D structure homology models of the wild-type and mutants of atACS. The mutated residues are colored in tan. The change in binding pocket size and shape is highlighted with a red box. The nitrogen and oxygen atoms of the targeted residues are shown in blue and red, respectively. The orange circle indicates the cut away opening to reveal the inside of the carboxylate binding pocket. The chemo-physical nature of the cavity surface is color-coded as tan for hydrophobic, red for negatively changed polarity, and blue for positively charged

399 427 polarity. The binding pocket model was generated using PyMOL. The Ala and Gly residues in the double and triple mutant images are not in the view as they are situated behind the expanded carboxylate binding cavity.

Figure 6: Comparison of the catalytic efficiency (kcat/Km) of the wild-type atACS with single and double site mutants (A, B), and triple site mutants (C), assayed with straight- and branched- 42 chain carboxylate substrates. The variant enzymes in panel A were purified using the Tris-HCl extraction buffer, and variants in panel C were purified using the HEPES extraction buffer (see

Materials and Methods). Data are the average of triplicate determination + standard error.

Tables

Table 1: Sequence identity (%) and RMSD (Å) values of all atom structural alignment of 3D structures of atACS and several acyl- CoA synthetases

1PG4# 1RY2# pcICS* atACS* 3NYR# 1MD9# mtACS* atMCS* sePCS* % Å % Å % Å % Å % Å % Å % Å % Å % Å 1PG4# 45.5 10.5 30 13.7 53.9 1.8 23.4 6.7 19.9 13.1 48 1.5 22.9 8.6 37 2.4 1RY2# 30.5 8.8 45.9 10.5 21 16.9 18 6.6 43 10.6 21.8 11.7 37.3 10.8 pcICS* 29.4 14 24.5 14.7 19.0 10.4 31.4 14 21.6 12.6 26.0 13.5 atACS* 22 8.3 21 13.7 46.4 3.7 22 6.8 35.8 3.2 3NYR# 27.6 12.5 23.3 6.5 39.8 4.6 21 9.2 1MD9# 22 13.1 23.9 11.8 21.5 12.4 mtACS* 22.7 11.6 37.3 2.8 4

atMCS* 18 10.2 3

sePCS*

1PG4: Salmonella enterica acetyl-CoA synthetase; 1RY2: Saccharomyces cerevisiae acetyl-CoA synthetase 1; pcICS: Pseudomonas chlororaphis isobutyryl-CoA synthetase; atACS: Arabidopsis thaliana acetyl-CoA synthetase; 3NYR: Streptomyces coelicolor malonyl-CoA ligase; 1MD9: Bacillus subtilis adenylation domain of NRPS to activate 2,3-dihydroxybenzoate; mtACS: Methanobacter thermoautotrophicus acetyl-CoA synthetase; atMCS: Arabidopsis thaliana malonyl-CoA synthetase; sePCS: Salmonella enterica propionyl-CoA synthetase. *Indicates proteins whose tertiary structures were computationally generated. #Indicates proteins whose tertiary structures were experimentally determined. 44

Table 2: Michaelis-Menten kinetic parameters of ATP and CoA for wild-type and variant atACS enzymes

ATP CoA

Carboxylate Enzyme Vmax Vmax a substrate Km (mM) (µmol/min Km (mM) (µmol/min ) )

Wild-type Acetate 0.20 ± 0.03 1.47 ± 0.05 0.15 ± 0.02 1.44 ± 0.06 Val399Ala Acetate 0.06 ± 0.01 0.84 ± 0.02 0.09 ± 0.02 1.18 ± 0.06 Trp427Gly Butyrate 0.02 ± 0.003 2.2 ± 0.1 0.14 ± 0.05 3.2 ± 0.4 Val399Ala, Butyrate 0.06 ± 0.02 1.9 ± 0.2 0.09 ± 0.02 2.5 ± 0.2 Trp427Gly Wild-type* Acetate 0.050 ± 0.002 11.1 ± 0.1 0.11 ± 0.03 11.1 ± 0.6 Thr324Gly, Hexanoate Val399Ala, 0.26 ± 0.05 3.5 ± 0.2 0.14 ± 0.03 3.5 ± 0.1 Trp427Gly* Ile323Ala, Val399Ala, Hexanoate 0.11 ± 0.03 1.11 ± 0.06 0.04 ± 0.02 1.00 ± 0.08 Trp427Gly*

The variants marked with an asterisk (*) were purified using HEPES buffer, as described in the Materials and Methods.

Supporting information

Table S1. Kinetic parameters of atACS variants

-1 -1 -1 atACS variant Substrate Km (mM) kcat (s ) kcat/Km (s mM ) Acetate 0.27 ± 0.01 2.41 ± 0.03 8.9 ± 0.5 Wild-type Propionate 4.0 ± 1.2 0.66 ± 0.07 0.16 ± 0.05 Acetate 3.1 ± 0.2 3.04 ± 0.05 0.99 ± 0.07 Val399Ala Propionate 3.1 ± 0.4 2.65 ± 0.07 0.85 ± 0.10 Acetate 10.8 ± 1.8 14.3 ± 0.9 1.32 ± 0.24 Propionate 37.5 ± 5.0 6.2 ± 0.5 0.16 ± 0.03 Butyrate 2.9 ± 0.6 9.4 ± 0.5 3.3 ± 0.7 Trp427Gly Valerate 6.5 ± 0.7 13.8 ± 0.4 2.13 ± 0.25 Hexanoate 12.6 ± 1.4 13.0 ± 0.5 1.03 ± 0.12 Heptanoate 10.0 ± 1.8 11.0 ± 0.6 1.10 ± 0.20 Octanoate 48 ± 12 5.8 ± 0.8 0.12 ± 0.03 Acetate 44 ± 6 5.07 ± 0.31 0.11 ± 0.02 Propionate 31 ± 5 12.94 ± 1.07 0.41 ± 0.07 Butyrate 1.5 ± 0.1 14.87 ± 0.57 9.9 ± 1.0 Val399Ala, Trp427Gly Valerate 4.0 ± 0.6 12.25 ± 0.47 3.1 ± 0.5 Hexanoate 7.3 ± 0.9 9.20 ± 0.31 1.3 ± 0.2 Heptanoate 5.8 ± 0.9 9.12 ± 0.43 1.6 ± 0.2 Octanoate 6.3 ± 2 3.04 ± 0.30 0.48 ± 0.16 Acetate 0.042 ± 0.008 0.99 ± 0.02 24 ± 5 Propionate 0.62 ± 0.23 1.1 ± 0.1 1.8 ± 0.8 Thr324Val Isobutyrate 2.1 ± 0.8 0.33 ± 0.05 0.15 ± 0.08 3-Methylvalerate 125 ± 416 3.9 ± 1 0.03 ± 0.2 Acetate 0.18 ± 0.04 1.91 ± 0.08 10.4 ± 2.9 Propionate 3.2 ± 1.1 1.78 ± 0.26 0.56 ± 0.28 Butyrate 0.19 ± 0.09 0.13 ± 0.01 0.68 ± 0.34 Thr324Gly Isobutyrate 0.15 ± 0.30 0.23 ± 0.08 1.6 ± 2.4 3-Methylvalerate 0.04 ± 0.06 0.077 ± 0.009 1.9 ± 2.9 4-Methylvalerate 0.07 ± 0.05 0.17 ± 0.01 2.4 ± 2.1 Wild-type* Acetate 0.12 ± 0.04 5.7 ± 0.3 48 ± 7 Acetate 7.8 ± 17 0.6 ± 0.7 0.07 ± 0.2 Propionate 98 ± 45 3.3 ± 0.9 0.03 ± 0.02 Butyrate 20 ± 2 6.7 ± 0.2 0.33 ± 0.04 324 Thr Gly, Valerate 0.4 ± 0.1 8.8 ± 0.5 22 ± 6 Val399Ala, Trp427Gly* Hexanoate 0.13 ± 0.05 5.7 ± 0.4 43 ± 18 Heptanoate 0.71 ± 0.45 11.0 ± 1.9 15 ± 9.8 Octanoate 2.1 ± 0.7 5.2 ± 0.6 2.4 ± 0.9 Isobutyrate 219 ± 69 4.0 ± 0.9 0.018 ± 0.007 46

Table S1 contd 3-Methylvalerate 2.7 ± 0.6 7.7 ± 0.6 2.9 ± 0.6 Thr324Gly, Val399Ala, 4-Methylvalerate 0.22 ± 0.07 6.6 ± 0.4 29 ± 10 427 4- Trp Gly* 0.25 ± 0.05 5.9 ± 0.2 2 ± 4 Methylhexanoate Propionate 98 ± 42 3.3 ± 0.8 0.03 ±0.02 Valerate 2.3 ± 0.7 2.7 ± 0.3 1.20 ±0.38 Hexanoate 0.58 ± 0.05 2.67 ± 0.06 4.62 ±0.40 Heptanoate 1.05 ± 0.13 2.22 ± 0.08 2.11 ±0.27 Ile323Ala, Val399Ala, Octanoate 4.1 ± 2.4 6.1 ± 1.6 1.47 ±0.95 Trp427Gly* Isobutyrate 110 ± 42 3.2 ± 0.8 0.03 ±0.01 3-Methylvalerate 1.5 ± 0.4 1.9 ± 0.2 1.22 ±0.37 4-Methylvalerate 4.3 ± 0.5 6.7 ± 0.3 1.57 ±0.19 4- 1.0 ± 0.3 1.5 ± 0.1 1.6 ±0.51 Methylhexanoate Butyrate 4.5 ± 0.9 3.8 ± 0.4 0.9 ± 0.2 Valerate 2.1 ± 1.8 2.2 ± 0.7 1.0 ± 1.0 Ile323Ala, Hexanoate 7.8 ± 1.0 4.8 ± 0.4 0.62 ± 0.09 Val399Gly, Heptanoate 5.0 ± 1.3 5.4 ± 0.7 1.1 ± 0.3 Trp427Gly* Octanoate 2.4 ± 0.7 4.5 ± 0.5 1.9 ± 0.5 3-Methylvalerate 3.01 ± 0.4 2.5 ± 0.1 0.8 ± 0.1 4-Methylvalerate 1.2 ± 0.4 3.0 ± 0.3 2.4 ± 0.8 Thr324Ser, Butyrate 16 ± 5 19 ± 4 1.2 ± 0.5 Val399Ala, Valerate 1.5 ± 0.5 7.2 ± 0.9 4.9 ± 1.8 Trp427Gly* 4-Methylvalerate 1.2 ± 0.3 8.2 ± 0.6 6.7 ± 1.6 Acetate 1.9 ± 0.4 11.0 ± 0.8 6.0 ± 1.2 Trp427Ala* Propionate 23 ± 11 21 ± 7 0.9 ± 0.5 Butyrate 5.1 ± 0.7 10.8 ± 0.8 2.1 ± 0.3 Acetate 8.9 ± 3.8 7.2 ±1.8 0.81 ± 0.4 Trp427Ser* Isobutyrate 1.6 ± 1.4 1.5 ±0.5 1.0 ± 0.9 Trp427His* Acetate 14 ± 11 15 ± 7 1.1 ± 0.9 Acetate 10 ± 2 10.6 ± 1.4 1.0 ± 0.3 Trp427Val* Propionate 38 ± 27 2.5 ± 1.5 0.06 ± 0.06 Trp427Phe* Acetate 1.2 ±0.4 11.7 ± 1.1 15.2 ± 4.7 The variants marked with asterisk (*) were extracted with the purification protocol using

HEPES buffer, while all other enzymes were purified with Tris-HCl buffer (see Materials and

Methods). All enzyme variants were assayed with all 11 carboxylate substrates listed in the

Material and Methods, but data are only presented with the carboxylate substrates that exhibited activity. 47

CHAPTER 3. UNDERSTANDING EFFECTS OF ACETYLATION IN

REGULATING ARABIDOPSIS ACETYL COA SYNTHETASE

This chapter has been written in a format for submission to the journal Archives of

Biochemistry and Biophysics

Naazneen Sofeo1,2,3, Basil J. Nikolau1,2,3,4

Abstract

Acetyl-CoA synthetase (ACS) is known to be regulated by the post translational modification of reversible acetylation. This post translational modification is one of the most studied modifications. In organisms like bacteria and yeast, it has been shown that this modification leads to reduction of activity of the enzyme. A conserved lysine residue in the

A10 motif of homologous ACS has been shown to be the one involved in this process of reversible acetylation. In this study we have analyzed the effect of acetylation on Arabidopsis

ACS activity and the conserved lysine involved in this modification. According to our knowledge, this is the first report of the effect acetylation has on a plant ACS using in vitro studies.

1 Roy J. Carver Department of Biochemistry, Biophysics and Molecular Biology, Iowa

State University, Ames, IA 50011, USA, 2 Engineering Research Center for Biorenewable

Chemicals Iowa State University, Ames, IA 50011, USA, 3 Center for Metabolic Biology, Iowa

State University, Ames, IA 50011, USA,4 Corresponding Author

Introduction

The interplay between acetylation and deacetylation is a crucial process in regulating many biological processes including chromatin structure and transcriptional regulation and the regulation of cellular metabolism [1–3]. This post-translational modification is evolutionarily conserved over a range of phyla, and two enzymes play a role in this process: a) lysine acetyl transferases (KATs) that acetylate lysine residues; and b) lysine deacetylases (KDACs) that hydrolyze and remove the acetyl group from this side chain [1]. In addition, non-enzymatic auto-acetylation of lysine residues in proteins occurs at alkaline pH, particularly in sequences where a lysine residue is flanked by positively charged amino acids [4].

Acetyl-CoenzymeA (CoA) synthetase (ACS) is a ~72 kDa protein, which plays an important role in fatty acid and polyketide biosynthetic pathways. In a wide variety of organisms ACS activity is known to be regulated by reversible post-translational acetylation.

Specifically, the ACS from Mycobacterium tuberculosis has been shown to be auto-acetylated using an acetate containing buffer at alkaline pH [5]. The enzymatic acetylation of a lysine residue as a mechanism of ACS regulation has been extensively studied in bacterial systems

[3,6,7]. For example, the non-acetylated ACS of Salmonella enterica is about 480-fold more active than the acetylated enzyme [8]. The deacetylation of this enzyme is catalyzed by CobB, which is a homolog of the yeast Sir2 deacetylase [9]. The acetylation of this ACS occurs on the side chain of a Lys residue that is positioned in the middle of a conserved motif called A10.

Structural studies of the S. enterica ACS indicate that acetylation does not induce any conformational changes in the enzyme [10]. However, acetylation significantly affects the first half-reaction of the ACS-catalyzed reaction, presumably by aligning the acetate-substrate with 49 the active site residue(s) [10]; this acetylation reaction does not affect the second half reaction of the ACS-catalyzed reaction [8].

In plant systems, such as Arabidopsis, the metabolic function of ACS is not completely clear. Unlike other biological systems, ACS is not the physiological source of acetyl-CoA that is the precursor of de novo fatty acid biosynthesis in plastids [11]; this is in spite of the fact that ACS is plastid localized [12]. In this study, we explored if the acetylation-based regulation of ACS is extrapolatable from bacteria, yeast and algae to affect activity of this enzyme in plant systems.

Materials & Methods

Generation of atACS variants

The atACS ORF was cloned into the BamHI restriction site of the pET30f vector [13]. The pET30f vector is a modified pET30 vector (Takara, Clontech, Mountain View, CA) that will fuse an N- terminal poly-histidine coding sequence with a TEV cleavage site, to the atACS ORF sequence. Mutagenesis of targeted residues in the atACS sequence was performed using the

QuikChange Lightning site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA).

Sequence confirmed pET30f constructs were subsequently transformed into E. coli

ArcticExpress (DE3) cells (Agilent Technologies, Santa Clara, CA) for protein expression.

Expression and purification of wild-type and variant proteins

Arctic Express strains harboring plasmids that express either wild-type atACS or variants were grown overnight at 37 °C with shaking at 250 rpm for aeration, in 5-10 ml LB medium containing the antibiotics kanamycin (50 μg/ml) and gentamycin (20 μg/ml). Arabidopsis 50

SRT1 (AT5G55760) and SRT2 (AT5G09230) ORFs [14,15] were cloned into the BamHI restriction site of the pET30f vector. The sequence encoding the 50 amino acid long signal peptide was removed from the SRT2 construct. Sequence confirmed pET30f constructs were subsequently transformed into E. coli ArcticExpress (DE3) cells (Agilent Technologies, Santa

Clara, CA) for protein expression. Precultures of these strains were grown overnight at 37 °C with shaking at 250 rpm for aeration, in 10 ml LB medium containing the antibiotics kanamycin (50 μg/ml) and gentamycin (20 μg/ml). The overnight preculture was used to inoculate 0.5-1.0 L LB media containing no antibiotics, and these cultures were grown at 30

°C with shaking at 250 rpm, until an OD600 of ~0.6. Expression was then induced by the addition of IPTG to a final concentration of 0.1 mM, and the culture was grown at 13 °C with shaking at 250 rpm for 24-48 hours.

Cells were harvested by centrifugation at 10,000g for 10 minutes, and for atACS variants resuspended in a buffer containing 20 mM HEPES-KOH, pH 7.5, 0.5 M NaCl, 5 mM imidazole, 0.1% (v/v) Triton X-100, 0.1 mg/ml phenyl methyl sulfonyl fluoride (PMSF), and

10 μl/ml Protease Inhibitor Cocktail (Sigma-Aldrich Co., St. Louis, MO). The buffer used for SRT1 and SRT2 was 20 mM sodium phosphate buffer, pH 8.0, 0.5 M NaCl, 1 mM DTT,

5 mM imidazole, 0.1% (v/v) Triton X-100, 0.1 mg/ml PMSF, and 10 μl/ml Protease Inhibitor

Cocktail (Sigma-Aldrich Co., St. Louis, MO). Cells were disrupted by sonication on ice, and all subsequent steps were conducted at 4 °C. The cell extract was centrifuged at 20,000xg for

30 minutes, and the supernatant was retained, filtered through a 0.45 μm filter (Corning Inc.,

Corning, NY) and applied to a column containing 2 ml PerfectPro Ni-NTA agarose (5 Prime,

Inc., Gaithersburg, MD). The column was washed with increasing concentrations of imidazole, and the atACS, SRT1 or SRT2 proteins were eluted with 0.2 M imidazole. atACS 51 samples were immediately dialyzed into 10 mM HEPES-KOH, pH 7.5, 10 mM KCl, 2mM

TCEP, 10% glycerol. SRT1 and SRT2 proteins were immediately dialyzed into 10 mM sodium phosphate buffer, pH 8.0, 0.5 M NaCl, 1 mM DTT and 10% glycerol.

Gel Filtration Chromatography

Size exclusion gel filtration chromatography was conducted with an AKTA FPLC system

(GE Healthcare Life Sciences, Pittsburg, PA). A 100 µl aliquot of purified atACS protein (8-

10 mg/ml) was injected into a prepacked Superdex 200 Increase 10/300 GL gel filtration column (GE Healthcare Life Sciences, Pittsburg, PA). The buffer used for atACS variants was 10 mM HEPES-KOH, pH 7.5, 10 mM KCl, 2 mM TCEP, 10% glycerol. For SRT1 and

SRT2 the buffer was 10 mM sodium phosphate buffer, pH 8.0, 0.5 M NaCl, 1 mM DTT and

10% glycerol. Elution was with a flow rate of 0.4 ml/minute, and protein elution was monitored using a UV absorbance detector, at 280nm.

Acetylation and deacetylation of atACS

The non-enzymatic, autoacetylation of ACS (~130 µg 0f purified protein) was conducted in a

1-mL volume of acetylation buffer consisting of 10 mM potassium acetate, 10 mM MgCl2, 10 mM ATP, 50 mM Tris-HCl, pH 8.0 [5]. Following incubation at 37°C for 2 hours, the protein solution was dialyzed into 10 mM HEPES, pH 7.5, 10 mM KCl, 2 mM TCEP, 10% glycerol.

The ability of SRT1 or SRT2 proteins to catalyze the deacetylation of acetylated ACS was assayed in separate reactions. Acetylated ACS (130 µg of protein) was incubated overnight at 25 oC, with either purified SRT1 (25 µg) or purified SRT2 (25 µg) proteins in 10- mL of deacetylation buffer consisting of 137 mM NaCl, 2.7 mM KCl, 1 mM MgCl2, 400 uM

NAD+, 50 mM Tris-HCl, pH 8.5 [5]. 52

Protein analysis

Isolated protein preparations were evaluated by SDS PAGE, and the acetylation status of ACS was evaluated by Western blot analysis, using an antibody (1:1000 diluted) that reacts with N-ε-acetyl-L-lysine (Cell Signaling Technology, #9441).

The acetylation status of ACS was also evaluated by mass spectrometric analysis of rLysC (Promega Corporation, Madison, WI) digested proteins using Q Exactive™ Hybrid

Quadrupole-Orbitrap Mass Spectrometer (Thermo Fisher Scientific Inc., Waltham, MA), at the Iowa State University Protein Facility (http://www.protein.iastate.edu). The purified protein preparations of auto-acetylated ACS and wild-type ACS were subjected to SDS-PAGE analysis. Following staining with Coomassie Brilliant Blue, the identified protein band was excised and digested using an Investigator™ ProGest (Genomic Solutions, Digilab Inc,

Hopkinton, MA) in 0.5 mL buffer (50 mM Tris, pH 8) with LysC at a concentration of 2 µg/100

µg sample in the presence of 5 mM DTT and 15 mM iodoacetamide. The results were analyzed using SequestHT against the ACS protein sequence, with four possible missed cleavage sites.

The possible modifications analyzed were acetylation of Lys, carbamidomethylation of Cys, deamidation of Asn and Gln and oxidation of Met residues.

Mass-spectrometric analysis of isolated intact proteins was also conducted using a

Waters SYNAPT G2-Si High Definition Mass Spectrometer, at the Iowa State University

Protein Facility (http://www.protein.iastate.edu).

Spectrophotometric atACS Activity Assay

ACS activity was measured by coupling the acetate-dependent formation of AMP from ATP to the oxidation of NADH, using the enzymes myokinase, pyruvate kinase and lactate dehydrogenase [16] (Sigma-Aldrich Co., St. Louis, MO). The oxidation of NADH was 53

monitored by the rate of decrease in absorbance at 340 nm. The kcat of each atACS enzyme variant was determined by assaying each enzyme with different acetate concentrations, ranging between 0 and 5 mM. Specific activity of atACS was determined by using 5 mM acetate.

Assays were performed in 96-well micro-titer dishes at 37°C with a final well-volume of 100

μl. The change in A340 of each well was monitored with a BioTek ELx808TM Absorbance

Microplate Reader, using Gen5TM Data Analysis software (BioTek Instruments, Winooski,

VT). The standard reaction contained 50 mM Tris-HCl, pH 7.5, and 5mM Tris(2- carboxylethyl) phosphine hydrochloride (Thermo Fisher Scientific Inc., Waltham, MA); 5%

(v/v) ethanol; 5 mM ATP; 6mM MgCl2; 5 mM phospho(enol)pyruvate; 0.4 mM NADH; 2-20

U pyruvate kinase/lactate dehydrogenase mix (Sigma-Aldrich Co., St. Louis, MO); 10U of rabbit muscle myokinase or 1U chicken muscle myokinase (Sigma-Aldrich Co., St. Louis,

MO); and 5 μg of purified recombinant atACS protein.

Circular Dichroism Spectroscopy

CD spectra were obtained at the Iowa State University Chemical Instrumentation Facility

(https://www.cif.iastate.edu) using a Jasco J-710 CD (JASCO Analytical Instruments, Easton,

MD). Spectra were taken at wavelength range of between 190 nm and 260 nm, with the following settings parameter: Pitch point 1, Speed 20 nm/min, Response time 4 sec, Bandwidth

1 nm, Temperature 20 °C and Data mode CD-HT. Spectra where acquired with a 1 mm dichroically neutral quartz cuvette and the average of 3 scans was used for analysis. For acquisition of spectra, protein samples in 10 mM HEPES-KOH, pH 7.5, 10 mM KCl, 2 mM

TCEP and 10% glycerol were diluted into double distilled H2O to a final concentration of 0.1-

0.2 mg/ml. A baseline spectrum for the same dilution of the buffer was obtained and subtracted from the experimental protein spectra. After data collection, spectra were converted from 54

2 -1 millidegrees to molar ellipticity ([θ], degrees cm dmol ). Spectra were analyzed using a suite of algorithms collectively called CDPro [17], utilizing algorithms: SELCON3, CDSSTR, and

CONTIN/LL.

Results

Comparing the Arabidopsis ACS sequence to ACS sequences from a range of diverse biological phyla identifies 10 conserved motifs (A1 to A10) that are scattered evenly throughout the length of these sequences. Three of these motifs A3, A4 and A5, contain the residues that determine the carboxylate substrate specificity of this class of enzymes

[10,18,19], and the A10 motif contains a highly conserved lysine residue (at position 622 of the Arabidopsis enzyme) (Figure 1), which appears to be susceptible to reversible acetylation a mechanism by which many of these enzymes are regulated [8].

Several different strategies were used in this study to evaluate and deduce that Lys622 residue was the target of the acetylation in the Arabidopsis ACS. These experiments focused on the characterization of mutant variants, Lys622Ala, Lys622Gln and Lys622Arg. SDS PAGE analysis of the purified preparations of these variant proteins indicate that high-levels of expression were achieved, recovering 10-15 mg of purified protein/L of culture (Figure 2A).

Size exclusion gel filtration chromatography of the purified protein variants show that all proteins elute as a single symmetrical peak, at elution volumes consistent with the molecular weight of a monomer (Figure 2B). CD spectral analysis of these preparations show that all proteins are similarly folded as the wild-type enzyme (Figure 2C), and secondary structure composition calculations indicate no significant differences among the proportion of the 55 proteins that acquire alpha helices, beta sheet, turns and unordered secondary structures (Figure

2D).

Enzymatic assays of the three Lys622 variants demonstrate that compared to the wild- type enzyme, catalytic competence of these mutants is reduced by more than 30-fold.

622 Specifically, the kcat for the Lys Arg mutant is reduced by 30-fold (increasing Km by 25-fold), while the catalytic activity of the Lys622Ala and Lys622Gln mutants is even further reduced to such levels that catalytic constants could not be accurately determined (Figure 3A & B).

The FPLC purified wild-type and Lys622 mutant variants of ACS were chemically autoacetylated in vitro, and their acetylation status was determined by western blot analysis using the anti-acetyl-lysine antibody. While the wild-type enzyme is acetylated, all the Lys622 mutants are not acetylated (Figure 3C). Furthermore, mass-spectrometric analysis of the wild- type protein and the acetylated protein is consistent with the acetylation of the Lys622.

Namely, the predicted theoretical mass value of the wild-type protein is predicted to be

75876.000 Da. This value is increased by ~42 Da following acetylation of the wild-type ACS, consistent with the acetylation-modification of the protein at a single site. In addition, the mass-spectrometric analysis of the Lys622Ala and Lys622Gln variants are within 1.000 Da from the predicted changes associated with the individual mutations; compare 75876.000 Da for the wild-type enzyme with 75819.000 for the Lys622Ala mutation, and 75876.000 for the Lys622Gln mutation (Figure 4). Therefore, the western blot and mass spectrometric data indicate that in the Arabidopsis ACS Lys622 is the site of acetylation, as this acetylation is absent when the

Lys622 residue is mutated, and the mass changes of these mutants are consistent with the

Lys622Ala and the Lys622Gln mutations. 56

Direct demonstration that the auto-acetylation of ACS protein is targeted to residue

Lys622 was obtained by mass-spectrometric analysis of the LysC digested enzyme. In these analyses, the sequence-coverage of wild-type protein was 40.5% (Figure 5A), while for the acetylated ACS it is ~62% (Figure 5B). Although the non-acetylated LysC-generated peptide that contains Lys622 could not be identified from the wild-type protein, the acetylated form of this peptide was successfully detected in acetylated ACS. Specifically, the peptide

TRSGKAcIMRRILRK was identified with monoisotopic m/z value of 415.01083 Da (+1.22 mmu/+2.95 ppm), charge of +4 and theoretical mass MH+ ion of 1657.02151 Da using Sequest

HT (v1.17) (Figure 6).

Using these techniques, we also evaluated if the acetylated ACS protein could be deacetylated by one of two purified Arabidopsis sirtuin deacetylases, SRT1 and SRT2 [14,15].

Following acetylation, the purified ACS was incubated in vitro with purified SRT1 or SRT2.

Comparing the acetylation status of ACS by Western blot analysis with anti-acetyl-lysine antibody before and after this treatment, indicates that acetylated ACS is not a substrate for these two sirtuin deacetylases (Figure 7A).

The purified acetylated and non-acetylated Arabidopsis ACS were assayed for catalytic activity. The acetylation status of these preparations was confirmed by Western blot analysis with anti-acetyl-lysine antibody (Figure 7A & 8A), and by mass spectrometric analysis (Figure

5B). The specific activity of these two ACS preparations was compared. Figure 5B shows that there is no significant difference in the specific activity of the acetylated and non-acetylated enzyme. The Michaelis-Menten kinetic parameters (Km, kcat and kcat/Km) of these two forms of

ACS are also not affected by acetylation (Figure 8B & C). In parallel, we also compared the specific activity of acetylated ACS after treatment with SRT1 or SRT2. Statistical analysis 57

(using student t-test, at p-values <0.05) indicates that SRT1 treatment did not affect ACS activity, but a small, but significant decrease in specific activity was detected upon treatment with SRT2 (Figure 7B).

Thus, collectively these data establish that although acetylation of Lys622 occurs as with

ACS enzymes from a variety of organisms, this event is not a significant modulator of the

Arabidopsis ACS catalytic activity.

Discussion

Lysine acetylation-deacetylation cycles are important mechanisms for the post- translational regulation of a wide range of biological processes. These regulatory cycles occur over a wide range of biological phyla. These acetylation modifications can also occur on serine or threonine side-chains [20], but the focus of this discussion is on the acetylation of amine groups, and can occur at either at the N-terminus of proteins, or at the Nε -amine of lysine side chains [1–3]. In plants, 100s of Nε-amine lysine sites have been identified [21], and these post- translational modification effects occur on both histone and non-histone proteins; the latter are involved in a variety of metabolic processes including photorespiration, redox regulation, TCA cycle, and amino acid and protein metabolism [22]. These acetylation reactions mask the positive charge of lysine side chains, and thus potentially affect the nature of the interactions within the tertiary structure of the modified protein [4,23].

Previous studies of ACSs have shown reversible acetylation to be an important regulatory mechanism for modulating the activity of this enzyme. In diverse phyla, the acetylation of ACS inactivates the enzyme [6–8,24]. In contrast, the acetylation of a long chain acyl CoA synthetase in the diatom Phaeodactylum tricornutum appears to lead to enzymatic 58 activation [25]. Structural studies with members of the ACS superfamily have shown that the side chain of the Lys residue that is targeted for acetylation, electrostatically interacts with the

α-carboxylate group of the substrate. In the first half of the reaction, this lysine residue is projected into the active site even though it is located in the C terminal domain. In addition, this lysine residue hydrogen bonds with O-4’ and O-5’ of the ribose moiety of the ATP substrate [27]. Thus, the acetylation of the lysine residue can disrupt these interactions, which is a possible explanation for the loss of enzymatic activity upon this modification.

In this study we identified the acetylation site for the plant ACS as occurring at Lys622.

This identity is consistent with the position of this residue in a motif (A10) that is highly conserved among many ACS enzymes, and is the homologous residue that has been shown to be the target of acetylation in a bacterial ACS (i.e., Cryptococcus neoformans: PDB 5VPV).

The sequence homology-based identification of Lys622 as the site of acetylation of the plant enzyme was experimentally confirmed by a combination of site directed mutagenesis and mass-spectrometric analyses. Moreover, it appears that in vitro the acetylation of the plant

ACS is autocatalytic and does not require additional enzymatic components. This is similar to the M. tuberculosis ACS [5]. Finally, prior studies have established that this conserved Lys residue has an important role in catalysis by ACS, particularly in the first half reaction catalyzed by this enzyme [10]. Although this important functionality is also conserved in the plant ACS, the acetylation of this residue does not significantly affect catalysis.

References

[1] A. Drazic, L.M. Myklebust, R. Ree, T. Arnesen, The world of protein acetylation, Biochim. Biophys. Acta - Proteins Proteomics. 1864 (2016) 1372–1401. doi:10.1016/j.bbapap.2016.06.007. 59

[2] E. Verdin, M. Ott, 50 years of protein acetylation: from gene regulation to epigenetics, metabolism and beyond, Nat. Rev. Mol. Cell Biol. 16 (2015) 258–264. doi:10.1038/nrm3931.

[3] V. Bernal, S. Castaño-Cerezo, J. Gallego-Jara, A. Écija-Conesa, T. de Diego, J.L. Iborra, M. Cánovas, Regulation of bacterial physiology by lysine acetylation of proteins, N. Biotechnol. 31 (2014) 586–595. doi:10.1016/j.nbt.2014.03.002.

[4] T. Narita, B.T. Weinert, C. Choudhary, Functions and mechanisms of non-histone protein acetylation, Nat. Rev. Mol. Cell Biol. (2018) 1. doi:10.1038/s41580-018-0081- 3.

[5] R. Li, J. Gu, P. Chen, Z. Zhang, J. Deng, X. Zhang, Purification and characterization of the acetyl-CoA synthetase from Mycobacterium tuberculosis, Acta Biochim. Biophys. Sin. (Shanghai). 43 (2011) 891–899. doi:10.1093/abbs/gmr076.

[6] T. Noy, H. Xu, J.S. Blanchard, Acetylation of acetyl-CoA synthetase from Mycobacterium tuberculosis leads to specific inactivation of the adenylation reaction, Arch. Biochem. Biophys. 550–551 (2014) 42–49. doi:10.1016/j.abb.2014.04.004.

[7] J.G. Gardner, F.J. Grundy, T.M. Henkin, J.C. Escalante-Semerena, Control of acetyl- coenzyme A synthetase (AcsA) activity by acetylation/deacetylation without NAD+involvement in Bacillus subtilis, J. Bacteriol. 188 (2006) 6715. doi:10.1128/JB.01181-06.

[8] V.J. Starai, I. Celic, R.N. Cole, J.D. Boeke, J.C. Escalante-Semerena, Sir2-dependent activation of acetyl-CoA synthetase by deacetylation of active lysine, Science (80-. ). 298 (2002) 2390–2392. doi:10.1126/science.1077650.

[9] A.W. Tsang, J.C. Escalante-Semerenat, CobB, a new member of the SIR2 family of eucaryotic regulatory proteins, is required to compensate for the lack of nicotinate mononucleotide:5,6- dimethylbenzimidazole phosphoribosyltransferase activity in cobT mutants during cobalamin biosynthesis in Sal, J. Biol. Chem. 273 (1999) 31788– 31794. doi:10.1074/jbc.273.48.31788.

[10] A.M. Gulick, V.J. Starai, A.R. Horswill, K.M. Homick, J.C. Escalante-Semerena, The 1.75 Å crystal structure of acetyl-CoA synthetase bound to adenosine-5???- propylphosphate and coenzyme A, Biochemistry. 42 (2003) 2866–2873. doi:10.1021/bi0271603.

[11] J. Ke, R.H. Behal, S.L. Back, B.J. Nikolau, E.S. Wurtele, D.J. Oliver, The Role of Pyruvate Dehydrogenase and Acetyl-Coenzyme A Synthetase in Fatty Acid Synthesis in Developing Arabidopsis Seeds, Plant Physiol. 123 (2000) 497–508. doi:10.1104/pp.123.2.497.

[12] M. Lin, D.J. Oliver, The Role of Acetyl-Coenzyme A Synthetase in Arabidopsis, PLANT Physiol. 147 (2008) 1822–1829. doi:10.1104/pp.108.121269.

[13] F. Jing, D.C. Cantu, J. Tvaruzkova, J.P. Chipman, B.J. Nikolau, M.D. Yandeau- Nelson, P.J. Reilly, Phylogenetic and experimental characterization of an acyl-ACP thioesterase family reveals significant diversity in enzymatic specificity and activity, BMC Biochem. 12 (2011) 44. doi:10.1186/1471-2091-12-44.

[14] A.-C. Konig, M. Hartl, P.A. Pham, M. Laxa, P.J. Boersema, A. Orwat, I. Kalitventseva, M. Plochinger, H.-P. Braun, D. Leister, M. Mann, A. Wachter, A.R. Fernie, I. Finkemeier, The Arabidopsis Class II Sirtuin Is a Lysine Deacetylase and Interacts with Mitochondrial Energy Metabolism, PLANT Physiol. 164 (2014) 1401– 1414. doi:10.1104/pp.113.232496.

[15] F. Zhang, L. Wang, E.E. Ko, K. Shao, H. Qiao, Histone Deacetylases SRT1 and SRT2 Interact with ENAP1 to Mediate Ethylene-Induced Transcriptional Repression, Plant Cell. 30 (2018) 153–166. doi:10.1105/tpc.17.00671.

[16] A.R. Horswill, J.C. Escalante-Semerena, The prpE gene of Salmonella typhimurium LT2 encodes propionyl-CoA synthetase, Microbiology. 145 (1999) 1381–1388. doi:10.1099/13500872-145-6-1381.

[17] N. Sreerama, R.W. Woody, Estimation of Protein Secondary Structure from Circular Dichroism Spectra: Comparison of CONTIN, SELCON, and CDSSTR Methods with an Expanded Reference Set, Anal. Biochem. 287 (2000) 252–260. doi:10.1006/abio.2000.4880.

[18] V.J. Starai, J.C. Escalante-Semerena, Acetyl-coenzyme A synthetase (AMP forming)., Cell. Mol. Life Sci. 61 (2004) 2020–30. doi:10.1007/s00018-004-3448-x.

[19] M.A. Marahiel, T. Stachelhaus, H.D. Mootz, Modular Peptide Synthetases Involved in Nonribosomal Peptide Synthesis, Chem. Rev. 97 (1997) 2651–2674. doi:10.1021/cr960029e.

[20] S. Mukherjee, Y.-H. Hao, K. Orth, A newly discovered post-translational modification--the acetylation of serine and threonine residues., Trends Biochem. Sci. 32 (2007) 210–6. doi:10.1016/j.tibs.2007.03.007.

[21] R.S.P. Rao, J.J. Thelen, J.A. Miernyk, In silico analysis of protein Lys-N(?)- acetylation in plants, Front. Plant Sci. 5 (2014) 381. doi:10.3389/fpls.2014.00381.

[22] A.C. König, M. Hartl, P.J. Boersema, M. Mann, I. Finkemeier, The mitochondrial lysine acetylome of Arabidopsis, Mitochondrion. 19 (2014) 252–260. doi:10.1016/j.mito.2014.03.004.

[23] A.M. James, K. Hoogewijs, A. Logan, A.R. Hall, S. Ding, I.M. Fearnley, M.P. Murphy, Non-enzymatic N-acetylation of Lysine Residues by AcetylCoA Often Occurs via a Proximal S-acetylated Thiol Intermediate Sensitive to Glyoxalase II., Cell Rep. 18 (2017) 2105–2112. doi:10.1016/j.celrep.2017.02.018.

[24] W.C. Hallows, S. Lee, J.M. Denu, Sirtuins deacetylate and activate mammalian acetyl- CoA synthetases, Proc. Natl. Acad. Sci. 103 (2006) 10230–10235. doi:10.1073/pnas.0604392103.

[25] Z. Chen, L. Luo, R. Chen, H. Hu, Y. Pan, H. Jiang, X. Wan, H. Jin, Y. Gong, Acetylome Profiling Reveals Extensive Lysine Acetylation of the Fatty Acid Metabolism Pathway in the Diatom Phaeodactylum tricornutum., Mol. Cell. Proteomics. 17 (2018) 399–412. doi:10.1074/mcp.RA117.000339.

[26] Q. Wang, Y. Zhang, C. Yang, H. Xiong, Y. Lin, J. Yao, H. Li, L. Xie, W. Zhao, Y. Yao, Z.-B. Ning, R. Zeng, Y. Xiong, K.-L. Guan, S. Zhao, G.-P. Zhao, Acetylation of metabolic enzymes coordinates carbon source utilization and metabolic flux., Science. 327 (2010) 1004–7. doi:10.1126/science.1179687.

[27] E. Conti, T. Stachelhaus, M.A. Marahiel, P. Brick, Structural basis for the activation of phenylalanine in the non-ribosomal biosynthesis of gramicidin S., EMBO J. 16 (1997) 4174–83. http://www.ncbi.nlm.nih.gov/pubmed/9250661 (accessed February 10, 2019).

Figures

Figure 1: Comparison of the amino acid sequences of AMP-forming family of acyl-CoA synthetases from different organisms that utilize acetate (ACS) or propionate (PCS) as substrates. Alignment was performed using Clustal Omega. Asterisk indicate the conserved lysine residue proposed to be reversibly acetylated. White letters on a black background indicate identity, white letters on a gray background indicate similarity, and black letters on a white background indicate no conservation.

Figure 2: Characterization of ACS Lys622 mutants. A. SDS-PAGE analysis of purified wild- type and indicated Lys622 mutant variants. B. Size exclusion chromatography (FPLC) of wild- type and Lys622 mutant variants of ACS. C. CD spectra of wild-type and Lys622 mutant variants of ACS. Measurements are an average of three replicate scans. D. Composition of secondary structure composition calculated from the CD spectra of wild-type and Lys622 mutant variants of ACS. Errors bars indicate standard errors from 3 replicate CD spectra calculated by 3 algorithms, SELCON3, CDSSTR, and CONTIN/LL. 64

Figure 3: A. Michaelis-Menten kinetic analysis of ACS Lys622 variants. B. Expanded view to demonstrate that the Lys622 variants have very low activity, as compared to the wild-type enzyme. C. Acetylation of ACS Lys622 variants. Western blot analysis with anti-acetyl-lysine antibody of acetylated wild-type ACS, and the indicated ACS Lys622 variants. The top panel is the image from western blot analysis and bottom panel shows Ponceau Staining of the same membrane.

Figure 4: Mass spectrometric analysis of intact acetylated ACS. A. Acetylated wild-type ACS; major peak at mass 75918. B. Acetylated ACS Lys622Ala variant; major peak at mass 75818.

C. Acetylated ACS Lys622Gln variant; major peak at mass 75875.

Figure 5: Mass-spectrometric coverage map of LysC digested wild-type (A) and acetylated

(B) ACS showing detected peptides and modifications on specific amino acid. The green shade identifies sequences of detected peptides. The modifications are indicated as A: acetylation; C: carbamidomethylation; D: deamidation; O: oxidation. Lysine 622 is identified with an asterisk

(*).

Figure 6: LC-MS/MS spectra and MS2 fragmentation pattern (Inset) of the peptide containing lysine 622, confirming acetylation.

Figure 7: Activity and acetylation levels of Arabidopsis acetyl-CoA synthetase. A. Western blot analysis with anti-acetyl-lysine antibody of non-acetylated and acetylated ACS, and acetylated ACS after incubation with deacetylases, SRT1 or SRT2. The top panel is the image from western blot analysis and bottom panel shows Ponceau S staining of the same membrane.

B. Specific activity of non-acetylated and acetylated ACS, and acetylated ACS after treatment with deacetylases, SRT1 or SRT2. Asterisk indicates statistically significant difference from acetylated ACS specific activity (p-value < 0.05 for student t-test).

Figure 8: Comparison of acetylated and non-acetylated Arabidopsis thaliana acetyl-CoA synthetase. A. Western blot analysis of purified acetylated and non-acetylated ACS with anti- acetyl-lysine antibody. The top panel is the image from the western blot analysis and bottom panel shows the Ponceau S staining of the same membrane. B. Substrate dependence of acetylated and non-acetylated ACS activity. C. Kinetic parameters Km, kcat and catalytic efficiency (kcat/Km) of non-acetylated and acetylated ACS.

CHAPTER 4. SIRTUIN-MEDIATED REGULATION OF PLANT ACETYLOME

AND METABOLOME

Abstract

The post translational, reversible modification by acetylation of lysine residues is widespread across all biological clades. Among the deacetylases involved in removal of this acetyl group, this study characterized the two sirtuin-family deacetylases in Arabidopsis, SRT1 and SRT2.

In vitro characterizations of recombinantly expressed purified SRT1 and SRT2 proteins establish that they are sirtuin deacetylases. A comparative integrated biochemical-genetic study of the acetylome and metabolome of srt1 or srt2 single mutants and srt1srt2 double mutant plants indicate that although there are undetectable effects on the growth phenotype of the plants, there are considerable changes in the molecular features that were assayed. The changes induced by the double mutant indicate that these proteins deacetylate both histone and non-histone proteins. In addition, the mutant-induced changes in the metabolome indicate that the two SRT homologs are involved in regulating different metabolic processes, generating distinct metabolomes. The combined datasets can be used to identify individual lysine acetylation sites to mechanistically understand the role of each acetylation-deacetylation cycle in regulating metabolic processes.

Introduction

The post-translational modification of the lysine side chain amine by acetylation is a reversible process that affects a wide range of processes in both prokaryotes and eukaryotes

[1]. Two types of enzymes catalyze this reversible process, lysine acetyl transferases (KATs) that acetylate the protein side chain, and lysine deacetylases (KDACs) that remove the acetyl group [2]. In plants, 100s of lysine acetylation sites have been identified [3], and these include 71 histone proteins that modulate nucleosome structure and thus regulate transcription, and the acetylation of non-histone proteins that affect other biological processes [4] are also acetylated.

In plants, typified by Arabidopsis, three types of KDACs have been characterized [5].

Type I proteins are homologous to the yeast Reduced Potassium Deficiency 3 (RPD3) protein, which is also widely found in other eukaryotes. Type II KDACs are plant-specific HD-tuins, and they are characterized by a conserved HD-2 type HDAC domain in the N terminus of the protein [5,6]. Type III KDACs are sirtuin proteins, which are homologous to the yeast Sir2 protein [5], and all sirtuins share the common feature that they are NAD+-dependent enzymes and they are inhibited by nicotinamide [7]. The Arabidopsis genome encodes two sirtuin proteins, SRT1 and SRT2 [8–10] and 14 additional Type I and Type II KDACs [5].

In contrast to plants, that express 2 sirtuins, other eukaryotes, such as humans have 7 sirtuin proteins (called SIRT). Phylogenetically, the human SIRT1, SIRT2 and SIRT3 are similar to the yeast Sir2 sirtuin, while SIRT4 and SIRT5 are mitochondrial, and are homologos to bacterial sirtuins, and SIRT6 and SIRT7 are related to archaeal-type sirtuins [11]. Sirtuins can deacetylate both histones and also non-histone proteins [12], and this deacetylation reaction can either decrease [13] or increase [14] the functionality of the substrate protein.

In plants, such as Arabidopsis, SRT1 and SRT2 can also deacetylate histones and non- histone proteins domains [15] affecting epigenetic regulation particularly during ethylene signaling [8], and also affect energy metabolism [9]. In this study we integrated in vitro characterization of the SRT1 and SRT2 proteins, with the in vivo characterization of the effect of genetically removing sirtuin activities and assaying the effect on the plant acetylome and metabolome.

Results

In vitro characterization of Arabidopsis SRT proteins

The two sirtuin proteins encoded by the Arabidopsis genome (SRT1 and SRT2) were individually expressed in E. coli and purified with the His-tag that was genetically fused to the

N-terminus of each protein. SDS-PAGE analysis of the purified proteins (Figure 1A) indicate that each preparation was over 95% pure and yields of between 1-4 mg of SRT proteins were achieved from 1-L of bacterial culture. These analyses in combination with size exclusion gel chromatography (Figure 1B) indicate that each protein is purified primarily as a monomer, with molecular weights of 54 kDa for SRT1 and 31 kDa for SRT2. SRT1 and SRT2 share 23% and 26% sequence identity, respectively, with the S. cerevisiae sirtuin, SIR2 (data not shown).

Each of the purified preparations was assayed for the ability to catalyze a deacetylation reaction using a commercial synthetic substrate which is an acetylated, luminogenic peptide.

The deacetylation of this substrate results in increased luminescence. The results shown in

Figure 1C indicate that both the SRT1 and SRT2 are deacetylating enzymes, and both enzymes show similar activity against this substrate. Moreover, both the SRT1- and SRT2-catalyzed reactions are susceptible to inhibition by nicotinamide, with an IC50 of 3.75 mM and 1.27 mM, respectively (Figure 1D); these values being close to values reported for classical sirtuins (~

0.1 mM) [16]. These in vitro properties of the purified SRT1 and SRT2 proteins are shared characteristics that define sirtuin proteins [7].

Characterization of sirtuin Arabidopsis mutants

Arabidopsis lines that carry both srt1 and srt2 mutant alleles were generated by crossing plants that carried either the srt1 or srt2 single mutant alleles. These lines were propagated by selfing, and progeny were genotyped to confirm that they are either homozygous mutant or 73 homozygous wild-type at these loci. The subsequent characterizations of these lines were conducted on pooled tissue collected from 21-day old sibling seedlings.

RT-PCR analysis of the expression of the SRT1 and SRT2 mRNA establish that each mutant individually eliminated the expression of each transcript, and that there is no compensatory change in the expression of the other homolog when either gene is mutated

(Figure 2). Side-by-side comparison of the growth morphology of plants ranging from 6-days old seedlings to 54-days old flowering plants, indicate that the individual srt1 and srt2 mutations, and the srt1srt2 double mutant do not affect a growth associated phenotypic difference from the wild-type plants (Figure 3).

The effect of srt1 srt2 double mutant on the Arabidopsis acetylome

Compared to wild type plants, we detected 211 differentially acetylated proteins in the srt1srt2 double mutant (p-value < 0.05) (Supplemental Table 2). The majority of these changes in the acetylome showed decreased acetylation, with only 75 proteins showing increased acetylation levels. These modulations in acetylation status ranged between a 5-fold increase for the protein encoded by AT1G76160, to 75% decrease for the protein encoded by

AT5G66530.

A combination of database queries was used to gain a broad understanding of the functional consequence of the srt1srt2-induced changes in the acetylome. These databases include GO slim analysis [17,18], Reactome Pathways in Panther [19,20], and AraCyc [21,22].

Figure 4 shows the GO terms and Reactome pathways associated with proteins that show significant increases and decreases in acetylation levels in the srt1srt2 double mutant. These queries reveal altered acetylation status of proteins associated with gene expression

(transcription), translation, DNA repair, amino acid metabolism, response to heat stress, lipid 74 metabolism, ATP binding or biosynthesis, oxidoreduction and transfer reactions. Use of

AraCyc identified changes in the acetylation status of functionalities associated with binding of histones to DNA, ascorbate utilization, flavin biosynthesis, photosynthesis light reactions, hydrolysis of triacylglycerols (e.g., tributyrin), oxidoreduction and cadmium utilization reactions (Figure 5).

All but two of the functional categories identified by these analyses display decreased acetylation status. This is somewhat surprising because the srt1srt2 mutation eliminates sirtuin-catalyzed deacetylating capabilities, which would be expected to increase the acetylation status of proteins. A potential explanation for these observations is that the observed changes in the acetylation status are not the result of the direct sirtuin-catalyzed deacetylation reactions, rather that sirtuins affect a core capability that cascades to a broader range of acetylating activities, thus increasing the acetylation status of these proteins. This explanation fits the model that sirtuins catalyze the deacetylation of histones that alters chromatin structure and hence initiating cascades of changes in gene expression, which this study reveals is associated with changes in acetylation status of down-stream proteins.

Effect of the sirtuin mutations on the Arabidopsis metabolome

Untargeted metabolomic analysis was conducted on wild-type seedlings and compared to the metabolomes of the srt1srt2 double mutant, and to the srt1 and srt2 single mutants.

These analyses detected 356 analytes, and 102 of these could be chemically identified

(Supplemental Table 1). The Volcano-plot visualization of the changes in the metabolome indicate that collectively these mutations significantly (p-value < 0.05) alter the accumulation

75 levels of 205 metabolites from the wild-type levels (Figure 6A). Thus, ~60% of the detected metabolome is altered by at least one srt mutation, and these alterations include both increased

(86) and decreased (155) accumulation of the metabolites.

Most changes in the metabolome are caused by the srt1 mutation, which affected the accumulation of 153 metabolites, and ~65% of these occurred uniquely in the srt1 mutation

(Figure 6B). The majority (65%) of these srt1-specific metabolite changes are decreases in levels of accumulation. In comparison, the srt2 mutation uniquely affects the accumulation of only 63 metabolites, with an additional 17 metabolites being affected by both srt1 and srt2 mutations. These changes in the metabolome indicate that the SRT1 and SRT2 gene products both share certain functions (represented by 42 metabolites that are affected by both mutations), but primarily affect distinct processes, generating metabolomes that are distinct from the wild-type and from each mutation.

The srt1srt2 double mutant expresses a metabolome that is more wild type-like than each of the srt1 and srt2 single mutants; furthermore, ~60% of the double mutant metabolite changes are shared with either srt1 or srt2 single mutants. Collectively, therefore these changes indicate that although the two SRT homologs have shared functions in determining a common metabolome, the two homologs are also involved in processes that can generate distinct metabolomes and judged by the number of metabolites that are affected, these latter singular changes appear to be more significant that the shared processes.

The interpretation of the results for the chemically identified metabolome (102 metabolites) is more revealing of potential function. Statistical analysis indicates that 67 of these chemically defined metabolites show significant (p-value < 0.05) differences between the mutants and the wild-type. Figures 7 and 8, identify the changes in abundance of the polar 76 and non-polar metabolites as affected by the srt1 and srt2 single mutants and the srt1srt2 double mutant lines.

Comparing these data, there is a distinct pattern difference in the effect of these mutations between these two classes of metabolites. Specifically, the effect of the srt1srt2 double mutant is near equivalent to that of the srt2 single mutant for all non-polar metabolites analyzed; namely, both of these mutations greatly reduce the accumulation of these metabolites to about 3-10% of wild-type levels. In contrast, whereas the srt1 mutation also reduces the accumulation of these non-polar metabolites, the impact of this mutation is not as dramatic.

This result indicates that the SRT2 homolog primarily regulates metabolic processes that support the biosynthesis of the non-polar metabolites. The common metabolic features of the metabolic pathways that supports the accumulation of these metabolites is the use of acetyl-

CoA as the precursor, and the use of NADP(P)H to biochemically reduce the oxidation state of the carbon that constitute these metabolites. This latter attribute maybe a consequence of the dependence of sirtuins on NAD, linking deacetylation reactions to the energy status of the organism [23,24].

In contrast, the effect of these mutations on the polar metabolites are not as generalizable. Specifically, the effect on the polar metabolome is discrete, and does not appear to be focused on a class of metabolites. This however may be a reflection of the fact that in contrast to the non-polar metabolites, the metabolic origins of the polar metabolites are considerably more diverse. For example, the srt1 mutation has dramatic effects on the accumulation of glutamine and ascorbic acid, and these effects are in opposite directions, inducing increased accumulation of glutamine and reducing the accumulation of ascorbic acid.

These two major metabolic alterations are uniquely induced by the srt1 mutation and appear 77 to be suppressed by combining the srt1 mutation with the srt2 mutation. Thus, in regulating these two metabolic processes, the two SRT homologs appear to be genetically epistatic in their actions. However, this is not generalizable for other polar metabolites (e.g., ornithine and

2-piperidone). In these examples, all three mutant lines examined equally increased accumulation of these metabolites.

Specific changes in metabolite abundance are also observed which demonstrate that sirtuins are involved in regulating specific pathways. For example, organic acids (e.g., fumaric acid, malic acid, oxalic acid and glyceric acids) are significantly lowered in srt1 and/or srt2 single mutants, while the level of these metabolites is restored to near wild-type amounts in the srt1srt2 double mutant. Additionally, higher levels of glutamine occur in the srt1 single mutant (Figure 7). These alterations may indicate that sirtuins affect the acetylation-based regulation of proteins associated with the TCA cycle.

Relationship between metabolome and acetylome of sirtuin mutants

A number of potential causative correlations are notable when one compares changes in the acetylome with changes in the metabolome of the srt1srt2 double mutant. For example, the level of the amino acid, alanine is significantly increased in the double mutants (Figure 7), and this correlates with higher acetylation status of two alanine aminotransferase isozymes encoded by At1g23310 (GGT1) and At1g70580 (GGT2) (Supplemental Table 2). These enzymes catalyze transamination reactions using alanine as either a substrate or product, in biosynthetic or catabolic process [25,26]. Similar correlations are detectable in the two datasets for other amino acids metabolic processes. These include increased ornithine levels

(Table 5) and decreased acetylation of ornithine transcarbamoylase (AT1G75330)

(Supplemental Table 2), which are involved in arginine metabolism. 78

Figure 4B and 4C indicate that a number of proteins related to pyruvate metabolism, gluconeogenesis, glycolysis and glucose metabolism experience significant decreases in acetylation levels. Some of these are glyceraldehyde 3-phosphate dehydrogenase

(AT1G12900), phosphoglycerate kinase family protein (AT1G56190.2), triosephosphate isomerase (AT2G21170.1) and phosphoglucose isomerase (AT4G24620.1). Specifically, these mutations caused increased levels of a number of 3-carbon metabolites that are either intermediates of glycolysis or are derived from these intermediates (e.g., pyruvate and glycerol

3-phosphate acyl ester) (Figure 7). The latter metabolically links glycolysis and glycerophospholipids [27].

Another correlation between the two datasets is the reduction in the accumulation of non-polar metabolites, particularly of fatty acids (Figure 8), and the change in the acetylation status of enzymes involved in these metabolic processes. This includes the significant difference in the acetylation levels of the biotin carboxylase subunit of acetyl-CoA carboxylase

(AT5G35360) and 3-ketoacyl-acyl carrier protein synthase III (AT1G62640) (Supplemental

Table 2), both of these enzymes catalyze reactions that commit acetyl-CoA metabolism to de novo fatty acid biosynthesis [28,29].

Discussion

Reversible acetylation regulates a plethora of biological processes, many of which are mediated by changes in gene expression. A well characterized mechanism that facilitates some of these modulations is the acetylation of lysine residues in histones, which relaxes the interactions that facilitate nucleosome assembly enabling access to DNA leading to transcriptional activation [1,30]. In addition to histone acetylation, lysine acetylation of non- histone proteins also regulates biological processes [9,31,32]. Maintaining and regulating the 79 processes of acetylation and deacetylation of lysine residues is therefore critical to the regulatory mechanism.

In plants, large number of histone and non-histone lysine acetylation sites have been identified, and these have been associated with regulation of gene expression and other biological processes [3,32]. Prior such studies have identified the importance of deacetylases in maintaining the regulation of the acetylation status of proteins, and perturbing these activities by mutations have indicated that they are involved in plant development and stress responses [33].

The Arabidopsis genome encodes at least 16 histone deacetylases [5], two of which appear to be sirtuin-type deacetylases, SRT1 and SRT2. Both SRT1 and SRT2 have histone deacetylase domains, and in vitro characterization has demonstrated that SRT2 has sirtuin characteristics i.e., inhibition by nicotinamide and NAD+ dependence [9,15]. In vivo studies have shown that SRT1 affects the ethylene signaling pathway [8] and is linked to stress response and interacts with transcriptional repressors [8,34]. SRT2 has been shown to be involved in energy metabolism affecting levels of sugars and amino acids [9]. Moreover, genetic complementation studies have shown that only SRT2, and not SRT1 of Arabidopsis compensates for the high salt sensitivity of yeast sirtuin-mutant [8]. Our in vitro studies demonstrate that Arabidopsis SRT1 and SRT2, both are definitively active classical sirtuins deacetylases. Furthermore, the dataset presented herein expands the role of sirtuin deacetylases in affecting plant metabolic processes.

The acetylome of plants lacking the 2 sirtuin deacetylases indicate that SRT1 and SRT2 affects the acetylation of both histone and non-histone proteins. These non-histone acetylated proteins can be related to regulation of a number of metabolic processes such as glucose 80 metabolism, fatty acid metabolism, amino acid metabolism, and the TCA cycle. The parallel non-targeted metabolomic analyses of the sirtuin mutant plants reinforce these conclusions and identify key proteins whose acetylation status potentially effects specific metabolic processes in planta.

The comparative metabolomic studies illustrate an interesting effect that will require additional insights to fully comprehend. Specifically, individual srt1 and srt2 mutations affect distinct portions of the metabolome, however, the combination of the two mutations are not additive, but rather produces a different metabolome that appears to be more wild-type like.

Some of these metabolomic effects may indicate that SRT1 and SRT2 potentially show genetic epistasis. Another potential complexity that will need to be further explored is the response of other acetylation-deacetylation enzymes in response to the srt1 and srt2 mutations [5]. For example, prior studies have indicated that the histone deacetylase, HDA14 affects about 10% of the Arabidopsis metabolome [32]. Further detailed analyses of individual lysine acetylation sites will allow mechanistic understanding of the role each acetylation-deacetylation cycle, which can potentially contribute to the enhancement of plant systems, which could in turn be used to improve the value, quality and sustainability of crops.

Methods

Plant material and growth condition

Arabidopsis thaliana, Colombia [Col-0] was used as the wild type plant. Plants lines carrying either srt1 (SALK_001493) or srt2 (SALK_149295) mutant alleles were obtained from the Arabidopsis Biological Resource Center (http://abrc.osu.edu). Homozygous srt1 and srt2 lines were crossed to obtain the double knockout, srt1 srt2 line. Sequences of DNA primers used for genotyping these plants are listed in Table 1. 81

Plants were grown at 22°C with continuous illumination (photosynthetic photon flux density of 100 µmol m-2 s-1). Seeds were sterilized and sown on Murashige and Skoog agar medium as described previously [35]. At 21-days after sowing, seedlings were harvested, pooled and immediately flash frozen in liquid nitrogen and stored at -80°C prior to lyophilization.

Isolation and characterization of plant nucleic acids

DNA was extracted from plant tissues and PCR-based genotyping was performed as described previously [36]. RNA was extracted from 21-day old pulverized seedling tissues using the TRIzol™ reagent as per manufacturer’s instructions (Invitrogen). The isolated RNA was treated with DNase (Ambion), and reverse transcribed using Double Primed RNA to cDNA EcoDry Premix (Takara, Clontech, Mountain View, CA). Gene specific mRNA presence was determined by RT-PCR using primers listed in Table 1; the ubiquitin

(At4g05320) mRNA was used as the positive control.

SRT1 and SRT2 cloning, expression, purification and chromatography

Putative Arabidopsis SRT1 (AT5G55760) and SRT2 (AT5G09230) ORFs were cloned into the BamHI restriction site of the pET30f vector. The pET30f vector [37] is a modified pET30 vector (Takara, Clontech, Mountain View, CA). The sequence encoding the 50 amino acid long signal peptide at the N-terminus of SRT2 was excluded from the construct. Sequence confirmed pET30f constructs were subsequently transformed into E. coli ArcticExpress (DE3) cells (Agilent Technologies, Santa Clara, CA) for protein expression. These strains were grown overnight at 37°C with shaking at 250 rpm for aeration, in 10 ml LB medium containing the antibiotics kanamycin (50 μg/ml) and gentamycin (20 μg/ml). The overnight culture was used to inoculate 0.5 L LB media containing no antibiotics, and these cultures were grown at 82

° 30 C with shaking at 250 rpm, until an OD600 of ~0.6. Expression was then induced by the addition of IPTG to a final concentration of 0.1 mM, and the culture was grown at 13°C with shaking at 250 rpm for 24-48 hours.

Cells were collected by centrifugation at 5000g for 20 minutes and resuspended in a

20 mM sodium phosphate buffer, pH 8.0 containing 0.5 M NaCl, 1 mM DTT, 5 mM imidazole, 0.1% (v/v) Triton X-100, 0.1 mg/ml phenyl methyl sulfonyl fluoride (PMSF), and

10 μl/ml Protease Inhibitor Cocktail (Sigma-Aldrich Co., St. Louis, MO). Cells were disrupted by sonication on ice, and all subsequent steps were conducted at 4°C. The cell extract was centrifuged at 20,000xg for 30 minutes, and the supernatant was retained, filtered through a 0.45 μm filter (Corning Inc., Corning, NY) and applied to a column containing 2ml

PerfectPro Ni-NTA agarose (5 Prime, Inc., Gaithersburg, MD). The column was washed with increasing concentrations of imidazole, and the proteins were eluted with 0.2 M imidazole.

Samples were immediately dialyzed into 10 mM sodium phosphate buffer, pH 8.0 containing

0.5 M NaCl, 1 mM DTT and 10% glycerol.

Chromatography

Size exclusion gel filtration chromatography was conducted with an AKTA FPLC system (GE Healthcare Life Sciences, Pittsburg, PA). Protein samples (100 µl of ~2 mg/ml) were injected into a prepacked Superdex 200 Increase 10/300 GL gel filtration column (GE

Healthcare Life Sciences, Pittsburg, PA). The buffer used for FPLC was 10 mM sodium phosphate buffer, pH 8.0 containing 0.5 M NaCl, 1 mM DTT and 10% glycerol. The flow rate of 0.4 ml/minute was used and protein elution was monitored using an UV absorbance detector at 280 nm.

SRT1 and SRT2 kinetic assay

The SRT1 and SRT2 activity assays were performed using SIRT-Glo assay (Promega

Corporation, Madison, WI) as per manufacturer protocol. 2.43 µg of protein was used per reaction for both SRT1 and SRT2. The commercial synthetic substrate is an acetylated, luminogenic peptide.

Non-targeted Metabolite Profile Analysis by GC-MS

Metabolite extracts were prepared from 2-10 mg of pooled lyophilized tissues as reported previously [38]. Extracts were prepared. Internal standards, Ribitol (25 µg) and nonadecanoic acid (25 µg) were spiked into the extracts as internal standards for the polar and non-polar fractions, respectively. In parallel, blank samples were prepared with the identical technique, without including Arabidopsis tissue.

Hot methanol (0.35 ml at 60°C) was added to each sample, and incubated at 60°C for

10 min. The sample was then sonicated for 10 min at full power. Chloroform (0.35 ml) and

HPLC grade water (0.3 ml) was added to each sample, and the mixture vortexed for 1 to 3 min.

After centrifugation at 13,000 g for 5 min, 200 ml of the upper phase (polar fraction) and

200 ml of the lower phase (non-polar fraction) were separately removed into separate 2 ml GC-

MS vials. The polar sample was dried in a Speed-Vac concentrator (model SVC 100H, Savant,

NY) and the non-polar sample was dried by evaporation facilitated by blowing nitrogen gas over the solvent.

Samples were then double derivatized by methoximation and silylation.

Methoximylated was accomplished by adding 50 µl of methoxyamine hydrochloride (20 mg/mL dissolved in dry pyridine), and the reaction mixture was incubated at 30 °C for 1.5 hours. Silylation was performed by adding 70-µL of N, O- 84

Bis(trimethylsilyl)trifluoroacetamide (BSTFA) with 1% trimethyl chlorosilane (TMCS) and the mixture incubated at 65 °C for 30 min.

GC-MS analysis was performed using an Agilent 6890 GC connected to an Agilent

5973 quadrupole mass-spectrometer. Chromatography was with an Agilent 19091S-433, HP-

5MS (5%-Phenyl)-methyl siloxane column (30 m x 0.25 mm x 0.25 mm). One µL of the derivatized sample was injected in splitless mode. An initial oven temperature of 70 °C was maintained for 2 min, and the temperature was then increased from 70 °C to 320 °C at the rate of 5 °C per min, and maintained at this temperature for 5 min. The inlet temperature was set at 280 °C and helium flow rate at 1 mL per min. The ionization energy was set at 70 eV and the interface was temperature at 280 °C.

The GC-MS data files were deconvoluted and queried against an in-house MS-library and the NIST 14 Mass Spectral Library using NIST AMDIS software [39].

Sample Processing for acetylome analysis and Protein Extraction

Whole plant samples were harvested and ground in liquid nitrogen using a pre-chilled mortar and pestle. Sample preparation methods were loosely based on results and methods from previous published work [40]. For each individual sample, five volumes (v:w) of Tris- buffered phenol (pH 8) was added to 500 mg of ground tissue and then mixed with 5 volumes

(buffer:tissue, v:w) of extraction buffer (50mM Tris pH 7.5, 1mM EDTA pH 8, 0.9 M sucrose).

The tissue and buffers were then spun down at 13,000 x g for 10 minutes at 4C, the phenol phase moved to a new tube, and a second phenol extraction was performed on the aqueous phase using the same volume of phenol as before. The two phenol phases were combined in a new tube with 5 volumes (v:v) of pre-chilled 0.1 M ammonium acetate in methanol. After mixing well, it was kept at -80C overnight prior to centrifugation at 4,500 x g for 10 minutes 85 at 4C. Two more additional precipitation steps were performed both by resuspending the pellet in 0.1 M ammonium acetate in methanol and incubating each time for 1 hour at -20C followed by centrifugation at 4,500 x g for 10 minutes. Following the two additional precipitation steps, the pellet was resuspended in 70% Methanol and kept at -20C overnight prior to centrifugation at 4,500 x g for 10 minutes at 4C. The supernatant was removed, and the protein pellet was placed in a vacuum concentrator until near dry.

Protein Preparation and Digestion

Two volumes (buffer:pellet, v:v) of protein digestion buffer (8M urea, 50mM Tris pH

7, 5mM Tris(2-carboxyethyl)phosphine hydrochloride (TCEP)) was added to the pellet. The samples were then resuspended via probe sonication. The protein concentration for each sample was then determined using a Bradford protein assay (ThermoScientific). The solubilized protein was then added to an Amicon Ultracel -30K Centrifugal Filter (Cat #

UFC903024) and centrifuged at 4,000 x g for 30 minutes, repeat once more. Four mls of urea solution (8M urea, 0.1M Tris pH 8) with 2mM TCEP was added to the filter unit and then centrifuged at 4,000 x g for 30 minutes. Two mls of iadoacetamide (IAM) solution (50mM

IAM, 8M urea, 0.1M Tris pH 8) was added and units were incubated in the dark at room temperature for 30 minutes before being centrifuged at 4,000 x g for 20 minutes. Two mls of urea solution (8M urea, 0.1M Tris pH 8) was added to the unit and centrifuged at 4,000 x g for

20 minutes, this step was repeated once more. Two mls of 0.05M NH4CO3 was added to the filter unit and centrifuged at 4,000 x g for 20 minutes, this step was repeated once more. A new collection tube was added and 2ml of the 0.05M NH4CO3 solution with trypsin (enzyme to protein ratio 1:100) was added and the samples were incubated overnight at 37C. Undigested protein amounts were determined via Bradford assays then trypsin (1μg/μl) and an equal 86 volume of LysC (0.1μg/μl) were added at a ratio of 1:100 (enzyme to protein). The units were then incubated for 4 hours at 37C. Units were centrifuged at 4,000 x g for 20 minutes. One ml of 0.05M NH4CO3 solution was added and the centrifuge step was performed again. The previous step was repeated once more for a total of ~4mL of eluted peptides. The peptides were then concentrated in a vacuum concentrator, resuspended in water and peptide concentration was measured using the Pierce BCA Protein Assay Kit (ThermoScientific).

C18 Desalting and Dimethyl Labeling

Around 5mg of peptides from each sample was desalted with C18 columns (Waters) and labeled with light (WT) /heavy (Mutant) dimethyl labels. The C18 column was rinsed with

3 ml methanol, followed by 3ml 0.1% formic acid twice. The peptides (resuspended with 0.1% formic acid) for each sample were loaded to a C18 column and ran through the column. The resulting flow through of peptide solutions were reloaded to the column once, washed with

3ml of 0.1% formic acid twice and then 3ml of water twice. After that, 8ml of dimethyl labeling solutions (5 ml of 50 mM NaH2PO4 pH 7.5, 17.5ml of 50mMNa2HPO4 pH 7.5, 1.25 ml of 4%

(v/v) formaldehyde in water (CH2O for light label, 13CD2O for heavy label) and 1.25ml of 0.6

M cyanoborohydride in water (NaBH3CN for light label, or NaBD3CN for heavy label) mixed together fresh before labeling) were flushed through the C18 column, and the flow through was re-run through the column once more. The columns were then washed with 2.5ml of water a total of 3 times. The dimethyl labeled peptides were eluted from the C18 column with 0.25ml of 20% acetonitrile followed immediately by 0.25ml 40% acetonitrile and finally 0.5ml 80% acetonitrile (in that order). The peptides solutions were speed vacuumed to almost dry in a vacuum concentrator and stored at -80C.

Acetylated Peptide Enrichment

The dimethyl labeled peptides for each pair of samples (one mutant/heavy and one

WT/light) were mixed together and resuspended with 3.5ml of water, and the pH was adjusted by 1M tris-HCl pH 7.5. Antibody-based acetyl peptide enrichment was performed based on the methods of [41]. Two mgs of anti-acetyl lysine antibody (cat # ICP0388) was used for acetyl-peptide enrichment of each pair of labeled samples. Solution content from the tubes of the antibody were moved to a 0.2μm spin filter unit and washed with 4ml of 50mM Tris-HCl pH 7.5 a total of 3 times. The dimethyl labeled peptide solution was then added onto the same filter unit and used to resuspend the antibody. The fully resuspended peptide-antibody solution was transferred to a 15ml tube and incubated at 4C for 1 hour with rotation. After the incubation, the peptides-antibody solution was moved back to the 0.2μm spin filter unit and centrifuged to separate the antibody-bound peptides and any unbound peptides. The antibody- bound peptides were washed with 4ml of 50mM Tris-HCl pH 7.5 a total of 3 times. The acetyl- peptides were then eluted from the antibody with two separate 1.5ml 0.1% TFA elutions. The antibody was washed twice with 4ml of 50mM Tris-HCL pH 7.5 and a second acetyl-peptides enrichment was done with the peptide flow through from the first elution step. The acetyl- peptides from the 2 steps of enrichment were combined together and desalted with a C18 column (Waters). Final peptide amounts for each pair of KO(Heavy)/WT(Light) labeled peptides were quantified using the Pierce BCA Protein assay kit (Thermo Scientific).

LC/MS-MS

An Agilent 1260 quaternary HPLC was used to deliver a flow rate of ~600 nL min-1 via a splitter. All columns were packed in house using a Next Advance pressure cell and the nanospray tips were fabricated using fused silica capillary that was pulled to a sharp tip using 88 a laser puller (Sutter P-2000). Twenty-five g of peptides was loaded onto 20cm capillary columns packed with 5μM Zorbax SB-C18 (Agilent), which was connected using a zero dead volume 1μm filter (Upchurch, M548) to a 5cm long strong cation exchange (SCX) column packed with 5μm PolySulfoethyl (PolyLC). The SCX column was then connected to a 20cm nano-spray tip packed with 2.5μM C18 material (Waters). The 3 sections were joined and mounted on a custom electrospray source for on-line nested peptide elution. A new set of columns was used for every sample. Peptides were eluted from the loading column onto the

SCX column using a 0 to 80% acetonitrile gradient over 60 minutes. Peptides were then fractionated from the SCX column using a series of salt steps at the following concentrations:

10, 30, 32.5, 35, 37.5, 40, 42.5, 45, 50, 55, 65, 75, 85, 90, 95, 100, 150, and 1000mM. For these analyses, buffers A (99.9% H2O, 0.1% formic acid), B (99.9% ACN, 0.1% formic acid),

C (100mM ammonium acetate, 2% formic acid), and D (1M ammonium acetate, 2% formic acid) were utilized. For each salt step, a 150-minute gradient program comprised of a 0–5 minute increase to the specified ammonium acetate concentration, a 5–10 minutes hold, 10–14 minutes at 100% buffer A, 15–120 minutes 10–35% buffer B, 120–140 minutes 35–80% buffer

B, 140–145 minutes 80% buffer B, and 145–150 minutes buffer A was employed.

Eluted peptides were analyzed using a Thermo Scientific Q-Exactive Plus high- resolution quadrupole Orbitrap mass spectrometer, which was directly coupled to the HPLC.

Data dependent acquisition was obtained using Xcalibur 4.0 software in positive ion mode with a spray voltage of 2.00 kV and a capillary temperature of 275°C and an RF of 60. MS1 spectra were measured at a resolution of 70,000, an automatic gain control (AGC) of 3e6 with a maximum ion time of 100ms and a mass range of 400-2000 m/z. Up to 15 MS2 were triggered at a resolution of 17,500. An AGC of 1e5 with a maximum ion time of 50ms, an isolation 89 window of 1.5 m/z, and a normalized collision energy of 28 were utilized. Charge exclusion was set to unassigned, 1, 5–8, and >8. MS1 that triggered MS2 scans were dynamically excluded for 25s.

Database Search and FDR Filtering

The raw data were analyzed using MaxQuant version 1.6.3.3 [42]. Spectra were searched against the Arabidopsis thaliana Tair10 protein list which was complemented with reverse decoy sequences and common contaminants by the MaxQuant software. For group specific parameters the type was set to “Standard”, the multiplicity set to “2”, Max Labeled set to “3”, and for the light label we selected “DimethNter0” and “DimethLys0”. Additionally, the heavy label was set to “DimethLys8” and “DimethNter8”. “Carbamidomethyl cysteine” was set as a fixed modification while “methionine oxidation” and “protein N-terminal acetylation” and “acetyl (K)” were set as variable modifications. Digestion parameters were set to “specific” and “Trypsin/P;LysC” max missed set to “5”. Up to two missed cleavages were allowed. A false discovery rate less than 0.01 and protein identification level was required. The “second peptide” option was used to identify co-fragmented peptides. The

“match between runs” feature of MaxQuant was not utilized

Statistical Analysis for acetylome data

The ratio of the heavy label to the light label (H/L) from the MaxQuant output file titled

“Acetyl (K)Sites” was used for all statistical analysis on the proteomics data. First, the ratios were multiplied by 100 to both transform the ratios to integer values for downstream analysis and increase the spread of the ratio distribution to improve differential protein abundance detection. Then, the null hypothesis that the ratio is equal to 1, i.e., the heavy and light labels are equal was tested (on the transformed data, this is equivalent to testing that the ratio is equal 90 to 100). To test this null hypothesis on the transformed ratios, 3 biological replicates were simulated from a Poisson distribution with mean 100 which represent control samples.

PoissonSeq [43] was then used to compare the experimental transformed ratios to the simulated control ratios. Since the simulated control samples are random, differential abundance analysis was repeated on 100 different simulated control samples (i.e. 100 runs) to control for any random effects. Differentially abundant proteins were defined as proteins whose ratios had a q-value < 0.1 in at least one run or a p-value < 0.05 in at least 1/3 of the runs and ratio ≤ 0.8 or ratio ≥ 1.25 (i.e., fold change in label intensity of at least 1.25).

Statistical Analysis of metabolomics data and GO ontology analysis

Relative concentrations of the metabolites between the knockout lines and the wild-type was compared for statistical significance using student t-test. Significant differences in metabolites was visualized using a volcano plot of the log2 fold difference between the knockouts and wild- type against the negative log10 of calculated respective p-values.

For the list of proteins which had significantly different acetylation levels between the wild- type and the srt1srt2 double knockout, a Gene Ontology (GO) analysis [17,18] was performed using PANTHER database [19,20]. GO terms for biological processes and molecular functions and Reactome pathways associated with the list of proteins were analyzed. A Fisher’s exact test for enrichment of GO terms was also performed and p-values were corrected for FDR. The list of proteins was also analyzed against the AraCyc database [21,22].

References

[1] E. Verdin, M. Ott, 50 years of protein acetylation: from gene regulation to epigenetics, metabolism and beyond, Nat. Rev. Mol. Cell Biol. 16 (2015) 258–264. doi:10.1038/nrm3931.

[2] V. Bernal, S. Castaño-Cerezo, J. Gallego-Jara, A. Écija-Conesa, T. de Diego, J.L. Iborra, M. Cánovas, Regulation of bacterial physiology by lysine acetylation of proteins, N. Biotechnol. 31 (2014) 586–595. doi:10.1016/j.nbt.2014.03.002.

[3] R.S.P. Rao, J.J. Thelen, J.A. Miernyk, In silico analysis of protein Lys-N€-acetylation in plants, Front. Plant Sci. 5 (2014) 381. doi:10.3389/fpls.2014.00381.

[4] A.C. König, M. Hartl, P.J. Boersema, M. Mann, I. Finkemeier, The mitochondrial lysine acetylome of Arabidopsis, Mitochondrion. 19 (2014) 252–260. doi:10.1016/j.mito.2014.03.004.

[5] R. Pandey, A. Müller, C.A. Napoli, D.A. Selinger, C.S. Pikaard, E.J. Richards, J. Bender, D.W. Mount, R.A. Jorgensen, Analysis of histone acetyltransferase and histone deacetylase families of Arabidopsis thaliana suggests functional diversification of chromatin modification among multicellular eukaryotes., Nucleic Acids Res. 30 (2002) 5036–55. http://www.ncbi.nlm.nih.gov/pubmed/12466527 (accessed January 26, 2019).

[6] A. Lusser, G. Brosch, A. Loidl, H. Haas, P. Loidl, Identification of maize histone deacetylase HD2 as an acidic nucleolar phosphoprotein., Science. 277 (1997) 88–91. http://www.ncbi.nlm.nih.gov/pubmed/9204905 (accessed March 2, 2019).

[7] J. Gallego-Jara, A. Écija Conesa, T. de Diego Puente, G. Lozano Terol, M. Cánovas Díaz, Characterization of CobB kinetics and inhibition by nicotinamide., PLoS One. 12 (2017) e0189689. doi:10.1371/journal.pone.0189689.

[8] F. Zhang, L. Wang, E.E. Ko, K. Shao, H. Qiao, Histone Deacetylases SRT1 and SRT2 Interact with ENAP1 to Mediate Ethylene-Induced Transcriptional Repression, Plant Cell. 30 (2018) 153–166. doi:10.1105/tpc.17.00671.

[9] A.-C. Konig, M. Hartl, P.A. Pham, M. Laxa, P.J. Boersema, A. Orwat, I. Kalitventseva, M. Plochinger, H.-P. Braun, D. Leister, M. Mann, A. Wachter, A.R. Fernie, I. Finkemeier, The Arabidopsis Class II Sirtuin Is a Lysine Deacetylase and Interacts with Mitochondrial Energy Metabolism, PLANT Physiol. 164 (2014) 1401–1414. doi:10.1104/pp.113.232496.

[10] X. Liu, W. Wei, W. Zhu, L. Su, Z. Xiong, M. Zhou, Y. Zheng, D.-X. Zhou, Histone Deacetylase AtSRT1 Links Metabolic Flux and Stress Response in Arabidopsis, Mol. Plant. 10 (2017) 1510–1522. doi:10.1016/j.molp.2017.10.010. 92

[11] R.A. Frye, Phylogenetic Classification of Prokaryotic and Eukaryotic Sir2-like Proteins, (2000). doi:10.1006/bbrc.2000.3000.

[12] T. Finkel, C.-X. Deng, R. Mostoslavsky, Recent progress in the biology and physiology of sirtuins, Nature. 460 (2009) 587–591. doi:10.1038/nature08197.

[13] V.J. Starai, I. Celic, R.N. Cole, J.D. Boeke, J.C. Escalante-Semerena, Sir2-dependent activation of acetyl-CoA synthetase by deacetylation of active lysine, Science (80-. ). 298 (2002) 2390–2392. doi:10.1126/science.1077650.

[14] Q. Wang, Y. Zhang, C. Yang, H. Xiong, Y. Lin, J. Yao, H. Li, L. Xie, W. Zhao, Y. Yao, Z.-B. Ning, R. Zeng, Y. Xiong, K.-L. Guan, S. Zhao, G.-P. Zhao, Acetylation of metabolic enzymes coordinates carbon source utilization and metabolic flux., Science. 327 (2010) 1004–7. doi:10.1126/science.1179687.

[15] C. Hollender, Z. Liu, Histone deacetylase genes in Arabidopsis development., J. Integr. Plant Biol. 50 (2008) 875–85. doi:10.1111/j.1744-7909.2008.00704.x.

[16] M.T. Schmidt, B.C. Smith, M.D. Jackson, J.M. Denu, Coenzyme Specificity of Sir2 Protein Deacetylases, J. Biol. Chem. 279 (2004) 40122–40129. doi:10.1074/jbc.M407484200.

[17] M. Ashburner, C.A. Ball, J.A. Blake, D. Botstein, H. Butler, J.M. Cherry, A.P. Davis, K. Dolinski, S.S. Dwight, J.T. Eppig, M.A. Harris, D.P. Hill, L. Issel-Tarver, A. Kasarskis, S. Lewis, J.C. Matese, J.E. Richardson, M. Ringwald, G.M. Rubin, G. Sherlock, Gene Ontology: tool for the unification of biology, Nat. Genet. 25 (2000) 25– 29. doi:10.1038/75556.

[18] The Gene Ontology Consortium, Expansion of the Gene Ontology knowledgebase and resources., Nucleic Acids Res. 45 (2017) D331–D338. doi:10.1093/nar/gkw1108.

[19] P.D. Thomas, M.J. Campbell, A. Kejariwal, H. Mi, B. Karlak, R. Daverman, K. Diemer, A. Muruganujan, A. Narechania, PANTHER: a library of protein families and subfamilies indexed by function., Genome Res. 13 (2003) 2129–41. doi:10.1101/gr.772403.

[20] P.D. Thomas, A. Kejariwal, N. Guo, H. Mi, M.J. Campbell, A. Muruganujan, B. Lazareva-Ulitsky, Applications for protein sequence-function evolution data: mRNA/protein expression analysis and coding SNP scoring tools, Nucleic Acids Res. 34 (2006) W645–W650. doi:10.1093/nar/gkl229.

[21] S.Y. Rhee, P. Zhang, H. Foerster, C. Tissier, AraCyc: Overview of an Arabidopsis Metabolism Database and its Applications for Plant Research, in: Plant Metabolomics, Springer-Verlag, Berlin/Heidelberg, 2006: pp. 141–154. doi:10.1007/3-540-29782- 0_11. 93

[22] P. Zhang, H. Foerster, C.P. Tissier, L. Mueller, S. Paley, P.D. Karp, S.Y. Rhee, MetaCyc and AraCyc. Metabolic Pathway Databases for Plant Research, PLANT Physiol. 138 (2005) 27–37. doi:10.1104/pp.105.060376.

[23] J. Landry, A. Sutton, S.T. Tafrov, R.C. Heller, J. Stebbins, L. Pillus, R. Sternglanz, The silencing protein SIR2 and its homologs are NAD-dependent protein deacetylases, Proc. Natl. Acad. Sci. 97 (2000) 5807–5811. doi:10.1073/pnas.110148297.

[24] S. Imai, C.M. Armstrong, M. Kaeberlein, L. Guarente, Transcriptional silencing and longevity protein Sir2 is an NAD-dependent histone deacetylase, Nature. 403 (2000) 795–800. doi:10.1038/35001622.

[25] A.H. Liepman, L.J. Olsen, Alanine Aminotransferase Homologs Catalyze the Glutamate:Glyoxylate Aminotransferase Reaction in Peroxisomes of Arabidopsis, PLANT Physiol. 131 (2003) 215–227. doi:10.1104/pp.011460.

[26] N. Ohkama-Ohtsu, S. Radwan, A. Peterson, P. Zhao, A.F. Badr, C. Xiang, D.J. Oliver, Characterization of the extracellular γ-glutamyl transpeptidases, GGT1 and GGT2, in Arabidopsis, Plant J. 49 (2007) 865–877. doi:10.1111/j.1365-313X.2006.03004.x.

[27] Y. Mugabo, S. Zhao, A. Seifried, S. Gezzar, A. Al-Mass, D. Zhang, J. Lamontagne, C. Attane, P. Poursharifi, J. Iglesias, E. Joly, M.-L. Peyot, A. Gohla, S.R.M. Madiraju, M. Prentki, Identification of a mammalian glycerol-3-phosphate phosphatase: Role in metabolism and signaling in pancreatic β-cells and hepatocytes., Proc. Natl. Acad. Sci. U. S. A. 113 (2016) E430-9. doi:10.1073/pnas.1514375113.

[28] J. Sun, J. Ke, J.L. Johnson, B.J. Nikolau, E.S. Wurtele, Biochemical and molecular biological characterization of CAC2, the Arabidopsis thaliana gene coding for the biotin carboxylase subunit of the plastidic acetyl-coenzyme A carboxylase., Plant Physiol. 115 (1997) 1371–83. http://www.ncbi.nlm.nih.gov/pubmed/9414551 (accessed March 1, 2019).

[29] K. Dehesh, H. Tai, P. Edwards, J. Byrne, J.G. Jaworski, Overexpression of 3-ketoacyl- acyl-carrier protein synthase IIIs in plants reduces the rate of lipid synthesis., Plant Physiol. 125 (2001) 1103–14. http://www.ncbi.nlm.nih.gov/pubmed/11161065 (accessed March 1, 2019).

[30] K. Struhl, Histone acetylation and transcriptional regulatory mechanisms, Genes Dev. 12 (1998) 599–606. doi:10.1101/gad.12.5.599.

[31] F. Hosp, I. Lassowskat, V. Santoro, D. De Vleesschauwer, D. Fliegner, H. Redestig, M. Mann, S. Christian, M.A. Hannah, I. Finkemeier, Lysine acetylation in mitochondria: From inventory to function, Mitochondrion. 33 (2017) 58–71. doi:10.1016/j.mito.2016.07.012.

[32] M. Hartl, M. Füßl, P.J. Boersema, J.-O. Jost, K. Kramer, A. Bakirbas, J. Sindlinger, M. Plöchinger, D. Leister, G. Uhrig, G.B. Moorhead, J. Cox, M.E. Salvucci, D. Schwarzer, M. Mann, I. Finkemeier, Lysine acetylome profiling uncovers novel histone deacetylase substrate proteins in Arabidopsis., Mol. Syst. Biol. 13 (2017) 949. doi:10.15252/msb.20177819.

[33] Y. Shen, W. Wei, D.-X. Zhou, Histone Acetylation Enzymes Coordinate Metabolism and Gene Expression, Trends Plant Sci. 20 (2015) 614–621. doi:10.1016/j.tplants.2015.07.005.

[34] X. Liu, W. Wei, W. Zhu, L. Su, Z. Xiong, M. Zhou, Y. Zheng, D.-X. Zhou, Histone Deacetylase AtSRT1 Links Metabolic Flux and Stress Response in Arabidopsis, Mol. Plant. 10 (2017) 1510–1522. doi:10.1016/j.molp.2017.10.010.

[35] H. Jin, Z. Song, B.J. Nikolau, Reverse genetic characterization of two paralogous acetoacetyl CoA thiolase genes in Arabidopsis reveals their importance in plant growth and development, Plant J. 70 (2012) 1015–1032. doi:10.1111/j.1365- 313X.2012.04942.x.

[36] X. Li, H. Ilarslan, L. Brachova, H.-R. Qian, L. Li, P. Che, E.S. Wurtele, B.J. Nikolau, Reverse-genetic analysis of the two biotin-containing subunit genes of the heteromeric acetyl-coenzyme A carboxylase in Arabidopsis indicates a unidirectional functional redundancy., Plant Physiol. 155 (2011) 293–314. doi:10.1104/pp.110.165910.

[37] F. Jing, D.C. Cantu, J. Tvaruzkova, J.P. Chipman, B.J. Nikolau, M.D. Yandeau-Nelson, P.J. Reilly, Phylogenetic and experimental characterization of an acyl-ACP thioesterase family reveals significant diversity in enzymatic specificity and activity, BMC Biochem. 12 (2011) 44. doi:10.1186/1471-2091-12-44.

[38] M.A. Schmidt, W.B. Barbazuk, M. Sandford, G. May, Z. Song, W. Zhou, B.J. Nikolau, E.M. Herman, Silencing of soybean seed storage proteins results in a rebalanced protein composition preserving seed protein content without major collateral changes in the metabolome and transcriptome., Plant Physiol. 156 (2011) 330–45. doi:10.1104/pp.111.173807.

[39] S.E. Stein, An integrated method for spectrum extraction and compound identification from gas chromatography/mass spectrometry data, J. Am. Soc. Mass Spectrom. 10 (1999) 770–781. doi:10.1016/S1044-0305(99)00047-1.

[40] G. Song, P.Y. Hsu, J.W. Walley, Assessment and Refinement of Sample Preparation Methods for Deep and Quantitative Plant Proteome Profiling, Proteomics. 18 (2018) 1800220. doi:10.1002/pmic.201800220.

[41] J.W. Walley, Z. Shen, M.R. McReynolds, E.A. Schmelz, S.P. Briggs, Fungal-induced protein hyperacetylation in maize identified by acetylome profiling., Proc. Natl. Acad. Sci. U. S. A. 115 (2018) 210–215. doi:10.1073/pnas.1717519115.

[42] S. Tyanova, T. Temu, J. Cox, The MaxQuant computational platform for mass spectrometry-based shotgun proteomics., Nat. Protoc. 11 (2016) 2301–2319. doi:10.1038/nprot.2016.136.

[43] J. Li, D.M. Witten, I.M. Johnstone, R. Tibshirani, Normalization, testing, and false discovery rate estimation for RNA-sequencing data., Biostatistics. 13 (2012) 523–38. doi:10.1093/biostatistics/kxr031. 96

Figures

Figure 1: Characterization of SRT1 and SRT2 recombinant proteins. (A) Purified recombinant proteins resolved on SDS-PAGE gel. SRT2 protein indicated with white arrow and SRT1 indicated with blue arrow. (B) Time-course enzymatic assay of SRT1 and SRT2 protein.

Protein activity indicated by rise in the luminescence signal (RLU). (C) Size exclusion gel chromatography runs for SRT1 and SRT2. The purple downward arrows indicate elution volumes for molecular weight standards. (D) Nicotinamide inhibition assay of SRT1 and

SRT2. The IC50 values for SRT1 is 3.75 mM (R2 = 0.93), and SRT2 is 1.27 mM (R2 = 0.91).

Figure 2: RT-PCR analysis of SRT1, SRT2 and UBQ mRNAs in RNA isolated from wild-type, srt1 and srt2 single mutants and srt1 srt2 double mutant 21-day old seedlings. Primers used for each mRNA are listed in Table 1.

Figure 3: Phenotypic characterization of wild-type, srt1, srt2 and srt1srt2 Arabidopsis lines in various stages of growth. (A) 6 days old seedlings (B) 21 days old seedlings (C) 34 days old plants (D) 54 days old plants.

Figure 4: Functional classification of proteins that are differentially acetylated between wild- type and srt1 srt2 double mutant lines of Arabidopsis (p-value < 0.05). GO Slim Molecular

Function categories (A), GO Slim Biological Processes categories (B) were obtained from

Gene Ontology Consortium. Reactome Pathways categories (C) were obtained from the

Panther database. Red bars indicate proteins which had decreased abundance of acetylation in srt1srt2 and blue bars indicate proteins which had decreased abundance of acetylation in srt1srt2. The numbers on the x-axis indicates number of proteins belonging to the specific category.

100

Figure 5: List of reactions/ pathways found to be differentially acetylated in wild-type vs srtdKO lines of Arabidopsis (p-value < 0.05) using AraCyc Cellular Overview Tool.

101

Figure 6: A. Volcano plot representation of changes in the Arabidopsis metabolome induced by mutations in the srt1 and srt2 genes. Data points above the horizontal dotted line represent metabolites that show statistically significant changes in accumulation (p-value < 0.05) in either srt1 (black data points) and srt2 (red data points) single mutants and the srt1 srt2 double mutant (blue data points) as compared to wild-type seedlings. Error bars indicate standard error from biological replicate (n=3), with each replicate consisting of pooled tissue of 21-day old seedlings.

B. Venn diagram set-representation of number of metabolites that are significantly different in the srt1, srt2 single mutants and in the srt1 srt2 double mutant seedlings. Inside the bracket and the table, the red color number indicates number of metabolites that reduced in abundance compared to wild-type and green indicates the number that increased in abundance compared to wild-type.

102

Figure 7: Quantitative changes in the metabolome of chemically defined analytes for polar metabolites. Error bars indicate standard error from biological replicates (n = 3). The metabolites listed have significant changes in abundance (p-value < 0.05) in one or more of the mutants, srt1 and srt2 single mutants or srt1 srt2 double mutant lines compared to wild-type.

103

Figure 8: Quantitative changes in the metabolome of chemically defined analytes for non-polar metabolites. Error bars indicate standard error from biological replicates (n = 3). The metabolites listed have significant changes in abundance (p-value < 0.05) in one or more of the mutants, srt1 and srt2 single mutants or srt1 srt2 double mutant lines compared to wild-type.

104

Tables

Table 1: List of primers

Sequence Description Border primer (At T-DNA insertion) for 5’- ATTTTGCCGATTTCGGAAC -3’ DNA genotyping 5’- GAGAACAGCACGAAACGAAAC -3’ Right primer for SRT1 DNA genotyping 5’- GTTTCGCTTGACACATGTTCC -3’ Left primer for SRT1 DNA genotyping 5’- TTCCACATTCTGTGCTAACCC -3’ Right primer for SRT2 DNA genotyping 5’- CGCAGAGAGAGAACAAAATCG -3’ Left primer for SRT2 DNA genotyping 5’- TCCAGGGCTCGGATCCATGTCTTTAGGT TACGCAGAGAAATT -3’ Primers to clone SRT1 (at Bam HI site of 5’- pet30f vector) GCTCGAATTCGGATCCCTATGCCTTGGT TTCTTCTGCCA -3’ 5’- TCCAGGGCTCGGATCCTTCGATCCCTGG TGGTTCT -3’ Primers to clone SRT2 (at Bam HI site of 5’- pet30f vector) GCTCGAATTCGGATCCGCTAGAGAGCT GGGACACT -3’ 5’- CGTGCAATGCCGAGTATGAC -3’ For RT-PCR confirmation of SRT1 5’- GCCTTCGATTTTCGCTTCCC -3’ knockout plants 5’- GGATGGAGGAGGTTCACTGC -3’ For RT-PCR confirmation of SRT2 5’- AGCGATCCCACGTCAAGAAC -3’ knockout plants 5’- AGACCATCACCCTTGAAGTGGA -3’ Ubiquitin primers for sirtuin knockout 5’- TAGAAACCACCACGAAGACGCA -3’ plants

Supplemental Information

The supplemental table 1 and 2 has been added as supplementary documents with this thesis to

Proquest with the following names.

Supplemental Table 1: Chapter4_SupplementalTable1_TotalMetabolomeAnalysis

Supplemental Table 2: Chapter4_SupplementalTable2_TotalAcetylomeAnalysis

105

CHAPTER 5. GENERAL CONCLUSION

Conclusions

Acetyl-CoA Synthetase catalyzes the adenosine-triphosphate (ATP) dependent activation of acetate to acetyl-CoA. It is a member of the AMP-forming acyl-CoA synthetase family, which can utilize a broad range of other carboxylate substrates. Acetyl-CoA acts as the starter molecule for fatty acid biosynthesis and various polyketide biosynthesis pathway. Engineering ACS to utilize novel carboxylate substrates would lead to generation of alternative acyl-CoA substrates for fatty acid or polyketide biosynthesis would provide the means to build a technology platform to produce a diversity of chemical products with wide ranging applications [1]. The

Arabidopsis ACS, that was the focus of this study only utilizes acetate efficiently as a substrate, and propionate to a much lesser extent. In Chapter 2, we extrapolated from experimentally determined 3-D structural data of two phylogenetically distantly related ACS enzymes, and modelled the potential carboxylate substrate binding pockets of two enzymes that also assemble thioester bonds but display different specificities toward the carboxylate substrate

(acetate or isobutyrate).

Rational mutagenesis experiments guided by this modeling were successful in initially introducing substrate promiscuity to the enzyme, and ultimately altering its substrate specificity. We have been particularly successful in shifting the substrate specificity of this eukaryotic ACS away from its native substrate (acetate), to be equally efficient in utilizing novel branched chain carboxylates that are not substrates of any of the guiding model enzymes.

The variant enzymes that we have generated are distinct in that they are not promiscuous, but 106 they show shifted substrate specificities, being equally efficient with the new substrates, as the wild-type enzyme is with its native substrate.

In a wide variety of organisms ACS activity is known to be regulated by reversible post-translational acetylation. In diverse phyla, the acetylation of ACS inactivates the enzyme

[2–5]. In contrast, the acetylation of a long chain acyl CoA synthetase in the diatom

Phaeodactylum tricornutum appears to lead to enzymatic activation [6]. In plant systems, such as Arabidopsis, the metabolic function of ACS is not completely clear. Unlike other biological systems, ACS is not the physiological source of acetyl-CoA that is the precursor of de novo fatty acid biosynthesis in plastids [7]; this is in spite the fact that ACS is plastid localized [8].

In Chapter 3, we explored if the acetylation-based regulation of ACS affects activity of this enzyme in plant systems. We identified the acetylation site for the plant ACS as occurring at Lys622. This identity is consistent with the position of this residue in a motif (A10) that is highly conserved among many ACS enzymes. The sequence homology-based identification of

Lys622 as the site of acetylation of the plant enzyme was experimentally confirmed by a combination of site directed mutagenesis and mass-spectrometric analyses. Moreover, it appears that in vitro the acetylation of the plant ACS is autocatalytic and does not require additional enzymatic components. Although the presence of this lysine is important in enzyme catalysis in the plant ACS, the acetylation of this residue does not significantly affect catalysis.

Among the deacetylases involved in removal of the acetyl group during reversible acetylation, there are two sirtuin family proteins in Arabidopsis (Arabidopsis thaliana) SRT1 and SRT2 [9–11]. Understanding the role of sirtuins in Arabidopsis would help identify the target proteins regulated through reversible acetylation. Specifically, in Chapter 4, we analyzed how change in the acetylation state of proteins led to metabolic shifts. This is crucial to 107 understanding the role, acetyl-CoA plays in plant regulatory processes. We found that both sirtuins, SRT1 and SRT2, were active proteins comparable in enzymatic activity and could be inhibited by nicotinamide. Hence, they are definitively sirtuin family deacetylase proteins.

Untargeted metabolomic analysis revealed 67 chemically identifiable compounds to be significantly different in abundance in sirtuin knockout mutants compared to wild-type. Total acetylome analysis identified 211 different proteins to be significantly differentially acetylated in sirtuin double knockout compared to wild-type. Specifically, individual srt1 and srt2 mutations affect distinct portions of the metabolome, however, the combination of the two mutations are not additive, but rather produces a different metabolome that appears to be more wild-type like. Some of these metabolomic effects may indicate that SRT1 and SRT2 potentially show genetic epistasis. The acetylome of plants lacking the 2 sirtuin deacetylases indicate that SRT1 and SRT2 affects the acetylation of both histone and non-histone proteins.

These non-histone acetylated proteins can be related to regulation of processes associated with a number of metabolic processes such as glucose metabolism, fatty acid metabolism, amino acid metabolism, and the TCA cycle. The parallel non-targeted metabolomic analyses of the sirtuin mutant plants reinforce these conclusions and identify key proteins whose acetylation status potentially effects specific metabolic processes in planta.

Hence, this dissertation this is an exemplary study of a research path to generate novel enzymes. Moreover, the novel acyl-CoA synthetase enzymes that were generated herein can serve as genetically encoded catalytic elements for bioengineering multiple metabolic pathways and potentially producing useful novel chemicals [1]. This dissertation also establishes that, plant acetyl-CoA synthetase is different in terms of how it reacts to the post translational modification of reversible acetylation. This could potentially provide a key to the 108 question, as to why this plastid localized protein is evolutionarily conserved in plants, even though the major player responsible for generation of acetyl-CoA in plastids, is pyruvate dehydrogenase complex [7]. Lastly, this dissertation investigates the role of two sirtuin deacetylase proteins in Arabidopsis. The dataset presented herein thus expands the role of sirtuin deacetylases in affecting plant metabolic processes.

References

[1] B.J. Nikolau, M.A.D.N. Perera, L. Brachova, B. Shanks, Platform biochemicals for a biorenewable chemical industry., Plant J. 54 (2008) 536–45. doi:10.1111/j.1365- 313X.2008.03484.x.

[2] V.J. Starai, I. Celic, R.N. Cole, J.D. Boeke, J.C. Escalante-Semerena, Sir2-dependent activation of acetyl-CoA synthetase by deacetylation of active lysine, Science (80-. ). 298 (2002) 2390–2392. doi:10.1126/science.1077650.

[3] T. Noy, H. Xu, J.S. Blanchard, Acetylation of acetyl-CoA synthetase from Mycobacterium tuberculosis leads to specific inactivation of the adenylation reaction, Arch. Biochem. Biophys. 550–551 (2014) 42–49. doi:10.1016/j.abb.2014.04.004.

[4] W.C. Hallows, S. Lee, J.M. Denu, Sirtuins deacetylate and activate mammalian acetyl- CoA synthetases, Proc. Natl. Acad. Sci. 103 (2006) 10230–10235. doi:10.1073/pnas.0604392103.

[5] J.G. Gardner, F.J. Grundy, T.M. Henkin, J.C. Escalante-Semerena, Control of acetyl- coenzyme A synthetase (AcsA) activity by acetylation/deacetylation without NAD+involvement in Bacillus subtilis, J. Bacteriol. 188 (2006) 6715. doi:10.1128/JB.01181-06.

[6] Z. Chen, L. Luo, R. Chen, H. Hu, Y. Pan, H. Jiang, X. Wan, H. Jin, Y. Gong, Acetylome Profiling Reveals Extensive Lysine Acetylation of the Fatty Acid Metabolism Pathway in the Diatom Phaeodactylum tricornutum., Mol. Cell. Proteomics. 17 (2018) 399–412. doi:10.1074/mcp.RA117.000339.

109

[7] J. Ke, R.H. Behal, S.L. Back, B.J. Nikolau, E.S. Wurtele, D.J. Oliver, The Role of Pyruvate Dehydrogenase and Acetyl-Coenzyme A Synthetase in Fatty Acid Synthesis in Developing Arabidopsis Seeds, Plant Physiol. 123 (2000) 497–508. doi:10.1104/pp.123.2.497.

[8] M. Lin, D.J. Oliver, The Role of Acetyl-Coenzyme A Synthetase in Arabidopsis, PLANT Physiol. 147 (2008) 1822–1829. doi:10.1104/pp.108.121269.

[9] F. Zhang, L. Wang, E.E. Ko, K. Shao, H. Qiao, Histone Deacetylases SRT1 and SRT2 Interact with ENAP1 to Mediate Ethylene-Induced Transcriptional Repression, Plant Cell. 30 (2018) 153–166. doi:10.1105/tpc.17.00671.

[10] A.-C. Konig, M. Hartl, P.A. Pham, M. Laxa, P.J. Boersema, A. Orwat, I. Kalitventseva, M. Plochinger, H.-P. Braun, D. Leister, M. Mann, A. Wachter, A.R. Fernie, I. Finkemeier, The Arabidopsis Class II Sirtuin Is a Lysine Deacetylase and Interacts with Mitochondrial Energy Metabolism, PLANT Physiol. 164 (2014) 1401–1414. doi:10.1104/pp.113.232496.

[11] X. Liu, W. Wei, W. Zhu, L. Su, Z. Xiong, M. Zhou, Y. Zheng, D.-X. Zhou, Histone Deacetylase AtSRT1 Links Metabolic Flux and Stress Response in Arabidopsis, Mol. Plant. 10 (2017) 1510–1522. doi:10.1016/j.molp.2017.10.010.

110

APPENDIX A. USING A STRATEGY THAT EXPANDS THE GENETIC CODE TO

GENERATE IN VIVO ACETYLATED LYS-622 MODIFICATION OF THE

ARABIDOPSIS ACETYL-COA SYNTHETASE

Introduction

The genetic code contains 61 codons that encode 20 canonical amino acids, and 3 stop codons. After translation, proteins are sometimes chemically modified at specific amino acid sites, and these post-translational modifications provide proteins with new functions and are often mechanisms for altering biological functions (i.e., regulation). A strategy for expanding the genetic code can be used to study such post-translational modification (PTM) of proteins

[1]. In this strategy, an already modified amino acid is inserted into a protein sequence at a targeted position during the in vivo synthesis of that protein. A number of different non- canonical amino acids have been incorporated using this strategy, including -acetyl-lysine

[2]. The advantage offered by this strategy is the ability to generate homogeneous preparations of modified proteins, and thus is a powerful technique to study PTMs.

The strategy uses an orthogonal pair of amino-acyl tRNA synthetase and its respective tRNA, and this pair are genetically programmed for expression in a host cell, typically

Escherichia coli. This pair of amino-acyl tRNA synthase and tRNA recognize a specific non- canonical amino acid, which is incorporated into a specific position of a protein during in vivo protein synthesis. The position of incorporation is defined by positioning a stop codon at the designated position of the mRNA, which is recognized by the anticodon of the suppressor tRNA that carries the non-canonical amino acid [3]. This technique has been previously used for the site-specific incorporation of acetyl-lysine in different proteins from different organisms, and extensively reviewed [1]. 111

One drawback of this strategy is that the acetylated lysine containing protein that is generated, can be deacetylated in vivo or during extraction and purification by native deacetylases present in the host organism. Although this can be somewhat overcome by the addition of nicotinamide (or other deacetylase inhibitors), as the host cells are being propagated or in the extraction buffer to inhibit deacetylase activity during purification of the acetylated protein. Another approach is to use analogs of acetyl-lysine, which cannot be deacetylated; for example thio-acetyl-lysine [5] and trifluoro-acetyl-lysine [6].

In this study, the strategy for the in vivo incorporation of acetyl-lysine was used to study the post-translational modification of ε-amine lysine by acetylation of Arabidopsis acetyl-CoA synthetase (atACS). Specifically, the non-canonical -acetyl-lysine was incorporated at position 622 (As identified in chapter 2 of this thesis) using this technique, and the acetylated protein (ACS-AcK) was expressed, purified and analyzed.

Methods

Wild-type atACS was expressed and purified as described in Chapter 1 of this thesis.

For the purpose of incorporating acetyl-lysine into the atACS sequence, the atACS ORF was sub-cloned into the pCDF-1b vector with a C-terminal His6 tag. The codon encoding for lysine-622 was mutated to the “TAG” stop codon, using Quikchange Lightning Site mutagenesis kit (Agilent Technologies, Santa Clara, CA) as per manufacturer’s protocol. The mutagenized clone was sequence confirmed. The orthogonal tRNA containing pTech plasmid was obtained from Dr. Chenguang Fan (Department of Chemistry & Biochemistry, University of Arkansas) and co-transformed with the pCDF-1b construct into BL21(DE3) cells [7,8].

The strain was grown overnight at 37 °C, with shaking at 250 rpm for aeration, in 5-

10 ml LB medium containing the antibiotics streptomycin (50 μg/ml) and chloramphenicol (50 112

μg/ml). The overnight culture was used to inoculate 0.25 L LB media containing the same antibiotics. These cultures were grown at 37 °C with shaking at 250 rpm, until an OD600 of

~0.6. Expression was then induced by the addition of IPTG to a final concentration of 0.4 mM.

The media was also supplemented with final concentrations of 20 mM nicotinamide and 2 mM Nε-acetyl-lysine (Sigma-Aldrich Co., St. Louis, MO). The culture was then grown at 22°C with shaking at 250 rpm for 24-48 hours.

Cells were harvested by centrifugation at 10,000g for 10 minutes, and resuspended in a buffer consisting of 20 mM HEPES-KOH, pH 7.5, 0.5 M NaCl, 20 mM nicotinamide, 5 mM imidazole, 0.1% (v/v) Triton X-100, 0.1 mg/ml phenyl methyl sulfonyl fluoride (PMSF), and

0.45 μm filter (Corning Inc., Corning, NY) and applied to a column containing 2ml PerfectPro

Ni-NTA agarose (5 Prime, Inc., Gaithersburg, MD). The column was washed with increasing concentrations of imidazole, and the atACS protein was eluted with 0.2 M imidazole. All the wash buffers had 20 mM nicotinamide. Samples were immediately dialyzed into 10 mM

HEPES-KOH, pH 7.5, 10 mM KCl, 2mM TCEP, 10% glycerol.

Chromatography

Size exclusion gel filtration chromatography was conducted with an AKTA FPLC system (GE Healthcare Life Sciences, Pittsburg, PA). A 100 µl aliquot of purified protein (4-

5 mg/ml) was injected into a prepacked Superdex 200 Increase 10/300 GL gel filtration column (GE Healthcare Life Sciences, Pittsburg, PA). The buffer used for FPLC was 10 mM

HEPES-KOH, pH 7.5, 10 mM KCl, 2mM TCEP, 10% glycerol. The flow rate of 0.4 113 ml/minute was used and protein elution was monitored using an UV absorbance detector at

280nm.

Protein analysis

Isolated protein preparations were evaluated by SDS PAGE and subjected to Western blot analysis, using an antibody (1:1000 diluted) that reacts with N-ε-acetyl-L-lysine (Cell

Signaling Technology, #9441). The spectrophotometric assays to measure ACS activity was as described in Chapter 1 of this thesis.

The acetylation status of ACS-Ack was also evaluated by mass spectrometric analysis of rLysC (Promega Corporation, Madison, WI) digested proteins using Q Exactive™ Hybrid

Following staining with Coomassie Brilliant Blue, the identified protein band was excised and digested using an Investigator™ ProGest (Genomic Solutions, Digilab Inc, Hopkinton, MA) in 0.5 mL buffer (50 mM Tris, pH 8) with LysC at a concentration of 2 µg/100 µg sample in the presence of 5 mM DTT and 15 mM iodoacetamide. The results were analyzed using

SequestHT against the ACS protein sequence, with four possible missed cleavage sites. The possible modifications analyzed were acetylation of Lys, carbamidomethylation of Cys, deamidation of Asn and Gln and oxidation of Met residues. cobB- strain

Bacterial cells such as E. coli express a native sirtuin deacetylases encoded by the cobB gene [9]. To eliminate the potential deacetylation by this endogenous sirtuin, cobB- mutant strain (JW1106) were obtained the E. coli Genetic Resources at Yale CGSC, The Coli Genetic 114

Stock Center. The JW1106 strain was co-transformed with pCDF-1b with atACS (with TAG stop codon at Lys622) plasmid and the pTECH plasmid. Because the resulting strain could not be successful lysogenized with Lambda DE3, the expression of ACS-AcK could not be conducted in this cobB- strain.

Results and Discussion

In order to incorporate acetyl-lysine (AcK) in to the atACS protein at position #622, a stop codon was introduced into the ORF sequence at this position. Thus, in the absence of

AcK incorporation, a truncated atACS protein would be expressed. However, when AcK is incorporated at position 622, the resultant ACS-AcK variant would also incorporate a C- terminal His6-tag, which would ensure protein purification via affinity chromatography with a nickel-column.

Based on this strategy, the purified protein recovered by Ni-column chromatography is assumed to be the ACS-AcK variant. SDS PAGE analysis of the recovered preparations identified two proteins bands, close to the expected 72-kDa band of atACS (Figure A1. A).

Size exclusion gel filtration chromatography of the protein preparation identified two UV- absorbing peaks (A and B). (Figure A1. B). These peaks correspond to molecular weights of

425 kDa and 105 kDa for Peak A and B respectively. The fractions under the two peaks were pooled, concentrated for further analysis by SDS-PAGE and western blot analyses, with anti- acetyl-lysine antibody and anti-His-tag antibody. The SDS-PAGE analysis showed the molecular weight of the proteins in the two peaks to be very close to each other, approximately

72kDa. Hence, based on this and the FPLC molecular weight analysis, protein in Peak A is in oligomeric state, whereas the protein in Peak B maybe either a monomer or dimer. 115

The ~72-kDa protein band in Peak B reacts with the anti His-tag antibody, whereas the protein band in Peak A does not react with this antibody. Additionally, these two proteins preparations were subjected to chymotrypsin digestion and resulting peptides analyzed for similarity to atACS protein sequence by mass-spectrometry. The Peak B peptide sequences showed 55% coverage of the atACS sequence, whereas Peak A peptides showed only 4.5% coverage of the atACS sequence (Figure A2 & A3). Additionally, enzymatic assays of Peak A and B preparations indicate that Peak A does not support ACS activity, whereas Peak B does support the catalysis of acetyl-CoA synthesis. Based on these results therefore, we conclude that Peak B is the ACS-AcK variant.

The LysC digested peptide analysis established that ACS-Ack (Peak B) is acetylated at lysine 622. The sequence coverage for wild-type ACS protein is 40.5% (Figure A5. A) while for the ACS-Ack Peak B it is ~62% (Figure A5. B). Although the lysine-622 containing peptide could not be identified from the wild-type protein, it was successfully detected from the ACS-

Ack protein. The acetylated lysine-622 peptide sequence was identified as,

TRSGKAcIMRRILRK based on the monoisotopic m/z value of 415.01038 Da (+0.76 mmu/+1.84 ppm), with a charge of +4 and theoretical mass of the MH+ ion of 1657.01967 Da

(Figure A6). In addition to the acetylated lysine-622 residue, an addition acetylated lysine residue was detected (Lys 201), but this was acetylated in both the wild-type ACS and the

ACS-Ack variant.

The specific activity of the recovered ACS-AcK is considerably less than the wild-type atACS enzyme. This result indicates that acetylation of Lys-622 of atACS results in the inactivation of catalytic capability. This conclusion is not consistent with the results reported in Chapter 2 of this thesis, which indicates that the auto-acetylated enzyme maintains catalytic 116 activity similar to that of the non-acetylated enzyme. Reduction in enzymatic activity has been observed in other proteins expressed and characterized using the genetic code expansion strategy [1,8]. Also, there is one potential explanation for this inconsistency, and further enzymatic characterizations are needed to test this possibility. Specifically, in the current experiment the non-canonical AcK residue was incorporated into a protein that carries a C- terminal His-tag, whereas the control wild-type atACS carries an N-terminal His-tag. Hence, comparison with the appropriate wild-type atACS control with a C-terminal His-tag needs to be conducted to get a definitive conclusion on whether acetylation reduces catalytic activity.

References

[1] H. Chen, S. Venkat, P. McGuire, Q. Gan, C. Fan, Recent Development of Genetic Code Expansion for Posttranslational Modification Studies., Molecules. 23 (2018). doi:10.3390/molecules23071662.

[2] J.A. Johnson, Y.Y. Lu, J.A. Van Deventer, D.A. Tirrell, Residue-specific incorporation of non-canonical amino acids into proteins: recent developments and applications, Curr. Opin. Chem. Biol. 14 (2010) 774–780. doi:10.1016/j.cbpa.2010.09.013.

[3] J. Xie, P.G. Schultz, A chemical toolkit for proteins — an expanded genetic code, Nat. Rev. Mol. Cell Biol. 7 (2006) 775–782. doi:10.1038/nrm2005.

[4] Z. Chen, L. Luo, R. Chen, H. Hu, Y. Pan, H. Jiang, X. Wan, H. Jin, Y. Gong, Acetylome Profiling Reveals Extensive Lysine Acetylation of the Fatty Acid Metabolism Pathway in the Diatom Phaeodactylum tricornutum., Mol. Cell. Proteomics. 17 (2018) 399–412. doi:10.1074/mcp.RA117.000339.

117

[5] S. Venkat, D.T. Nannapaneni, C. Gregory, Q. Gan, M. McIntosh, C. Fan, Genetically encoding thioacetyl-lysine as a non-deacetylatable analog of lysine acetylation in Escherichia coli, FEBS Open Bio. 7 (2017) 1805–1814. doi:10.1002/2211-5463.12320.

[6] F. Zhang, Q. Zhou, G. Yang, L. An, F. Li, J. Wang, A genetically encoded 19F NMR probe for lysine acetylation., Chem. Commun. (Camb). 54 (2018) 3879–3882. doi:10.1039/c7cc09825a.

[7] S. Venkat, C. Gregory, J. Sturges, Q. Gan, C. Fan, Studying the Lysine Acetylation of Malate Dehydrogenase, J. Mol. Biol. 429 (2017) 1396–1405. doi:10.1016/j.jmb.2017.03.027.

[8] S. Venkat, H. Chen, A. Stahman, D. Hudson, P. McGuire, Q. Gan, C. Fan, Characterizing Lysine Acetylation of Isocitrate Dehydrogenase in Escherichia coli, J. Mol. Biol. 430 (2018) 1901–1911. doi:10.1016/j.jmb.2018.04.031.

[9] V.J. Starai, I. Celic, R.N. Cole, J.D. Boeke, J.C. Escalante-Semerena, Sir2-dependent activation of acetyl-CoA synthetase by deacetylation of active lysine, Science (80-. ). 298 (2002) 2390–2392. doi:10.1126/science.1077650.

118

Figures

Figure A1: Purification of atACS-AcK variant A. Coomassie stained SDS-PAGE of the purified protein preparations. Both atACS-AcK and the wild-type atACS were with the His- tag which was fused to the C-terminus and N-terminus of these proteins, respectively. B. FPLC size exclusion gel filtration chromatography of three preparations of atACS-Ack variants.

Aliquots of each peak were analyzed by SDS-PAGE and stained with Coomassie Blue (shown under the profile). The indicated fractions were pooled as Peak A (Fraction A7-A10) and Peak

B (Fraction B10-B7) and subjected to further analysis.

119

Figure A2: Mapping of detected peptides from Chymotrypsin digest of Peak A protein, to

Arabidopsis ACS sequence. Coverage of detected peptides is 4.53% against Arabidopsis ACS sequence.

120

Figure A3: Mapping of detected peptides from Chymotrypsin digest of Peak B protein, to

Arabidopsis ACS sequence. Coverage of detected peptides is 55.14% against Arabidopsis ACS sequence.

121

Figure A4: Characterization of purified atACS-AcK preparations. A. SDS-PAGE analysis of

Peak A and Peak B, stained by Coomassie Brilliant Blue, or by western blot analysis using anti-acetyl-lysine antibody or anti-His-tag antibody. B. Acetate dependence of the ACS activity of the FPLC purified Peak C of the atACS-AcK variant.

122

Figure A5: Coverage map of LysC digested wild-type ACS (A) and ACS-Ack Peak B (B) showing detected peptides and modifications on specific amino acid. The green shade shows sequences of detected peptides. The modifications are indicated as acetylation: A, carbamidomethylation: C, deamidation: D and oxidation: O. Lysine 622 is indicated with an asterisk (*).

123

Figure A6: LC-MS/MS spectra and MS2 fragmentation pattern (Inset) of the peptide containing lysine 622 confirming acetylation for ACS-Ack protein.

124

Tables

Table A1: Michaelis-Menten kinetic parameters of atACS-AcK variant recovered in FPLC fractions encompassing Peak B compared to wild-type atACS, and the contaminating Peak A protein.

Wild-type ACS ACS Ack Peak B Peak A Km 0.18 ± 0.03 2.75 ± 1.46 NA kcat 15.05 ± 0.65 0.44 ± 0.12 NA Vmax 9.82 ± 0.43 0.29 ± 0.08 NA

Table A2: Sequences of DNA primers used in this study.

Primer Sequence (5’-3’) Experimental purpose AGGAGATATACCATGATGGCATCG ACSpCD1bF To sub-clone atACS construct GAGGAGAATGATC from pET24b into pCDF-1b CAGCAGCCTAGGTTATCAGTGGTG ACSpCD1bR vector GTGGTGGTGG TGCCAAAGACGAGAAGCGGATAG K622TAGF ATAATGAGGAGA To mutate atACS Lys622 to the TCTCCTCATTATCTATCCGCTTCTC stop codon TAG K622TAGR GTCTTTGGCA ACS_F4 GTTTCGAACTCTGTTTGGGG For DNA sequencing of ACS_R5 ATCCGCGAGTGTGCTAGTATC mutated atACS construct

125

APPENDIX B. SUBCELLULAR LOCALIZATION OF THE SRT1 PROTEIN IN

ARABIDOPSIS THALIANA

Introduction

The Arabidopsis genome contains two genes that encode for sirtuin deacetylase proteins, SRT1 and SRT2 [1–3]. SRT2 has been shown to dual localized in the nucleus [3] and the mitochondria [1]. In contrast, SRT1 has been recently shown to be localized in the nucleus [4].

Transgenic expression of 35S::SRT1-GFP construct in 6 day old Arabidopsis roots or by infiltrating leaves of N. benthamiana, indicates that SRT1 is localized solely in the nucleus.

We generated transgenic Arabidopsis plants carrying a 35S::SRT1-GFP to further analyze the sub-cellular localization of this protein.

Method

Construction of 35S::SRT1-GFP vector

The plasmid vector, pEARLEYGATE 103 was used to make the GFP constructs. SRT1

ORF sequence was amplified from SRT1-pET30f construct (described in Chapter 2 of this thesis). This SRT1 fragment (Table B1, Figure B1) was cloned into the XhoI site of the plasmid pEARLEYGATE 103 (Figure B2), using In-fusion cloning (Takara, Clontech, Mountain

View, CA). The final vector construct was identified via a PCR assay (Figure B3) and subsequently confirmed by sequencing of the construct. A 35S::GFP plasmid, in the pEARLEYGATE 103 vector was obtained from Xinyu Fu (Nikolau-group). These vectors were transformed into Agrobacterium tumefaciens, strain C58C1, which were subsequently used to perform Agrobacterium mediated transformation of the Arabidopsis wild-type Col-0 line [5]. Transgenic plants were identified based upon BAR-resistance and confirmed by PCR- 126 based genotyping. The sequences of the PCR primers used in these confirmation experiments are identified in (Table B1).

Co-localization control plants were generated by using vectors obtained from ABRC

(https://abrc.osu.edu); these being the organelle markers CD3-975 (Tonoplast) and CD3-1007

(Plasma membrane) (Table B2). These vectors were transformed into the either wild-type plants or plants already transformed with the 35S::SRT1-GFP vector. These transformed plants were then screened on MS-Agar medium with the appropriate resistances as mentioned in

Table B2. Seeds were collected from individual confirmed transgenic plants and propagated by single-seed descent and used for confocal microscopy. Confocal microscopy was performed using 15-20 days old seedling roots and leaves of T3 generation transgenic plants. The confocal imaging was done using Leica SP5 X MP confocal/multiphoton microscope system at the ISU Roy J. Carver High Resolution Microscopy Facility

(http://www.biotech.iastate.edu/biotechnology-service-facilities/hrmf/). The laser wavelengths [excitation (nm)/emission (nm)] were as following: 489/500 to 575 for GFP and

587/600 to 650 for chlorophyll autofluorescence.

Result and Conclusion

The GFP localization from the 35S::SRT1-GFP construct is localized to both the nucleus

(Figure B4. C-D) and the plasma membrane (Figure B4. E-F, Figure B5. A-C) of the leaf epidermal cells. However, the localization of GFP from the control 35S::GFP transgene also shows nuclear localization (Figure B4. A-B), which is consistent with prior reports that the non-targeted expression of the GFP protein from the 35S promoter localizes to both the cytosol and nucleus [6]. Hence additional experiments are needed to confirm that the SRT1-directed localization of GFP to the nucleus is because of the SRT1 sequence or is a result of the GFP 127 protein sequence. However, our results indicate that the SRT1-directed localization of GFP is uniquely directed to the plasma membrane, which is distinct from the localization of the control, untargeted GFP construct (Figure B4. A-B). Therefore, from these results we can conclude that SRT1 might localize to the plasma membrane.

References

[1] A.-C. Konig, M. Hartl, P.A. Pham, M. Laxa, P.J. Boersema, A. Orwat, I. Kalitventseva, M. Plochinger, H.-P. Braun, D. Leister, M. Mann, A. Wachter, A.R. Fernie, I. Finkemeier, The Arabidopsis Class II Sirtuin Is a Lysine Deacetylase and Interacts with Mitochondrial Energy Metabolism, PLANT Physiol. 164 (2014) 1401–1414. doi:10.1104/pp.113.232496.

[2] F. Zhang, L. Wang, E.E. Ko, K. Shao, H. Qiao, Histone Deacetylases SRT1 and SRT2 Interact with ENAP1 to Mediate Ethylene-Induced Transcriptional Repression, Plant Cell. 30 (2018) 153–166. doi:10.1105/tpc.17.00671.

[3] C. Wang, F. Gao, J. Wu, J. Dai, C. Wei, Y. Li, Arabidopsis putative deacetylase AtSRT2 regulates basal defense by suppressing PAD4, EDS5 and SID2 expression., Plant Cell Physiol. 51 (2010) 1291–9. doi:10.1093/pcp/pcq087.

[4] X. Liu, W. Wei, W. Zhu, L. Su, Z. Xiong, M. Zhou, Y. Zheng, D.-X. Zhou, Histone Deacetylase AtSRT1 Links Metabolic Flux and Stress Response in Arabidopsis, Mol. Plant. 10 (2017) 1510–1522. doi:10.1016/j.molp.2017.10.010.

[5] S.J. Clough, A.F. Bent, Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana., Plant J. 16 (1998) 735–43. http://www.ncbi.nlm.nih.gov/pubmed/10069079 (accessed February 13, 2019).

[6] C. Böhm, N.M. Seibel, B. Henkel, H. Steiner, C. Haass, W. Hampe, SorLA signaling by regulated intramembrane proteolysis., J. Biol. Chem. 281 (2006) 14547–53. doi:10.1074/jbc.M601660200.

128

Figures

Figure B1: Agarose gel electrophoresis of the PCR products obtained from the amplification of the SRT1 ORF to be cloned into pEARLEYGATE103 plant transformation vector. Lane 1 is molecular weight ladder. Lane 2, 3 and 4 are PCR products for SRT1 amplification. 129

Figure B2: Agarose gel electrophoresis of linearized pEARLEYGATE103 vector. Lane 1: molecular weight ladder. Lane 2 is pEARLEYGATE103 linearized with restriction enzyme

XhoI. Lane 3 is undigested control pEARLEYGATE103.

130

Figure B3: 1% Agarose gel for In-fusion cloning products for construction of 35S::SRT1-GFP

(pEARLEYGATE 103). First and last lane is molecular weight ladder. 2nd lane is negative control with no DNA. 3rd lane is empty pEARLEYGATE103. 4th lane is SRT1 positive control.

5th -14th lane is distinct colonies tested for successful cloning of SRT1 into pEARLEYGATE103.

131

Figure B4: Confocal microscopic examination of GFP localization in root and leaf tissues of

Arabidopsis seedlings. A-B. Control transgenic roots showing GFP localizing from the

35S::GFP constructs. C-D. GFP expression (green) in leaf epidermal cells from the

35S::SRT1-GFP transgene; green GFP-fluorescence is detected in the nuclei of leaf epidermal cells and yellow fluorescence indicates overlay of GFP fluorescence with red auto- fluorescence. E-F. 35S::SRT1-GFP transgenic plants showing localization of SRT1-GFP

(Green) in the plasma membrane of leaf epidermal cells. Red region indicates auto- fluorescence. Yellow region indicates overlay of GFP fluorescence with auto-fluorescence.

132

Figure B5: Confocal microscopic examination of fluorescence co-localization in root and leaf tissues of Arabidopsis seedlings. A-C. GFP (green) fluorescence from 35S::SRT1-GFP transgenic plants showing localization in the plasma membrane of leaf epidermal cells.

Red/Yellow region indicates auto-fluorescence. D-F. GFP (green) (E) and mCherry (red) (F) fluorescence in transgenic plants carrying both the 35S::SRT1-GFP and CD-1007 (the plasma membrane marker) transgenes. Panel D is the overlay of the green and mCherry markers. White arrows indicate regions where the SRT1-GFP and plasma membrane marker co-localize.

133

Tables

Table B1: DNA primer sequences used in this study

Name Sequence Description 5’- SRT1GfpF CATTTGGAGAGGACACGCATGTCTTT AGGTTACGCAGAG-3’ To amplify SRT1 5’- SRT1GfpR CTAGTCAGATCTACCATCTGCCTTGG TTTCTTCTGC-3’ At GFP at protein C terminal pEG103R 5’-TGGGACAACTCCAGTGAAAAGT-3’ region SRT1Fwd 5’-ACCCTGCGTGTTGCGAGTGTT-3’ Gene specific primer for SRT1

Table B2: List of organelle markers and SRT1-GFP constructs used to study the sub-cellular localization of SRT1

Name of construct Description Resistance Tonoplast-mCherry marker, CD3-975 Kanamycin 35S promoter Plasma membrane-mCherry CD3-1007 Kanamycin marker, 35S promoter SRT1protein with GFP 35S::SRT1-GFP Bar marker, 35S promoter 35S::GFP Only GFP marker Bar