CURRICULUM VITAE Saddef Haq, Ph.D. Program in Toxicology University of Maryland Baltimore School of Medicine

Home , Akashiwo sanguinea, Amoebophyra

Beyond the Dinoflagellate Transcriptome: Validation of Protein Production via Biochemical Analysis and Mass Spectrometry

Item Type dissertation

Authors Haq, Saddef

Publication Date 2018

Abstract Dinoflagellates are members of the Alveolata (meaning “with cavities”), a monophyletic group of single cell protists which includes apicomplexans and ciliates that exhibit a diverse mode of nutrition, ranging from predation to photo autotrophy to int...

Keywords bioluminescence; dinoflagellates; toxin synthesis; Acetyl-CoA Carboxylase; Dinoflagellida; Polyketide Synthases; Proteomics

Download date 27/09/2021 16:23:13

Link to Item http://hdl.handle.net/10713/8970

CURRICULUM VITAE Saddef Haq, Ph.D. Program in Toxicology University of Maryland Baltimore School of Medicine

Institute of Marine and Environmental Technology 701 E. Pratt Street Baltimore, Maryland 21202 Email: [email protected]

EDUCATION 2013-2018 University of Maryland, Baltimore, MD Program in Toxicology Doctor of Philosophy

2006-2008 Rutgers University, New Brunswick, NJ Bachelor of Science in Biochemistry

2003-2006 Union County College, Cranford, NJ Associates of Science in Biology

EMPLOYMENT HISTORY July 2013-Present Graduate Research Assistant Graduate Program in Life Sciences, Program in Toxicology University of Maryland, Baltimore, Baltimore, MD Lab of Dr. Allen R. Place at the Institute of Marine and Environmental Technology, Baltimore, MD.

Sep 2010-May 2013 Senior Research Technician Center for Infection and Immunity Columbia University, School of Public Health New York, NY

Mar 2009-Dec 2009 Laboratory Technician Unigene Laboratories Inc. Boonton, NJ

Oct 2008-Feb 2009 Sampling Laboratory Technician Imclone Systems Inc. Branchburg, NJ

HONORS AND AWARDS June 2016-June 2017 Ratcliffe Environmental Entrepreneurship Fellowship Stipend, tuition, and technology support Additional award to start company and develop product prototype

November 2015 Travel Award for US Harmful Algal Bloom Symposium

PUBLICATIONS Haq S., Oyler B.L., Williams E.P., Khan M.M., Goodlett D.R., Nita-Lazar A., Bachvaroff T.R., Place A.R. Investigating a Multi-Modular Polyketide Synthase in Amphidinium carterae. 2018. In Preparation.

Haq, S.; Bachvaroff, T. R.; Place, A. R. Characterization of Acetyl-CoA Carboxylases in the Basal Dinoflagellate Amphidinium carterae. Mar. Drugs 2017, 1–10.

Haq, S.; Bachvaroff, T. R.; Place, A. R. Phylogenetic Analysis of Acetyl CoA Carboxylases in Dinoflagellates. In 17th ICHA Proceedings; 2017, 138-141.

Rodriguez, J. D.; Haq, S.; Bachvaroff, T.; Nowak, K. F.; Nowak, S. J.; Morgan, D.; Cherny, V. V.; Sapp, M. M.; Bernstein, S.; Bolt, A.; Decoursey, T. E.; Place, A. R.; Smith, S. M. E. Identification of a Vacuolar Proton Channel that Triggers the Bioluminescent Flash in Dinoflagellates. PLoS One 2017, 12, 1–24.

Cai, P.; He, S.; Zhou, C.; Place, A. R.; Haq, S.; Ding, L.; Chen, H.; Jiang, Y.; Guo, C.; Xu, Y.; Zhang, J.; Yan, X. Two new karlotoxins found in Karlodinium veneficum (strain GM2) from the East China Sea. Harmful Algae 2016.

Bhat, N.; Tokarz, R.; Jain, K.; Haq, S.; Weatherholtz, R.; Chandran, A.; Karron, R.; Reid, R.; Santosham, M.; O’Brien, K. L.; Lipkin, W. I. A Prospective Study of Agents Associated with Acute Respiratory Infection among Young American Indian Children. Pediatr. Infect. Dis. J. 2013.

Tokarz, R.; Haq, S.; Sameroff, S.; Howie, S. R. C.; Lipkin, W. I. Genomic analysis of coxsackieviruses A1, A19, A22, enteroviruses 113 and 104: Viruses representing two clades with distinct tropism within enterovirus c. J. Gen. Virol. 2013, 94, 1995–2004.

Tokarz, R.; Hirschberg, D. L.; Sameroff, S.; Haq, S.; Luna, G.; Bennett, A. J.; Silva, M.; Leguia, M.; Kasper, M.; Bausch, D. G.; Lipkin, W. I. Genomic analysis of two novel human enterovirus C genotypes found in respiratory samples from Peru. J. Gen. Virol. 2013, 94, 120–127.

Tokarz, R.; Briese, T.; Morris, G.; Ideh, R.; Chimah, O.; Ebruke, B.; Desai, A.; Haq, S.; Sameroff, S.; Howie, S. R. C.; Lipkin, W. I. Serotype analysis of streptococcus pneumoniae in lung and nasopharyngeal aspirates from children in the Gambia by Masstag PCR. J. Clin. Microbiol. 2013, 51, 995–997.

Tokarz, R.; Firth, C.; Madhi, S. A.; Howie, S. R. C.; Wu, W.; Sall, A. A.; Haq, S.; Briese, T.; Ian Lipkin, W. Worldwide emergence of multiple clades of enterovirus 68. J. Gen. Virol. 2012, 93, 1952–1958.

PRESENTATIONS Haq S, Bachvaroff TR, Oyler B, Place AR. Acetyl-CoA Carboxylases in Dinoflagellates: Fueling the Polyketide Synthase Pathways. US Symposium on Harmful Algae. Baltimore, MD. November 11-17, 2017. Speed Talk and Poster Presentation.

Haq S, Bachvaroff TR, Oyler B, Place AR. Acetyl-CoA Carboxylases in Dinoflagellates: Fueling the Polyketide Synthase Pathways. Gordon Research Conference: Mycotoxins and Phycotoxins. Easton, MA. June 18-23, 2017. Poster Presentation.

Haq S, Bachvaroff TR, Oyler B, Place AR. Acetyl-CoA Carboxylases in Dinoflagellates: Fueling the Polyketide Synthase Pathways. Gordon Research Seminar: Mycotoxins and Phycotoxins. Easton, MA. June 17-18, 2017. Oral Presentation.

Haq S, Bachvaroff TR, Goodlett DR, Place AR. Acetyl-CoA Carboxylases in Dinoflagellates: Fueling the Polyketide Synthase Pathways. International Conference on Harmful Algae. Florianopolis, Brazil. October 9-14, 2016. Poster Presentation.

Haq S, Bachvaroff TR, Goodlett DR, Place AR. Acetyl-CoA Carboxylases in Dinoflagellates: Fueling the Polyketide Synthase Pathways. International Marine Biotechnology Conference. Baltimore, MD. August 29-September 2, 2016. Oral Presentation.

Haq S, Bachvaroff TR, Goodlett DR, Place AR. Acetyl-CoA Carboxylases in Dinoflagellates: Fueling the Polyketide Synthase Pathways. Oceans and Human Health Grantee Meeting. Research Triangle Park, NC. April 13-14, 2016. Poster Presentation.

Haq S, Bachvaroff TR, Williams E, Goodlett DR, Place AR. One Pathway-Two Products: A Polyketide Synthase in Dinoflagellates Results in Both Toxin and Fatty Acid Production. Mass Spectrometry School in Biotechnology and Medicine. Dubrovnik, Croatia. July 5-11, 2015. Poster Presentation.

Haq S, Bachvaroff TR, Williams E, Goodlett DR, Place AR. One Pathway-Two Products: A Polyketide Synthase in Dinoflagellates Results in Both Toxin and Fatty Acid Production. US Symposium on Harmful Algae. Long Beach, CA. November 15-19, 2015. Speed Talk and Poster Presentation.

Abstract

Title: Beyond the Dinoflagellate Transcriptome: Validation of Protein Production via

Biochemical Analysis and Mass Spectrometry

Saddef Haq, Doctor of Philosophy, 2018

Dissertation Directed By: Allen R. Place, PhD, Professor

University of Maryland, Center for Environmental Science

Institute of Marine and Environmental Technology

Dinoflagellates are members of the Alveolata (meaning “with cavities”), a monophyletic group of single cell protists which includes apicomplexans and ciliates that exhibit a diverse mode of nutrition, ranging from predation to photo autotrophy to intracellular parasitism. Dinoflagellates are both primary producers and consumers in the food web, sometimes at the same time, and best known for their dominant role in causing harmful algal blooms. In keeping with their unusual nuclear chromatin architecture

(condensed throughout the cell cycle) the degree to which dinoflagellates use transcriptional responses to alter mRNA expression levels appears limited. In addition, it appears dinoflagellates have extensive gene duplications for most biological processes.

For these reasons, the goal for this dissertation was to measure protein levels directly for important cellular phenotypes. We began with characterizing a proton selective channel

(Hv1) that was hypothesized to be involved in bioluminescence in Lingulodinium polyedrum with goals of elucidating function, location, and production of this protein.

Next, we focused on acetyl CoA carboxylases (ACC) which are responsible for the rate limiting step in fatty acid (FAS) and polyketide synthases (PKS). We find the presence of these proteins in thirteen dinoflagellates surveyed and show there is distribution in the

plastid and cytosol. In Amphidinium carterae we verify protein production with high performance liquid chromatography-coupled to tandem mass spectrometry (LC-MS/MS), otherwise known as shotgun proteomics as well as compare transcript and protein levels across a diel cycle. Lastly, we focused on a multi-modular polyketide protein in A. carterae with implications in toxin and fatty acid production due to the domain arrangement which is indicative of a noniterative trans-AT PKS process. We show successful inhibition of 14C acetate incorporation into toxins and fatty acids with the addition of the ketosynthase specific (KS) fatty acid inhibitor cerulenin. Partial protein production is verified by western blotting and LC-MS/MS analysis of KS peptides reveals a strong covalent bond with cerulenin addition. Taken together the work discussed in this thesis has resulted in a better understanding of three different cellular processes in dinoflagellates using a shotgun proteomic approach. Future work continuing this trend of protein characterization in dinoflagellates will help elucidate many uncharacterized pathways in dinoflagellate biology.

Beyond the Dinoflagellate Transcriptome: Validation of Protein Production via Biochemical Analysis and Mass Spectrometry

by Saddef Haq

Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, Baltimore in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2018

DEDICATION

To my mother Shaheena Haq, my grandmother Ghulam Sakina, and my husband Joseph Villari.

iii

ACKNOWLEDGEMENTS

I would first like to thank my PI Dr. Allen Place for letting me join his lab which allowed me to work on this interesting project. I would also like to thank him for always being available and letting me stop by whenever I had questions or needed suggestions, I couldn’t have gotten through this process without you. I would also not have been able to do this without Dr. Tsvetan Bachvaroff who is not only on my committee but also has helped me with every publication including this dissertation. Thank you for always being there to help me and answer my questions no matter how swamped your schedule was.

Dr. David Goodlett for letting me join his group where I always received great feedback and I was able learn about mass spectrometry. I would like to thank the rest of my committee members Dr. William Randall and Dr. Mariusz Karbowski for their feedback.

Special thanks for my readers Dave and Mariusz for putting in the time to go through this whole dissertation. I would also like to thank Dr. Squibb who served as the director of the

Toxicology program when I first started and helped guide me through classes and rotations.

A huge thanks to my lab members Ernest Williams, Mary Larkin, and Ben Oyler for always keeping lab conversations fun and interesting and always being there to answer any lab or life related questions. Past lab member Grant Jones for teaching me all the techniques I learned and giving me a crash course in dinoflagellate biology. The Goodlett lab members Mohsin Khan, Will Fondrie, Shivangi Awasthi, Tao Liang, Lisa Leung and former members Young Ah Goo and Bao Tran for my proteomics training.

I was lucky enough to conduct my research at the Institute for Marine and

Environmental Technology (IMET) which has been the greatest of experiences. I would like to thank Nick Hammond and Russell Hill for implementing the REEF program which provided a year of funding and invaluable training in entrepreneurship. Carol

Ratcliffe for generously providing the funding to make this program possible. The administrative staff here that has been wonderful in taking care of all our paperwork and travel Monica Chacon, Sam Horn, and Chris Wilson.

At UMB I would like to thank Linda Horne for being our program coordinator for the first 4 years of my PhD, you made everything so easy for us! I would also like to thank my classmates Stephanie Lehman and Jimena Dancy we started together in core course and I am so glad we are still friends and I hope we continue our tradition of pedicures and dinners! I would also like to thank Jasmine Jacobs you went from being my trainer who I didn’t care much for to a close friend and you kept me from being a lazy bum. Mary Larkin gets a second shout out for becoming one of my closest friends in addition to my lab-mate and somehow, we reminded each other to just barely get our paperwork in on time to make it this far.

My friends who I have known for most of my life Liora Engel-Smith and Nadia

Chaudhry for always being encouraging and not minding my venting sessions. My

Nikovic girls, Dija, Selvie, and Flora for always being a fun time when I come and visit.

Lastly, I would like to thank all of my family. My husband Joseph Villari for not only encouraging me through this whole process but enduring a torturous commute for years while I finished classes. I love you! To the Villari family for always being supportive. A huge thanks to my dad Ikram Haq and my siblings Roaida Haq, Furqan v

Haq, and Faizan Haq and the new siblings I acquired along the way Sabaa Alvi and

Bismarck Koranteng. The newest family member Samir Koranteng for being the best and cutest distraction during this writing process. An extra special thanks for my mom

Shaheena Haq and my grandma Ghulam Sakina for all your sacrifice in raising us and I hope this makes you proud!

Table of Contents CHAPTER 1: INTRODUCTION AND BACKGROUND ...... 1 1.1 Dinoflagellates ...... 1 1.2 Dinoflagellates and Harmful Algal Blooms ...... 6 1.3 Amphidinium carterae as a Model Dinoflagellate ...... 11 1.4 Acetyl Co-A Carboxylases ...... 12 1.5 Polyketide Synthases ...... 14 1.6 Bioluminescence ...... 19 1.7 Project Goals ...... 20 CHAPTER 2: IDENTIFICATION OF A VACUOLAR PROTON CHANNEL THAT TRIGGERS THE BIOLUMINESCENT FLASH IN DINOFLAGELLATES ..... 22 2.1 Abstract ...... 22 2.2 Introduction ...... 23 2.3 Results ...... 24 2.4 Discussion ...... 46 2.5 Materials and Methods ...... 52 CHAPTER 3: PHYLOGENETIC ANALYSIS OF ACETYL COA CARBOXYLASES IN DINOFLAGELLATES ...... 61 3.1 Abstract ...... 61 3.2 Introduction ...... 62 3.3 Results and Discussion ...... 62 3.4 Materials and Methods ...... 66 CHAPTER 4: CHARACTERIZATION OF ACETYL COA CARBOXYLASES IN THE BASAL DINOFLAGELLATE AMPHIDINIUM CARTERAE ...... 69 4.1 Abstract ...... 69 4.2 Introduction ...... 69 4.3 Results ...... 71 4.4 Discussion ...... 76 4.5 Materials and Methods ...... 79 CHAPTER 5: INVESTIGATING A MULTI-DOMAIN POLYKETIDE SYNTHASE IN AMPHIDINIUM CARTERAE ...... 86 5.1 Abstract ...... 86 5.2 Introduction ...... 87 5.3 Results ...... 90 5.4 Discussion ...... 105 5.5 Materials and Methods ...... 109 CHAPTER 6: SUMMARY AND IMPLICATIONS OF THE RESEARCH ...... 117 6.1 Main Findings ...... 117 6.2 Future work characterizing acetyl CoA carboxylases ...... 120 6.3 Future work characterizing PKSs and PKS/NRPSs in Amphidinium carterae ...... 120

vii

6.4 Future work characterizing interacting partners of PKSs in Amphidinium carterae ...... 122 6.5 Closing Remarks ...... 124 CHAPTER 7: REFERENCES ...... 125

viii

List of Tables

Table 2.1: Primers and predicted product sizes for qPCR...... 44 Table 2.2: Sequence of longest open reading frame of contig 26784 in MMETSP1032 assembly, and sequence of codon-optimized gene of LpHV1...... 45 Table 4.1: Primer sequences for mRNA quantification experiments...... 82 Table 5.1: Percent incorporation of 14C acetate into total lipids, sulpho-amphidinol, and amphidinol...... 94 Table 5.2: Primer sequences used in RT-PCR experiments...... 112 Table 5.3: Epitope sequence of antibodies synthesized by GenScript as well as concentrations of primary and secondary antibodies used in western blotting experiments...... 114 Table 5.4: Experimental set up of synthetic peptide experiments analyzed by mass spectrometry...... 115

List of Figures

Figure 1.1: Model of Evolutionary Characteristics of Dinoflagellates ...... 4 Figure 1.2: Distinct features of dinoflagellate nuclear, plastid, and mitochondrial genomes...... 5 Figure 1.3: Suite of toxins produced by HAB species...... 10 Figure 1.4: Structures of Karlotoxins and Amphidinol 3 ...... 11 Figure 1.5: Fatty Acid Synthesis Pathway ...... 14 Figure 1.6: Schematic of a Type I Polyketide Synthase ...... 16 Figure 1.7: Iterative and Noniterative Polyketide Synthases ...... 18 Figure 1.8: Luciferase Binding Protein and mRNA Rates over Time ...... 20 Figure 2.1: LpHV1 is a voltage gated proton channel...... 26 Figure 2.2: LpHV1 is a proton selective channel...... 28 Figure 2.3: LpHV1 exhibits classical pH dependent gating...... 30 Figure 2.4: LpHV1 exhibits pH dependent gating...... 32 Figure 2.5: Voltage and pH dependence of LpHV1 activation kinetics...... 34 Figure 2.6: LpHV1 distribution is consistent with scintillon localization...... 36 Figure 2.7: LpHV1 localizes to the scintillon...... 38 Figure 2.8: Isolated scintillon preparations contain peptides matching the LpHV1 sequence...... 40 2+ Figure 2.9: Zn inhibits LpHV1 proton currents and scintillon luminescence...... 41 Figure 2.10: Control Fluorescence Images for Figure 2.6 ...... 42 Figure 2.11: Control Fluorescence Images for Figure 2.7 ...... 43 Figure 2.12: Alignment of predicted amino acid sequences of putative dinoflagellate HV1s...... 43 Figure 3.1: Phylogenetic Analysis of Acetyl CoA Carboxylases in Thirteen Dinoflagellate Species...... 65 Figure 3.2: Western blotting of natively biotinylated proteins in three dinoflagellate species...... 66 Figure 4.1: The expressed acetyl CoA carboxylases (ACCs) in Amphidinium carterae (CCMP 1314)...... 73 Figure 4.2: Quantification of ACC mRNA in Amphidinium carterae...... 74 Figure 4.3: Streptavidin Western blotting of Amphidinium carterae...... 75 Figure 4.4: Shotgun proteomic analysis of ACCs in Amphidinium carterae...... 76 Figure 5.1: Structures of three polyenoic fatty acids synthesized by dinoflagellates...... 87 Figure 5.2: Sterolysin compounds synthesized by different dinoflagellate species...... 89 Figure 5.3: A multi-modular polyketide synthase (PKS) in Amphidinium carterae...... 91 Figure 5.4: Thin layer chromatography (TLC) of extracted lipids...... 92 Figure 5.5: Quantitation of 14C acetate incorporation into lipids and amphidinol over a 2 hour period after cerulenin addition...... 94 Figure 5.6: HPLC UV absorbance at 280 nm displaying amphidinol and sulpho-amphidinol (red trace) overlaid with 14C acetate incorporation dpm (black trace)...... 96 Figure 5.7: Total toxin and lipid levels remain steady in A. carterae and A. sanguinea over a 2 hour period...... 97 x

Figure 5.8: Multi-Modular PKS transcript abundance across diel cycle...... 98 Figure 5.9: Western blotting analysis of proteins in a multi-domain polyketide synthase (PKS) in Amphidinium carterae...... 99 Figure 5.10: Mass spectra of synthetic peptide, DTACSS, alone and coupled with cerulenin, a cerulenin analog, and iodoacetamide in vitro...... 101 Figure 5.11: Tandem mass spectra (CID) of KS peptide reacted with cerulenin and the shorter cerulenin analog...... 102 Figure 5.12: Sequence alignment of the three KS domains in the multi-modular PKS in A. carterae...... 103 Figure 5.13: Sequence alignment of the three KS domains in the multi-modular PKS in A. sanguinea...... 104 Figure 5.14: Structure of cerulenin analog that has a shorter side chain than cerulenin. 105 Figure 6.1: Multi-Modular PKS Model in Amphidinium carterae ...... 122 Figure 6.2: Distribution of ATs in Amphidinium carterae...... 124

LIST OF ABBREVIATIONS

ACC, Acetyl CoA carboxylase ACP, Acyl carrier protein AT, Acyl transferase BC, Biotin carboxylase BCCP, Biotin carrier protein BSA, Bovine serum albumin C-CTX, Caribbean ciguatoxin cDNA, Complementary DNA CID, Collision induced dissociation CT, Carboxyl transferase cTP, Chloroplast transit peptide DAD, Diode array detector DAPI, 4',6-Diamidino-2-Phenylindole, Dihydrochloride DEPC, Diethyl pyrocarbonate DH, Dehydratase DHA, Docosahexaenoic acid DTT, Dithiothreitol

EH, Nernst potential for H+ EPA, Eicosapentaenoic acid ER, Enoyl reductase ESAW, Enriched seawater artificial water ESI, Electrospray ionization FAS, Fatty acid synthase GST, Glutathione s-transferase H+, Hydrogen ion HAB, Harmful algal bloom HEK, Human embryonic kidney HLP, Histone like protein HPLC, High performance liquid chromatography HRP, Horseradish peroxidase Hv1, Proton selective channel IT-TOF MS, ion trap-time of flight mass spectrometer ITS, Internal transcribed spacer kDa, Kilo Dalton kHv1, K. veneficum proton selective channel KmTx, Karlotoxin KR, Ketoreductase xii

KS, ketosynthase LBP, Luciferin binding protein LCF, Luciferase enzyme LC-MS/MS, high performance liquid chromatography – tandem mass spectrometry LpHv1, Lingulodinium polyedrum proton selective channel LSU, Ribosomal large subunit mRNA, Messenger RNA MS, Mass spectrometry MS/MS, Tandem mass spectrometry mV, millivolt NRPS, Non-ribosomal peptide synthetase ORF, Open reading frame P-CTX, Pacific ciguatoxin PBS, Phosphate Buffered Saline PCR, Polymerase chain reaction PFA, Paraformaldehyde PKS, Polyketide synthase PUFA, Polyunsaturated fatty acid PVDF, Polyvinylidene difluoride qRT-PCR, Reverse transcription quantitative polymerase chain reaction RPL7, Ribosomal protein L7 RT-PCR, Reverse transcription polymerase chain reaction SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis SDS, Sodium dodecyl sulfate SL, Spliced leader TCEP, Tris(2-carboxyethyl)phosphine TE, Thioesterase TLC, Thin layer chromatography TM, Melting point TMA, Tetramethylammonium TOFA, 5-(Tetradecyloxy)-2-furoic acid UTR, Untranslated region

xiii

CHAPTER 1: INTRODUCTION AND BACKGROUND

1.1 Dinoflagellates

Dinoflagellates (Dinophyta) are a largely planktonic division of motile unicellular microalgae that have two flagella. They can be found in both freshwater and marine environments and include autotrophs, heterotrophs, parasites, and mixotrophs. Generally, dinoflagellates are a more significant component of the phytoplankton in warmer waters and are a major source of eukaryotic marine carbon fixing [1–3].

Dinoflagellates are most famous for their harmful algal blooms (HAB) also known as red tides which result in beach and fisheries closures yearly [4]. However, they have equally beneficial qualities. The dinoflagellate species Symbiodinium exists in a symbiotic relationship with corals. The loss of this symbiosis results in coral bleaching and is often caused by warming water temperatures [5]. In the commercial market you may have encountered products purified from dinoflagellates. Docosahexaenoic acid

(DHA), is an important additive to baby formulas and is often purified from the dinoflagellate Crypthecodinium [6].

Nearly half of the 4000 described dinoflagellate species are photosynthetic [7]. The majority of photosynthetic species contain a plastid with three chloroplast membranes and the characteristic pigment peridinin, these are referred to as peridinin plastids [8].

While chlorophyll c is found in other microalgae, peridinin is a carotenoid produced only by dinoflagellates. Peridinin has a 37 carbon backbone, rather than the typical 40 carbons of most carotenoids. Other dinoflagellate species contain plastids derived from green algae [9], and haptophytes [10], while some dinoflagellates also contain entire diatoms

[11], and cryptophyte plastids that are only maintained for a span of months [12,13].

Interestingly, peridinin dinoflagellates moved most of their plastid genome to the nucleus only retaining only a handful of genes that are found on plasmid-like minicircles [14,15].

Dinoflagellates along with apicomplexans and ciliates are a members of the superphylum Alveolata based on the shared character of cortical alveoli [16] . Within alveolates, apicomplexans and dinoflagellates are sister to one another. Dinoflagellates and related ‘Marine Alveolates’ are extremely diverse and abundant in environmental clone libraries [17]. Within the large clade of dinoflagellates there are two major groups, the core dinoflagellates and the syndineans, with several smaller groups such as the X- cells, chromerids and others appearing outside of these two major groups [18–22]. The feature differentiating core dinoflagellates from other eukaryotes is the unusual nucleus, often referred to as the dinokaryon. With features such as condensed chromosomes during interphase and lack of nucleosomes, the dinokaryon is distinctly different from typical eukaryotic nuclei (Figure 1.1) [2,23]. In the dinokaryon the chromosomes exist in a condensed, liquid crystal state throughout the cell cycle. Instead of packaging the DNA in nucleosomes, dinoflagellates appear to have replaced the bulk histone function with two non-eukaryotic proteins, the dino-viral like protein (DVNP) and a histone like protein (HLP) [19,24–26]. The chromosome interior is densely packed with DNA, with very low levels of protein, some RNA, and a large number of Mg and Ca ions [23]. The chromosomes appear to be too dense for transcription to occur on the interior, instead genes located on the extrachromosomal loops associated with HLPs are probably transcriptionally active (Figure 1.2) [27–29].

Other than these structural anomalies of the nucleus, dinoflagellates further distinguish themselves from other alveolates with high DNA content. Core

dinoflagellates have a large number of chromosomes and haploid DNA content ranging from 1.5 pg of DNA per cell in the coral symbiont Symbiodinium to 189 pg in the bioluminescent species Lingulodinium polyedrum [23,30]. Due to a DNA content often orders of magnitude larger than the human genome (3 pg of DNA per haploid cell), full genome sequencing is very difficult. To date only draft genomes of Symbiodinium and the syndinean Hematodinium have been published [31–33]. Most sequencing work has been focused on transcriptomes, although with the advances in sequencing technologies as well as lower costs more dinoflagellate genomes could be sequenced.

These anomalous features continue down to the most basic unit of information,

DNA. In dinoflagellates up to 70% of thymine bases are replaced with 5 hydroxymethyl uracil normally found in eukaryotic cells only during oxidative stress but otherwise more commonly found in bacteria and bacteriophage [2,34–36]. At the transcript level, a 22 nucleotide spliced leader (SL) sequence is added to the 5’ end of all messenger RNA

(mRNA) transcripts in dinoflagellates (Figure 1.2) [37,38]. This feature can confirm transcripts are complete and do in fact belong to dinoflagellates rather than contaminants or bacteria that may be in the culture. Some highly expressed genes are encoded in large tandem arrays with short intergenic spacers and are trans-spliced while moderately expressed genes tend to be intron rich and are encoded by a single gene [39,40].

In addition to the unusual nuclei, dinoflagellates also have a number of genetic features that make it difficult to apply techniques from genetically tractable organisms.

Figure 1.1: Model of Evolutionary Characteristics of Dinoflagellates

Figure denotes character states for dinoflagellates mapped onto a phylogeny. Some of the key features seen in the core dinoflagellates are spliced leader trans-splicing of mRNA, loss of nucleosomes, presence of histone like proteins (HLP), liquid crystalline chromosomes, and the presence of 4-methyl sterols and dinosterol. Representative morphotypes are pictured in the upper right. Figure from Janouskovec et al. 2016 [41].

Figure 1.2: Distinct features of dinoflagellate nuclear, plastid, and mitochondrial genomes.

Model showing the liquid crystalline chromosome structure and peripheral loops on which transcription occurs. B) Model showing trans-splicing of mRNA messages as well as tandem gene arrays, and poly- adenylation which completes the mature mRNA transcript. C) Model showing fragmented mitochondrial genome. D) Plastid minicircles found in dinoflagellate plastids and the genes associated with them. Figure from Wisecaver and Hackett 2011 [2].

1.2 Dinoflagellates and Harmful Algal Blooms

Harmful algal blooms (HABs) occur when large numbers of algae in the water produce a toxic substance or create a noxious environment for other life. Diatoms and cyanobacteria can be responsible for toxic blooms, but perhaps the largest contributor to marine harmful algal blooms worldwide are dinoflagellates [2,4,42]. A large contributor to HABs and the increasing number of observed HABs is eutrophication, the input of excessive nutrients into water systems often as a result from runoff from urban and agricultural environments.

The excess nutrients in the water enable HAB species to grow rapidly resulting in high biomass algal blooms [43]. These blooms can cause economic and public health disruptions as a result of fish kills, or human and animal illness from consuming contaminated seafood

[1,44,45]. Not all HAB species result in a high biomass bloom, Gambierdiscus toxicity is a result of bioaccumulation in the food chain, not high cell numbers. This is important to keep in mind since Gambierdiscus is implicated in the largest number of human health cases [45].

Ciguatera is a common algal-derived human morbidity. The compound ciguatoxin produced by the dinoflagellate species Gambierdiscus bioaccumulates in the food chain and when humans or other organisms consume contaminated fish ciguatera poisoning may occur [45]. Gambierdiscus grow attached to microalgae and corals in tropical and sub- tropical climates. Fish feed on the macroalgae and ingest the Gambierdiscus, thus accumulating ciguatoxin from the dinoflagellates. The toxins can pass up the food chain to larger carnivorous fish such as Barracuda, Snapper, and Amberjacks at levels high enough to cause human illness [45]. Ciguatoxins are polycyclic polyether compounds (Figure 1.3) that activate human voltage gated sodium channels (VGSC) leading to gastrointestinal,

neurological, and cardiovascular disturbances [46–48]. Due to bioaccumulation in humans, ciguatera symptoms can persist from days to months depending on the amount of toxin ingested [45]. The number of congeners of not only ciguatoxin, but most dinoflagellate toxins, are difficult to describe as new congeners are constantly being discovered with slightly different structures and varying potencies [49]. For example Pacific ciguatoxin P-

CTX-1 can cause illness at levels of 0.1 g/kg in the flesh of fish while Caribbean ciguatoxin C-CTX-1 is 10 fold less toxic because it is more polar and easier to detoxify

[50]. In areas where ciguatoxins are prevalent monitoring programs are in place to prevent contaminated fish from reaching markets for consumption [49,50].

Harmful algal blooms are often referred to as ‘red tides’ describing the red tinge the water takes on when a high biomass bloom occurs. Karenia brevis causes red tides along the Florida Gulf of Mexico coast in the United States [51]. It can also be found in the

Caribbean Sea and along the coasts of New Zealand [52]. Karenia brevis produces brevetoxins, which is also a cyclic polyether compound (Figure 1.3) that accumulates in shellfish and can induce neurotoxic shellfish poisoning when the contaminated shellfish is consumed. Levels above 80 g/kg per 100g of shellfish tissue is considered toxic and results in closures of shellfish beds [52]. Brevetoxins bind to site 5 on VGSCs resulting in neuro-excitation which leads to gastrointestinal symptoms such as vomiting and diarrhea and neurological symptoms such as confusion and slurring of speech [52,53]. In addition to toxicity via consumption, brevetoxins are also known to cause respiratory effects when brevetoxins become aerosolized in the surf during red tides irritating the throat and upper respiratory tract of beachgoers during an active bloom. Animal illnesses are also common

during Karenia blooms with recorded deaths of bottlenose dolphins and manatees due to brevetoxin ingestion of contaminated organisms [52].

Not all HAB species cause human illness, but that doesn’t mean they are any less disruptive. Karlodinium veneficum, a bloom forming dinoflagellate species with a worldwide temperate distribution has been found in the British Isles, East China Sea, the

Chesapeake Bay, and Australia [1]. While this species is not implicated in human illness it can cause large-scale fish kills leading to economic disruption in the fishing industry. A local example occurred at HyRock Fish Farm where 15,000 reciprocal hybrid striped bass were lost to a bloom that was attributed largely to K. veneficum with additional events the following year with a 8,000 adult fish and 5,000 fingerling fish loss [1]. The karlotoxins

(KmTx) K. veneficum produces (Figure 1.4) are ichthyotoxic compounds that complex preferentially with membrane sterols such as cholesterol and ergosterol resulting in cell lysis due to uncontrolled movement of ions and water across the membrane. These toxins are termed sterolysins and are found in several dinoflagellate genera. Self-protection is provided by gymnodinosterol, a 4-methyl sterol, which does not interact with karlotoxins

(Figure 1.1) [54]. Fish are more susceptible to these toxins because their respiratory organs, the gills, are on the outside of the fish. Lysis of gills can cause suffocation, often compounded by hypoxia during bloom decline [55]. At least 15 karlotoxin congeners have been described with more variations being discovered consistently. Karlotoxins as a general structure contain a polar polyol arm, a hinge region, and a lipophilic arm (Figure

1.4). Toxin potency seems to be mediated by the length of the lipophilic arm. KmTx-1, with an 18-carbon lipophilic arm, is 10-fold more potent than KmTx-2 which has a 16 carbon lipophilic arm [1]. KmTx2, the more commonly observed congener has a 63 C

backbone and a Cl at the end of the lipophilic arm (Figure 1.4) [56]. A study of 11 KmTx2 producing Karlodinium strains had toxin levels varying from 0.04-4.38 pg/cell showing that strains can not only have varying congeners of toxin but also levels [57]. Sheepshead minnow showed mortality within an hour at KmTx2 concentrations of 0.5 g/mL while juvenile mortalities were observed between 0.1-0.5 g/mL of KmTx2 [58]. Mouse bioassays showed no mortality at 500 g/kg of KmTx2 via intraperitoneal injection, showing that fish are more sensitive to these karlotoxins due to their respiratory anatomy

[59].

Figure 1.3: Suite of toxins produced by HAB species.

Structures of a variety of HAB toxins produced by dinoflagellates, diatoms, bacteria, and cyanobacteria. Brevetoxin produced by Karenia brevis is responsible for red tide events in Florida and the Gulf of Mexico. Ciguatoxin produced by Gambierdiscus is responsible for Ciguatera fish poisoning in tropic and sub-tropic climates. Figure from Kalaitzis et al. 2010 [60].

Figure 1.4: Structures of Karlotoxins and Amphidinol 3

Structures of two different karlotoxins and amphidinol 3 with three distinct regions, the polyol arm, the hinge region, and the lipophilic arm that varies in carbon length. A) Structure of karlotoxin 1 B) Structure of karlotoxin 2 C) Structure of amphidinol 3. Figure from Deeds et al. 2015 [55]. A majority of HAB toxins are synthesized through polyketide synthases, which have been studied to date mostly through isotope labeling studies, however the exact molecular mechanisms of toxin synthesis have still not been elucidated [61].

1.3 Amphidinium carterae as a Model Dinoflagellate

In biology, model organisms such as mice, yeast, and mammalian cell lines are often used as a model to infer the biology of more complex, expensive, or ethically challenging

organisms. In dinoflagellates Amphidinium carterae is often used as a model to study other dinoflagellates [62]. A. carterae is a dinoflagellate with worldwide distribution that falls at the base of the group referred to as the ‘core dinoflagellates’ meaning it has all the nuclear characteristics dinoflagellates discussed in section 1.1 [2,41]. It produces cytotoxic compounds including amphidinols, amphidinolides, and macrolide compounds [63].

Amphidinols are structurally almost identical to karlotoxins and have the same mechanism of toxicity (Figure 1.4) [64]. Self-protection in Amphidinium however, is obtained with the

4-methyl sterol amphisterol rather than gymnodinosterol found in Karlodinium [23,54].

They are able to synthesize these structurally similar toxins, despite having different plastid types. Karlodinium contains a haptophyte plastid while Amphidinium has a peridinin plastid [2]. Despite these differences there is enough genetic and toxin similarity between these two species that we use Amphidinium as a model species to study Karlodinium.

Smaller genome size, fewer gene duplications, ability to grow axenic cultures, quick doubling time, and high final cell density are additional reasons why A. carterae is utilized as a model species. These features allow for more options than would be possible if working on a slower growing species allowing us to gain insights applicable broadly to other dinoflagellates.

1.4 Acetyl Co-A Carboxylases

Acetate is the 2-C building block used not only to build polyketides but primary metabolites like fatty acids as well. An often rate-limiting step to incorporating acetate into lipids or toxins is acetyl CoA carboxylase (ACC). Acetyl CoA carboxylases are biotinylated proteins responsible for the initial rate-limiting conversion of acetyl CoA to malonyl CoA in both FAS and PKS pathways. Once this conversion occurs, the activated

malonyl CoA enters either the FAS or PKS pathways (Figure 1.5) [65]. ACC requires biotin carboxylase (BC), carboxyl transferase (CT), and the biotin carrier protein (BCCP) domains. In eukaryotic homomeric ACCs these domains are found as a large polypeptide including a central domain, while typical eubacterial ACC are composed of several individual proteins, each with one or more domains. Cellular location of these proteins varies depending on the organism. In plants and algae ACC is either compartmentalized to the plastid, found in the cytosol, or is found in both locations [65–67]. During lipid synthesis in plastid containing organisms, the initial conversion of acetyl CoA to malonyl

CoA, by ACC, as well as elongation of the acyl chain to C16 occur in the plastid. Later steps of either elongation beyond C16, formation of triacylglycerides (TAG), or desaturation of fatty acids occurs in the endoplasmic reticulum of algae, but has not been as thoroughly researched as lipid synthesis in animals and plants [67]. Toxins and other

PKS products are also thought to be synthesized in the plastid where there is an abundance of acetyl CoA [67].

Figure 1.5: Fatty Acid Synthesis Pathway

A single cycle of FAS elongation starts with acetyl CoA carboxylase (ACC) which converts acetyl CoA to malonyl CoA. The carboxy group is then removed as the acetate goes through a cycle of FAS elongation. The cycle of FAS involves reactions with the ketoreductase (KR), enoylreductase (ER), and acyltransferase (AT) domain until the product reaches 16-C at which point it is released by the thioesterase (TE) domain. Figure from Cantu et al. 2011 [68]. 1.5 Polyketide Synthases

Polyketide synthases are multi-domain enzyme complexes, similar to fatty acid synthases (FAS), where individual domains sequentially and processively add two acetate- derived carbons to a growing carbon chain [61,69]. PKS processes are mechanistically similar to FAS and polyunsaturated fatty acid (PUFA) synthesis so that FAS and PUFA synthase can be viewed as subsets of the more general PKS pathway [70,71]. The PKS have been divided into three categories based on arrangement and function. Type I PKS enzymes have multiple catalytically active domains, on one large polypeptide such as the human fatty acid synthase (FAS) system or the bacterial DEBS system that synthesizes erythromycin [72,73]. Type II PKS pathways have individual domains as separate monofunctional proteins and they function in trans to produce their product [72]. Type III

PKSs are smaller and involved in processes such as chalcone synthase in plants or melanin production in bacteria [73–75].

In polyketide synthesis, initially an acyl monomer is loaded onto a phosphopantetheine “arm” of the acyl carrier protein (ACP) by acyl transferase (AT) domains. The monomer that is used will vary, with the starting monomer often being different from the monomer used in the remaining 2-C additions. The ACP is the site of monomer attachment as well as the site of the remaining reactions that occur in PKS biosynthesis. Condensation reactions are carried out by ketosynthase (KS) domains while reduction reactions are facilitated by ketoreductase (KR), dehydratase (DH), and enoylreductase (ER) domains (Figure 1.5). The thioesterase (TE) domain is responsible for the release of the finished product (Figure 1.6) [72,73,76].

Figure 1.6: Schematic of a Type I Polyketide Synthase

The organization of a type I polyketide synthase that synthesizes the antibiotic pikromycin. Each of the 6 modules consists of individual domains where the arrangement varies resulting in different product. This figure depicts an iterative process where synthesis begins with the starter molecule, methylmalonyl CoA, and ends with the release of the product off the TE domain. Ketosynthase KS, acyltransferase AT, dehydratase DH, enoyl reductase ER, ketoreductase KR and acyl carrier protein ACP; KSQ is a decarboxylase; KR* is inactive. Figure from Dutta et al. 2014 [69].

Initially in dinoflagellates PKS transcripts were found that were phylogenetically similar to type I PKS, but structurally were monomers like type II PKSs [73,76]. However, with an increasing number of transcriptomes becoming available it appears that dinoflagellates do possess multi-modular type I PKSs and PKS/NRPS hybrids [77]. Within

type I PKSs there are iterative and noniterative processes. In iterative type I PKS the same set of domains (a module) are used repeatedly for elongation. In noniterative processes each active site is only used once and multiple modules often arranged in a single polypeptide are common [71,78]. Within noniterative processes the AT domain can act in cis, meaning within the module, or trans, as a free standing protein (Figure 1.7).

Phylogenetic analysis has revealed that trans-ATs are common in algae and it seems that integrated ATs have been displaced by trans-ATs in algal PKS evolution [71].

To date most of the information known about dinoflagellate polyketide synthesis has been obtained through isotope labeling studies [61,79–81]. Dinoflagellates are fully invested in polyketide production as their method of secondary metabolite production.

Studies have revealed that dinoflagellates often break polyketide assembly rules that are more consistently followed by bacterial or fungal species [61]. In other words, what is considered an exception in bacterial or fungal PKS is likely the rule in dinoflagellate polyketide biosynthesis. Some examples of what make dinoflagellate polyketide synthesis different include frequent carbon deletions, methylations, and methyl groups separated by an even number of carbon atoms [61]. These features are rarely found in polyketides in other organisms but are found in almost every dinoflagellate polyketide product studied to date. It seems that dinoflagellates make polyketide products of greatly varying structure and function from only three precursor units, acetate, glycolate, and glycine [61].

While polyketide synthase pathways have been well defined in other species such as fungi and bacteria there is very little known about protist PKSs, at the genetic and protein level [73,76]. These pathways are normally studied with a combination of genetic manipulation and synthetic biology to probe for the functional pathways responsible for

synthesizing the product of interest [82]. Unfortunately due to the genetic complexity of the dinoflagellate genome, traditional methods to study polyketide natural products are not applicable at this time [76].

Figure 1.7: Iterative and Noniterative Polyketide Synthases

Type II PKS are depicted with monofunctional proteins that work together to form a complex. Type I iterative PKS use the same module in a cyclic fashion for elongation of the product. Type I noniterative PKS use distinct modules in a large polypeptide where the AT domains can be cis or within the module, or trans as a stand-alone protein. Figure from Shelest et al. 2015 [71].

1.6 Bioluminescence

Dinoflagellates are also one of the many marine bioluminescent organisms.

Bioluminescent bays in Puerto Rico serve as tourist hot spots with people traveling worldwide to see this event. Aside from being a magnificent sight, the study of the mechanism of bioluminescence has taught us much about the circadian clock and protein synthesis in dinoflagellates [83].

Lingulodinium polyedrum is a photosynthetic dinoflagellate well studied for its bioluminescence. Bioluminescence occurs in organelles called scintillons where luciferin, the protein responsible for the light production is located [84]. Under normal conditions luciferin is bound to luciferin binding protein (LBP). When there is shear stress the pH drops below 7 in the scintillon, the luciferase enzyme (LCF) oxidizes the luciferin causing an emission of photons which are observed as a blue light [83,85,86].

In L. polyedrum bioluminescence is controlled by a circadian clock where bioluminescence is observed during the dark and is negligible during the day [87,88]. In L. polyedrum LCF, LBP, luciferin, and scintillons are destroyed at the end of every night and resynthesized before the next night. The mRNA levels of these proteins remain constant during the diurnal cycle suggesting translational rather than transcriptional control of gene expression (Figure 1.8) [83,89,90]. Not all bioluminescent dinoflagellates follow this pattern of diurnal degradation and synthesis. Pyrocystis lunula keeps a constant levels of mRNA, protein, and number of scintillons and instead moves the scintillons to the periphery of the cell at night and chloroplasts to the periphery of the cell during the day

[91].

Figure 1.8: Luciferase Binding Protein and mRNA Rates over Time

Luciferase binding protein (LBP) synthesis rates (squares), LBP abundance (diamonds), and LBP mRNA levels (circles) shown over time. LBP synthesis and abundance change over time while mRNA levels remain relatively constant. Figure from Hastings 2013 [90].

The circadian clock and translational control observed in bioluminescent species holds true for other dinoflagellates as well [90]. Translational control and constant mRNA levels under differing conditions have been observed in many other species as reviewed by Morse

[92].

1.7 Project Goals

Advances in sequencing technologies have had a huge impact in dinoflagellate studies.

While only two species have genome sequences, there are many transcriptome datasets with more being added at a rapid rate [31,32,92]. While these studies have yielded abundant transcript level information about dinoflagellates, corresponding information on cellular control is still lacking. Transcript levels are not always correlated with protein production as was described above for bioluminescence in Lingulodinium polyedrum [92] [83,90].

While some species show transcriptional changes in response to stimuli such thermal stress

in Symbiodinium, other species keep constant mRNA levels despite external stressors. For example, L. polyedrum showed no changes in mRNA expression when subjected to a 24 hour light period as opposed to its normal 12:12 light:dark cycle [92].

In Chapter 1,the goal was use these valuable transcriptome datasets as a starting point for questions to probe, but to move beyond transcripts to study the proteins involved in specific pathways. We began our work by studying a proton channel in L. polyedrum. We wanted to characterize this proton channel and confirm that 1) it was in fact a proton channel as the sequence suggests and has all the characteristics of a proton channel and 2) that the protein is in fact produced at the correct size and in the correct location in the cell.

In Chapter 2, we moved onto acetyl CoA carboxylases because of their involvement in fatty acid as well as polyketide product synthesis. We characterized the phylogenetic distribution across different dinoflagellate species and showed the protein is synthesized in three different dinoflagellates, Akashiwo sanguinea, Amphidinium carterae, and

Karlodinium veneficum. The A. carterae ACCs were further characterized by mass spectrometry-based proteomics as well as sequence analysis to assign protein location.

Chapter 3 focused on a multi-modular polyketide protein with implicated in toxin production. Beginning with cerulenin inhibition of the ketosynthase (KS) domain, involved in fat and toxin synthesis, we conducted an isotope labeling study to show effects on fat and toxin synthesis. Antibodies designed to different domains in the sequence were used to show evidence of protein production and to see whether a full sized protein was present.

CHAPTER 2: IDENTIFICATION OF A VACUOLAR PROTON CHANNEL THAT TRIGGERS THE BIOLUMINESCENT FLASH IN DINOFLAGELLATES1 2.1 Abstract

In 1972, J. Woodland Hastings and colleagues predicted the existence of a proton selective channel (HV1) that opens in response to depolarizing voltage across the vacuole membrane of bioluminescent dinoflagellates and conducts protons into specialized luminescence compartments (scintillons), thereby causing a pH drop that triggers light emission. HV1 channels were subsequently identified and demonstrated to have important functions in a multitude of eukaryotic cells. Here we report a predicted protein from

Lingulodinium polyedrum that displays hallmark properties of bona fide HV1, including time-dependent opening with depolarization, perfect proton selectivity, and characteristic

pH dependent gating. Western blotting and fluorescence confocal microscopy of isolated L. polyedrum scintillons immunostained with antibody to LpHV1 confirm

LpHV1’s predicted organellar location. Proteomics analysis demonstrates that isolated

2+ scintillon preparations contain peptides that map to LpHV1. Finally, Zn inhibits both

LpHV1 proton current and the acid-induced flash in isolated scintillons. These results implicate LpHV1 as the voltage gated proton channel that triggers bioluminescence in L. polyedrum, confirming Hastings’ hypothesis. The same channel likely mediates the action potential that communicates the signal along the tonoplast to the scintillon.

1 Rodriguez, J. D.; Haq, S.; Bachvaroff, T.; Nowak, K. F.; Nowak, S. J.; Morgan, D.; Cherny, V. V.; Sapp, M. M.; Bernstein, S.; Bolt, A.; Decoursey, T. E.; Place, A. R.; Smith, S. M. E. Identification of a vacuolar proton channel that triggers the bioluminescent flash in dinoflagellates. PLoS One 2017, 12, 1–24. 22

2.2 Introduction

The first postulation that a depolarization-activated, proton selective channel (HV1) should exist was published in 1972 by J. Woodland Hastings and colleagues [93]. A decade later, Thomas and Meech reported the first voltage-gated proton conductance to be identified by voltage-clamp studies, in snail neurons [94]. Subsequent electrophysiological studies have elucidated the defining characteristics of a family of voltage-gated proton-selective channels, HV1, which have been found in amphibia [95], rat [96], human [97–100], insects [101], and both multicellular [102] and unicellular marine species [103,104]. HV1 are exquisitely selective for protons [105], resulting from a critical Asp residue in the S1 transmembrane helix [104,106] that is thought to interact

3 with an arginine in S4 [107,108]. HV1 have a single-channel conductance 10 smaller than most ion channels [109], reflecting the 106 lower concentration of permeant ions.

They open with depolarization but their voltage-dependence is strongly influenced by both external and internal pH, pHo and pHi [110], such that a one unit change in either pHo or pHi (or in the pH gradient pH = pHo - pHi) shifts the gH-V relationship by 40 mV

[111]. Predicted HV1 genes are nearly ubiquitous in eukaryotic genomes and the protein has multiple demonstrated functions in various eukaryotic cells [112].

Dinoflagellate bioluminescence is a striking phenomenon, producing brilliant blue flashes in the ocean water when the organisms are stimulated mechanically [113] in the form of shear stress [114,115]. The biochemical basis of dinoflagellate light production was elucidated in seminal work over several decades by Hastings and colleagues, who focused mainly on Lingulodinium polyedrum (formerly Gonyaulax polyedra). This work established that the light originates in organelles called scintillons arising from

evaginations of the central vacuole membrane [116–120]. Scintillons contain a breakdown product of chlorophyll [121] called luciferin that is bound to luciferin binding protein (LBP) at pH > 7, while at pH < 7 LBP releases luciferin, thereby enabling it to interact with the enzyme luciferase (LCF) that catalyzes the reaction with oxygen and luciferin to produce light [93,116,118,122–124]. Low pH activates the LCF in bioluminescent dinoflagellates [118,125] providing a second mechanism by which protons trigger the flash.

Fogel and Hastings [93] reasoned that a hypothetical voltage-sensitive proton channel should respond to a stimulus-induced depolarization of the L. polyedrum vacuole membrane [126], and that such a channel would transport protons from the acidic vacuole into the scintillons, thus providing the pH change that triggers bioluminescence. HV1 proton channels have been reported in several non-bioluminescent unicellular marine species [103,104] where they function in calcium fixation (coccoliths) and possibly in feeding (dinoflagellates). Here we identify a bona fide HV1 in the bioluminescent species studied by Hastings and colleagues, L. polyedrum, and demonstrate its scintillon localization. In addition to confirming a longstanding prediction we authenticate a unique mode for HV1 in the vacuole membrane as the control step in the signal transduction pathway that leads to dinoflagellate bioluminescence.

2.3 Results

Sequence similarity searches of RNA-seq assemblies [127,128] from the bioluminescent organism L. polyedrum revealed a sequence (gi: 346282507) with the signature sequence pattern [104] that has proven to be diagnostic of HV1 (longest open reading frame from this assembly transcript is shown in Table 2.2. The cDNA libraries

prepared from L. polyedrum populations sampled at mid-light and mid-dark and probed with PCR primers designed using the RNA-seq data produced PCR products with expected sizes from this putative HV1 Table 2.1. qRT-PCR shows that the RNA for the putative HV1 was expressed in L. polyedrum (data not shown).

LpHV1 is a bona fide voltage-gated proton channel

We ordered the synthesis of a mammalian codon-optimized gene (sequence shown in Table 2.2) corresponding to the predicted L. polyedrum HV1 gene. When the gene was expressed in a human cell line (HEK-293) the gene product produced voltage- and time- dependent currents in voltage-clamped cells Figure 2.1. Non-transfected HEK-293 cells sometimes had small native proton currents, but otherwise had no significant conductances that exhibited time dependent activation or tail currents under our recording conditions. The proton current during depolarizing voltage pulses turned on more rapidly than in mammalian species in which activation time constants may be seconds at room temperature [4,5,6]. LpHV1 current also differed from most species studied to date in turning on much more rapidly at more positive voltages (discussed below). Similar to

HV1 in all other species, LpHV1 exhibited strong sensitivity to pH. As in the cell in

Figure 2.1A-C, when pHo was decreased, the voltage range over which the gH was activated shifted positively.

LpHV1 can produce inward current

In the vacuole or scintillon membrane, the proton channel is expected to be oriented with its external side facing the vacuole. In order to mediate the flash-triggering action potential in the vacuole membrane and enable H+ influx into the scintillon to activate luciferase, the channel in situ must be able to conduct inward current. At symmetrical

pH, activation occurred well positive to 0 mV (e.g., Figure 2.1) and thus only outward current was produced, as in all other species except for K. veneficum [104]. In fact, the average “threshold” voltage at which LpHV1 was clearly activated was 46 ± 1.8 mV

(mean ± SEM, n = 38; 23 cells and 15 patches), well positive to +23 mV reported for a variety of native proton currents, mostly in mammalian cells, or to -10 to +10 mV in hHV1 (Table 3 in [112]). However, Figure 2.1 shows that when there was a large inward pH gradient (simulating the low pH vacuole and high pH cytoplasm that exist in L. polyedrum), inward current was activated. In vivo, LpHV1 is exposed to an enormous inward pH gradient (pH is 3.5), because the flotation vacuole has very low pH 4.5

[120], compared with cytoplasmic pH ~8. HEK-293 cells did not tolerate such a large gradient, but with a moderate inward gradient (pH 2.0), we observed inward H+ current

(Figure 2.1D). Inward currents were detected in ten cells with 1.0 to 3.0 U gradients.

Extrapolated to in vivo conditions, LpHV1 should conduct inward current when activated.

Figure 2.1: LpHV1 is a voltage gated proton channel.

(A-C) Families of whole-cell proton currents at different pHo in a cell transfected with LpHV1, with pHi 7.0. Voltage pulses were applied from a holding potential of -60 mV (A, B), or -40 mV (C), in 10 mV increments up to the voltage indicated. (D) Inward H+ currents can be seen with large inward pH gradients. Currents are shown during pulses to 30, 50, 60, and 70 mV as indicated, in an inside-out patch with pHo 7 (pipette) and pHi 9 (bath), according to the standard convention in which downward deflections indicate inward current flow. From the tail currents upon repolarization to the holding potential of -40 mV, it is clear that the gH was already activated detectably by the pulse to 50 mV, with small inward current evident during the pulse to 60 mV, and larger inward current at 70 mV.

LpHV1 is proton selective

Measurement of the reversal potential (Vrev) of the current in cells expressing LpHV1 over a wide range of pHo and pHi confirmed that these currents were proton selective,

+ because Vrev was close to the Nernst potential for H (EH) (dashed green line in Figure

+ 2.2). When tetramethylammonium was replaced by other small cations Vrev did not

+ + change, confirming H selectivity. The mean change in Vrev when TMA in the bath was exchanged with Na+ or K+, respectively, was 0.4 ± 0.4 mV (mean ± SEM, n = 4) or 1.2 ±

1.8 mV (n = 4), after correction for the measured liquid junction potentials. The expressed gene product is clearly a highly proton-selective voltage-gated channel, so we named it LpHV1, and explored its similarities and differences from HV1 in other species.

Figure 2.2: LpHV1 is a proton selective channel.

The reversal potential (Vrev) was measured, usually by tail currents (as shown in the examples in the insets), in both whole-cell (n = 10) and excised, inside-out patch configurations (n = 10) over a wide range of pH (pHo 4.5-9.0; pHi 4.5-10.0). Measurements at multiple pH in individual cells or patches are connected by lines. The heavy dashed green line indicates Vrev = EH, which would indicate perfect proton selectivity. Whole-cell data are plotted as triangles, diamonds, or hexagons, and pink Xs, connected by solid lines; other symbols are from inside-out patches, connected by dashed lines. Insets show tail current measurements from the same inside-out patch, with pHo 7 in the pipette, and pHi 8 or 7, as indicated, in the bath. Vrev shifts from -2 mV at pHi 7 to 53 mV at pHi 8, a change of 55 mV, near the Nernst expectation of 58 mV for perfect H+ selectivity. LpHV1 exhibits pH dependent gating

In all species studied so far, HV1 exhibits pH dependent gating, in which the voltage range of channel opening is defined by the pH gradient, pH (pHo - pHi). The position of the proton conductance-voltage, gH-V, relationship generally shifts 40 mV/unit change in either pHo or pHi (positively for increasing pHi and negatively for increasing pHo) [105,110,111]. Figure 2.3 shows gH-V relationships in a cell studied in whole-cell configuration with changes in pHo (A) and in an excised, inside-out patch of membrane in which pHi was changed (B). It is evident from Figure 2.3A that increasing

pHo shifts the voltage dependence of LpHV1 opening negatively, as observed in HV1 from all species studied to date. It is further evident that there was little shift between pHo 8 and pHo 9, indicating that this effect saturates at high pHo, as reported recently in human HV1, hHV1, kHV1, and EhHV1 [129]. The data in Figure 2.3B show that changing pHi similarly shifts the gH-V relationship, with decreasing pHi shifting the curve negatively, as occurs in all other HV1.

Figure 2.3: LpHV1 exhibits classical pH dependent gating.

The gH-V relationships calculated from LpHV1 currents in whole cell measurements (A) or inside-out patches (B) are strongly affected by pHo or pHi, respectively. The proton conductance, gH, was usually calculated from the measured reversal potential, Vrev, and the amplitude of a single rising exponential fitted to the current. In some cases, for example with test pulses near Vrev, the amplitude of the tail current was used, after appropriate scaling. 30

The effects of pHo and pHi on the position of the gH-V relationship were assessed quantitatively by plotting the voltage at which the gH was 10% of its maximal value, gH,max, as a function of pHo or pHi (Figure 2.4). This parameter was chosen because it is clearly defined, it does not require forcing data to fit a Boltzmann distribution, and it does not require estimating the threshold of activation, which has been used for this purpose previously by us and others but is imprecise and arbitrary. These considerations are discussed at greater length elsewhere (see Methods of [129]). When pHo or pHi was varied, the gH-V relationship shifted by ~40 mV/unit change in pH, matching the reference line showing this slope (Figure 2.4). For both pHo and pHi the shift tended to saturate above pH 8, similar to HV1 from rat [111], human, the dinoflagellate

Karlodinium veneficum, and the coccolithophore Emiliania huxleyi [129]. The saturation occurs above typical environmental pH, and likely reflects the approach of pH to the pKa of one or more pH sensing sites.

Figure 2.4: LpHV1 exhibits pH dependent gating.

The position of the gH-V relationship was established from gH-V relationship plots by measuring the voltage at which gH was 10% of gH,max. Measurements in individual cells or patches at several pH are connected by lines. Color coding indicates pHi for whole cell measurements (A) or pHo for inside-out patch measurements (B). As a reference, the arbitrarily positioned dashed line in each panel shows the slope that corresponds to a shift of 40 mV/unit change in either pHo or pHi. Except at high pHo or pHi, the data are roughly parallel to the reference lines, indicating a slope of 40 mV/unit pH. The slope decreases at pH>8, indicating saturation of pH dependence.

LpHV1 channel opening kinetics

The rising current during a depolarizing pulse was fitted with a single exponential function to determine the time constant of activation, act. At symmetrical pH 8 (pHo = pHi = 8), act ranged 45-600 msec at +60 mV (n = 6); at symmetrical pH 7, act ranged 30-

470 msec at +80 mV (n = 8). Examples of the dependence of act on pH are shown for whole-cell measurements (Figure 2.5A) in which pHo was varied, and for inside-out patches of membrane (Figure 2.5B) in which pHi was varied. In both configurations, changes in pH appeared to simply shift the act -V relationship along the voltage axis. As in all species, current activation (turn-on) became faster with greater depolarization

(Figure 2.1 A-C). However, in LpHV1 this property was markedly exaggerated.

Activation became much faster (smaller act) with larger depolarizing pulses; typically,

act was 100 times faster at voltages 60 mV more positive.

Figure 2.5: Voltage and pH dependence of LpHV1 activation kinetics.

A. Voltage and pHo dependence of LpHV1 activation kinetics. Currents were fitted by single rising exponentials to obtain the time constant of channel opening (activation, act). These measurements were made in the same cell with pHi 7.0, studied at three different pHo. B. Voltage and pHi dependence of LpHV1 activation kinetics. These measurements were made in the same inside-out patch of membrane with pHo 7.0, studied at six different pHi.

LpHV1 stains multiple membranes in intact cells

If LpHV1 functions to allow protons across the vacuole into the scintillons as predicted, it should be localized in scintillon membranes. The other bioluminescence proteins LCF and LBP were previously demonstrated to localize to scintillons in L. polyedrum [116,119,122]. We immunostained PFA-fixed, methanol-dehydrated L. polyedrum cells with chicken anti-LCF, rabbit anti-LBP, or rabbit anti-LpHV1. We visualized their localization with organism-specific secondary antibodies, each labeled with a different fluorophore. Western blotting of L. polyedrum whole cell lysates and purified recombinant LCF, LBP, and LpHV1 probed with these primary antibodies detected proteins of the expected sizes and no cross-reactivity (data not shown; but see

Figure 2.7). The confocal microscopy images in Figure 2.6 demonstrate that LCF and

LBP are distributed in a punctate pattern as previously observed in dark-harvested L. polyedrum [117]. Fluorescence intensities of labeled cells were significantly different than both pre-serum and no-primary-antibody controls (see also Figure 2.10 for negative control images). As expected from previous studies [122,130], total fluorescence from antibody-labeled LCF and LBP decreases significantly in cells fixed at mid-light phase compared to mid-dark phase (data not shown).

Figure 2.6: LpHV1 distribution is consistent with scintillon localization.

Fixed whole L. polyedrum were probed with antibodies to LBP, LCF, and LpHV1, stained with fluorescently labeled secondary antibodies to appropriate IgG, and visualized via confocal microscopy. Maximum projection of a representative Z-stack for each primary antibody is shown. Scale bars in all panels = 10 M. Images were analyzed for per area fluorescence from each secondary antibody using Zen software tools. Bars represent means +/- S.D. of fluorescence from 20-30 individual cells from 2-5 separate preparations; significant differences from no-antibody or pre-serum controls are indicated with asterisks. Images of negative controls are presented in Figure 2.10. The proposed function of HV1 in bioluminescence requires localization in scintillon membranes (evaginations of the vacuole membrane); nonetheless, no cellular membrane could be excluded a priori as a potential site for HV1. As seen in Figure 2.6, a significant fraction of the LpHV1 localizes around the periphery of the organism, consistent with an additional plasma membrane location.

LpHV1 co-localizes with LCF and LBP in isolated scintillons

To further explore the cellular localization of HV1, we isolated scintillons from L. polyedrum using density gradient centrifugation [122,131]. Scintillon isolation was confirmed initially by a bioluminescence activity assay [122]. Western blots of proteins

extracted from isolated scintillons and probed with antibodies to LBP and LCF demonstrate the presence of these known scintillon markers and also show no detectable cross-reactivity of these antibodies with recombinant LpHv1 protein (Figure 2.7B). As has been described previously, presumed proteolysis products for LBP and LCF are visible in these Western blots. As seen in Figure 2.7A, Western blotting of proteins extracted from isolated scintillon preparations shows that the antibody to LpHV1 detects a scintillon protein with an apparent size of about 30 kDa; this antibody detects purified recombinant LpHV1 at the expected size of 37 kDa. Many possibilities, including different post-translational processing in bacteria compared to dinoflagellates, could explain the apparent size difference between recombinant and native LpHV1. Both in purified recombinant protein preparations and in preparations of scintillon protein, our antibody to LpHV1 frequently detects protein bands at ~60 and at ~80 kDa, which are consistent with the size of a truncated (60 kDa) or full length (80 kDa) LpHV1 dimer.

LpHV1 has a strongly predicted coiled-coil region in its C-terminus, so it likely dimerizes like HV1 from several other species [132–135]. Our antibodies to scintillon proteins do not cross react (Figure 2.7B). Western blotting from separate preparations consistently shows more LpHV1 (~2 fold) in scintillons isolated during the day phase than the night phase (as in Figure 2.7A), although the difference is not statistically significant.

Figure 2.7: LpHV1 localizes to the scintillon.

(A) Total protein from isolated L. polyedrum scintillons were Western blotted and probed with anti-LpHV1. (B) Total protein from isolated L. polyedrum scintillons, and also purified recombinant LCF, LpHV1, and GST-labeled LBP, were Western blotted and probed with the antibody indicated. (C) Isolated scintillons were fixed and immunostained as in Figure 2.6. Scale bars in all panels = 2 m. Scintillons in different treatments were identified by their native luciferin fluorescence; the percentage of scintillons in each treatment that exhibited secondary antibody fluorescence is shown. Number of scintillons scored for each treatment: LBP-LCF, n=55; LpHV1-LCF n=55; LCF preserum, n=41; LpHV1 pre-serum, n=15; no primary antibody, n=35. Images of negative controls are presented in Figure 2.11.

Scintillon isolation was further confirmed by confocal microscopy of fixed scintillon preparations, using native luciferin fluorescence [136] to positively identify scintillons

[117] (Figure 2.7C). As previously described [117], we observed a low level of contaminating chlorophyll fluorescence. On the slide, chlorophyll fluorescence rarely overlapped with luciferin fluorescence, and many structures with luciferin fluorescence but no nearby chlorophyll fluorescence was visible. Immunostaining of fixed scintillon

preparations confirm the presence of LBP and LCF in isolated scintillons (Figure 2.7C).

Immunostaining of fixed scintillon preparations show that LpHV1 localizes to luciferin- containing structures that also contain LCF (Figure 2.7C). Because both luciferin and

LCF are known markers of the scintillon structure, this is strong evidence that LpHV1 localizes to scintillon membranes.

Mass spectrometry detects LpHV1 peptides in isolated scintillon preparations

To further confirm the presence of LpHV1 in scintillons, we subjected proteins from scintillon preparations to tandem mass spectrometry (MS/MS) analysis. We excised five prominent bands from an Imperial stained gel: one at 33 kDa, close to the size expected for LpHV1, and others ranging from ~25 to ~50 kDa. We detected 17 different peptides from the 33 kDa band, and 2 different peptides from the 50 kDa band, each with more than one independent peptide spectrum, that met the probability threshold for matching the predicted sequence of LpHV1. The sequences of these peptides are mapped onto the predicted sequence of LpHV1 in Figure 2.8. Many distinct peptides from the excised bands met the probability threshold for scintillon marker proteins LBP and LCF. In addition, a large number of peptides with significant matches to known chloroplast proteins (e.g., ribulose-bisphosphate carboxylase, glyceraldehyde-3-phosphate dehydrogenase, peridinin-chlorophyll a-binding protein, etc.) were detected, consistent with the known chloroplast contamination of the scintillon preparation.

Figure 2.8: Isolated scintillon preparations contain peptides matching the LpHV1 sequence.

The sequence and predicted secondary structure acid of LpHV1 is shown (Johns S.J., TOPO2, Transmembrane protein display software, http://www.sacs.ucsf.edu/TOPO2/). Acidic residues in the transmembrane helices are shown in red, basic residues in dark blue, and aromatic residues in gray. Brown diamonds indicate the overlap of peptide sequences found by mass spectrometry analysis of isolated scintillons and the epitope to which the antibody against LpHV1 was raised; otherwise peptide sequences are shown in orange squares and the epitope is shown in green stars.

2+ Zn inhibits LpHV1 proton currents and the flash induced by acid in isolated scintillons 2+ Few HV1 inhibitors have been identified; the most potent in many species is Zn

2+ [137,138]. Figure 2.9 shows that Zn inhibited LpHV1 currents detectably at 10 M, and

substantially at 100 M. At similar concentrations, Zn2+ inhibited luminescence elicited by acid exposure in scintillons isolated from L. polyedrum (Figure 2.9B and Figure 2.9C).

No flash was elicited when detergent was present, suggesting that intact scintillons were required for the response. These results strongly support the hypothesis that LpHV1 is the proton channel that triggers the bioluminescent flash in L. polyedrum.

2+ Figure 2.9: Zn inhibits LpHV1 proton currents and scintillon luminescence.

2+ (A) Proton currents at +60 mV at pHo 7.0 were reduced by Zn . The mean reduction of current by 100 M Zn2+ was 63 ± 11 % (mean ± SEM, n = 4). (B) Luminescence of L. polyedrum scintillons stimulated by 50 mM acetate and measured in a plate reader was inhibited by Zn2+. (C) Zn2+ sensitivity of luminescence of L. polyedrum scintillons stimulated by 50 mM acetate and measured with a photometer generously provided by J. W. Hastings.

Figure 2.10: Control Fluorescence Images for Figure 2.6

Fluorescence images from pre-serum or no-antibody control treatments of fixed whole cells. Cells and images were prepared as for Figure 2.6.

Figure 2.11: Control Fluorescence Images for Figure 2.7

Fluorescence images from pre-serum or no-antibody control treatments of fixed isolated scintillons. Scintillons and images were prepared as for Figure 2.7.

Figure 2.12: Alignment of predicted amino acid sequences of putative dinoflagellate HV1s.

Sequences were found by BLAST searches of RNA-seq projects [127,128]and were aligned with MSA-Probs [139]. Dinoflagellate sequences are shown aligned with sequence logos of individual transmembrane helices obtained from an alignment of animal Hv1s

[104]. The sequence logos are numbered for hHv1. Overlapping partial sequences from

Alexandrium monilatum are not shown.

SEQUENCE PRODUCT SIZE TM (°C) HV1_1_F CTTCAAAGCACGAGGAGCAT 157 60.54

HV1_1_R AGGTAGTGCGTCTCCAGGAC 157 59.33

HV1_2_F ACTGCAAGGCCTACGTGGA 185 61.81

HV1_2_R GTGCCCAGCTAGGAAGAGG 185 59.96

HV1_3_F GGCATCCTCGTGATCTTCAT 141 60.04

HV1_3_R CAGGTTCGTCACCAGGATCT 141 60.11

HV1_4_F AGAAGCTGATGGTGCTGGAC 200 60.42

HV1_4_R GGAGAGCTCCTTGGGAAAAG 200 60.32

HV1_5_F CTTCAGGGAGAAGCTGATGG 196 59.94

HV1_5_R GGGAAAAGTTCCCAGGACA 196 59.89

Table 2.1: Primers and predicted product sizes for qPCR.

Sequences of LpHv1

LpHv1 sequence from transcriptome >MMETSP1032_1_(paired)_contig_26874 CCGCGGCCGCGGCCGCCGGCCGATGGCTGGGCATCATGGAGCGCCTTCAA AGCACGAGGAGCATGCCGCAACGGCGCCGCACGGCATCAAGCACGGCCTG CAGCTGTACAACAGCAAGGCCTGCCTGGTGGTCCTGTTCTTCCTGCTGATCC TGGACGTGTGCATCGTCGTGGCATCGGGCGTCCTGGAGACGCACTACCTGA TCTCCAAGGCCGACGACTGCAAGGCCTACGTGGACGCCTGCCCCCACGGCC ACGGCCGCCGCCTCGACGCCCCGCGCCCGTGGAGCGGCCTCGGCGCCGCC GGCGCAGAAGCCGCGCCCTCGCCCGGGCGCCGACTGGACGGCCTCGACGC CGCTGACGCAGAAGCGAGCAGCGGCGGCGGCCTCTTCCTAGCTGGGCACG ACGCGGGGCGGCAGCTGAGCTCTTCGGACGGCGACCAGATCGACTGCCAC

CACCCGCACTTCGGCAACCACAGCCTGCACGACGCCGAGAAGATCCTGGC CTACATCTCCATCGGCATCCTCGTGATCTTCATCATCGAGCAGCTGCTGCTC ATCGCCGCGATGAGGGGCGCGTACTTCAGGGAGAAGCTGATGGTGCTGGA CGGCTTCGTGATCACCCTCTCGCTGCTGCTGGAGATCCTGGTGACGAACCT GCCGCTGGGCGGCCTCCTCGTCGTCGCGCGCATCTGGCGCTTCGCCCGGGC GGGGCACGGCACCATGGAGGGCTCGCACGACGTGCACAAGGTGCACCCTG TCCTGGGAACTTTTCCCAAGGAGCTCTCCGAGCAGGTCTGGGCCCACATGT CAGGCGAGAAGTGGGAGGCCATGCTCCTGCGCAACGGCACGGAGAAGCTG GAGATGGACGTGGCAAAGGAGGAGCAGCGCATCGCGGCCCAGATCGCGCA GGCCCACCCCAGCGTCGTGCTCCGCGCCCTGGCCGCGGAGCGCCAGCGCCA GGCCGCCCGCGAGCAGCTGCAGGCGCAGGAGAGCAAGGCCTCGGGCCGGC CGGCCGTTTGAGGCCCGCGCCGCCCCCCCCGCGCCCCTGGAGGGGGCGCAC GCCGCCACGCCTT

LpHv1 DNA sequence, optimized for mammalian codons >LpHv1 optimized for mammalian expression ATGGCAGGACATCACGGAGCACCAAGCAAGCACGAAGAGCACGCCGCAAC CGCCCCTCACGGAATCAAGCACGGACTGCAGCTGTATAACAGCAAGGCCT GCCTGGTGGTCCTGTTCTTTCTGCTGATTCTGGATGTCTGTATCGTGGTCGC CTCTGGCGTGCTGGAGACACATTACCTGATTAGTAAGGCTGACGATTGCAA AGCCTATGTGGATGCTTGTCCACACGGACATGGCCGGAGACTGGACGCACC CAGACCTTGGTCAGGACTGGGAGCAGCTGGAGCTGAGGCAGCACCAAGCC CCGGGAGGCGCCTGGACGGACTGGATGCTGCAGACGCAGAAGCTAGCTCC GGAGGAGGACTGTTCCTGGCAGGACATGATGCCGGCAGGCAGCTGTCTAG TTCAGACGGCGATCAGATTGATTGCCACCATCCCCACTTCGGGAACCACTC CCTGCATGACGCTGAAAAGATCCTGGCATACATCTCTATTGGCATCCTGGT CATCTTCATCATTGAGCAGCTGCTGCTGATCGCCGCTATGCGAGGGGCCTA TTTCAGGGAAAAACTGATGGTCCTGGACGGATTTGTGATTACCCTGAGTCT GCTGCTGGAGATCCTGGTGACAAATCTGCCTCTGGGAGGACTGCTGGTGGT CGCAAGAATCTGGAGGTTCGCACGAGCAGGACATGGAACTATGGAAGGCT CTCACGATGTGCATAAGGTCCACCCTGTGCTGGGGACCTTTCCAAAAGAGC TGAGCGAACAAGTGTGGGCCCACATGTCCGGAGAGAAATGGGAAGCTATG CTGCTGCGGAATGGCACCGAGAAGCTGGAAATGGACGTGGCCAAAGAGGA ACAGCGCATTGCAGCCCAGATCGCTCAGGCACACCCCTCCGTGGTCCTGCG GGCCCTGGCTGCAGAGCGACAGCGGCAGGCCGCAAGAGAACAGCTGCAGG CTCAGGAAAGCAAGGCATCAGGAAGACCCGCCGTGTGA

Table 2.2: Sequence of longest open reading frame of contig 26784 in MMETSP1032 assembly, and sequence of codon-optimized gene of LpHV1.

2.4 Discussion

In 1972, J. Woodland “Woody” Hastings predicted the existence of voltage-activated proton-selective channels that trigger the bioluminescent flash in Gonyaulax polyedra

(now Lingulodinium polyedrum) [93]. An action potential in the vacuole membrane was known to initiate the flash [126]. The scintillon contains luciferase, its substrate luciferin, and luciferin-binding protein. A drop in pH both activates luciferase [125] and releases luciferin from its binding protein [140]. Because the vacuole pH was 3.5-4.5 [120,141], a proton-permeable channel that was opened by depolarization during the action potential would allow rapid proton flux into the scintillon, releasing luciferin, activating luciferase, and triggering the flash [93]. The present results strongly support this hypothesis.

Luciferin, LBP, and LCF from several species cross react [118,142] supporting a common role of pH, and presumably HV1, in the signal transduction pathway. We used

RNA-seq data for L. polyedrum to identify a putative HV1 gene. We confirmed that the organism expresses the RNA (Table 2.1) and protein (Figure 2.7) predicted from this gene.

Structural comparison of LpHV1 with other proton channels

LpHV1 is the tenth HV1 gene to be identified and confirmed by voltage-clamp studies in a heterologous expression system. Globally, LpHV1 resembles all other HV1 in having four transmembrane helical segments. Compared with hHV1, it has a short N terminus, ~33 amino acids vs. ~100 in hHV1. LpHV1 has a very long S1-S2 linker of

~100 residues vs. ~10 in hHV1. S1-S2 linkers over 70 residues long appear in two electrophysiologically confirmed HV1 from coccolithophores, and five of the twelve high confidence dinoflagellate HV1 shown in Figure 2.12. These long linkers do not share

significant similarity to each other, and BLAST searches of the linker region reveal little similarity to proteins of known function. A region of about 20 amino acids (351-373) in the LpHV1 linker is predicted to be helical by several secondary structure prediction programs. LpHV1 most likely exists as a dimer, because it has a strongly predicted coiled- coil region in its C-terminus. HV1 from several other species is thought to dimerize, in large part enforced by C terminal coiled-coil interactions [132–135,143].

The sequence of LpHV1 includes the amino acids thus far identified to play critical roles in the function of the molecule in other species [144]. LpHV1 shares the “signature sequence” of all other HV1 that includes an Asp in S1 and the motif RxWRxxR in S4

[104,106]. The Asp in the middle of the S1 segment is crucial to establishing proton selectivity [101,104,106,108], the Arg in S4 are thought to confer voltage sensing as in other VSDs, and the Trp affects multiple properties of HV1 [129]. LpHV1 has residues

(Phe171, Leu42 and Ile199 ) that we propose form the hydrophobic gasket that has been identified in other voltage gated ion channels [145–148]. These three hydrophobic residues are aligned horizontally near the middle of the membrane, where they separate internally and externally accessible aqueous vestibules. The position occupied by Phe171, identified as a delimiter of the charge transfer center [149], is conserved almost universally among VSD-containing molecules.

A number of acidic amino acids are thought to stabilize the channel in closed, open, and intermediate states by electrostatic interactions with the cationic charges in S4.

112 119 123 Likely countercharges in the S1 helix in hHV1 include Asp , Glu , Asp , and

125 45 52 56 possibly Lys [144]. The corresponding positions in LpHV1 are Asp , Ser , Glu ,

58 and His . When externally-accessible acidic groups are neutralized by mutation, the gH-

V relationship shifts positively, because the open state is disfavored [144]. The somewhat

119 125 more positive activation of LpHV1 may reflect the absence of Glu and Lys . On the other hand, the other identified dinoflagellate HV1, kHV1 activates 60 mV more

119 negatively than all other species, despite kHV1 lacking Glu (with Gly instead). The important countercharges in S2 and S3, are conserved in all identified HV1 including

174 195 185 LpHV1, with its Glu and Asp . Another acid in S3 unique to HV1 is Asp (hHv1 numbering), which is absent in other VSDs [104,144]. In LpHV1 this position is occupied by the conservative substitution Glu206. In summary, with the exception of one acidic residue in S1, the charges in the transmembrane region of LpHV1 are quite similar to those in other HV1.

The electrophysiological properties of LpHV1 are consistent with its proposed function in triggering the flash We measured several biophysical properties of the LpHV1 channel to determine whether the LpHV1 gene product has properties consistent with its proposed function of triggering the bioluminescent flash. These properties include 1) proton selectivity, 2) activation by depolarization, 3) opening kinetics comparable to that of the flash, and 4) the ability to conduct inward current (from vacuole to cytoplasm), all of which should occur in vivo. Our experiments were done at room temperature (20-25°C) which is within the range of oceanic temperatures, but other possible differences between our experimental conditions and those in vivo cannot all be evaluated so easily. The mammalian cells used as an expression system (HEK-293 cells) may process proteins differently than dinoflagellates and likely have a different membrane composition than L. polyedrum. Salt concentrations appropriate for mammalian cells, ~300 mOsm, are about half those of sea water. A critical factor for the HV1 protein is the pH gradient. The

cytoplasmic pH in dinoflagellates is estimated to be pH 8.0 [93], but the vacuolar pH in bioluminescent species L. polyedrum [120] and Noctiluca [141] is 4.5 and 3.5, respectively. HEK-293 cells did not survive such a large pH gradient. The membranes of mammalian cells tolerate neither extreme pH per se, nor extreme pH gradients. Our experiments address the posed questions with these constraints.

LpHV1 meets the first two requirements: it is clearly highly proton selective (Figure

2.2) and it opens with depolarization (Figure 2.1). The next question is whether the kinetics of LpHV1 match that of the bioluminescent flash. The flash recorded from individual L. polyedrum had a latency of 15-22 msec [150] and the flash in individual scintillons isolated from L. polyedrum or from entire organisms had a time-to-peak of

~100-200 msec [117,151]. LpHV1 opens faster than HV1 of mammalian species (where

act is measured in seconds), but slower than that of snail neurons with act of a few milliseconds [110]. As in all species, LpHV1 channels open faster with increasing depolarization. Differing sharply from other species, the activation kinetics of LpHV1 depends quite steeply on voltage. As evident in Figure 2.5, act was ~1 s just above

Vthreshold but became ~100 times faster within 50-60 mV, changing e-fold in just ~10 mV.

In contrast, in several mammalian or amphibian cells, the voltage required to change act e-fold ranges 40-80 mV [95,152–156]. As a result of the steep voltage dependence of act, a large depolarization could activate LpHV1 within ~10 msec. The kinetics of LpHV1 thus seems consistent with the kinetics of the bioluminescent response. Density of LpHV1 expression in the native membrane and the presence of other ion channels would also modulate the response.

+ We also addressed whether LpHV1 could produce inward H current during an action potential in situ. At symmetrical pH (e.g., pHo 7, pHi 7) inward current was not observed. With an inward pH gradient of 1-2 units (e.g., pHo 7, pHi 9), however, inward current was observed. An example is shown in Figure 2.1D. Once inward H+ current is activated in the tonoplast membrane, the resulting H+ influx would further depolarize the membrane regeneratively, opening more channels and driving the membrane potential toward EH, which for pH 4.5//8.0 is in the vicinity of +200 mV. Correspondingly, action potential peaks of 200 mV have been recorded in situ in Noctiluca [113]. We were at first surprised that LpHV1 activated relatively positively, compared with other HV1 [112] and especially when compared with kHV1 that activates well negative to EH [104]. Perhaps, given the enormous inward H+ gradient across the L. polyedrum tonoplast, it would be perilous for the cell to allow LpHV1 activation except when triggering a flash. K. veneficum is not bioluminescent, and the function of kHV1 is uncertain, but is likely different from that in bioluminescent species.

The cellular localization of LpHV1 is consistent with its proposed role in bioluminescence Three lines of evidence demonstrate convincingly that LpHV1 is expressed in the vacuole membrane surrounding the scintillons, confirming its predicted role in bioluminescence: 1) Western blots of isolated scintillons show antibody staining of a protein with the predicted size and no detectable cross-reactivity with LCF or LBP

(Figure 2.7); 2) immunostaining of isolated scintillons with the same antibodies shows that LpHV1 colocalizes with LCF, a marker of the scintillon organelles (Figure 2.7); and

3) proteomics analysis of proteins extracted from isolated scintillons shows the expected presence of LBP and LCF proteins (Data not shown) and the presence of LpHV1 protein

(Figure 2.8) in these structures. Further confirmation that LpHV1 is present in scintillons is the inhibition of the flash in isolated scintillons by Zn2+ at concentrations that inhibit

LpHV1 proton currents (Figure 2.9).

Upon stimulation by shear stress, mechanosensor (probably stretch activated) channels at the surface of L. polyedrum [151,157] are thought to relay a signal through intracellular calcium signaling via G-proteins [86,158], but the molecular identities of the signaling components that result in the action potential at the vacuole membrane and subsequent luminescence are unknown. It is likely that LpHV1 mediates the action potential in the vacuole membrane that triggers the flash, because H+ is the only ion with a sufficiently positive Nernst potential to generate an action potential that peaks at +200 mV [113]. Voltage gated Na+ channels have been reported in the outer membrane of

Noctiluca, but their reversal potential in sea water is only +33 mV [159]. Although

Hastings’ original proposal required a proton channel to open only during the action potential in scintillon membranes, it was later realized that the same channel could also mediate the action potential in the tonoplast [120,160]. We have identified additional putative HV1 sequences in RNA-seq data from two other bioluminescent dinoflagellates,

Alexandrium tamarense and Noctiluca scintillans, excellent evidence that HV1, like LCF and LBP, is a conserved component of the signal transduction pathway that leads from shear stress at the organism surface to the light flash.

Immunostaining of whole cells suggests that membranes other than those surrounding the scintillons may contain LpHV1 (Figure 2.6). The non-bioluminescent dinoflagellate K. veneficum expresses a HV1 in feeding populations at night [104]; we have identified putative HV1 genes in several additional dinoflagellates (Figure 2.12)

based on the presence of an Asp in the middle of S1 crucial to proton selectivity

[101,104,106] and a signature sequence in S4 involved in gating (10). In other organisms

HV1 functions in many processes [112], raising the likelihood that in L. polyedrum

LpHV1 also serves purposes in addition to the control of bioluminescence. The dinoflagellates (bioluminescent or not) in which we found HV1 gene sequences span a large fraction of the dinoflagellate phylogenetic tree [161,162]. Taken together, these data suggest that primordial HV1 functions have been co-opted by the bioluminescent species for light production. Intriguingly, antibody to LCF stained trichocysts in L. polyedrum [120], a tantalizing hint of possible non-bioluminescent functions for the bioluminescence enzyme.

2.5 Materials and Methods

Sequence searching and alignments

The K. veneficum voltage-gated proton channel (kHV1, NCBI accession JN255155) was used as a BLAST probe with e-value cutoff of 10-2 against the Marine Microbial

Eukaryote Transcriptome Sequence Project (MMETSP; [128]) which yielded a single contig (26874) in MMETSP1032. BLAST searches of the non-redundant database at

NCBI using the translation of this contig yielded HV1 from diatoms and coccolithophores

-11 with e-values of 10 . Using various HV1 probes, searches of MMETSP also yielded full and partial sequences for putative HV1s from Akashiwo sanguinea, Alexandrium monilatum, Alexandrium tamarense, Amphidinium carterae, Azadinium spinosum,

Karenia brevis, Noctiluca scintillans, Symbiodinium, and Scrippsiella trocoida. We included a recently described putative HV1 from Prorocentrum minimum [163] . These sequences were aligned with a set of high confidence animal HV1s using MSAprobs

[139] (Figure 2.12). The predicted secondary structure for LpHV1 was drawn using

TOPO2 (Johns S.J., TOPO2, Transmembrane protein display software, http://www.sacs.ucsf.edu/TOPO2/).

Gene and antibody synthesis

Genes for LpHV1 (based on longest open reading frame of RNA-seq contig 19215 from the L. polyedrum library) and LCF (NCBI accession AF085332.1), codon-optimized for mammalian (LpHV1) or E. coli (LCF) expression, were synthesized by Genscript

Corp. and subcloned into pcDNA and pEGFP (LpHv1) or pQE-30 (LCF). Peptides corresponding to chosen epitopes from LpHV1 (CDAGRQLSSDGDQ) and LCF

(CLDYPKKRDGWLEKN) were synthesized by Genscript and polyclonal antibodies to these peptides were raised and affinity purified from rabbit (LpHV1, final protein concentration 1.1 mg/ml) or chicken (LCF, final protein concentration 0.88 mg/ml).

Preserum from animals was also provided by Genscript. The native gene for LBP in pGEX-4T, and rabbit antibody raised to LBP, were the generous gifts of Dr. David Morse

(University of Montreal).

Electrophysiology

HEK-293 cells were grown to ~80% confluence in 35 mm cultures dishes. Cells were transfected with 0.4-0.5 µg of cDNA using Lipofectamine 2000 (Invitrogen) or PEI

(polyethylenimine, Sigma). After 6 h at 37ºC in 5% CO2, cells were trypsinized and re- plated onto glass cover slips at low density for patch clamp recording the following day.

We selected green cells under fluorescence for recording. Whole-cell or excised inside- out patch configurations of the patch-clamp technique were carried out as described in detail previously [164]. Bath and pipette solutions were used interchangeably. They

+ – contained (in mM) 2 MgCl2, 1 EGTA, 80-100 buffer, 75-120 TMA CH3SO3 (to bring the osmolality to ~300 mOsm), titrated using TMAOH. Buffers with pKa near the desired pH were used: Homopipes for pH 4.5-5.0, MES for pH 5.5-6.0, BisTris for pH 6.5, BES for pH 7.0, HEPES for pH 7.5, Tricine for pH 8.0, CHES for pH 9.0, and CAPS for pH

10. Experiments were done at 21ºC or at room temperature (20-25°C). Current records are shown without leak correction.

Reversal potentials (Vrev) in most cases were determined from the direction and amplitude of tail current relaxation over a range of voltages, following a prepulse that activated the proton conductance, gH. When the gH activated negative to Vrev the latter could be determined directly from families of currents. Currents were fitted with a single exponential to obtain the activation time constant (act) and the fitted curve was extrapolated to infinite time to obtain the “steady-state” current amplitude, from which the gH was calculated. The voltage at which gH was 10% of gH,max (VgH,max/10 ) was determined after defining gH,max as the largest gH measured.

Cell culture

L. polyedrum (CCMP 1932, obtained from National Center for Marine Algae and

Microbiota-Bigelow Laboratory for Ocean Sciences) cultures were grown in three locations with minor differences in conditions. Cultures were grown in L1 minus Si

[165] or F/2 minus Si [166] medium prepared in artificial seawater (Instant Ocean,

Blacksburg) and maintained in 12:12 or 14:10 light:dark cycle (photon flux 100

μmoles/m2/s) at 18-20°C. Cultures were allowed to grow until reaching a cell density of

4,000-10,000 cells/ml at which point they were collected at mid-dark and mid-day time points.

RNA extraction, qPCR analysis, and cDNA cloning

Primer sets covering the entire predicted LpHv1 coding sequence (Table 2.1) were used for qPCR analysis performed on cells harvested at mid-light or mid-dark. Cells from 50 ml aliquots of culture were collected by centrifugation. The pellet was used immediately or flash frozen in an ethanol and dry ice bath and stored at -80°C. Samples from each timepoint were resuspended and Dounce homogenized in 1 ml TRI reagent

(Zymoresearch), and RNA was extracted according to the manufacturer's protocol. RNA was reverse transcribed using Superscript II reverse transcriptase (Invitrogen by Life

Technologies) with random primers (Invitrogen) according to the manufacturer's protocol. Quantitative real-time PCR was performed in triplicate using an Applied

Biosystems (Life Technologies) Fast 7500 thermal cycler with primers listed Table 2.1.

Thermal cycling conditions consisted of an initial denaturation at 95°C for 2 minutes followed by 40 cycles of denaturation at 95°C for 15 seconds, annealing and fluorescent data collection at 60°C for 15 seconds, and extension at 72°C for 30 seconds. Cycle thresholds and baselines were determined manually, and cycle thresholds were averaged and compared across time points.

Recombinant protein expression and purification

Recombinant LBP and LpHV1 proteins tagged with glutathione s-transferase (GST) were expressed in E. coli by induction with isopropyl b-d-1 thiopyranogalactoside (IPTG) at 23°C (LBP) or 17°C (LpHV1) for 8-24 hours. Recombinant LCF protein tagged with

6-His was expressed in E. coli by induction with IPTG at 17 C for 16-24 hours. Proteins were purified from E. coli according to the manufacturer’s instructions (GE-Healthcare for LBP and LpHV1, Qiagen for LCF). GST tag was cleaved from LpHV1 by digestion with Prescission protease according to the instructions of the vendor. In one experiment,

GST tag was cleaved from LBP by digestion with thrombin for 16 hours; the insoluble

LBP precipitate was pelleted by centrifugation and solubilized in SDS-PAGE loading buffer.

Scintillon isolation, luminescence assays, and gel analysis

Scintillons were isolated from cultures grown in two different locations and isolated using a sucrose [131] or a Percoll [122] density gradient. For luminescence assays to test metal sensitivity, the 0.5 ml fraction from a sucrose gradient with the highest luminescence was kept separate and considered to comprise “pure scintillons”. This fraction was pelleted, washed with buffer, resuspended and then diluted to about 1 ml in the extraction buffer. 100 l of the diluted scintillon preparation was added to wells of a

96 well plate. 50 l of 5x metal ion/drug solution was added to the same well. The reaction was started by injecting 150 l of 0.05 M acetic acid and luminescence was measured immediately using pClamp software (Molecular Devices). No flash was detected in the presence of detergents (0.03-0.1% TWEEN-20 or 0.1% SDS).

For gel analysis, total scintillon protein was extracted by heating for 10 min at 95°C in SDS-PAGE sample buffer. Total scintillon proteins, and preparations of purified recombinant proteins, were separated by SDS-PAGE, Western blotted onto PVDF, blocked with 5% milk or Licor blocking buffer (Licor) and probed with diluted antibodies to LCF (1:1000), LBP (1:10000), and LpHV1 (1:3000). Blots were visualized with secondary antibodies to appropriate animal IgG conjugated either to HRP or to Licor fluorescent tags. Western blotting results from different scintillon sources were directly comparable.

Fixation and Immunostaining of cells and isolated scintillons

We used the procedure of [167] with modifications. Preliminary experiments indicated that low speed centrifugation was optimum to preserve cell integrity, so all centrifugation steps were performed at 300 x g. Cells were pelleted from 50-100 ml of culture by centrifugation, washed with seawater, and fixed in 4% Paraformaldehyde in seawater at room temperature. Fixed cells were washed with phosphate buffered saline

(PBS), permeabilized in 100% methanol at 4°C, washed at room temperature with PBS supplemented with 0.1% tween and 1 % BSA (PBST-BSA), then blocked by incubation in PBST-BSA at room temperature. Primary antibodies, or serum controls diluted to the same protein concentrations, or vehicle controls, were added and incubated overnight at

4°C. Cells were collected by centrifugation and washed at room temperature with PBST-

BSA. For detection, Alexa555-Fluor-conjugated goat anti-rabbit IgG (H+L) and/or

Alexa555-Fluor-conjugated goat anti-chicken IgG (ThermoFisher, Waltham MA) was added and incubated in the dark at room temperature. Cells were washed with PBST-

BSA at room temperature, incubated with 4',6-diamidino-2-phenylindole (DAPI) 5µg/mL in PBS, and washed again. Cells were incubated in 2 drops of Vectashield overnight at

4°C in the dark and 15 to 50 μL were mounted on slides. Fixation and immunostaining of isolated scintillons were performed as described, but without the methanol permeabilization step. Cells and scintillons were visualized using a Zeiss LSM 700 confocal microscope, equipped with a 20x (whole cells) and 40x (scintillons) 1.2 NA C-

Apochromat objective. For immunofluorescent localization, all channel pinholes were set to 1 Airy Unit. Isolated scintillons identified by luciferin fluorescence were scored for the presence or absence of secondary antibody fluorescence. Confocal slices or maximum

intensity projections of the Z-stack were rendered using Zeiss Zen software, and processed using Adobe Photoshop.

Preparation of scintillon proteins for mass spectrometry

Purified scintillon preparations were concentrated by ultra-centrifugation at 4°C.

The pellet was re-suspended in SDS sample buffer, heated for 5 minutes at 95 °C, and cleared by centrifugation at 10,000 × g. 50 μL of sample was loaded and run on Novex

NuPAGE 4-12% bis-tris gels according to the manufacturer’s protocol. Gel bands corresponding to the location of the presumptive proteins were excised with a clean scalpel. Samples were processed using the Thermo Scientific In-Gel Tryptic Digestion

Kit according to manufacturer’s protocol. Gel bands were destained twice with 200 μL destaining solution (~25 mM ammonium bicarbonate in 50% acetonitrile) and incubated at 37°C with shaking for 30 minutes. Samples were reduced by incubation at 60°C for 10 min in 50mM TCEP (tris(2-carboxyethyl)phosphine) in 25 mM ammonium bicarbonate buffer. Free sulfhydryl groups were alkylated by incubation in 100 mM iodoacetamide at room temperature for 1 hour in the dark. Gel pieces were shrunk in acetonitrile. For the initial proteomic analysis samples were treated overnight with 100 ng trypsin at 30°C.

The second samples were treated with 100ng trypsin and digestion was performed at

50°C with high pressure using the PBI Barocyler (Pressure Biosciences Inc.) according to manufacturer’s protocol. Samples were dried in a SpeedVac.

Mass Spectrometry Analysis and data processing

Scintillon samples were analyzed by electrospray ionization on an Elite tandem orbitrap mass spectrometer (Thermo Scientific Inc). Nanoflow HPLC was performed by using a Waters NanoAcquity HPLC system (Waters Corporation). Peptides were trapped

on a fused-silica pre-column (100 μm i.d. 365 μm o.d.) packed with 2 cm of 5 μm (200

Å) Magic C18 reverse-phase particles (Michrom Bioresources, Inc). Subsequent peptide separation was conducted on a 75 μm i.d. x 180 mm long analytical column constructed in-house and packed with 5 μm (100 Å) Magic C18 particles, using a Sutter Instruments

P-2000 CO2 laser puller (Sutter Instrument Company). The mobile phase A was 0.1% formic acid in water and mobile phase B was 0.1% formic acid in acetonitrile. Peptide separation was performed at 250 nL/min in a 95 min run, in which mobile phase B started at 5%, increased to 35% at 60min, 80% at 65min, followed by a 5 min wash at 80% and a

25 min re-equilibration at 5%. Ion source conditions were optimized by using the tuning and calibration solution recommended by the instrument provider. Data were acquired by using Xcalibur (version 2.8, Thermo Scientific Inc.). MS data was collected by top-15 data-dependent acquisition. Full MS scan of range 350 – 2000 m/z was performed with

60K resolution in the orbitrap followed by collision induced dissociation (CID) fragmentation of precursors in iontrap at normalized collision energy of 35. The MS/MS spectra of product ions were collected in rapid scan mode.

Acquired tandem mass spectra were searched for sequence matches against

UniprotKB database using COMET. The following modifications were set as search parameters: peptide mass tolerance at 10 ppm, trypsin digestion cleavage after K or R

(except when followed by P), one allowed missed cleavage site, carboxymethylated cysteines (static modification), and oxidized methionines (variable modification/differential search option). PeptideProphet and ProteinProphet, which compute a probability likelihood of each identification being correct, were used for statistical analysis of search results. PeptideProphet probability ≥ 0.9 and ProteinProphet

probability ≥ 0.95 were used for positive identification at an error rate of less than 1%.

Only proteins identified by more than one unique peptide sequence were included in the study.

Acknowledgements

This work is dedicated to the memory of Woody Hastings, who inspired and facilitated this project. We thank Dr. David Morse (University of Montreal) for his generous gifts of LBP cDNA and antibody to LBP, and for his very helpful advice and discussions; we also thank D. R. Goodlett (University of Maryland) for advice on the proteomics analysis.

CHAPTER 3: PHYLOGENETIC ANALYSIS OF ACETYL COA CARBOXYLASES IN DINOFLAGELLATES1 3.1 Abstract

Dinoflagellates are a diverse group of organisms that produce potent bioactive secondary metabolites. The molecular mechanisms by which these compounds are produced are not yet understood. Acetyl CoA carboxylase (ACC), is a necessary protein in polyketide and fatty acid synthesis. In this study, we survey thirteen dinoflagellate transcriptomes in addition to chlorophyte green algae, haptophytes, diatoms, and alveolates to characterize ACC in these organisms. Our analysis reveals that dinoflagellates express only ACC with homomeric, eukaryotic domain architecture, including the eukaryote specific central domain. Streptavidin western blotting shows many biotinylated proteins as well as bands close to the calculated ACC size in three dinoflagellate species. Two major clades of dinoflagellate ACC were found, both within the previously defined plastidial type of homomeric ACC. One clade exclusively includes peridinin pigmented dinoflagellates. The other clade contains all thirteen dinoflagellate species including those without peridinin chloroplasts as well as non-photosynthetic dinoflagellates. Based on phylogeny and chloroplast transit peptides, these two clades likely represent two different ACCs where one is putatively plastid while the other is cytosolic.

1 Haq S, Place AR, Bachvaroff TR. 2017. Phylogenetic Analysis of Acetyl CoA Carboxylases in Dinoflagellates. In: Proença LAO, Hallegraeff GM, editors. Marine and Freshwater Harmful Algae - Proceedings of the 17 International Conference on Harmful Algae 9-14 October 2016. Florianópolis, Brazil. p. 138–141. 61

3.2 Introduction

Dinoflagellates are a genetically complex phylum including many toxin producing species that are major contributors to harmful algal blooms (HABs) worldwide [1,168].

The mechanisms by which these products are synthesized are still not known. A necessary component, and rate limiting step, of both fatty acid and polyketide synthesis is

Acetyl CoA Carboxylase (ACC). ACC is responsible for the conversion of acetyl CoA to malonyl CoA which commits the substrate to go through the FAS or PKS cycles

[66,169]. Recently a description of the ACC in the basal dinoflagellate Amphidinium carterae based on domain arrangement, expression profiles, and shotgun proteomics has been made [170]. As with other carboxylating enzymes, a biotin carrier is first charged followed by transfer of the carbon onto the acetyl CoA. In eukaryotes, a homomeric form exists as a large polypeptide consisting of sequentially arranged functional domains connected by a central domain unique to eukaryotic ACC [171]. The chloroplast and cytosolic versions can have different evolutionary histories in different algal groups. For example, in land plants the cytosolic and plastid ACC are relatively closely related, while in most stramenopiles and apicomplexans the cytosolic and plastid ACC are found in two distinct clades [172]. Characterization of ACCs in dinoflagellates will prove to be an important step in elucidating FAS/PKS pathways.

3.3 Results and Discussion

The survey of thirteen dinoflagellate transcriptome datasets revealed between one and four full-length or near full-length putative ACC sequences per species. Searches against Perkinsus marinus, Hematodinium sp., and Amoebophrya sp. did not recover sequences with the correct domain architecture. In addition, sequences from the

apicomplexan Plasmodium and Cryptosporidium were difficult to align and were on very long branches, so they were excluded from the analysis.

The ACC phylogeny contains a large number of poorly supported nodes, but specific clades such as the chlorophyte green algae (100 % bootstrap), haptophytes (100%), diatoms (100%) and alveolates (78%) were well supported (Figure 3.1). The monophyletic alveolates, included the chromerids, apicomplexans and dinoflagellates and is generally consistent with overall organismal phylogeny. Within dinoflagellates, there were two major clades of ACC. One included all eight of the peridinin pigmented dinoflagellates (with duplications in three species) and none of the alternately pigmented or non-photosynthetic species. Of the 11 sequences in the peridinin-associated clade, 8 were full length sequences in the alignment, and 4 had a putative chloroplast transit peptide (cTP) with a cleavage site prior to the first conserved domain. Based on a restricted distribution to the peridinin-pigmented dinoflagellates and predicted cTP, this clade is proposed to contain the chloroplast-targeted ACC. Of the anomalously pigmented dinoflagellates only a single sequence from G. foliaceum had a cTP. Both the alternately pigmented diatom-containing species also had another ACC version imbedded with the diatoms, while the second ACC from the haptophyte pigmented Karenia brevis and Karlodinium veneficum did not group with the haptophytes in the phylogenetic analysis. A second clade contained sequences from all thirteen of the dinoflagellates surveyed. This second clade, included all the species surveyed whether non- photosynthetic or alternately pigmented, and is proposed as the cytosolic ACC. Of the 10 full length, only one, the A. sanguinea ACC had a score over cTP threshold but the cut site was inside the mature protein. Within the putative cytosolic clade, a smaller, but well

supported division is seen. Four relatively distantly related peridinin pigmented species,

Prorocentrum hoffmanianum, Symbiodinium sp., Akashiwo sanguinea, and

Gambierdiscus australes made up the smaller clade, while the remaining nine dinoflagellates, including all the alternately pigmented or non-photosynthetic species were in the larger clade. This division of the cytosolic clade into two parts is not congruent with current dinoflagellate phylogeny, nor in general is there good bootstrap support within the two cytosolic clades. However, within the peridinin-pigmented chloroplast ACC most of the nodes are well supported (all but one have bootstrap support

>90%), but the branching pattern is not consistent with current phylogenies of dinoflagellates using large numbers of concatenated proteins [41,162].

Overall, the phylogeny suggests dinoflagellates duplicated a homomeric ACC present in the common ancestor of apicomplexans, chromerids and dinoflagellates, and retained one version for chloroplast-associated functions. This peridinin chloroplast version was then lost from the alternately pigmented and non-photosynthetic species, as part of chloroplast replacement. A second duplication with differential loss is also possible within the cytosolic ACC, with one distinct version found exclusively within four species.

100 Crypthecodinium cohnii Haptophyte derived plastid 72 Scrippsiella trochoidea Diatom containing Glenodinium foliaceum

Levanderina fissa C

Non-photosynthetic 60 Gymnodinium catenatum y 87 t 100 Karenia brevis o Peridinin plastid Karenia brevis s

Karlodinium veneficum l * cTP predicted 100 100 i Oxyrrhis marina c Oxyrrhis marina 73 Amphidinium carterae 87 Prorocentrum hoffmanianum 75 Symbiodinium sp. B1 100 Akashiwo sanguinea Gambierdiscus sp 100 Gymnodinium catenatum 61 100 Gymnodinium catenatum C

99 P Levanderina fissa h

* e

o 100 Akashiwo sanguinea r

i 100 r Gambierdiscus australes d

* o

99 i

p 100 Scrippsiella trochoidea n

Symbiodinium sp. B1 a 78 n 99 100 Prorocentrum hoffmanianum s 100 Prorocentrum hoffmanianum* t Amphidinium carterae* Symbiodinium sp. B1 100 Apicomplexans 100 Chromera velia Vitrella brassicaformis 100 Diatoms 100 100 Glenodinium foliaceum* Diatom-containing dinoflagellates 65 96 Durinskia baltica Phaeodactylum tricornutum Guillardia theta Aureococcus anophagefferens 82 100 Green algae 93 Aureococcus anophagefferens 100 Haptophytes 75 Nannochloropsis oculata 0.3 Ectocarpus siliculosus Substitutions per site 100 Karlodinium veneficum Karenia brevis Haptophyte derived plastid Figure 3.1: Phylogenetic Analysis of Acetyl CoA Carboxylases in Thirteen Dinoflagellate Species.

Maximum likelihood phylogeny of ACC from dinoflagellates, chromerids, apicomplexans, heterokonts, cryptophyte, haptophyte, and green algae. The most likely tree is shown with support for 1000 rapid bootstrap replicates above the branches when over 50%. Species marked with * indicate a chloroplast transit peptide (cTP) was found while species in bold indicate a spliced leader sequence was found. The bulk of the dinoflagellates contain peridinin-pigmented plastids, but two alternately pigmented groups were included with haptophyte and diatom derived symbionts. In addition, there are two non-photosynthetic species included in the tree.

Protein expression and biotinylation was confirmed using western blotting with a streptavidin-HRP conjugate to visualize all natively biotinylated proteins resulting in several prominent bands in three species surveyed. Focusing on high molecular weight bands closest to the calculated protein mass from transcriptome data we putatively assign bands shown in red boxes as ACC (Figure 3.2). The molecular weight of these proteins for K. veneficum was 251 kDa, A. sanguinea at 276 kDa, and two bands for A. carterae at

276 and 262 kDa. These assignments were previously confirmed by shotgun proteomics for A. carterae and are to be confirmed in future by the same method for K. veneficum

and A. sanguinea [170]. This survey of dinoflagellate ACCs reveals that most species contain multiple ACC proteins that potentially localize to the plastid or cytosol.

Figure 3.2: Western blotting of natively biotinylated proteins in three dinoflagellate species.

Natively biotinylated proteins were visualized with western blotting using streptavidin-HRP conjugate. A. K. veneficum and A. sanguinea were run at concentration of 304,000 and 14,400 cell equivalents respectively. B. A. carterae was run at a concentration of 193,200 cell equivalents. The putatively assigned ACC region is shown in a red box for all species.

3.4 Materials and Methods

The previously identified Amphidinium carterae ACCs contained the conserved domains, carboxy loading domain (cd 06850), biotin carrier (COG0439), the central domain (pfam 08326), and a carboxy-transferase domain (pfam 01039), and were compared against twelve of the available dinoflagellate transcriptomes [173], including the non-photosynthetic Crypthecodinium cohnii and Oxyrrhis marina. In addition, two atypically pigmented groups were sampled, the diatom-containing species Glenodinium foliaceum and Durinskia baltica, and the haptophyte-pigmented Karenia brevis and

Karlodinium veneficum [174]. Initial analysis (data not shown) placed all the dinoflagellate ACC within the plastidial clade as defined by Huerlimann and Heimann in

2015; the plastidial sequences from additional outgroups included the apicomplexans,

heterokonts, cryptophytes and green algae [172]. The ACC were aligned using Clustal

Omega with the full-iteration mode and the alignment examined in Mesquite. The phylogenetic analysis was performed with RAxML using 1000 rapid bootstraps with the optimal JTT amino acid substitution matrix on an alignment of 2876 amino acids

(TreeBase accession TB2:S21383). Amino acid sequences were analyzed using the

ChloroP CBS predictive software to predict the presence of a chloroplast transit peptides.

Sequences were analyzed and evaluated by cutoffs recommended by the CBS group

[175].

The A. carterae (CCMP 1314) was grown in ESAW at a salinity of 32 [176], while

K. veneficum (CCMP 2936) and Akashiwo sanguinea were grown in ESAW at a salinity of 15 in polystyrene culture flasks [177]. The cultures were maintained under constant light at 50m cm-2 s-1 with a 14:10 light:dark schedule. Cultures were grown to a high cell density and collected at mid-day time points. The cells from 50 ml of culture were centrifuged at 350 x g for 10 minutes at room temperature in a conical tube. The supernatant was poured off and the pellet was flash frozen in a bath of dry ice and ethanol and stored at -80°C. Cell pellets of A. carterae, K. veneficum, and A. sanguinea were resuspended in 1ml of 1X SDS sample buffer (50mM Tris-Cl pH 6.8, 2% SDS, 0.1% bromophenol blue, 10% glycerol, 100mM dithiothreitol), boiled for 5 minutes at 95 °C, and centrifuged at 10,000 x g for 5 minutes. Samples were loaded at varying cell equivalents and separated on Novex NuPAGE 3-8% tris-acetate gels at 150 V (constant) for 1 hour. Proteins were transferred onto a PVDF membrane using the high molecular weight program on the Trans Blot Turbo Transfer System (Bio-Rad) for 14 minutes. The membrane was rinsed with deionized water and blocked with 1X iBind solution (Life

Technologies) for 1 hour at room temperature. The membrane was incubated with a streptavidin-HRP conjugate (Invitrogen) using the iBind Western Device (Life

Technologies) per manufacturers protocol at a dilution of 1:1000. The blot was incubated for 5 minutes in Clarity ECL Substrate (Bio-Rad) and imaged on the ChemiDoc Touch

Imaging System by Bio-Rad.

Acknowledgements

This is contribution #5339 from UMCES, #17-202 for IMET, and #884 from the

ECOHAB program. This research was funded in part by grants from OHH NIH

R01ES021949-01/NSFOCE1313888 and NOAA-NOS-NCCOS-2012-2002987 to ARP and the Ratcliffe Environmental Entrepreneur Fellowship. We thank Dr. Wayne Coats for providing A. sanguinea cultures.

CHAPTER 4: CHARACTERIZATION OF ACETYL COA CARBOXYLASES IN THE BASAL DINOFLAGELLATE AMPHIDINIUM CARTERAE 1

4.1 Abstract

Dinoflagellates make up a diverse array of fatty acids and polyketides. A necessary precursor for their synthesis is malonyl-CoA formed by carboxylating acetyl CoA using the enzyme acetyl-CoA carboxylase (ACC). To date, information on dinoflagellate ACC is limited. Through transcriptome analysis in Amphidinium carterae, we found three full- length homomeric type ACC sequences; no heteromeric type ACC sequences were found.

We assigned the putative cellular location for these ACCs based on transit peptide predictions. Using streptavidin Western blotting along with shotgun proteomics, we validated the presence of ACC proteins. Additional bands showing other biotinylated proteins were also observed. Transcript abundance for these ACCs follow the global pattern of expression for dinoflagellate mRNA messages over a diel cycle. This is one of the few descriptions at the transcriptomic and protein level of ACCs in dinoflagellates.

This work provides insight into the enzymes which make the CoA precursors needed for fatty acid and toxin synthesis in dinoflagellates.

4.2 Introduction

Marine dinoflagellates, a type of phytoplankton, are primary producers in the ecosystem that are of great environmental and economic importance. Toxin producing,

1 Haq, S.; Bachvaroff, TR.; Place, AR. Characterization of Acetyl-CoA Carboxylases in the Basal Dinoflagellate Amphidinium carterae. Mar. Drugs 2017, 15, 149. 69

harmful algal bloom (HAB) species are responsible for fish mortality as well as human and animal illness worldwide [1,178,179]. Economically, dinoflagellates and algae produce important compounds such as docosahexaenoic acid (DHA) and algal biofuels [180,181].

A necessary protein involved in the production of both fatty acid (FA) products as well as polyketide (PKS) products is acetyl CoA carboxylase (ACC), which produces “activated acetate” units as malonyl CoA [66,169].

ACCs are natively biotinylated proteins that convert acetyl CoA (two carbons) to malonyl CoA (three carbons). Once this conversion occurs, the substrate, malonyl CoA, is committed to go through fatty acid synthase (FAS) or polyketide synthase (PKS) cycles.

Within the ACC protein there are several domains. Biotin carboxylase (BC) is responsible for carboxylating the biotin cofactor, which is then transferred to the carboxyl transferase

(CT) domain via the biotin carrier protein (BCCP) which serves as an attachment point for the biotin. The CT domain then carboxylates the acetyl CoA, converting it to malonyl CoA

[66,67]. In addition, in the eukaryotic ACC protein there is a large middle domain, the central domain, with no currently defined function [171].

Two different ACC forms exist, a homomeric form found in most eukaryotes and a heteromeric form found in most prokaryotes. The homomeric form exists as one large polypeptide where all the subunits are on one protein with sizes typically in the 250 kDa range [182]. Heteromeric subunits exist as separate proteins that work together to carboxylate acetyl CoA. Sub-unit sizes and organization are species-dependent [65–67].

The ACC cellular location will vary depending on the organism. In animal cells, ACCs can be found in the cytosol and mitochondria, while plant and algal ACCs are located in the cytosol and plastid [183,184]. Plastid ACC blocking compounds are used as

commercially available herbicides, while mitochondrial and cytosolic versions are studied for human illnesses such as diabetes and obesity [65]. Nomenclature of ACCs also varies depending on the organism. For the purposes of our study, we will refer to plastid ACC as

ACC1 and cytosolic as ACC2 in keeping with previous studies involving algal ACC [67].

In dinoflagellates, information about ACC and its role in fatty acid, toxin, and other secondary metabolite production is limited. Amphidinium carterae is a polyketide producing dinoflagellate that produces amphidinol. This linear polyketide toxin forms pores in cell membranes after complexing with sterols, resulting in cell lysis [185,186].

With characteristics such as worldwide distribution, production of amphidinol toxin, and classification as a basal dinoflagellate, Amphidinium is often utilized as a model dinoflagellate genus [62,187]. In our transcriptome study, we utilized A. carterae (CCMP

1314) to characterize and annotate ACC in dinoflagellates.

4.3 Results

Transcriptome Analysis of Amphidinium carterae (CCMP 1314) ACCs

The transcriptome analysis of Amphidinium carterae revealed three homomeric type

ACCs containing all the domains normally present in eukaryotic homomeric ACCs in the correct order (Figure 4.1). We refer to these three ACC sequences as ACC1a, ACC1b, and

ACC2. GenBank accession numbers are ACC1a (MF043929), ACC1b (MF043930), and

ACC2 (MF043931). The ACC1a assembled transcript began with a 144 base 5′ untranslated region (UTR), a 6327 base open reading frame (ORF) and a 210 base 3′ UTR.

ACC1b contained a 144 base 5′ UTR, a 6240 base ORF and a 26 base 3′ UTR. The ACC2 transcript began with eight bases of spliced leader (SL) sequence, followed by a 67 base 5′

UTR, a 6237 base ORF, and ended with a 133 base 3′ UTR. Based on the pairwise

alignment of ACC1 and ACC2 using BlastP, the two proteins were 47% identical over

2047 amino acids. The two subtypes, ACC1a and ACC1b, were 98% identical over 2044 amino acids and 98% identical over 6374 nucleotides (Figure 4.1a).

Based on the Center for Biological Sequence Analysis (CBS) ChloroP prediction software, ACC1a and 1b contained a chloroplast transit peptide (cTP), with a cTP score of

0.567, while ACC2, with a cTP score of 0.442, did not [175,188,189]. The cTP length was predicted to be 37 amino acids for both ACC1a and 1b (Figure 4.1a). When the three ACC sequences were aligned, ACC1a and 1b started 48 residues before ACC2. This offset was demonstrated by the alignment of a conserved tyrosine at residue 60 of ACC1a and 1b, which corresponded to residue 12 of ACC2. For the initial 48 residues, Kyte-Doolittle hydrophilicity plots revealed a pattern of 18 hydrophobic residues immediately preceding the predicted cTP, followed by a seven amino acid hydrophilic region in ACC1a and 1b.

ACC2 did not have a similar pattern of residues (Figure 4.1b) [190].

Figure 4.1: The expressed acetyl CoA carboxylases (ACCs) in Amphidinium carterae (CCMP 1314).

(A) The three homomeric ACC sequences discovered through transcriptome assembly and analysis that contained the four domains of eukaryotic ACC shown above. The table shows the calculated molecular weight, pI, FPKM, cTP score, and predicted cellular location based on ChloroP. (B) Kyte-Doolittle hydrophilicity plot is shown for both ACC sequences. The sequence alignment is anchored by a conserved tyrosine residue shown in red boxes. A red arrow is drawn to show the predicted cleavage site of the chloroplast transit peptide.

Quantification of ACC Transcript Levels in Amphidinium carterae

ACC mRNA quantification was consistent with the global transcript patterns in dinoflagellates over a diel light cycle. Previously, in A. carterae, global changes in mRNA abundance were observed near the light:dark transition [191,192]. The ACC transcripts closely followed this pattern with decreasing abundance towards the end of the light period, followed by recovery in the dark period (Figure 4.2). Ribosomal protein L7 (RPL7) was used as a native reference gene. Cycle thresholds for ACC were normalized to RPL7 values, and all samples were run in triplicate. The transcript abundance was generally

ACC2 > ACC1b > ACC1a at most time points, albeit with some reordering during the very low abundance measures around the light:dark transition.

Profile of Natively Biotinylated Proteins in Amphidinium carterae

Streptavidin-HRP Western blot over the diel cycle shows the natively biotinylated proteins in A. carterae ranging from 40–280 kDa. Samples were collected 6 h before and after the light:dark transition, and 300,000 cell equivalents per lane were run for each time point. Two bands are observed at 272 kDa and 258 kDa (Figure 4.3), which are consistent with, but larger than the estimated sizes of the different ACCs observed in the transcriptome analysis (Figure 4.1). A prominent band is also observed at 160 kDa; we hypothesize this to be pyruvate carboxylase, but further experiments are required to verify this assignment. Densitometry results that were normalized to the maximum pixel density show a slight decrease in protein abundance at 6 h before and 6 h after the light:dark transition. The highest protein abundance was observed 2–4 h prior to the dark period

(Figure 4.3).

Figure 4.2: Quantification of ACC mRNA in Amphidinium carterae.

Cells were collected at one hour increments 6 h before and after the light:dark transition. The white background shows the light period (T − Hour) while the shaded region shows the dark period (T + Hour) with T0 representing when the lights are turned off. The mRNA was quantified using quantitative reverse transcriptase PCR to observe the fold change relative to RPL7 as a native reference gene. The apparent peak at −1 h is due to a greater drop in RPL7 transcript abundance than observed with the ACCs.

Figure 4.3: Streptavidin Western blotting of Amphidinium carterae.

(A) Biotinylated proteins were surveyed over a 12-h period using Western blotting with streptavidin horseradish peroxidase conjugate at 300,000 cell equivalents per time point in A. carterae. T0 represents lights off, with hours before and after represented as T − hour and T + hour, respectively. (B) Densitometry pixel densities were normalized to the maximum density peak and plotted against time for ACC1a, 1b and 2.

Peptide Coverage of ACCs in Amphidinium carterae

A shotgun proteomics experiment was conducted using high molecular weight protein bands (approximately 220, 160 kDa) that were separated on sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The gel bands were then digested using trypsin, producing peptide fragments. Many of these peptide fragments could be mapped back to the three ACCs in A. carterae and covered all the domains in these proteins. The highest number of peptide fragments to map back to ACC were present in the 220 kDa gel band. Peptide coverage of 20.3%, 20.4%, and 18.5% are observed for ACC1a, ACC1b, and

ACC2, respectively (Figure 4.4). The high similarity between ACC1a and 1b mean that some peptides mapped to both predicted proteins.

Figure 4.4: Shotgun proteomic analysis of ACCs in Amphidinium carterae.

Peptide sequences were determined using LC-MS/MS proteomic techniques. Peptide coverage is represented with red arrows, while blue arrows indicate different ACC domains. Black arrows denote the biotin attachment site in each sequence.

4.4 Discussion

ACCs are the rate-limiting commitment step in fatty acid and polyketide synthesis

[169,193]. Due to their involvement in essential processes, these proteins have been heavily studied in many organisms. However, little is known about ACCs in dinoflagellates, which are presumably involved in fatty acid and toxin synthesis. In this study, we describe ACCs present in a basal toxin-producing dinoflagellate Amphidinium carterae.

We began our investigation by searching our transcriptome database for ACCs in A. carterae [194]. This analysis uncovered three sequences that contained all the domains ordinarily found in eukaryotic ACCs; no heteromeric type ACCs were found. A biotin carboxylase domain, a biotin attachment site, a central domain, and carboxyl transferase domain were all present (Figure 4.1). Finding these domains on one polypeptide in addition to the inclusion of the central domain, a non-catalytic domain specific to eukaryotic ACCs, was consistent with a homomeric type ACC [171].

We predict that the two types of ACCs compartmentalize to different cellular locations where ACC1a and 1b are targeted to the plastid, while ACC2 is cytosolic. In addition to phylogenetic analysis, we observe in the ACC1 sequences a chloroplast transit peptide

(cTP) with a hydrophobicity pattern consistent with a signaling peptide (Figure 4.4)

[190,195]. This is similar to other plastid-containing organisms that are phylogenetically similar, such as Toxoplasma gondii, which has two different homomeric ACCs, one found in the plastid while the other is found in the cytosol [196]. This varies from plants such as

Arabidopsis thaliana, which contain two cytosolic versions of ACC [67,197].

This prediction is validated by high cTP scores for ACC1a and 1b when analyzed by the ChloroP predictive software (Figure 4.1). ChloroP outputs cTP scores between 0.4–0.6, where scores above 0.5 are predicted to contain a cTP. These sequences have a cTP score that exceeds the 0.55 threshold, which is considered a strong prediction for the presence of a cTP [175]. In addition, when these sequences are aligned to a tyrosine residue that is conserved across multiple dinoflagellate and algal species, ACC1a and 1b have an additional 48 amino acids at the N-terminus of the protein sequence (Figure 4.1) [190].

This 48 amino acid sequence in ACC1a and 1b contains a group of hydrophobic amino acids immediately preceding the predicted site of the transit peptide, a pattern that has been observed in plastid-targeted proteins in other dinoflagellates [195]. ACC2 did not have a predicted transit peptide, nor was there a pattern of hydrophobic residues at the N-terminus of the sequence. Taken together, we reason that ACC1a and 1b are located in the plastid, while ACC2 is located in the cytosol of A. carterae.

We further validated the presence of ACCs in A. carterae by Western blotting and shotgun proteomics to verify the presence of these proteins, as dinoflagellates exhibit high

levels of translational control [191,192]. Using streptavidin-HRP Western blotting on axenic A. carterae, we obtained a profile of all natively biotinylated proteins over a diel cycle (Figure 4.3). Two bands observed at 272 kDa and 258 kDa were putatively assigned as our ACCs of interest, ACC2 and ACC1, respectively. We also observed a slight decrease in protein abundance at the time points furthest from the end of the light:dark transition at

−6 and +6 h (Figure 4.3).

Results from shotgun proteomics showed extensive peptide coverage of all three ACCs observed in A. carterae (Figure 4.4). Due to the high peptide coverage (Figure 4.4) and high expression level (fragments per kilobase per million mapped reads) (Figure 4.1), these proteins are abundantly expressed and translated in the cell [194]. These proteins are present in large amounts likely because they catalyze a limiting step in essential fatty acid synthesis and the production of secondary metabolites through polyketide synthases [61].

These proteins now represent an attractive target for further dissection of the PKS/FAS pathways in dinoflagellates.

Conclusions:

To date, few descriptions of acetyl CoA carboxylases in dinoflagellates exist [198].

Our current study shows the presence of two types of homomeric type ACCs in a basal dinoflagellate, A. carterae. We plan to further characterize these ACCs with phylogenetic comparisons to other species as well as functional studies to show their involvement in fatty acid and polyketide production.

4.5 Materials and Methods

Sequence Acquisition and Searching

The transcriptome used in this study was assembled from the 100 base paired end reads in the short read archive SRX722011 at NCBI, assembled de novo with Trinity [162]. The assembled data was compared to the NCBI reference sequence protein database by

BLASTx with e-value cut off of 1 x e-10 and the results were compiled into a relational database. Putative ACC sequences were extracted based on the sequence descriptions in the BLAST database and annotated using conserved domain search tool as a default of online (interactive) BLAST against the NCBI conserved domain database (CDD). Iterative searching using individual conserved domains was then conducted; each sequence extracted from A. carterae was compared to the CDD database using interactive BLAST to NCBI, tabulated, and then used to extract similar sequences from A. carterae (using

BLASTx with evalue cutoff of 1 x e-10). Any novel sequences were extracted, translated, and compared to the CDD, and added to the query set. When no new sequences with either the biotin carboxyl carrier protein (cd06850), ACC synthase central domain (pfam08326), or the carboxytransferase (pfam01039) conserved domains were found, the search was terminated. Only sequences with the domain architecture shown in Figure 4.1 were considered genuine ACC synthases. No single subunit sequences corresponding to heteromeric ACC synthases, and with matches to individual domains, were found.

ChloroP Analysis of Sequences

Amino acid sequences for ACCs in A. carterae were analyzed using the CBS ChloroP predictive software http://www.cbs.dtu.dk/services/ChloroP/. Sequences were pasted into

the input window in FASTA format and jobs were submitted using detailed output analysis. cTP scores were evaluated using cut-offs published by the CBS group [175].

Kyte-Doolittle Hydrophilicity Analysis

Kyte-Doolittle analysis was conducted on the first 100 amino acids of ACC protein sequences in A. carterae using Kyte-Doolittle hydrophilicity analysis in the protein analysis toolbox feature of MacVector analysis software (MacVector Inc., Version 15.1.5,

Apex, NC, USA).

Cell Culture and Harvest for Diel Quantification

Amphidinium carterae (CCMP 1314) was grown in enriched seawater artificial water

(ESAW) medium modified to contain 10 mM HEPES in a 20-L multiport polycarbonate carboy [176]. The cultures were maintained under constant light at 150 μm·cm−2·s−1 with a 14:10 light:dark schedule and bubbling of air infused with CO2 controlled by an

American Marine (Ridgefield, CT) pH controller set to turn on above pH 8.2 and turn off at pH 7.6. Two liters of medium was added to an actively growing culture on Monday,

Wednesday, and Friday, until a cell density of 172,000 cells/mL was reached on the morning of the experiment. Then, 250 mL of culture was aliquoted into 26 × 75 cm2 polystyrene culture flasks from Corning (Corning, NY, USA) and placed along the light bank to achieve equivalent light exposure as the stock culture in two rows of 13, with duplicates in an opposing direction. Each duplicate pair was taken for harvest according to the following schedule relative to the transition from light to dark in hours: −6.0, −4.0,

−2.0, −1.5, −1.0, −0.5, 0, 0.5, 1.0, 1.5, 2.0, 4.0, 6.0. The samples were each split into 50 mL and 200 mL aliquots. Each aliquot was centrifuged at 1000× g for 10 min at 20 °C. The

200 mL aliquot was frozen at −80 °C and the 50 mL aliquot was resuspended in 1 mL of

TRI Reagent® (Sigma-Aldrich, Saint Louis, MO, USA).

RNA Extraction and qPCR Analysis

Amphidinium carterae samples from each time point were extracted using TRI

Reagent® according to the manufacturer’s protocol using 1 mL of reagent. RNA was quantified on a Qubit™ Fluorometric Quantitation system (Thermo Fisher Scientific,

Waltham, MA, USA). RNA was reverse transcribed using SuperScript™ II reverse transcriptase (Invitrogen by Life Technologies) with Random Primers (Invitrogen,

Carlsbad, CA, USA) as per the manufacturer’s protocol. The cDNA was used as a template for quantitative real-time PCR using an Applied Biosystems (Life Technologies, Carlsbad,

CA, USA) Fast 7500 thermal cycler in triplicate, with the following reaction setup and primers listed in Table 1: 6 μL diethyl pyrocarbonate (DEPC) treated water, 2 μL of combined forward and reverse primers at 5 μm each, 10 μL of iTaq™ 2X master mix containing SYBR™ green and ROX (Bio-Rad, Hercules, CA, USA), and 2 μL of template cDNA at 10 ng/μL. Thermal cycling conditions consisted of 40 cycles of an initial denaturation at 95 °C for 2 min followed by 40 cycles of denaturation at 95 °C for 15 s, and annealing and fluorescent data collection at 60 °C for 1 min. Cycle thresholds and baselines were determined manually and were averaged and compared across time points.

Cycle thresholds of triplicate ACC samples were averaged, and those values were normalized to CT values of RPL7.

Table 4.1: Primer sequences for mRNA quantification experiments.

Cell Culture and Harvest for SDS-PAGE and Mass Spectrometry

Amphidinium carterae was cultured as described above. Cultures were grown to a high cell density (300,000 cells/mL or higher) as measured with a Coulter counter. Cells were collected at midday time points. Aliquots (50 mL) of culture were poured into a 50-mL conical tube and centrifuged at 350× g for 10 min at room temperature. The supernatant was poured off and the tube containing the pellet was placed in a bath containing ethanol and dry ice until frozen. The samples were then stored at −80 °C until analysis.

Western Blotting

Cell pellets of A. carterae were resuspended in 1 mL of 1X SDS sample buffer (50 mM Tris-Cl pH 6.8, 2% SDS, 0.1% bromophenol blue, 10% glycerol, 100 mM dithiothreitol), incubated for 5 min at 95 °C, and centrifuged at 10,000× g for 5 min.

Samples were loaded onto gel at 300,000 cell equivalents per lane and separated on

Novex™ NuPAGE 3%–8% tris-acetate gels at 150 V (constant) for 1 h. Proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane using the high molecular weight program on the Trans Blot® Turbo™ Transfer System (Bio-Rad) for 14 min. The membrane was rinsed with deionized water and blocked with 1X iBind™ solution (Life

Technologies) for 1 h at room temperature. The membrane was incubated with a

streptavidin-HRP conjugate using the iBind™ Western Device (Life Technologies) as per the manufacturer’s protocol at a concentration of 1:1000. Blot was incubated for 5 min in

Clarity ECL Substrate (Bio-Rad) and imaged on the ChemiDoc Touch Imaging System by

Bio-Rad.

Preparation of Proteins for Mass Spectrometry

The A. carterae cell pellet was resuspended in 1X SDS sample buffer (50 mM Tris-Cl pH 6.8, 2% SDS, 0.1% bromophenol blue, 10% glycerol, 100 mM dithiothreitol), incubated for 5 min at 95 °C, and centrifuged at 10,000× g for 5 min. 500,000 cell equivalents per lane were loaded and run on Novex™ NuPAGE 3–8% tris-acetate gels as per the manufacturer’s protocol. Gel bands corresponding to the location of the presumptive proteins were excised with a sterile scalpel. Samples were processed using the In-Gel

Tryptic Digestion Kit (Thermo Fisher Scientific) as per the manufacturer’s protocol. Gel bands were destained twice with 200 μL destaining solution (~25 mM ammonium bicarbonate in 50% acetonitrile) and incubated at 37 °C with shaking for 30 min. Samples were reduced by incubation at 60 °C for 10 min in 50 mM TCEP (tris(2- carboxyethyl)phosphine) in 25 mM ammonium bicarbonate buffer. Free sulfhydryl groups were alkylated by incubation in 100 mM iodoacetamide at room temperature for 1 h in the dark. Gel pieces were shrunk in 100% acetonitrile. The acetonitrile was removed, and the gel band was allowed to air dry for 10 min. The dried gel pieces were treated with 100 ng trypsin, in a 25 mM ammonium bicarbonate buffer, and incubated overnight at 30 °C with shaking. The supernatant was removed, placed in a clean tube, and then dried completely in a SpeedVac and stored at −20 °C until analysis.

Mass Spectrometry Analysis and Data Processing

Peptide samples were analyzed by electrospray ionization on a tandem Orbitrap Elite mass spectrometer (Thermo Scientific Inc.). Nanoflow HPLC was performed by using a

Waters NanoAcquity HPLC system (Waters Corporation, Milford, MA, USA). Peptides were trapped on a fused-silica pre-column (100 μm i.d. 365 μm o.d.) packed with 2 cm of

5 μm (200 Å) Magic C18 reverse-phase particles (Michrom Bioresources, Inc., Auburn,

CA, USA). Subsequent peptide separation was conducted on a 75 μm i.d. × 180 mm long analytical column constructed in-house and packed with 5 μm (100 Å) Magic C18 particles, using a Sutter Instruments P-2000 CO2 laser puller (Sutter Instrument Company, Novato,

CA, USA). The mobile phase A was 0.1% formic acid in water and the mobile phase B was 0.1% formic acid in acetonitrile. Peptide separation was performed at 250 nL/min in a

95-min run, in which mobile phase B started at 5%, increased to 35% at 60 min, 80% at 65 min, followed by a 5-min wash at 80% and a 25-min re-equilibration at 5%. Ion source conditions were optimized by using the tuning and calibration solution recommended by the instrument provider. Data were acquired by using Xcalibur (version 2.8, Thermo

Scientific Inc.). MS data was collected by top-15 data-dependent acquisition. Full MS scan of the range of 350–2000 m/z was performed with 60 K resolution in the Orbitrap followed by collision-induced dissociation (CID) fragmentation of precursors in an ion trap at the normalized collision energy of 35. The MS/MS spectra of product ions were collected in rapid scan mode.

Acquired tandem mass spectra were searched for sequence matches against the

UniprotKB database as well as a six-frame translation of the transcriptome data using

COMET. The following modifications were set as search parameters: Peptide mass tolerance at 10 ppm, trypsin digestion cleavage after K or R (except when followed by P),

one allowed missed cleavage site, carboxymethylated cysteines (static modification), and oxidized methionines (variable modification/differential search option). PeptideProphet and ProteinProphet, which compute a probability likelihood of each identification being correct, were used for statistical analysis of the search results. PeptideProphet probability

≥ 0.9 and ProteinProphet probability ≥ 0.95 were used for positive identification at an error rate of less than 1%. Only proteins identified by more than one unique peptide sequence were included in the study.

Acknowledgments:

This is contribution #5338 from UMCES, #17-201 for IMET, and #886 from the

ECOHAB program. This research was funded in part by grants from OHH NIH

R01ES021949-01/NSFOCE1313888 and NOAA-NOS-NCCOS-2012-2002987 to ARP and the Ratcliffe Environmental Entrepreneur Fellowship. The authors would like to gratefully acknowledge David R. Goodlett (University of Maryland), Young Ah Goo

(Northwestern University) and Bao Tran (Excet, Inc.) for their assistance with the proteomics data.

CHAPTER 5: INVESTIGATING A MULTI-DOMAIN POLYKETIDE SYNTHASE IN AMPHIDINIUM CARTERAE 1 5.1 Abstract

Dinoflagellates are unicellular organisms often implicated in harmful algal bloom

(HAB) events aggravated by potent toxins produced through polyketide synthase (PKS) pathways. However, to date we still do not know the exact mechanisms by which these toxins are synthesized. We first investigated how cerulenin, a fatty acid inhibitor, would affect 14C acetate incorporation into fats and toxins in the toxin producing dinoflagellate

Amphidinium carterae. Cerulenin inhibits acetate incorporation into amphidinol, sulpho- amphidinol, and all fat classes. Next, we focused on a multi-modular PKS sequence obtained through transcriptome sequencing that has been found in all dinoflagellates surveyed to date and is missing an AT domain in toxin producing species. This domain arrangement is similar to non-iterative with trans-AT PKSs that have been observed in other algal species. We find evidence of partial protein production via western blotting suggesting some protein processing is occurring. Tandem mass spectrometry analysis of synthetic peptides with the same sequence as the active site of all the KS domains in the multi-modular PKS reacted with cerulenin reveal a strong covalent bond that remains bound even with collision induced dissociation. This leads us to believe that this multi- modular PKS was inhibited by cerulenin and that trans-ATs are able to mediate the type of product produced with a non-iterative mechanism.

1 Haq S, Oyler BL, Williams EP, Khan MM, Goodlett DR, Nita-Lazar A, Bachvaroff TR, Place AR. 2018. Investigating a Multi-Modular Polyketide Synthase in Amphidinium carterae. In Preparation. 86

5.2 Introduction

Dinoflagellates are unicellular protists often implicated in harmful algal bloom (HAB) events which result in fish kills and human illnesses worldwide as a result of potent toxin production [4,199]. In vivo labeling studies have shown acetate incorporation patterns in dinoflagellate toxins across multiple species are consistent with a polyketide origin for these toxins [79–81].

Polyketide synthases (PKS) are similar to fatty acid synthases (FAS), where individual domains sequentially add activated acetate-derived 2-C units to a growing carbon chain

[61,69]. The PKS process begins when an acyl monomer is loaded onto a phosphopantetheine “arm” of the acyl carrier protein (ACP) by acyl transferase (AT) domains. The ACP is the site of monomer attachment as well as the site of condensation and reduction reactions that occur during PKS biosynthesis. Condensation reactions are carried out by ketosynthase (KS) domains while reduction reactions are mediated by ketoreductase (KR), dehydratase (DH), and enoylreductase (ER) domains. At the end of product synthesis the thioesterase (TE) domain is responsible for the release of the finished product [72,73,76,200].

Figure 5.1: Structures of three polyenoic fatty acids synthesized by dinoflagellates.

Dinoflagellates are known to make polyenoic fatty acids like octadecapentaenoic acid

(18:5n-3) [201], first identified from dinoflagellates (dinophytes) and, thereafter, found in raphidophytes and haptophytes (Figure 5.1). In these microalgae, 18:5 was shown to be

confined to galactolipids [202,203]. As suggested by Bell et al. (1997), 18:5n-3 might be more widespread in microalgae than thought, as 18:5n-3 might have been misidentified as

20:1n-9 in many studies. The most reasonable hypothesis about 18:5n-3 synthesis is the existence of a yet uncharacterized delta3-desaturase or a polyunsaturated fatty acid (PUFA) synthase.

In addition to PKSs, non-ribosomal peptide synthetases (NRPS) and NRPS/PKS hybrids are also present in dinoflagellates. These pathways have a mechanism similar to

PKSs, but amino acids are incorporated into the product. Adenylation (A) domains recognize specific amino acids to incorporate. The thiolation (T) domains form an aminoacylthioester intermediate while the condensation (C) domains form the peptide bond between substrates. Finally, the thioesterase (TE) domain releases the final product

[204]. PKS/NRPS hybrids combine domains from both of these pathways [61,200].

Until recently much of the research in dinoflagellate PKSs has been focused on chemical and biochemical studies [61]. With advances in sequencing technology and increasing transcriptome datasets becoming available, genetic information about dinoflagellate PKS, NRPS, and PKS/NRPS hybrids is rapidly growing [73,75,76,205,206].

Our goal for this study was to add to these existing data by focusing on multi-domain proteins containing keto-synthase domains.

We investigated a multi-modular PKS using genetic, biochemical, and mass spectrometry methods in the basal dinoflagellate Amphidinium carterae. This species is often used as a “model” dinoflagellate species due to its smaller genome, but still retaining the other qualities of HAB species [62]. Amphidinium carterae produces amphidinols, a sterolysin (Figure 5.2). These pore forming toxins produced through PKS pathways,

complex with specific cell membrane sterols resulting in cell lysis [80,185,207]. They also produce other cytotoxic amphidinolides and macrolide compounds of polyketide origin that have anti-fungal and anti-cancer properties [63].

In this paper we will focus on one multi-modular PKS, as described by Bachvaroff et al., which contained three KS domains (Personal correspondence with Dr. Tsvetan R.

Bachvaroff). We are focusing on this triple module PKS because transcriptome data have shown this domain arrangement is found in all dinoflagellates studied to date. The AT domain is missing in toxin producing species, but a cis AT, found within the transcript, is found in non-toxic species such as Akashiwo sanguinea, making it a strong candidate for involvement in toxin synthesis (Figure 5.3) (Personal correspondence with Dr. Tsvetan R.

Bachvaroff).

Figure 5.2: Sterolysin compounds synthesized by different dinoflagellate species.

Structurally similar sterolysin compounds synthesized by different dinoflagellate species including Symbiodinium, Karlodinium, Amphidinium, and Prorocentrum.

5.3 Results

Multi-Modular Polyketide Synthase in Amphidinium carterae.

The domain arrangement of the target multi-modular polyketide synthase included three consecutive KS domains, each followed by different modification domains. The estimated molecular weight of this protein is 478 kDa. The sequence begins with a spliced leader in K. veneficum and Amphidinium carterae and ends with a thioesterase domain

(TE). Sequences similar to this were found in all dinoflagellates surveyed via transcriptome analysis to date (Personal correspondence with Tsvetan R. Bachvaroff). However, in A. carterae the acyl transferase (AT) domain is not found in a cis configuration with the transcript as it is in Akashiwo sanguinea. The reactions performed by each ketosynthase

(KS) module are shown on the phosphopantotheinyl-binding site. After the first KS, the

DH and KR domains produce a double bond, while the second KS module produces a hydroxyl group, and the third KS module includes an ER domain producing a single bond.

Antibodies used in subsequent experiments were synthesized with complementarity to the regions depicted with asterisks (Figure 5.3).

* * *

Antibody Site * Phosphopantothenate KS Ketosynthase DH Dehydratase KR Ketoreductase AT Acyl Transferase ER Enoyl Reductase TE Thioesterase

Figure 5.3: A multi-modular polyketide synthase (PKS) in Amphidinium carterae.

The figure depicts the domain arrangement of a triple module PKS in A. carterae. Resultant products based on a purely processive synthesis are shown on the phosphopantotheinate site for each module. The AT domain is not present in A. carterae and site of presumed trans-AT action is depicted in pink. * Domains for which complementary polyclonal antibodies were synthesized. KS1, KS2, and KS3 are shown from left to right. Thin layer chromatography showing 14C acetate incorporation into lipids in A. carterae with cerulenin inhibition. Cerulenin, a KS specific fatty acid inhibitor, was added to A. carterae cultures to block

KS activity. After a 5 minute incubation 14C acetate was added to label newly synthesized lipids. The phosphorimage of a thin layer chromatography (TLC) plate of A. carterae lipids show that 14C acetate is not detectable in cerulenin treated samples. At 1 hour there is incorporation into uninhibited control samples but very little incorporation into samples treated with 100 M cerulenin (Figure 5.4A). After two hours there is extensive 14C acetate incorporation into all the lipid classes in the control samples while the samples treated with

100 M cerulenin show almost no 14C acetate incorporation (Figure 5.4B).

A Sterol esters

Triacylglycerols

Free Fatty acids 1,3-Diacylglycerols 1,2-Diacylglycerols Monoacylglycerols Phospholipids With cerulenin Without cerulenin Galactolipids 1 hr

B Sterol esters

Triacylglycerols

Free Fatty acids 1,3-Diacylglycerols 1,2-Diacylglycerols Monoacylglycerols Phospholipids Galactolipids With cerulenin Without cerulenin 2 hr

Figure 5.4: Thin layer chromatography (TLC) of extracted lipids.

Triplicate lipid samples showing 14C acetate incorporation patterns with and without 100 M cerulenin pre- incubation in Amphidinium carterae. A) Lipid profile of A. carterae 1 hour after 14C acetate addition with and without cerulenin incubation. B) Lipid profile of A. carterae 2 hours after 14C acetate addition with and without cerulenin incubation. Lipid classes corresponding with migration on the TLC plate are listed on the right of the image [208]. Quantitation of 14C acetate incorporation into lipids and amphidinols

The 14C acetate incorporation into lipids, amphidinol, and sulpho-amphidinol was reduced over time when cerulenin was added. Lipids were extracted and nCi of 14C acetate incorporation per mg lipid was calculated for A. carterae and A. sanguinea. At the 2 hour time point A. sanguinea showed 14C acetate incorporation of 1600 nCi/mg of lipid in the control samples while the cerulenin treated samples measured 450 nCi/mg of lipid. In A. carterae incorporation in control samples was 1200 nCi/mg lipid while cerulenin treated samples were 60nCi/mg of lipid. Control samples show increased 14C acetate incorporation

per mg of lipid over the two hour time period in both species. Cerulenin treated samples show a decrease in 14C acetate with a sharper drop of incorporation observed in A. carterae than in A. sanguinea (Figure 5.5A). Aqueous fractions from the lipid extraction were used to quantify toxin levels and label incorporation. In A. carterae nCi of 14C acetate per g amphidinol were found. At the 2 hour time point control samples result in 14C acetate incorporation of 0.35 nCi/g of amphidinol and 5.8 nCi/g sulpho-amphidinol. Cerulenin treated samples at the 2 hour time point resulted in incorporation values of 0.07 nCi/g amphidinol and 0.235 nCi/g of sulpho-amphidinol. Control samples showed an increase of 14C acetate incorporation over time with more incorporation into sulpho-amphidinol than amphidinol. Samples treated with 100 M cerulenin show a reduction of 14C acetate incorporation by the 1 and 2 hour time points with more incorporation observed into sulpho-amphidinol (Figure 5.5B). The percent of 14C label incorporation was calculated for total lipid and toxin fractions (Table 5.1). Less percent incorporation is observed in cerulenin sample treated samples when compared to control samples.

Figure 5.5: Quantitation of 14C acetate incorporation into lipids and amphidinol over a 2 hour period after cerulenin addition.

A) Quantitation of incorporation of 14C acetate into lipids in A. carterae and A. sanguinea at 0, 1, and 2 hour time points with and without cerulenin addition. B) Quantitation of incorporation of 14C acetate into amphidinol and sulpho-amphidinol at 0, 1, and 2 hour time points with and without cerulenin addition. Error bars show one standard deviation for triplicate samples.

Species Time Point Cerulenin (+/-) % in lipids % in sulpho- % in (hr) amphidinol amphidinol 0.110 0.020 0.0003 A. carterae 0 +

0.524 0.038 0.0005 A. carterae 1 +

0.058 0.002 0.0008 A. carterae 2 +

0.352 0.097 0.0004 A. carterae 0 -

0.702 0.014 0.0020 A. carterae 1 - 1.351 0.058 0.0035 A. carterae 2 -

0.002 A. sanguinea 0 + ------

0.198 A. sanguinea 1 + ------

0.325 A. sanguinea 2 + ------

0.230 A. sanguinea 0 ------

0.442 A. sanguinea 1 ------0.691 A. sanguinea 2 ------

Table 5.1: Percent incorporation of 14C acetate into total lipids, sulpho-amphidinol, and amphidinol. Table shows percent incorporation of Ci of 14C acetate incorporated into lipid and toxin fractions during the labeling experiment. Cerulenin addition is denoted by (+) and control samples are denoted with (-). Numbers are shown as percentages.

UV absorbance showing 14C acetate incorporation into amphidinol and sulpho- amphidinol

The UV absorbance chromatogram monitored at 280 nm overlaid with disintegrations per minute (dpm) in collected fractions shows 14C incorporation into sulpho-amphidinol and amphidinol for A. carterae. The sulpho-amphidinol eluted at 9 minutes and amphidinol eluted at 15 minutes respectively. The dpm from 14C acetate were counted for each fraction collected and overlaid with the absorbance spectra to show incorporation into toxin fractions. In the control samples there is 14C acetate incorporation into both forms of amphidinol with more incorporation observed in the sulpho-amphidinol form. The sample treated with 100 M cerulenin shows no incorporation of 14C acetate into the toxin but does show the UV absorbance peaks for existing sulpho-amphidinol and amphidinol at 9 and 15 minutes respectively (Figure 5.6A). Aqueous fraction samples for A. sanguinea were also run as a control. Akashiwo sanguinea does not make amphidinols so there are no peaks in the UV absorbance chromatograms at 280 nm or 14C acetate incorporation observed

(Figure 5.6B).

Figure 5.6: HPLC UV absorbance at 280 nm displaying amphidinol and sulpho-amphidinol (red trace) overlaid with 14C acetate incorporation dpm (black trace).

A) The red traces show the 280nm UV absorbance chromatograms for A. carterae extracts with amphidinol and sulpho-amphidinol peaks labeled. The black trace shows dpm of 14C acetate label incorporation into amphidinol and sulpho-amphidinol fractions without cerulenin addition. B) The red trace shows the UV absorbance chromatograms at 280 nm in A. sanguinea. The black trace shows the dpm of 14C acetate in this fraction.

Quantitation of total toxin and lipid levels in A. carterae and A. sanguinea

Toxin and lipid levels were quantified to show that total levels were unchanged during the 2 hour experiment. Relative sulpho-amphidinol and amphidinol levels were quantified using the area under the curve from LC analysis [209]. The analysis reveals that toxin levels remain relatively constant across the time course of the experiment. Total amphidinol levels remain around 2.5 g/mL of culture while sulpho-amphidinol levels remain between

5-6 g/mL of culture (Figure 5.7. Lipids were quantified using dry weight of the extracted lipids in A. carterae and A. sanguinea. Lipid levels in both species are similar and remain between 0.07-0.1 mg lipids across the 2 hour experiment (Figure 5.7B).

Figure 5.7: Total toxin and lipid levels remain steady in A. carterae and A. sanguinea over a 2 hour period.

A) Quantitation of amphidinol and sulpho-amphidinol in A. carterae during 14C acetate labeling study. B) Quantitation of lipids in A. carterae and A. sanguinea during 14C acetate labeling study. Multi-Modular PKS transcript abundance over diel in A. carterae

Transcript abundance of the multi-modular PKS shows the same pattern near the light:dark transition as has been observed in other dinoflagellate mRNA sequences studied in the past [170,191,192]. A decrease in transcripts is observed one hour before the lights are turned off with a sharp incline one hour after the lights are turned off. Transcript levels further from this time point remain relatively steady. The two ribosomal proteins, internal transcribed spacer (ITS) and ribosomal large subunit (LSU), were used as controls. The transcript levels of the multi-modular PKS are much lower in comparison to the ribosomal proteins with higher cT values between 20-30, while LSU and ITS have cT values between

3-9 which reveals that these are highly abundant transcripts (Figure 5.8).

Figure 5.8: Multi-Modular PKS transcript abundance across diel cycle.

Plot of cT values from a RT-PCR across a diel cycle 6 hours before and after lights are turned off at hr 0. Plot shows two primer pairs designed against multi-modular PKS sequence (purple and red) as well as 2 control proteins LSU and ITS (cyan and orange). PKS transcript levels lower than LSU and ITS as depicted by the higher cT values. All transcripts follow the same general decrease before and increase after lights are turned off.

Western blotting of three protein domains in the multi-modular PKS from A. carterae

Western blotting was performed using polyclonal antibodies designed against the KS1,

KR2, and TE domains of the multi-modular PKS protein (Table 5.3). None of the antibodies resulted in a band close to the calculated 478 kDa full size of the protein. The highest molecular weight bands observed in a cell pellet for anti-KS1, anti-KR2, and anti-

TE were 106 kDa, 295 kDa, and 290 kDa, respectively. Additional bands below the highest reported band were observed for all three antibodies (Figure 5.9A). The polyclonal

antibodies were blocked with the epitope peptide to show specificity of the antibody. All antibodies showed a reduction or complete absence of bands demonstrating specificity

(Figure 5.9B). Western blotting on a native protein lysate using the anti-TE antibody revealed the same band at 290 kDa as well as some additional bands at lower molecular weights, but not as many as were observed from the cell pellet (Figure 5.9C).

KS1 KR2 TE

A B A B A B C

300 kD 250 kD 180 kD

130 kD

100 kD

70 kD

50 kD 40 kD

Figure 5.9: Western blotting analysis of proteins in a multi-domain polyketide synthase (PKS) in Amphidinium carterae.

KS1: A) PKS proteins with the highest band at 106 kDa observed on a 300,000 cell equivalent pellet using anti-KS1 antibody. B) Peptide blocked with anti-KS1 to show specificity of the protein bands observed. KR2: A) PKS proteins with the highest band at 295 kDa observed on 250,000 cell equivalent pellet using anti-KR2 antibody. B) Peptide blocked anti-KR2 to show specificity of the protein bands observed. TE: A) PKS proteins with the highest band at 290 kDa observed on a 300,000 cell equivalent pellet using anti-TE antibody. B) Peptide blocked anti-TE to show specificity of the protein bands observed. C) PKS proteins with the highest band observed at 290 kDa in a 95,000 cell equivalent native protein lysate using anti-TE antibody. Mass spectra of adducts to the synthetic KS peptide

A synthetic peptide with the DTACSS sequence in the conserved active site of the three KS domains in the multi-modular PKS from A. carterae, was synthesized (Figure

5.12). The peptide was acetylated on the N-terminus and methylated at the C-terminus to keep the ends from reacting. The peptide was reacted with three compounds:

iodoacetamide, a widely used alkylating agent, cerulenin, and a shorter cerulenin analog with a truncated alkyl chain (Figure 5.14) then analyzed via tandem mass spectrometry.

The peptide formed adducts with iodoacetamide, cerulenin, and the short cerulenin analog respectively (Figure 5.10). Red boxes denote the inferred adduct structure with the peptide and an overlay showing the mass spectra (Figure 5.10).

100

Figure 5.10: Mass spectra of synthetic peptide, DTACSS, alone and coupled with cerulenin, a cerulenin analog, and iodoacetamide in vitro.

A) Structure of peptide reacted with cerulenin with corresponding ion observed at m/z 860.3463. B) Structure of peptide reacted with a cerulenin analog with a shorter side chain with corresponding ion observed at m/z 794.2979. C) Structure of peptide reacted with iodoacetamide with corresponding ion observed at m/z 694.2431. Ion detected for the peptide alone is observed at m/z 637.2200. Red boxes show structure of adduct.

101

Fragmentation of the synthetic KS peptide and cerulenin adducts

Tandem mass spectra show the peptide product ions from the cerulenin and short cerulenin reacted peptide. The spectra show multiple water losses as well as b- and y-ions that are covalently bound to cerulenin or the cerulenin analog and indicating the location of binding at the cysteine residue. Peptide product ions with a loss of cerulenin during collision induced dissociation (CID) are denoted be “-cerulenin” in Figure 5.11. The spectra also show ions corresponding to the full cerulenin-reacted peptide followed by ions corresponding to water losses in the 800-900 m/z range (Figure 5.11).

Figure 5.11: Tandem mass spectra (CID) of KS peptide reacted with cerulenin and the shorter cerulenin analog.

The black spectrum denotes product ions for cerulenin-peptide while the red spectrum denotes product ions for short cerulenin-peptide. Ion followed by – cerulenin denotes an ion that has lost its respective adduct during CID.

102

Figure 5.12: Sequence alignment of the three KS domains in the multi-modular PKS in A. carterae.

The red box denotes the sequence that was used to synthesize the synthetic peptide which is the conserved region that contains the active site cysteine.

103

Figure 5.13: Sequence alignment of the three KS domains in the multi-modular PKS in A. sanguinea.

The red box denotes the conserved residues around the active site cysteine in these proteins.

104

Figure 5.14: Structure of cerulenin analog that has a shorter side chain than cerulenin.

5.4 Discussion

Toxins made by polyketide synthases (PKS) in dinoflagellates often exacerbate harmful algal bloom (HAB) events, but the synthetic pathways for these toxins have still not been elucidated [4,61]. With an increasing number of genetic studies describing multi- domain PKS sequences obtained through transcriptome analysis we focused our studies on a triple module PKS. This sequence was pulled from other multi-domain PKS, PKS/NRPS, and NRPS sequences because of a crucial correlation with toxicity. The acyl-transferase

(AT) domain was found in non-toxin producing species but was missing in species that make toxins in a transcriptome survey of many dinoflagellate species (Figure 5.3) (Personal correspondence with Dr. Tsvetan R. Bachvaroff). This feature of AT domains not being included in type I PKSs is something that has been observed in other algae. Often the AT will act in trans to mediate the necessary acyl transferase reaction in a non-iterative with trans-AT type PKS [71].

Initially we looked at the overall effect of ketosynthase (KS) inhibition on metabolite synthesis, by using the KS inhibitor, cerulenin and 14C acetate label [210,211]. Cerulenin has been shown to inhibit all known fatty acid synthases [212]. In the dinoflagellate

Symbiodinium, cerulenin inhibition resulted in significant inhibition of free fatty acids as

105

well as phosphatidylethanolamines. These experiments were conducted on Amphidinium carterae, which makes the toxin amphidinol, and Akashiwo sanguinea, which does not make an amphidinol like compound but otherwise has the same sterol composition as A. carterae [23,213,214]. Labeling studies have shown that amphidinols are synthesized using glycolate as the starter molecule and acetate to provide the 2-C units for the rest of the molecule [80,215]. We measured the effects of cerulenin on lipids and toxins by quantifying 14C acetate incorporation into these compounds in the presence of cerulenin.

We began by studying effects on lipids which were expected to be inhibited with the cerulenin treatment. Phosphor imaging of a TLC plate reveals that by two hours there is almost complete cessation of 14C acetate incorporation into all classes of lipids including sterol esters, free fatty acids, triacylglycerols, diacylglycerols, and monoacylglycerols in

A. carterae. The control samples that did not receive a cerulenin treatment show an increase in 14C acetate incorporation over time (Figure 5.4). Quantitation of 14C into lipids shows the nCi of radiation detected from 14C acetate per mg of lipid decreases for A. carterae and

A. sanguinea lipid extracts over time, while control samples show an increase in incorporation by 1 and 2 hour marks (Figure 5.5A).

When we analyzed the toxins, there is a reduction of the nCi of 14C acetate per g of both sulpho-amphidinol and amphidinol over time while control samples continue to incorporate 14C acetate into both forms of amphidinol (Figure 5.5B). Interestingly, when counting the dpm in HPLC fractions, more 14C acetate was incorporated into the sulpho- amphidinol than amphidinol. This suggests to us that the toxin is synthesized in the sulpho form and then later processed to amphidinol (Figure 5.6). Quantitation of total lipid, sulpho-amphidinol, and amphidinol levels during the experiment remain steady suggesting

106

that cerulenin did not decrease overall levels during this two-hour experiment (Figure 5.7).

These experiments show that cerulenin is not only able to inhibit KS domains involved in lipid synthesis but also the KS domains involved in toxin synthesis.

The third module of the multi-domain PKS (Figure 5.3) has the same domain arrangement as a traditional fatty acid synthase (FAS), namely KS-DH-ER-KR-TE, with just the AT domain missing [216]. However, this arrangement is different from the arrangements observed in marine -proteobacteria PUFA synthases that also make some of the same PUFA compounds as dinoflagellates [217]. This leaves the possibility that the dinoflagellates are able to synthesize both of these products using the same machinery and the AT domains would be acting in trans configuration with the machinery in a noniterative fashion. The AT domain could be mediating product specificity in this scaffold.

Transcript abundance of the multi-modular PKS over a diel cycle reveals high cT values indicating the transcripts are not abundant which could suggest low protein abundance (Figure 5.8). Around the light:dark transition the decrease prior to lights off followed by a sharp increase one hour into the dark phase is an established pattern that has been observed in all mRNA transcripts studied to date [170,191,192]. The pattern is also observed when comparing between the PKS and the ribosomal control proteins ITS and

LSU with only significant differences observed in cT values.

To further characterize the multi-domain PKS in A. carterae, we synthesized antibodies to the KS1, KR2, and TE domains of this PKS sequence hypothesizing that we would see a protein at or near the calculated size of 478 kDa. None of the antibodies result in a band consistent with the full-sized protein, but anti-KR2 and anti-TE antibodies show similar size bands around 290 kDa, while anti-KS1 shows a much smaller band at 106 kDa

107

(Figure 5.9A). Incubation of antibodies with the peptides they were synthesized confirmed antibody specificity to the target epitopes (Figure 5.9B). Multiple bands are observed in the cell lysate samples with all 3 antibodies, as well as in a native protein lysate when blotted with the anti-TE antibody, which we believe is a controlled degradation process occurring in these cells (Figure 5.9C). The epitope sequences for all the antibodies were unique and not found in any other sequences in the A. carterae transcriptome (Data not shown).

In dinoflagellates, this is one of the first evidence demonstrating protein production of a multi-modular polyketide synthase. However, since we do not see evidence of the whole protein being produced there may be some processing or post-translational modifications occurring. We plan to continue to study this protein using shotgun proteomic methods to fully characterize the protein and any modifications that may be occurring as well as investigate its involvement in amphidinol synthesis.

To test whether cerulenin would be able to bind to the active site cysteine in the 3 KS domains in the multi-modular PKS, a peptide was synthesized with the sequence DTACSS which is the conserved KS active site in A. carterae (Figure 5.12) and A. sanguinea (Figure

5.13). When reacted with cerulenin or a cerulenin analog with a shorter side chain (Figure

5.14), adducts form through reaction with the sulfhydryl group on the cysteine residue (Fig

10A, B). Reaction with iodoacetamide, an alkylating agent, was done as a control and results in an adduct on the same sulfhydryl group (Figure 5.10C). Collision induced dissociation (CID) of the peptide shows the adduct is covalently bound to the cysteine and most product ions still have cerulenin or the cerulenin analog attached (Figure 5.11). Given the strong bond of the cerulenin with this KS active site peptide, we postulate that the three

108

KSs in the multi-domain PKS were all inhibited in our initial acetate labeling experiments, consequently inhibiting the 14C acetate incorporation into lipids in A. sanguinea and into lipids and amphidinols in A. carterae.

5.5 Materials and Methods

Cell culture conditions for A. carterae and A. sanguinea

Amphidinium carterae (CCMP 1314) was grown in ESAW medium modified to contain 10mM HEPES in polystyrene culture flasks from Corning (Corning NY) [176].

The cultures were maintained under constant light at 150m cm-2 s-1 with a 14:10 light:dark schedule. Akashiwo sanguinea was grown in the same conditions but ESAW medium was used at a salinity of 15 parts per thousand.

14C acetate labeling of A. carterae and A. sanguinea.

Cultures of A. carterae and A. sanguinea were cultured as described above. Cells were counted using a Coulter counter and had cell densities of 79,000 cells/mL for A. carterae and 1000 cells/mL for A. sanguinea. All experiments were done at a light intensity of 100

m cm-2 s-1 at 20 °C. For each experiment, 2 mL of culture were aliquoted to a clean test tube and experiments were conducted in triplicate. Treatment groups received cerulenin

(Sigma-Aldrich, Saint Louis, MO, USA) at a concentration of 100 M for A. carterae and

45M for A. sanguinea at 3 time points, 0 hour, 1 hour, and 2 hours. Control groups received 11L of DMSO at the same 3 time points, 0 hour, 1 hour, and 2 hours. Samples were incubated with cerulenin and DMSO for 5 minutes prior to acetic acid [1-14C] with a specific activity of 55.2 mCi/mmol (Lot# 184901) (Perkin Elmer, Waltham, MA) addition.

A total of 10 Ci of 14C acetic acid was added to each sample and was inverted three times to mix. Initial, T0 hour test tubes, were immediately placed in an ice bath to stop the

109

reaction. The T1hour and T2 hour test tubes were placed in ice baths after 1 and 2 hour incubations were complete respectively. After the cessation of all the reactions test tubes were spun down at 800 x g for 20 minutes. The supernatant was poured into a clean test tube and stored at 4° C. The pellet was stored at -20° C.

Lipid and toxin extraction from cell pellets of A. carterae and A. sanguinea

Lipid extraction followed the protocol as described by Adolf 2007 [218]. 1 mL of methylene chloride:methanol at a 2:1 ratio was added to the pelleted samples and vortexed for 10 seconds. 1 mL of methylene chloride at a ratio of 1:1 and 1 mL of methylene chloride at a ratio of 1:2 were added and the samples were vortexed for 10 seconds. Allow the phase separation to occur and separate the aqueous phase, which contains the toxins, and organic phases, which contains the lipids, were separated and placed in clean test tubes for further analysis.

Sample preparation for LC/MS analysis

The aqueous fraction from the lipid extraction as described above was used for toxin analysis. Samples were dried down in a Speed Vac to dryness. Samples were then resuspended in 1 mL of 30% methanol solution for injection into the LC/MS.

Thin layer chromatography (TLC) analysis of lipids in A. carterae and A. sanguinea

400 L of the extracted lipids were dried down in a speed vac at a low drying rate until completely dry. Samples were resuspended in 10 L of methylene chloride:methanol at a

1:1 ratio. Samples were dotted onto a silica HPTLC-GHL plate (Uniplate™, Newark, DE) in 3 L increments allowing samples to dry in between application. The plate was placed in a tank containing a migration solution of hexane: diethyl ether: formic acid at a ratio of

80:20:2 for 30 minutes. The TLC plate was dried in an oven at 70 C until completely dry.

110

The plate was exposed to a phosphor imager screen for 3 days and imaged on a Typhoon

Imager (GE Healthcare, Marlborough MA).

Cell culture and harvest for diel RNA quantification

Amphidinium carterae NCMA strain number 1314 was grown in ESAW medium

(Berges et al. 2001) modified to contain 10mM HEPES in a 20L multiport polycarbonate carboy. The cultures were maintained under constant light at 150um cm -2 s-1 with a 14:10 light:dark schedule and bubbling of air infused with CO2 controlled by an American Marine

(Ridgefield Ct) pH controller set to turn on above pH 8.2 and turn off at pH 7.6. Two liters of medium was added to an actively growing culture on Monday, Wednesday, and Friday until a cell density of 172,000 cells/mL on the morning of the experiment. 250ml of culture was aliquoted into 26 75 cm2 polystyrene culture flasks from Corning (Corning NY) and placed along the light bank to achieve equivalent light exposure as the stock culture in two rows of thirteen, with duplicates in an opposing direction. Each duplicate pair was taken for harvest according to the following schedule relative to the transition from light to dark in hours: -6.0, -4.0, -2.0, -1.5, -1.0, -0.5, 0, 0.5, 1.0, 1.5, 2.0, 4.0, 6.0. The samples were each split into 50mL and 200mL aliquots. Each aliquot was centrifuged at 1000 x g for 10 minutes at 20° C. The 200 mL aliquot was frozen at -80° C and the 50 ml aliquot was resuspended in 1mL of tri reagent (Sigma-Aldrich Saint Louis, MO).

RNA extraction and qPCR analysis

Duplicate samples from each timepoint were extracted using tri reagent according to the manufacturer's protocol using 1 mL of tri reagent. RNA was quantified on a Nanodrop

1000 (Thermo Fisher by Life Technologies Waltham, MA) and also on a Qubit 2.0 fluorometer (Life Technologies). RNA was reverse Transcribed using Superscript II

111

reverse transcriptase (Invitrogen by Life Technologies) with Random Primers (Invitrogen) according to the manufacturer's protocol. Generated cDNA was used as template for quantitative real-time PCR using an Applied Biosystems (Life Technologies) Fast 7500 thermal cycler in duplicate with the following reaction setup and primers listed in Table: 6

l diethyl pyrocarbonate (DEPC) treated water, 2 L of combined forward and reverse primers at 5 m each, 10 L of itaq 2X master mix containing sybr green and ROX (Bio-

Rad Hercules, CA), and 2 L of template cDNA at 10 ng/L. Thermal cycling conditions consisted of 40 an initial denaturation at 95° C for 2 minutes followed by 40 cycles of denaturation at 95° C for 15 seconds, annealing and fluorescent data collection at 60° C for

15 seconds, and extension at 72° C for 30 seconds. The reaction was completed with a melt curve to determine the presence of spurious PCR products. Cycle thresholds and baselines were determined manually, and relative quantities were determined across the diel time- course assuming a primer efficiency of 1.85 copies/cycle. This was done for normalization across the time-course and not to produce accurate quantities, thus primer efficiencies were not determined empirically for each primer pair.

Primer Name Forward Primer Reverse Primer A. carterae LSU GGCGATGAGGGATGAACCTA ACCACCGTCCTGCTGTCAGT

A. carterae ITS ATGGCGAATGAAAGGAGATG AGGGATGACAGATGCCAGAC A. carterae Triple KS GACTTCATTTGGTCGCAGGT TCAGCCAAGTCGTTTGTGAG begin A. carterae Triple KS ACAGGCCTTGTTGACAGCTT ACGTCGCACAGCTTTTTCTT end

Table 5.2: Primer sequences used in RT-PCR experiments.

112

Cell culture and harvest for SDS-PAGE and mass spectrometry

Amphidinium carterae was cultured as described above. Cultures were grown to a high cell density (300,000 cells/mL or higher) as measured with a hemocytometer. Cells were collected at mid-day time points. Aliquots (50 mL) of culture was poured into a 50 mL conical tube and centrifuged at 350 x g for 10 minutes at room temperature. The supernatant was poured off and the tube containing the pellet was placed in a bath containing ethanol and dry ice until frozen. The samples were then stored at -80° C until analysis.

Western blotting

Cell pellets of A. carterae were re-suspended in 1mL of 1X SDS sample buffer

(50mM Tris-Cl pH 6.8, 2% SDS, 0.1% bromophenol blue, 10% glycerol, 100mM dithiothreitol), incubated for 5 minutes at 95 °C, and centrifuged at 10,000 x g for 5 minutes. Samples were loaded onto gel at varying cell equivalents per lane and separated on Novex™ NuPAGE 3-8% tris-acetate gels at 150 V (constant) for 1 hour. Proteins were transferred onto a PVDF membrane using the high molecular weight program on the Trans Blot® Turbo™ Transfer System (Bio-Rad) for 14 minutes. The used transfer buffer was poured out and fresh cold transfer buffer was pipetted in without opening up the cassette. The high molecular weight program was run for an additional 7 minutes.

Membranes were washed briefly with TBS-T (50mM Tris base, 150mM NaCl, pH 7.4, and 0.05% (v/v) Tween 20). Membranes were incubated with primary antibody (Table

5.3) in TBS-T containing 5% non-fat dry milk overnight at 4 °C with gentle agitation.

The membrane was washed six times for 10 minutes each in TBS-T. Membranes were incubated with the appropriate HRP-conjugated secondary anti-body diluted in TBS-T

113

containing 5% non-fat milk for 1 hour at room temperature with gentle agitation. The membrane was washed six times for 10 minutes each in TBS-T. Blots were incubated for

5 minutes in Clarity Western ECL Substrate (Bio-Rad) according to manufacturer’s protocol and imaged on the ChemiDoc Touch Imaging System by Bio-Rad. Image processing and molecular weight analysis was conducted using Image Lab™ Software by

Bio-Rad.

Primary Peptide Sequence Dilution Secondary Antibody Dilution Antibody A. carterae Goat Anti-rabbit IgG-HRP CSFPGNAGGAQRY 1:2000 1:2500 KS1 Conjugate (Bio-Rad) A. carterae Goat Anti-rabbit IgG-HRP SRSGKVQPGFGLEGC 1:2000 1:2500 KR2 Conjugate (Bio-Rad) A. carterae Goat Anti-rabbit IgG-HRP KEVPVRQVPGGHFGC 1:2000 1:2500 TE Conjugate (Bio-Rad)

Table 5.3: Epitope sequence of antibodies synthesized by GenScript as well as concentrations of primary and secondary antibodies used in western blotting experiments. HPLC and LC–MS analysis of amphidinol

LC–MS was performed using an Agilent 1100 Series LC-MSD system, comprising a binary pump system, autosampler and diode array detector (DAD) with a micro high- pressure flow cell (6 mm pathlength, 1.7 μL volume), fraction collector and quadrupole mass spectrometer (G1956A SL) equipped with an electrospray ionization (ESI) interface.

Toxin samples in 30% methanol/water solutions were injected onto a C8 reversed phase

(LiChrosphere 125 mm × 4 mm, 5 μm bead size RP-8, Agilent) column and subjected to a

1 mL min-1 binary methanol/water gradient from 10 to 95% methanol over 25 min. The

DAD was collected from 190–950 nm wavelengths. Based on the UV–vis spectra, the absorption at 280 nm was used to detect amphidinol. The entire UV–vis spectra were saved for each detectable peak. The eluate from the DAD was split (1/3 to 1/6) using a graduated

114

micro-splitter valve (Upchurch Scientific). The major portion of the eluate was collected in multiple fractions while the remaining portion was subjected to MS analysis under the

-1 following spray chamber conditions: drying gas (N2) flow rate 10 L min , pressure 60 psi, temperature 350 °C, fragmentor voltage 350 V, capillary voltage 4000 V. A 1% (v/v) formic acid in water solution (0.1 mL min-1) was added post-column via a T-connector to provide lower pH conditions for enhanced positive mode ionization. A 5 mM ammonium acetate solution in water was used for negative mode ionization. Fractions of 20 s duration

(1/3 mL) during the first 32 min of the HPLC run were collected into scintillation vials to count disintegrations per minute (dpm) in fractions and quantify toxin concentrations

[209]. Total ion chromatograms for ions in the range from m/z 500 to 1500 were collected.

KS peptide and cerulenin adducts

A synthetic peptide was synthesized with the amino acid sequence, DTACSS with an acetylated N-terminus and an ester on the C-terminus, by Genscript® (Genscript®,

Piscataway, NJ). The peptide was dissolved in Optima® LC/MS water (Fisher Chemical,

Waltham, MA) to a concentration of 1 mg/mL. Reactions were done in 2 mL amber autosampler tubes from Agilent (Agilent, Santa Clara CA). All reaction components were added to the tube and allowed to incubate at room temperature for 2 hours and then placed at 4 °C until mass spectrometry analysis. Reactions are included in Table 5.4.

Peptide 50 mM Ammonium Adduct Compound Sample Name (µg) Bicarbonate pH 8 (µL) (µg) Peptide Alone 50 50 50 Peptide + IAA 50 50 50 Peptide + cerulenin 50 50 50 Peptide + short cerulenin 50 50 50

Table 5.4: Experimental set up of synthetic peptide experiments analyzed by mass spectrometry.

115

MS analysis of synthetic peptide with cerulenin

Peptides were dissolved in Optima® LC/MS water (Fisher Chemical, Waltham, MA) at a concentration of 1mg/mL. Peptide solutions were infused at a flow rate of 10 µL min-

1 through the electrospray ionization source of a LCMS IT-TOF mass spectrometer

(Shimadzu Scientific Instruments, Columbia, MD, USA). The CDL and heat block were set at 200 °C. Nebulization gas was set to 1.0 L min-1 and drying gas was set to “on”.

Accurate mass spectra were collected in positive ionization mode, from m/z 200-2000.

Tandem mass spectra were acquired using ultra-pure argon (Airgas, Alexandra, VA) as collision gas at an arbitrary collision energy setting of 40% and an arbitrary flow rate of

50%. Precursor ion accumulation time in the ion trap was set at 10 ms. Raw data files were processed in LCMSSolution software version 3.7 (Shimadzu Scientific Instruments,

Columbia, MD, USA) and, when necessary, converted to mzXML format using the

LCMSSolution convert tool. Peak picking and theoretical product ion m/z generation were performed in mMass v5.5 [219].

Acknowledgements

This research was funded in part by grants from OHH NIHR01ES021949-01/

NSFOCE1313888 and NOAA-NOS-NCCOS-2012-2002987 to ARP. Craig Townsend,

Ph.D. from Johns Hopkins University for providing the cerulenin analog.

116

CHAPTER 6: SUMMARY AND IMPLICATIONS OF THE RESEARCH 6.1 Main Findings

The main goal for this work was to move beyond sequencing data to characterize proteins in different dinoflagellate pathways at the protein level. High through-put RNA sequencing technologies have provided a plethora of previously unavailable transcriptome data for dinoflagellates [220]. However as was discussed in section 1.1, this only provides part of the information about these pathways. As transcriptional regulation seems to be lacking in dinoflagellates we focused on characterizing protein sequences focusing on function, location, and at the most basic level that they are translated [2].

A hypothesis posed by J. Woodland Hastings was that the pH shift resulting in bioluminescence is mediated by a proton selective channel in L. polyedrum. This work was integral to closing the gaps in the mechanism of light production in these bioluminescent organisms. In chapter 1 we investigated a protein in Lingulodinium polyedrum that had hallmarks of a proton selective channel (Hv1) thought to be responsible for bioluminescence. We began with a transcript that had sequence pattern similar to other proton channels [104]. Using this sequence, the gene was expressed in mammalian HEK cells to confirm that this was in fact a proton channel based on electrophysiological properties. These experiments confirmed that the channel was proton selective, showed time dependent depolarization, and pH dependent gating. To further characterize the protein, western blotting of purified scintillons with antibodies specific to the Hv1 sequence shows a single protein band with a molecular weight 30 kDa. These assignments were further verified using shotgun proteomic analysis of purified scintillon fractions resulting in many peptides mapping back to the soluble portion of the protein outside of the transmembrane region. Cellular localization of Hv1 to the scintillons was confirmed by 117

performing immunocytochemistry on whole Lingulodinium cells as well as purified scintillons. These results show that Hv1 in addition to luciferase and luciferase binding protein can all be localized to the scintillons in L. polyedrum.

In chapters 3 and 4 we shifted focus to acetyl CoA carboxylases, the rate limiting step in fatty acid and polyketide processes. These biotinylated proteins had not been well characterized in dinoflagellates, so we began with a phylogenetic analysis to look at the distribution of ACCs across thirteen dinoflagellate species, also including chlorophyte green algae, haptophytes, diatoms, and alveolates. Dinoflagellates possess a eukaryotic homomeric ACC, including a diagnostic central domain, with most species having more than one copy of this protein. We observed distinct clades based on the type of plastid with the peridinin, diatom, and haptophyte plastid-containing dinoflagellates clustering separately. We also observed one clade that contains all thirteen species, which we assign as cytosolic ACCs. Based on these observations, we assigned the ACCs to plastid and cytosol. To verify protein production of these ACCs we used streptavidin western blotting, as these proteins are natively biotinylated, to observe the presence of a high molecular weight band consistent with the calculated molecular weight of ACCs in Amphidinium carterae, Karlodinium veneficum, and Akashiwo sanguinea. We observe bands in the range of 250 kDa to 275 kDa for all 3 organisms. These bands have been putatively assigned as

ACCs and were confirmed by shotgun proteomics for A. carterae. This work showed that dinoflagellates contain more than one ACC protein with distribution in the cytosol and plastids.

In chapter 4 we continued characterizing ACCs this time focusing in on A. carterae.

The domain arrangement was biotin carboxylase, biotin, central domain, and carboxyl

118

transferase in 2 different ACCs, ACC 1 and 2, with a duplication in ACC 1. ChloroP transit peptide analysis revealed that ACC 1a and 1b are likely targeted to the plastid while ACC

2 is cytosolic [188]. Measure of mRNA transcripts across the diel cycle revealed a pattern consistent with most dinoflagellate transcripts analyzed to date. Next, protein expression using streptavidin western blotting across the diel cycle showed little change in protein expression with a small drop off at the +6 hr and -6 hr time points from the light:dark transition. Finally, shotgun proteomics was used to verify the western blotting assignments.

Peptides that mapped to ACC 1a, 1b, and 2 confirmed the presence of these proteins in A. carterae. This study resulted in a more in depth analysis of ACCs in A. carterae.

In chapter 5 we explored a multi-modular polyketide synthase (PKS) in A. carterae.

The focus was on a PKS transcript with 3 distinct modules found in all dinoflagellates surveyed to date and had a cis-AT in species that are not toxin producers, but the AT domain is missing in toxin producing species. We began by looking at the overall effects of a KS inhibitor, cerulenin, on fat and toxin production using 14C acetate labeling. We find that when KSs are inhibited with cerulenin, incorporation of 14C acetate into fats and toxins stops while overall toxin and fat levels remain steady. An interesting outcome of the labeling experiments was that more 14C acetate incorporation was observed in sulpho- amphidinol than amphidinol suggesting it is synthesized in the sulpho-form and then converted to amphidinol. To dissect the reaction between cerulenin and the KS domains we used a synthetic peptide with the sequence, DTACSS, the conserved active site in the multi-modular PKS. When reacted with cerulenin and a cerulenin analog the peptide formed a covalent bond to the cysteine residue as observed by tandem MS. We postulate the three KS domains in the multi-modular PKS were all inhibited in our initial radio-

119

labeling experiments but will require verification with proteomic experiments. To characterize multi-module PKS we performed western blotting against three different domains on this protein; KS1, KR2, and the TE domains. None of the antibodies resulted in a protein near the calculated full size of the transcript leading us to believe there is some type of cellular processing of the protein occurring. This was one of the first studies that characterized a multi-modular PKS in dinoflagellate beyond transcriptome data and is an important start to characterizing pathways by which toxins are synthesized.

6.2 Future work characterizing acetyl CoA carboxylases

To round out the characterization of the ACCs in A. carterae we want to look at ACC inhibition of acetate incorporation use the same experiment as in chapter 5, except instead of using cerulenin we would use 5-(Tetradecyloxy)-2-furoic acid (TOFA), an ACC specific inhibitor [221]. This would allow us to see 14C acetate incorporation in toxins and fats.

Since this protein is presumably involved in both pathways we would expect to see the synthesis of both product types being inhibited by TOFA in a dose dependent manner.

6.3 Future work characterizing PKSs and PKS/NRPSs in Amphidinium carterae

While we focused characterizing one multi-modular PKS protein in A. carterae we still need to address the role for single module KASs in toxin synthesis. The next step is to conduct shotgun proteomics on protein lysates of A. carterae with and without cerulenin addition [222]. In addition to finding peptide sequences consistent with transcript sequences, we would also look for cerulenin adducts to these proteins to verify that this multi-domain PKS was in fact inhibited in our radiolabeling experiments. With this work we hope to be able to get sequence data for the multi-modular PKS based on peptide

120

mapping. The protein is still enigmatic as western blotting never showed a band consistent with the calculated molecular weight for the full transcript.

PKSs often exist as dimers, so native protein gels and western blotting would confirm if these proteins existing as dimers as depicted in Figure 6.1 [223]. This will likely require anti-body pull down to isolate and concentrate these proteins as attempts to do these experiments on cell lysates in the past have failed.

The combination of the protein sequence in addition to protein arrangements will be a crucial start to fully characterizing this multi-modular PKS and then can be applied to other multi-modular PKS and PKS/NRPS hybrids in the future.

121

Figure 6.1: Multi-Modular PKS Model in Amphidinium carterae

Model depicts possible domain arrangement of the multi-modular PKS in A. carterae as a dimer. Putative location of where ATs are acting in trans are depicted in purple. 6.4 Future work characterizing interacting partners of PKSs in Amphidinium carterae Our current transcriptome data has revealed that AT domains missing in the multi- modular PKS described in chapter 5 in species that make toxins. We hypothesize that ATs

122

act in trans (Figure 6.1) allowing for the production of more than one type of product

[224,225]. To test this hypothesis, we need to troubleshoot antibody pulldown techniques as in the past they were not successful. After a successful pulldown we would analyze the protein by proteomics and see if we see any proteins interacting with this scaffold. This would be an important finding because of the implications in toxin production.

Finally, we would also need to determine which ATs may be interacting with the scaffold. Transcriptome sequencing has revealed that there are 11 ATs present in A. carterae (Personal correspondence with Dr. Tsvetan Bachvaroff). This suite of ATs is interesting because 1) there are 5 ATs with one of more ankyrin repeat which could possibly be used to anchor itself onto a protein scaffold [226]. 2) there is only one AT that has sequence homology to the FabD gene which is involved in fatty acid biosynthesis

[227]. These two observations along with the fact that the domain arrangement of the third module of the multi-modular PKS, KS-AT-DH-ER-KR-TE, is the same arrangement that is observed in FAS pathways [216], leads us to wonder whether this scaffold is used to produce both fatty acid products and toxins. Perhaps specific ATs act in trans, with the

FabD like AT coming in to make fatty acid products, while other ATs are involved in the synthesis of other products or intermediates. Characterizing these ATs would lead to some very interesting information as to how dinoflagellates synthesize these primary and secondary metabolites.

123

Figure 6.2: Distribution of ATs in Amphidinium carterae.

Phylogenetic tree shows the distribution of acyl-transferase domains in A. carterae. Variants with ankyrin repeats are labeled. The sequence that is most like the FabD involved in fatty acid synthesis is labeled. 6.5 Closing Remarks

The current work in dinoflagellate cellular biology needs to push past transcriptome sequencing data. These datasets provide the valuable information needed to characterize unknown proteins and their involvement in different pathways and can be used to generate testable hypotheses. The next step is to work with proteins themselves since transcripts do not necessarily mean the protein is synthesized or synthesized at the same size that is calculated from the transcript [192]. In agreement with Morse I believe that proteomic technologies will really help push this field forward [92,222]. High-throughput protein data is able to rapidly identify hundreds of proteins in one run and I believe it will be utilized as a main tool in characterizing proteins and protein pathways in dinoflagellates.

124

CHAPTER 7: REFERENCES

1. Place, A. R.; Bowers, H. A.; Bachvaroff, T. R.; Adolf, J. E.; Deeds, J. R.; Sheng, J. Karlodinium veneficum—The little dinoflagellate with a big bite. Harmful Algae 2012, 14, 179–195. 2. Wisecaver, J. H.; Hackett, J. D. Dinoflagellate Genome Evolution. Annu. Rev. Microbiol. 2011, 65, 369–387. 3. Falkowski, P. G.; Raven, J. a Aquatic Photosynthesis. Freshw. Biol. 1997. 4. Okaichi, T.; Anderson, D. M.; Nemoto, T. Red Tides: Biology, Environmental Science, and toxicology; Okaichi, T.; Anderson, D. M.; Nemoto, T., Eds.; Elsevier, 1989. 5. Hackett, J. D.; Anderson, D. M.; Erdner, D. L.; Bhattacharya, D. Dinoflagellates: A remarkable evolutionary experiment. Am. J. Bot. 2004, 91, 1523–1534. 6. Metz, J. G. Production of Polyunsaturated Fatty Acids by Polyketide Synthases in Both Prokaryotes and Eukaryotes. Science (80-. ). 2001, 293, 290–293. 7. Taylor, F. J. R.; Hoppenrath, M.; Saldarriaga, J. F. Dinoflagellate diversity and distribution. Biodivers. Conserv. 2008. 8. Saldarriaga, J. F.; Taylor, F. J. R.; Keeling, P. J.; Cavalier-Smith, T. Dinoflagellate Nuclear SSU rRNA Phylogeny Suggests Multiple Plastid Losses and Replacements. 9. Wilcox, L. W.; Wedemayer, G. J. Dinoflagellate with Blue-Green Chloroplasts Derived from an Endosymbiotic Eukaryote. Source Sci. New Ser. 1985, 227, 192–194. 10. Bjørnland, T.; Tangen, K. Pigmentation and morphology of a marine Gyrodinium (Dinophyceae) with a major carotenoid different from peridinin and fucoxanthin. J. Phycol. 1979, 15, 457–463. 11. Dodge, J. D. A dinoflagellate with both a mesocaryotic and a eucaryotic nucleus I. Fine structure of the nuclei. Protoplasma 1971, 73, 145–157. 12. Ian A. N. Lucas, M. V. The Fine Structure of Two Photosynthetic Species Of Dinophysis (Dinophysiales, Dinophyceae). J. Phycol. 1990, 26, 345–357. 13. Moestrup, O.; Daugbjerg, N. On Dinoflagellate Phylogeny and Classification. In Unravelling the Algae, the Past, Present, and Future of Algal Systematics; Brodie, J.; Lewis, J., Eds.; CRC Press: London, 2007; pp. 215–230. 14. Bachvaroff, T. R.; Concepcion, G. T.; Rogers, C. R.; Herman, E. M.; Delwiche, C. F. Dinoflagellate expressed sequence tags data indicate massive trasfer of chloroplast genes to the nuclear genome. Protist 2004, 155, 65–78. 15. Koumandou, V. L.; Nisbet, R. E. R.; Barbrook, A. C.; Howe, C. J. Dinoflagellate chloroplasts - Where have all the genes gone? Trends Genet. 2004, 20, 261–267. 16. Gajadhar, A. A.; Marquardt, W. C.; Hall, R.; Gunderson, J.; Ariztia-Carmona, E. V.; Sogin, M. L. Ribosomal RNA sequences of Sarcocystis muris, Theileria annulata and Crypthecodinium cohnii reveal evolutionary relationships among apicomplexans, dinoflagellates, and ciliates. Mol. Biochem. Parasitol. 1991, 45, 147–154. 17. Bork, P.; Bowler, C.; De Vargas, C.; Gorsky, G.; Karsenti, E.; Wincker, P. Tara Oceans studies plankton at Planetary scale. Science (80-. ). 2015, 348, 873. 18. Fast, N. M.; Xue, L.; Bingham, S.; Keeling, P. J. Re-examing alveolate evolution using multiple protein molecular phylogenies. J. Eukaryot. Microbiol. 2002, 49, 30–37. 19. Hackett, J. D.; Yoon, H. S.; Butterfield, N. J.; Sanderson, M. J.; Bhattacharya, D. Plastid Endosymbiosis Sources and Timing of the Major Events. In Evolution of Primary Producers in the Sea; Academic Press, 2007; pp. 109–132.

125

20. Bhattacharya, D.; Yoon, H. S.; Hedges, S.; Hackett, J. D. In The Timetree of Life; Hedges, S.; Kumar, S., Eds.; Oxford Univ. Press: Oxford, UK, 2009. 21. Freeman, M. A.; Fuss, J.; Kristmundsson, Á.; Bjorbækmo, M. F. M.; Mangot, J.-F.; Del Campo, J.; Keeling, P. J.; Shalchian-Tabrizi, K.; Bass, D. X-Cells Are Globally Distributed, Genetically Divergent Fish Parasites Related to Perkinsids and Dinoflagellates. Curr. Biol. 2017, 27, 1645–1651.e3. 22. Moore, R. B.; Obornik, M.; Janouskovec, J.; Chrudimsky, T.; Vancova, M.; Green, D. H.; Wright, S. W.; Davies, N. W.; Bolch, C. J.; Heimann, K.; Slapeta, J.; Hoegh- Guldberg, O.; Logsdon, J. M.; Carter, D. A. A photosynthetic alveolate closely related to apicomplexan parasites. Nature 2008, 451, 959–963. 23. Spector, D. L. Dinoflagellates; Spector, D. L., Ed.; Cell Biology; Academic Press, 1984. 24. Gautier, A.; Michel-Salamin, L.; Tosi-Couture, E.; McDowall, A. W.; Dubochet, J. Electron microscopy of the chromosomes of dinoflagellates in situ- confirmation of Bouligand’s liquid crystal hypothesis.pdf. J. Ultrastruct. Mol. Struct. Res. 1987, 10–30. 25. Rizzo, P. J.; Noodén, L. D. Chromosomal Proteins in the Dinoflagellate Alga Gyrodinium cohnii. Source Sci. New Ser. 1972, 176, 796–797. 26. Gornik, S. G.; Ford, K. L.; Mulhern, T. D.; Bacic, A.; Mcfadden, G. I.; Waller, R. F. Article Loss of Nucleosomal DNA Condensation Coincides with Appearance of a Novel Nuclear Protein in Dinoflagellates. Curr. Biol. 2012, 22, 2303–2312. 27. Chan, Y. H.; Wong, J. T. Y. Concentration-dependent organization of DNA by the dinoflagellate histone-like protein HCc3. Nucleic Acids Res. 2007, 35, 2573–2583. 28. Sala-Rovira, M.; Geraud, M. L.; Caput, D.; Jacques, F.; Soyer-Gobillard, M. O.; Vernet, G.; Herzog, M. Molecular cloning and immunolocalization of two variants of the major basic nuclear protein (HCc) from the histone-less eukaryote Crypthecodinium cohnii (Pyrrhophyta). Chromosoma 1991, 100, 510–518. 29. Sigee, D. C. Structural DNA and genetically active DNA in dinoflagellate chromosomes. BioSystems 1983, 16, 203–210. 30. LaJeunesse, T. C.; Lambert, G.; Andersen, R. A.; Coffroth, M. A.; Galbraith, D. W. Symbiodinium (Pyrrhophyta) genome sizes (DNA content) are smallest among dinoflagellates. J. Phycol. 2005, 41, 880–886. 31. Lin, S.; Cheng, S.; Song, B.; Zhong, X.; Lin, X.; Li, W.; Li, L.; Zhang, Y.; Zhang, H.; Ji, Z.; Cai, M.; Zhuang, Y.; Shi, X.; Lin, L.; Wang, L.; Wang, Z.; Liu, X.; Yu, S.; Zeng, P.; Hao, H.; Zou, Q.; Chen, C.; Li, Y.; Wang, Y.; Xu, C.; Meng, S.; Xu, X.; Wang, J.; Yang, H.; Campbell, D. A.; Sturm, N. R.; Dagenais-Bellefeuille, S.; Morse, D. The Symbiodinium kawagutii genome illuminates dinoflagellate gene expression and coral symbiosis. Science 2015, 350, 691–4. 32. Gornik, S. G.; Cassin, A. M.; MacRae, J. I.; Ramaprasad, A.; Rchiad, Z.; McConville, M. J.; Bacic, A.; McFadden, G. I.; Pain, A.; Waller, R. F.; by John Archibald, E. M. Endosymbiosis undone by stepwise elimination of the plastid in a parasitic dinoflagellate. Proc. Natl. Acad. Sci. USA 2015, 112, 5767–5772. 33. Aranda, M.; Li, Y.; Liew, Y. J.; Baumgarten, S.; Simakov, O.; Wilson, M. C.; Piel, J.; Ashoor, H.; Bougouffa, S.; Bajic, V. B.; Ryu, T.; Ravasi, T.; Bayer, T.; Micklem, G.; Kim, H.; Bhak, J.; LaJeunesse, T. C.; Voolstra, C. R. Genomes of coral dinoflagellate symbionts highlight evolutionary adaptations conducive to a symbiotic lifestyle. Sci. Rep. 2016.

126

34. Rae, P. M. M. 5-Hydroxymethyluracil in the DNA of a Dinoflagellate (density gradient centrifugation/Gyrodinium cohnii/physical effects of base composition/satellite DNA). 1973, 70, 1141–1145. 35. Rae, P. M. M. Hydroxymethyluracil in Eukaryote DNA: A Natural Feature of the Pyrrophyta. Source Sci. New Ser. 1976, 194, 1062–1064. 36. Hoet, P. P.; Coene, M. M.; Cocito, C. G. Replication Cycle of Bacillus subtilis hydroxymethyluracil-Containing Phages. Annu. Rev. Microbiol. 1992, 95–116. 37. Lidie, K. B.; Van Dolah, F. M. Spliced Leader RNA-Mediated trans-Splicing in a Dinoflagellate, Karenia brevis. J. Eukaryot. Microbiol. 2007, 54, 427–435. 38. Zhang, H.; Hou, Y.; Miranda, L.; Campbell, D. A.; Sturm, N. R.; Gaasterland, T.; Lin, S. Spliced Leader RNA Trans-Splicing in Dinoflagellates. Source Proc. Natl. Acad. Sci. United States Am. 2007, 104, 4618–4623. 39. Bachvaroff, T. R.; Place, A. R. From stop to start: Tandem gene arrangement, copy number and Trans-splicing sites in the dinoflagellate Amphidinium carterae. PLoS One 2008, 3. 40. Liu, L.; Hastings, J. W. Novel and rapidly diverging intergenic sequences between tandem repeats of the luciferase genes in seven dinoflagellate species. J. Phycol. 2006, 42, 96–103. 41. Janouškovec, J.; Gavelis, G. S.; Burki, F.; Dinh, D.; Bachvaroff, T. R.; Gornik, S. G.; Bright, K. J.; Imanian, B.; Strom, S. L.; Delwiche, C. F.; Waller, R. F.; Fensome, R. A.; Leander, B. S.; Rohwer, F. L.; Saldarriaga, J. F. Major transitions in dinoflagellate evolution unveiled by phylotranscriptomics. Proc. Natl. Acad. Sci. 2016, 201614842. 42. Cheung, M. Y.; Liang, S.; Lee, J. Toxin-producing cyanobacteria in freshwater: A review of the problems, impact on drinking water safety, and efforts for protecting public health. J. Microbiol. 2013, 51, 1–10. 43. Anderson, D. M.; Cembella, A. D.; Hallegraeff, G. M. Progress in Understanding Harmful Algal Blooms: Paradigm Shifts and New Technologies for Research, Monitoring, and Management. Annu. Rev. Mar. Sci 2012, 4, 143–76. 44. Camacho, F. G.; Rodríguez, J. G.; Mirón, a. S.; García, M. C. C.; Belarbi, E. H.; Chisti, Y.; Grima, E. M. Biotechnological significance of toxic marine dinoflagellates. Biotechnol. Adv. 2007, 25, 176–194. 45. Friedman, M.; Fernandez, M.; Backer, L.; Dickey, R.; Bernstein, J.; Schrank, K.; Kibler, S.; Stephan, W.; Gribble, M.; Bienfang, P.; Bowen, R.; Degrasse, S.; Flores Quintana, H.; Loeffler, C.; Weisman, R.; Blythe, D.; Berdalet, E.; Ayyar, R.; Clarkson- Townsend, D.; Swajian, K.; Benner, R.; Brewer, T.; Fleming, L. An Updated Review of Ciguatera Fish Poisoning: Clinical, Epidemiological, Environmental, and Public Health Management. Mar. Drugs 2017, 15, 72. 46. Arias, H. R. Marine Toxins Targeting Ion Channels. Mar. Drugs 2006, 4, 37. 47. Nicholson, G. M.; Lewis, R. J. Ciguatoxins: Cyclic Polyether Modulators of Voltage- gated Iion Channel Function. Mar. Drugs 2006, 4, 82. 48. Friedman, M. A. Ciguatera Fish Poisoning: Treatment, Prevention and Management. Mar. Drugs 2008, 6, 456–479. 49. Hokama, Y. Ciguatera Fish Poisoning : a Foodborne Disease. J. Toxicol. 2001, 20, 85–139. 50. Lehane, L.; Lewis, R. J. Ciguatera: Recent advances but the risk remains. Int. J. Food Microbiol. 2000, 61, 91–125.

127

51. Sengco, M. R. Prevention and control of Karenia brevis blooms. Harmful Algae 2009, 8, 623–628. 52. Watkins, S. M.; Reich, A.; Fleming, L. E.; Hammond, R. Neurotoxic shellfish poisoning. Mar. Drugs 2008, 6, 431–55. 53. Dechraoui, M. Y.; Naar, J.; Pauillac, S.; Legrand, A. M. Ciguatoxins and brevetoxins, neurotoxic polyether compounds active on sodium channels. Toxicon 1999. 54. Deeds, J.; Place, A. Sterol-specific membrane interactions with the toxins from Karlodinium micrum (Dinophyceae) — a strategy for self-protection? African J. Mar. Sci. 2006, 28, 421–425. 55. Deeds, J. R.; Hoesch, R. E.; Place, A. R.; Kao, J. P. Y. The cytotoxic mechanism of karlotoxin 2 (KmTx 2) from Karlodinium veneficum (Dinophyceae). Aquat. Toxicol. 2015, 159, 148–155. 56. Peng, J.; Place, A. R.; Yoshida, W.; Anklin, C.; Hamann, M. T. Structure and Absolute Configuration of Karlotoxin-2, an Ichthyotoxin from the Marine Dinoflagellate Karlodinium veneficum. 57. Bachvaroff, T. R.; Adolf, J. E.; Place, A. R. Strain variation in karlodinium veneficum (dinophyceae): Toxin profiles, pigments, and growth characteristics. J. Phycol. 2009, 45, 137–153. 58. Deeds, J. R.; Reimschuessel, R.; Place, A. R. Histopathological effects in fish exposed to the toxins from Karlodinium micrum. J. Aquat. Anim. Health 2006, 18, 136– 148. 59. Place, A. R.; Munday, R.; Munday, J. S. Acute toxicity of karlotoxins to mice. Toxicon 2014, 90, 184–190. 60. Kalaitzis, J. a.; Chau, R.; Kohli, G. S.; Murray, S. a.; Neilan, B. a. Biosynthesis of toxic naturally-occurring seafood contaminants. Toxicon 2010, 56, 244–258. 61. Van Wagoner, R. M.; Satake, M.; Wright, J. L. C. Polyketide biosynthesis in dinoflagellates: what makes it different? Nat. Prod. Rep. 2014, 31, 1101–37. 62. Murray, S.; Jørgensen, M. F.; Daugbjerg, N.; Rhodes, L. Amphidinium revisited. II. Resolving species boundaries in the Amphidinium operculatum species complex (Dinophyceae), including the descriptions of Amphidinium trulla sp nov. and Amphidinium gibbosum. comb. nov. J. Phycol. 2004, 40, 366–382. 63. Kobayashi, J.; Kubota, T. Bioactive macrolides and polyketides from marine dinoflagellates of the genus Amphidinium. J. Nat. Prod. 2007, 70, 451–460. 64. Deeds, J. R.; Hoesch, R. E.; Place, A. R.; Kao, J. P. Y. The cytotoxic mechanism of karlotoxin 2 (KmTx 2) from Karlodinium veneficum (Dinophyceae). Aquat. Toxicol. 2015, 159, 148–55. 65. Tong, L. Acetyl-coenzyme A carboxylase: Crucial metabolic enzyme and attractive target for drug discovery. Cell. Mol. Life Sci. 2005, 62, 1784–1803. 66. Tong, L. Structure and function of biotin-dependent carboxylases. Cell. Mol. Life Sci. 2013, 70, 863–891. 67. Huerlimann, R.; Heimann, K. Comprehensive guide to acetyl-carboxylases in algae. Crit. Rev. Biotechnol. 2013, 33, 49–65. 68. Cantu, D. C.; Chen, Y.; Lemons, M. L.; Reilly, P. J. ThYme: A database for thioester- active enzymes. Nucleic Acids Res. 2011. 69. Dutta, S.; Whicher, J. R.; Hansen, D. a; Hale, W. a; Chemler, J. a; Congdon, G. R.; Narayan, A. R. H.; Håkansson, K.; Sherman, D. H.; Smith, J. L.; Skiniotis, G. Structure

128

of a modular polyketide synthase. Nature 2014, 510, 512–7. 70. Keatinge-Clay, A. T. The structures of type I polyketide synthases. Nat. Prod. Rep. 2012, 29, 1050. 71. Shelest, E.; Heimerl, N.; Fichtner, M.; Sasso, S. Multimodular type I polyketide synthases in algae evolve by module duplications and displacement of AT domains in trans. BMC Genomics 2015, 16, 1–15. 72. Meier, J. L.; Burkart, M. D. The chemical biology of modular biosynthetic enzymes. Chem. Soc. Rev. 2009, 38, 2012–2045. 73. Monroe, E. A.; Johnson, J. G.; Wang, Z.; Pierce, R. K.; Van Dolah, F. M. Characterization and expression of nuclear-encoded polyketide synthases in the brevetoxin-producing dinoflagellate karenia brevis. J. Phycol. 2010, 46, 541–552. 74. Yu, D.; Xu, F.; Zeng, J.; Zhan, J. Type III polyketide synthases in natural product biosynthesis. IUBMB Life 2012, 64, 285–295. 75. Van Dolah, F. M.; Zippay, M. L.; Pezzolesi, L.; Rein, K. S.; Johnson, J. G.; Morey, J. S.; Wang, Z.; Pistocchi, R. Subcellular localization of dinoflagellate polyketide synthases and fatty acid synthase activity. J. Phycol. 2013, 49, 1118–1127. 76. Kohli, G. S.; John, U.; Figueroa, R. I.; Rhodes, L. L.; Harwood, D. T.; Groth, M.; Bolch, C. J. S.; Murray, S. a. Polyketide synthesis genes associated with toxin production in two species of Gambierdiscus (Dinophyceae). BMC Genomics 2015, 16, 1–10. 77. Van Dolah, F. M.; Kohli, G. S.; Morey, J. S.; Murray, S. A. Both modular and single- domain Type I polyketide synthases are expressed in the brevetoxin-producing dinoflagellate, Karenia brevis (Dinophyceae). J. Phycol. 2017, 53, 1325–1339. 78. Chen, H.; Du, L. Iterative polyketide biosynthesis by modular polyketide synthases in bacteria. Appl. Microbiol. Biotechnol. 2016, 100, 541–57. 79. Lee, M. S.; Repeta, D. J.; Nakanishi, K.; Zagorski, M. G. Biosynthetic Origins and Assignments of 13C NMR Peaks of Brevetoxin B. J. Am. Chem. Soc. 1986, 108, 7855– 7856. 80. Houdai, T.; Matsuoka, S.; Murata, M.; Satake, M.; Ota, S. Acetate labeling patterns of dinoflagellate polyketides, amphidinols 2, 3 and 4. Tetrahedron 2001, 57, 5551–5555. 81. Yamazaki, M.; Tachibana, K.; Satake, M. Complete 13C-labeling pattern of yessotoxin a marine ladder-frame polyether. Tetrahedron 2011, 67, 877–880. 82. Riva, E.; Wilkening, I.; Gazzola, S.; Li, W. M. A.; Smith, L.; Leadlay, P. F.; Tosin, M. Chemical Probes for the Functionalization of Polyketide Intermediates. Angew. Chemie Int. Ed. 2014, 53, 11944–11949. 83. Valiadi, M.; Iglesias-Rodriguez, D. Understanding Bioluminescence in Dinoflagellates—How Far Have We Come? Microorganisms 2013, 1, 3–25. 84. DeSa, R.; Hastings, J. W. The characterization of scintillons. Bioluminescent particles from the marine dinoflagellate, Gonyaulax polyedra. J. Gen. Physiol. 1968, 51, 105–122. 85. Sarmiento, F. O.; Frolich, L. M. Shear-Stress Dependence of Dinoflagellate Bioluminescence. 2007, 22, 278–287. 86. Chen, A. K.; Latz, M. I.; Sobolewski, P.; Frangos, J. a Evidence for the role of G- proteins in flow stimulation of dinoflagellate bioluminescence. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2007, 292, R2020-7. 87. Hastings, J. W. The Gonyaulax clock at 50: Translational control of circadian expression. Cold Spring Harb. Symp. Quant. Biol. 2007, 72, 141–144. 88. Dunlap, J. C.; Hastings, J. W. The biological clock in Gonyaulax controls luciferase

129

activity by regulating turnover. J. Biol. Chem. 1981, 256, 10509–10518. 89. Morse, D.; Milos, P. M.; Roux, E.; Hastings, J. W. Circadian regulation of bioluminescence in Gonyaulax involves translational control. Proc. Natl. Acad. Sci. U. S. A. 1989, 86, 172–6. 90. Hastings, J. Circadian Rhythms in Dinoflagellates: What Is the Purpose of Synthesis and Destruction of Proteins? Microorganisms 2013, 1, 26–32. 91. Knaust, R.; Urbig, T.; Li, L.; Taylor, W.; Hastings, J. W. The circadian rhythm of bioluminescence in Pyrocystis is not due to differences in the amount of luciferase: A comparative study of the three bioluminescent marine dinoflagellates. J. Phycol. 1998, 34, 167–172. 92. Morse, D.; Tse, S. P. K.; Lo, S. C. L. Exploring dinoflagellate biology with high- throughput proteomics. Harmful Algae 2018, 75, 16–26. 93. Fogel, M.; Hastings, J. W. Bioluminescence: mechanism and mode of control of scintillon activity. Proc. Natl. Acad. Sci. U. S. A. 1972, 69, 690–3. 94. Thomas, R. C.; Meech, R. W. Hydrogen ion currents and intracellular pH in depolarized voltage-clamped snail neurones. Nature 1982, 299, 826–828. 95. Barish, M. E.; Baud, C. A voltage-gated hydrogen ion current in the oocyte membrane of the axolotl, Ambystoma. J. Physiol. 1984, 352, 243–263. 96. DeCoursey, T. E. Hydrogen ion currents in rat alveolar epithelial cells. Biophys. J. 1991, 60, 1243–1253. 97. Kapus, A.; Romanek, R.; Qu, A. Y.; Rotstein, O. D.; Grinstein, S. A pH-sensitive and voltage-dependent proton conductance in the plasma membrane of macrophages. J. Gen. Physiol. 1993, 102, 729–760. 98. DeCoursey, T. E.; Cherny, V. V Potential, pH, and arachidonate gate hydrogen ion currents in human neutrophils. Biophys. J. 1993, 65, 1590–1598. 99. Krause, R. M.; Bernheim, L.; Bader, C.-R. Human skeletal muscle has a voltage- gated proton current. Neuromuscul. Disord. 1993, 3, 407–411. 100. Ramsey, I. S.; Moran, M. M.; Chong, J. a; Clapham, D. E. A voltage-gated proton- selective channel lacking the pore domain. Nature 2006, 440, 1213–6. 101. Chaves, G.; Derst, C.; Franzen, A.; Mashimo, Y.; Machida, R.; Musset, B. Identification of an HV 1 voltage-gated proton channel in insects. FEBS J. 2016, 283, 1453–64. 102. Sasaki, M.; Takagi, M.; Okamura, Y. A Voltage Sensor-Domain Protein Is a Voltage-Gated Proton Channel. Science (80-. ). 2006, 312, 589–592. 103. Taylor, A. R.; Chrachri, A.; Wheeler, G.; Goddard, H.; Brownlee, C. A voltage- gated H+ channel underlying pH homeostasis in calcifying coccolithophores. PLoS Biol. 2011, 9, e1001085. 104. Smith, S. M. E.; Morgan, D.; Musset, B.; Cherny, V. V; Place, A. R.; Hastings, J. W.; DeCoursey, T. E. Voltage-gated proton channel in a dinoflagellate. Proc. Natl. Acad. Sci. USA 2011, 108, 18162–18167. 105. DeCoursey, T. E. Voltage-gated proton channels and other proton transfer pathways. Physiol. Rev. 2003, 83, 475–579. 106. Musset, B.; Smith, S. M. E.; Rajan, S.; Morgan, D.; Cherny, V. V; Decoursey, T. E. Aspartate 112 is the selectivity filter of the human voltage-gated proton channel. Nature 2011, 480, 273–7. 107. Kulleperuma, K.; Smith, S. M. E.; Morgan, D.; Musset, B.; Holyoake, J.;

130

Chakrabarti, N.; Cherny, V. V; DeCoursey, T. E.; Pomès, R. Construction and validation of a homology model of the human voltage-gated proton channel hHV1. J. Gen. Physiol. 2013, 141, 445–65. 108. Dudev, T.; Musset, B.; Morgan, D.; Cherny, V. V; Smith, S. M. E.; Mazmanian, K.; DeCoursey, T. E.; Lim, C. Selectivity Mechanism of the Voltage-gated Proton Channel, HV1. Sci. Rep. 2015, 5. 109. Cherny, V. V; Murphy, R.; Sokolov, V.; Levis, R. A.; DeCoursey, T. E. Properties of Single Voltage-gated Proton Channels in Human Eosinophils Estimated by Noise Analysis and by Direct Measurement. J. Gen. Physiol. 2003, 121, 615–628. 110. Byerly, L.; Meech, R.; Moody Jr., W. Rapidly activating hydrogen ion currents in perfused neurones of the snail, Lymnaea stagnalis. J. Physiol. 1984, 351, 199–216. 111. Cherny, V. V; Markin, V. S.; DeCoursey, T. E. The voltage-activated hydrogen ion conductance in rat alveolar epithelial cells is determined by the pH gradient. J. Gen. Physiol. 1995, 105, 861–896. 112. DeCoursey, T. E. Voltage-Gated Proton Channels: Molecular Biology, Physiology, and Pathophysiology of the HV Family. Physiol. Rev. 2013, 93, 599–652. 113. Eckert, R.; Sibaoka, T. The flash-triggering action potential of the luminescent dinoflagellate Noctiluca. J. Gen. Physiol. 1968, 52, 258–82. 114. Latz, M. I.; Case, J. F.; Gran, R. L. Excitation of bioluminescence by laminar fluid shear associated with simple Couette flow. Limnol. Oceanogr. 1994, 39, 1424–1439. 115. Latz, M. I.; Rohr, J. Luminescent response of the red tide dinoflagellate Lingulodinium polyedrum to laminar and turbulent flow. Limnol. Oceanogr. 1999, 44, 1423–1435. 116. Hastings, J. W.; Dunlap, J. C. Cell free componenets in dinoflagellate bioluminescence. The particulate activity:Scintillons; the soluble components: Luciferase, luciferin, luciferin-binding protein. Methods Enzymol. 1986, 133, 307–327. 117. Johnson, C. H.; Inoue, S.; Flint, A.; Hastings, J. W. Compartmentalization of algal bioluminescence: autofluorescence of bioluminescent particles in the dinoflagellate Gonyaulax as studied with image-intensified video microscopy and flow cytometry. J Cell Biol 1985, 100, 1435–1446. 118. Schmitter, R. E.; Njus, D.; Sulzman, F. M.; Gooch, V. D.; Hastings, J. W. Dinoflagellate bioluminescence: a comparative study of invitro components. J. Cell. Physiol. 1976, 87, 123–34. 119. Nicolas, M. T.; Sweeney, B. M.; Hastings, J. W. The ultrastructural localization of luciferase in three bioluminescent dinoflagellates, two species of Pyrocystis, and Noctiluca, using anti-luciferase and immunogold labelling. J. Cell Sci. 1987, 87 ( Pt 1), 189–96. 120. Nicolas, A. M.; Nicolas, G.; Johnson, C. H.; Bassot, J.; Nicolas, M.; Hastings, J. W. Characterization of the Bioluminescent Organelles in Gonyaulax polyedra (Dinoflagellates) after Fast-Freeze Fixation and Antiluciferase immunogold Staining. J. Cell Biol. 1987, 105, 723–735. 121. Topalov, G.; Kishi, Y. Chlorophyll catabolism leading to the skeleton of dinoflagellate and krill luciferins: Hypothesis and model studies. Angew. Chemie - Int. Ed. 2001, 40, 3892–3894. 122. Desjardins, M.; Morse, D. The polypeptide components of scintillons, the bioluminescence organelles of the dinoflagellate Gonyaulax polyedra. Biochem. cell Biol.

131

1993, 71, 176–182. 123. Morse, D.; Pappenheimer, A. M.; Hastings, J. W. Role of a luciferin-binding protein in the circadian bioluminescent reaction of Gonyaulax polyedra. J. Biol. Chem. 1989, 264, 11822–6. 124. Lecuyer, B.; Arrio, B.; Fresneau, C.; Volfin, P. Dinoflagellate luciferases: purification of luciferases from Gonyaulax polyedra, Pyrocystis lunula, and Pyrocystis fusiformis. Arch. Biochem. Biophys. 1979, 196, 371–84. 125. Krieger, N.; Hastings, J. W. Bioluminescence: pH Activity Profiles of Related Luciferase Fractions. Science (80-. ). 1968, 161, 586–589. 126. Eckert, R.; Reynolds, G. T. The subcellular origin of bioluminescence in Noctiluca miliaris. J. Gen. Physiol. 1967, 50, 1429–58. 127. Beauchemin, M.; Roy, S.; Daoust, P.; Dagenais-Bellefeuille, S.; Bertomeu, T.; Letourneau, L.; Lang, B. F.; Morse, D. Dinoflagellate tandem array gene transcripts are highly conserved and not polycistronic. Proc. Natl. Acad. Sci. 2012, 109, 15793–15798. 128. Keeling, P. J.; Burki, F.; Wilcox, H. M.; Allam, B.; Allen, E. E.; Amaral-Zettler, L. a.; Armbrust, E. V.; Archibald, J. M.; Bharti, A. K.; Bell, C. J.; Beszteri, B.; Bidle, K. D.; Cameron, C. T.; Campbell, L.; Caron, D. a.; Cattolico, R. A.; Collier, J. L.; Coyne, K.; Davy, S. K.; Deschamps, P.; Dyhrman, S. T.; Edvardsen, B.; Gates, R. D.; Gobler, C. J.; Greenwood, S. J.; Guida, S. M.; Jacobi, J. L.; Jakobsen, K. S.; James, E. R.; Jenkins, B.; John, U.; Johnson, M. D.; Juhl, A. R.; Kamp, A.; Katz, L. a.; Kiene, R.; Kudryavtsev, A.; Leander, B. S.; Lin, S.; Lovejoy, C.; Lynn, D.; Marchetti, A.; McManus, G.; Nedelcu, A. M.; Menden-Deuer, S.; Miceli, C.; Mock, T.; Montresor, M.; Moran, M. A.; Murray, S.; Nadathur, G.; Nagai, S.; Ngam, P. B.; Palenik, B.; Pawlowski, J.; Petroni, G.; Piganeau, G.; Posewitz, M. C.; Rengefors, K.; Romano, G.; Rumpho, M. E.; Rynearson, T.; Schilling, K. B.; Schroeder, D. C.; Simpson, A. G. B.; Slamovits, C. H.; Smith, D. R.; Smith, G. J.; Smith, S. R.; Sosik, H. M.; Stief, P.; Theriot, E.; Twary, S. N.; Umale, P. E.; Vaulot, D.; Wawrik, B.; Wheeler, G. L.; Wilson, W. H.; Xu, Y.; Zingone, A.; Worden, A. Z. The Marine Microbial Eukaryote Transcriptome Sequencing Project (MMETSP): Illuminating the Functional Diversity of Eukaryotic Life in the Oceans through Transcriptome Sequencing. PLoS Biol. 2014, 12. 129. Cherny, V. V; Morgan, D.; Musset, B.; Chaves, G.; Smith, S. M.; DeCoursey, T. E. Tryptophan 207 is crucial to the unique properties of the human voltage-gated proton channel, hHV1. J Gen Physiol 2015, 146, 343–356. 130. Fritz, L.; Morse, D.; Hastings, J. W. The circadian bioluminescence rhythm of Gonyaulax is related to daily variations in the number of light-emitting organelles. J. Cell Sci. 1990, 95 ( Pt 2), 321–8. 131. Fogel, M.; Schmitter, R. E.; Hastings, J. W. On the physical identity of scintillons: bioluminescent particles in Gonyaulax polyedra. J. Cell Sci. 1972, 11, 305–317. 132. Koch, H. P.; Kurokawa, T.; Okochi, Y.; Sasaki, M.; Okamura, Y.; Larsson, H. P. Multimeric nature of voltage-gated proton channels. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 9111–6. 133. Lee, S.-Y.; Letts, J. A.; Mackinnon, R. Functional reconstitution of purified human Hv1 H+ channels. J. Mol. Biol. 2010, 387, 1055–1060. 134. Smith, S. M. E.; DeCoursey, T. E. Consequences of dimerization of the voltage- gated proton channel. Prog. Mol. Biol. Transl. Sci. 2013, 117, 335–360. 135. Tombola, F.; Ulbrich, M. H.; Isacoff, E. Y. The voltage-gated proton channel Hv1

132

has two pores, each controlled by one voltage sensor. Neuron 2008, 58, 546–556. 136. Dunlap, J. C.; Hastings, J. W. Biochemistry of Dinoflagellate Bioluminescence : Purification and Characterization of Dinoflagellate Luciferin from Pyrocystis Zunulaf. Biochemistry 1981, 20, 983–989. 137. Mahaut-Smith, M. P. The effect of zinc on calcium and hydrogen ion currents in intact snail neurones. J. Exp. Biol. 1989, 145, 455–64. 138. Cherny, V. V.; DeCoursey, T. E. pH-dependent inhibition of voltage-gated H(+) currents in rat alveolar epithelial cells by Zn(2+) and other divalent cations. J. Gen. Physiol. 1999, 114, 819–838. 139. Liu, Y.; Schmidt, B.; Masekll, D. MSAProbs: multiple sequence alignment based on pair hidden Markov models and partition function posterior probabilities. Bioinformatics 2010, 26, 1958–1964. 140. Fogel, M.; Hastings, J. W. A Substrate-Binding Protein in the Gonyaulax Bioluminescence Reaction. Arch. Biochem. Biophys. 1971, 142, 310–321. 141. Nawata, T.; Sibaoka, T. Ionic composition and pH of the vacuolar sap in marine dinoflagellate Noctiluca. Plant Cell Physiol. 1976, 17, 265–272. 142. Hamman, J. P.; Seliger, H. H. The mechanical triggering of bioluminescence in marine dinoflagellates: Chemical basis. J. Cell. Physiol. 1972, 80, 397–408. 143. Fujiwara, Y.; Kurokawa, T.; Okamura, Y. Long α helices projecting from the membrane as the dimer interface in the voltage-gated H(+) channel. J. Gen. Physiol. 2014, 143, 377–86. 144. DeCoursey, T. E.; Morgan, D.; Musset, B.; Cherny, V. V. Insights into the structure and function of H V 1 from a meta-analysis of mutation studies. J. Gen. Physiol. 2016, 148, 97–118. 145. Lacroix, J. J.; Hyde, H. C.; Campos, F. V; Bezanilla, F. Moving gating charges through the gating pore in a Kv channel voltage sensor. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, E1950-9. 146. Li, Q.; Wanderling, S.; Paduch, M.; Medovoy, D.; Singharoy, A.; McGreevy, R.; Villalba-Galea, C. a; Hulse, R. E.; Roux, B.; Schulten, K.; Kossiakoff, A.; Perozo, E. Structural mechanism of voltage-dependent gating in an isolated voltage-sensing domain. Nat. Struct. Mol. Biol. 2014. 147. Li, Q.; Shen, R.; Treger, J. S.; Wanderling, S. S.; Milewski, W.; Siwowska, K.; Bezanilla, F.; Perozo, E. Resting state of the human proton channel dimer in a lipid bilayer. Proc. Natl. Acad. Sci. 2015, 112, E5926–E5935. 148. DeCoursey, T. E. Structural revelations of the human proton channel. Proc. Natl. Acad. Sci. 2015, 112, 13430–13431. 149. Tao, X.; Lee, A.; Limapichat, W.; Dougherty, D. A. A gating Charge Transfer Center in voltage Sensors. NIH Public Access 2010, 328, 67–73. 150. Latz, M. I.; Bovard, M.; VanDelinder, V.; Segre, E.; Rohr, J.; Groisman, A. Bioluminescent response of individual dinoflagellate cells to hydrodynamic stress measured with millisecond resolution in a microfluidic device. J. Exp. Biol. 2008, 211, 2865–2875. 151. Jin, K.; Klima, J. C.; Deane, G.; Dale Stokes, M.; Latz, M. I. Pharmacological investigation of the bioluminescence signaling pathway of the dinoflagellate Lingulodinium polyedrum : evidence for the role of stretch-activated ion channels. J. Phycol. 2013, 49, 733–745.

133

152. Cherny, V. V; Thomas, L. L.; DeCoursey, T. E. Voltage-gated proton currents in human basophils. Biol. Membr. 2001, 18, 458–465. 153. DeCoursey, T. E.; Cherny, V. V Temperature dependence of voltage-gated H+ currents in human neutrophils, rat alveolar epithelial cells, and mammalian phagocytes. J. Gen. Physiol. 1998, 112, 503–522. 154. DeCoursey, T. E.; Cherny, V. V Voltage-activated proton currents in human THP-1 monocytes. J. Membr. Biol. 1996, 152, 131–140. 155. DeCoursey, T. E.; Cherny, V. V; DeCoursey, A. G.; Xu, W.; Thomas, L. L. Interactions between NADPH oxidase-related proton and electron currents in human eosinophils. J. Physiol. 2001, 535, 767–781. 156. Villalba-Galea, C. A. Hv1 proton channel opening is preceded by a voltage- independent transition. Biophys. J. 2014, 107, 1564–1572. 157. Tesson, B.; Latz, M. I. Mechanosensitivity of a Rapid Bioluminescence Reporter System Assessed by Atomic Force Microscopy. Biophys. J. 2015, 108, 1341–1351. 158. von Dassow, P.; Latz, M. I. The role of Ca(2+) in stimulated bioluminescence of the dinoflagellate Lingulodinium polyedrum. J. Exp. Biol. 2002, 205, 2971–86. 159. Oami, K.; Naitoh, Y.; Sibaoka, T. Voltage-gated ion conductances corresponding to regenerative positive and negative spikes in the dinoflagellateNoctiluca miliaris. J. Comp. Physiol. A 1995, 176, 625–633. 160. Nawata, T.; Sibaoka, T. Coupling between action potential and bioluminescence inNoctiluca: Effects of inorganic ions and pH in vacuolar sap. J. Comp. Physiol. 1979, 134, 137–149. 161. Orr, R. J. S.; Murray, S. A.; Stüken, A.; Rhodes, L.; Jakobsen, K. S. When Naked Became Armored: An Eight-Gene Phylogeny Reveals Monophyletic Origin of Theca in Dinoflagellates. PLoS One 2012, 7, e50004. 162. Bachvaroff, T. R.; Gornik, S. G.; Concepcion, G. T.; Waller, R. F.; Mendez, G. S.; Lippmeier, J. C.; Delwiche, C. F. Dinoflagellate phylogeny revisited: Using ribosomal proteins to resolve deep branching dinoflagellate clades. Mol. Phylogenet. Evol. 2014, 70, 314–322. 163. Pozdnyakov, I. A.; Skarlato, S. O. Analysis of the Dinoflagellate Prorocentrum Minimum Transcriptome: Identifying the Members of the Voltage-Gated Cation Channels Superfamily. Tsitologiia 2015, 57, 533–543. 164. Morgan, D.; Cherny, V. V; Murphy, R.; Katz, B. Z.; DeCoursey, T. E. The pH dependence of NADPH oxidase in human eosinophils. J. Physiol. 2005, 569, 419–431. 165. Guillard, R. R. L.; Hargraves, P. E. Stichochrysis immobilis is a diatom, not a chrysophyte. Phycologia 1993, 32, 234–236. 166. Guillard, R. R. L.; Ryther, J. H. Studies of marine planktonic diatoms. I. Cyclotella nana Hustedt and Detonula confervacea Cleve. Can. J. Microbiol 1962, 8, 229–239. 167. Brunelle, S. A.; Hazard, E. S.; Sotka, E. E.; Dolah, F. M. Van Characterization of a Dinoflagellate Cryptochrome Blue-Light Receptor With a Possible Role in Circadian Control of the Cell Cycle. J. Phycol. 2007, 43, 509–518. 168. Gómez, F. A quantitative review of the lifestyle, habitat and trophic diversity of dinoflagellates (Dinoflagellata, Alveolata). Syst. Biodivers. 2012, 10, 267–275. 169. Hopwood, D. A.; Sherman, D. H. Molecular Genetics of Polyketides and its Comparision to Fatty Acid Biosynthesis. Annu. Rev. Genet. 1990, 24, 37–66. 170. Haq, S.; Bachvaroff, T.; Place, A. Characterization of Acetyl-CoA Carboxylases in

134

the Basal Dinoflagellate Amphidinium carterae. Mar. Drugs 2017, 15, 149. 171. Hunkeler, M.; Stuttfeld, E.; Hagmann, A.; Imseng, S.; Maier, T. The dynamic organization of fungal acetyl-CoA carboxylase. Nat. Commun. 2016, 1–11. 172. Huerlimann, R.; Zenger, K. R.; Jerry, D. R.; Heimann, K. Phylogenetic analysis of nucleus-encoded acetyl-CoA carboxylases targeted at the cytosol and plastid of algae. PLoS One 2015, 10, 1–10. 173. del Campo, J.; Sieracki, M. E.; Molestina, R.; Keeling, P.; Massana, R.; Ruiz-Trillo, I. The others: our biased perspective of eukaryotic genomes. Trends Ecol. Evol. 2014, 29, 252–259. 174. Delwiche, C. F. Tracing the Thread of Plastid-Diversity through the Tapestry of Life. Am. Nat. 1999, 154. 175. Emanuelsson, O.; Brunak, S.; von Heijne, G.; Nielsen, H. Locating proteins in the cell using TargetP, SignalP and related tools. Nat. Protoc. 2007, 2, 953–971. 176. Berges, J. A.; Franklin, D. J.; Harrison, P. J. Evolution of an artificial seawater medium: Improvements in enriched seawater, artificial water over the last two decades. J. Phycol. 2001, 37, 1138–1145. 177. Gunderson, J. H.; Goss, S. H.; Coats, D. W. The Phylogenetic Position of Amoebophyra sp. Infecting Gymnodinium sanguineum. J. Euk. Microbiol. 1999, 46, 194–197. 178. Cembella, A. D. Chemical ecology of eukaryotic microalgae in marine ecosystems. Phycologia 2003, 42, 420–447. 179. Smayda, T. J. Bloom dynamics : Physiology , behavior , trophic e : ffects. Limonaology Oceanogr. 1997, 42, 1132–1136. 180. Pulz, O.; Gross, W. Valuable products from biotechnology of microalgae. Appl. Microbiol. Biotechnol. 2004, 65, 635–648. 181. Schenk, P. M.; Thomas-Hall, S. R.; Stephens, E.; Marx, U. C.; Mussgnug, J. H.; Posten, C.; Kruse, O.; Hankamer, B. Second Generation Biofuels: High-Efficiency Microalgae for Biodiesel Production. BioEnergy Res. 2008, 1, 20–43. 182. Brownsey, R. W.; Zhande, R.; Boone, a N. Isoforms of acetyl-CoA carboxylase: structures, regulatory properties and metabolic functions. Biochem. Soc. Trans. 1997, 25, 1232–1238. 183. Abu-Elheiga, L.; Brinkley, W. R.; Zhong, L.; Chirala, S. S.; Woldegiorgis, G.; Wakil, S. J. The subcellular localization of acetyl-CoA carboxylase 2. Proc. Natl. Acad. Sci. U. S. A. 2000, 97, 1444–1449. 184. Sasaki, Y.; Nagano, Y. Plant Acetyl-CoA Carboxylase: Structure, Biosynthesis, Regulation, and Gene Manipulation for Plant Breeding. Biosci. Biotechnol. Biochem. 2004, 68, 1175–84. 185. Houdai, T.; Matsuoka, S.; Matsumori, N.; Murata, M. Membrane-permeabilizing activities of amphidinol 3, polyene-polyhydroxy antifungal from a marine dinoflagellate. Biochim. Biophys. Acta - Biomembr. 2004, 1667, 91–100. 186. Morsy, N.; Houdai, T.; Konoki, K.; Matsumori, N.; Oishi, T.; Murata, M. Effects of lipid constituents on membrane-permeabilizing activity of amphidinols. Bioorganic Med. Chem. 2008, 16, 3084–3090. 187. Murray, S. A.; Garby, T.; Hoppenrath, M.; Neilan, B. A. Genetic diversity, morphological uniformity and polyketide production in dinoflagellates (amphidinium, dinoflagellata). PLoS One 2012, 7, 1–14.

135

188. Emanuelsson, O.; Nielsen, H.; von Heijne, G. ChloroP, a neural network-based method for predicting chloroplast transit peptides and their cleavage sites. Protein Sci. 1999, 8, 978–984. 189. Emanuelsson, O.; Nielsen, H.; Brunak, S.; von Heijne, G. Predicting subcellular localization of proteins based on their N-terminal amino acid sequence. J.Mol.Biol. 2000, 300, 1005–1016. 190. Haq, S.; Bachvaroff, T. R.; Place, A. R. Phylogenetic Analysis of Acetyl CoA Carboxylases in Dinoflagellates. In MARINE AND FRESHWATER HARMFUL ALGAE - Proceedings of the 17 International Conference on Harmful Algae 9-14 October 2016; Proença, L. A. O.; Hallegraeff, G. M., Eds.; Florianópolis, Brazil, 2017; pp. 138–141. 191. Williams, E.P., Jagus, R., and Place, A. R. Discovery of non-coding RNAs in Amphidinium carterae differentially expressed over a diel cycle. In 16th ICHA Proceedings; 2015; pp. 233–237. 192. Jones, G. D.; Williams, E. P.; Place, A. R.; Jagus, R.; Bachvaroff, T. R. The alveolate translation initiation factor 4E family reveals a custom toolkit for translational control in core dinoflagellates. BMC Evol. Biol. 2015, 15, 14. 193. Wakil, S. J.; Stoops, J. K.; Joshi, V. C. Fatty acid synthesis and its regulation. Annu. Rev. Biochem. 1983, 52, 537–79. 194. Bachvaroff, T. R.; Handy, S. M.; Place, A. R.; Delwiche, C. F. Alveolate Phylogeny Inferred using Concatenated Ribosomal Proteins. J. Eukaryot. Microbiol. 2011, 58, 223– 233. 195. Patron, N. J.; Waller, R. F.; Archibald, J. M.; Keeling, P. J. Complex protein targeting to dinoflagellate plastids. J. Mol. Biol. 2005, 348, 1015–1024. 196. Zuther, E.; Johnson, J. J.; Haselkorn, R.; McLeod, R.; Gornicki, P. Growth of Toxoplasma gondii is inhibited by aryloxyphenoxypropionate herbicides targeting acetyl- CoA carboxylase. Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 13387–13392. 197. Yanai, Y.; Kawasaki, T.; Shimada, H.; Wurtele, E. S.; Nikolau, B. J.; Ichikawa, N. Genomic organization of 251 kDa acetyl-CoA carboxylase genes in Arabidopsis: tandem gene duplication has made two differentially expressed isozymes. Plant Cell Physiol. 1995, 36, 779–87. 198. Huerlimann, R.; Zenger, K. R.; Jerry, D. R.; Heimann, K. Phylogenetic analysis of nucleus-encoded acetyl-CoA carboxylases targeted at the cytosol and plastid of algae. PLoS One 2015, 10, 1–10. 199. Tibbetts, J. Toxic Tides. Environ. Health Perspect. 1998, 106, A326–A331. 200. Fischbach, M. A.; Walsh, C. T. Assembly-line enzymology for polyketide and nonribosomal peptide antibiotics: Logic machinery, and mechanisms. Chem. Rev. 2006, 106, 3468–3496. 201. Joseph, J. D. Identification of 3, 6, 9, 12, 15-octadecapentaenoic acid in laboratory- cultured photosynthetic dinoflagellates. Lipids 1975, 10, 395–403. 202. Bell, M. V.; Dick, J. R.; Pond, D. W. Octadecapentaenoic acid in a raphidophyte alga, Heterosigma akashiwo. Phytochemistry 1997, 45, 303–306. 203. Leblond, J. D.; Chapman, P. J. Lipid Class Distribution of Highly Unsaturated Long Chain Fatty Acids in Marine Dinoflagellates. J. Phycol. 2000, 36, 1103–1108. 204. Felnagle, E. A.; Jackson, E. E.; Chan, Y. A.; Podevels, A. M.; Berti, D.; Mcmahon, M. D.; Thomas, M. G. Nonribosomal Peptide Synthetases Involved in the Production of Medically Relevant Natural Products. 2011, 5, 191–211.

136

205. Kohli, G. S.; John, U.; Van Dolah, F. M.; Murray, S. A. Evolutionary distinctiveness of fatty acid and polyketide synthesis in eukaryotes. ISME J. 2016. 206. Kohli, G. S.; John, U.; Van Dolah, F. M.; Murray, S. a. Functional and evolutionary distinctiveness of fatty acid and polyketide synthesis in protists. Pnas 2013, I. 207. Paul, G. K.; Matsumori, N.; Murata, M.; Tachibana, K. Isolation and chemical structure of amphidinol 2, a potent hemolytic compound from marine dinoflagellate Amphidinium klebsii. Tetrahedron Lett. 1995, 36, 6279–6282. 208. Skipski, V. P.; Smolowe, A. F.; Sullivan, C. R.; Barclay, M. Separation of Lipid Classes by Thin-Layer Chromatography. Biochim. Biophys. Acta 1965, 106, 386–396. 209. Bachvaroff, T. R.; Adolf, J. E.; Squier, A. H.; Harvey, H. R.; Place, A. R. Characterization and quantification of karlotoxins by liquid chromatography-mass spectrometry. Harmful Algae 2008, 7, 473–484. 210. Omura, S. The antibiotic cerulenin, a novel tool for biochemistry as an inhibitor of fatty acid synthesis. Bacteriol. Rev. 1976, 40, 681–697. 211. Funabashi, H.; Kawaguchi, A.; Tomoda, H.; Omura, S.; Okuda, S.; Iwasaki, S. Binding Site of Cerulenin in Fatty Acid Synthetasel Hiroshi Funabashi. Biochemistry 1989, 105, 751–755. 212. Vance, D.; Goldberg, I.; Mitsuhashi, O.; Bloch, K.; Omura, S.; Nomura, S. Inhibition of Fatty Acid Synthetases by the Antibiotic Cerulenin. Biochem. Biophys. Res. Commun. 1972, 48, 649–656. 213. Amo, M.; Suzuki, N.; Kawamura, H.; Yamaguchi, A.; Takano, Y.; Horiguchi, T. Sterol composition of dinoflagellates: Different abundance and composition in heterotrophic species and resting cysts. Geochem. J. 2010, 44, 225–231. 214. Scheuer, P. J. Marine Natural Products: Chemical and Biological Perspectives; Elsevier Science, 2013. 215. Cutignano, A.; Nuzzo, G.; Sardo, A.; Fontana, A. The Missing piece in biosynthesis of amphidinols: First evidence of glycolate as a starter unit in New Polyketides from Amphidinium carterae. Mar. Drugs 2017, 15, 1–13. 216. Smith, S. The animal fatty acid synthase: one gene, one polypeptide, seven enzymes. FASEB J. 1994, 8, 1248–1259. 217. Shulse, C. N.; Allen, E. E. Widespread Occurrence of Secondary Lipid Biosynthesis Potential in Microbial Lineages. PLoS One 2011, 6. 218. Adolf, J. E.; Place, A. R.; Stoecker, D. K.; Harding, L. J. W. Modulation of polyunsaturated fatty acids in mixotrophic Karlodinium veneficum (Dinophyceae) and its prey, Storeatula major (Cryptophyceae). J. Phycol. 2007, 43, 1259–1270. 219. Strohalm, M.; Hassman, M.; Košata, B.; Kodíček, M. mMass data miner: an open source alternative for mass spectrometric data analysis. Rapid Commun. Mass Spectrom. 2008, 22, 905–908. 220. Akbar, M.; Ahmad, A.; Usup, G.; Bunawan, H. Current Knowledge and Recent Advances in Marine Dinoflagellate Transcriptomic Research. J. Mar. Sci. Eng. 2018, 6, 13. 221. Mccune, A.; Harris, A. Mechanism of Hepatic Responsible Lipogenesis * for 5- ( Tetradecyloxy ) -24uroic Acid Inhibition. J. Biol. Chem. 1979, 254, 10095–10101. 222. Zhang, Y.; Fonslow, B. R.; Shan, B.; Baek, M.-C.; Yates, J. R.; III Protein analysis by shotgun/bottom-up proteomics. Chem. Rev. 2013, 113, 2343–94. 223. Khosla, C.; Herschlag, D.; Cane, D. E.; Walsh, C. T. Assembly Line Polyketide

137

Synthases: Mechanistic Insights and Unsolved Problems. 2014. 224. Shelest, E.; Heimerl, N.; Fichtner, M.; Sasso, S. Multimodular type I polyketide synthases in algae evolve by module duplications and displacement of AT domains in trans. BMC Genomics 2015, 16, 1015. 225. Piel, J. Biosynthesis of polyketides by trans-AT polyketide synthases. Nat. Prod. Rep. 2010, 27, 996. 226. Bennett, V.; Stenbuck, P. J. Identification and partial purification of ankyrin, the high affinity membrane attachment site for human erythrocyte spectrin. J. Biol. Chem. 1979, 254, 2533–41. 227. Verwoert, I. I.; Verbree, E. C.; van der Linden, K. H.; Nijkamp, H. J.; Stuitje, A. R. Cloning, nucleotide sequence, and expression of the Escherichia coli fabD gene, encoding malonyl coenzyme A-acyl carrier protein transacylase. J. Bacteriol. 1992, 174, 2851–7.

138