Genome-Enabled Investigation of the Minimal Growth Requirements and Phosphate Metabolism for Pelagibacter Marine Bacteria

AN ABSTRACT OF THE DISSERTATION OF

Paul J. Carini for the degree of Doctor of Philosophy in Microbiology presented on May 23, 2013. Title: Genome-Enabled Investigation of the Minimal Growth Requirements and Phosphate Metabolism for Pelagibacter Marine Bacteria

Abstract approved:

Stephen J. Giovannoni

Members of the SAR11 clade of heterotrophic α-proteobacteria are ubiquitous and abundant in the world’s oceans where they are thought to play a pivotal role in the global carbon cycle. The first SAR11 bacterium cultivated in vitro, ‘Candidatus

Pelagibacter ubique’ HTCC1062 (Ca. P. ubique), was isolated by dilution into sterile natural seawater, from which undefined dissolved organic carbon supplied nutrients for heterotrophic growth. However, variation in the composition of such native dissolved organic matter hindered efforts to identify the specific nutrient requirements of Ca. P. ubique and elucidate its metabolism. For this dissertation work, genome-enabled metabolic reconstruction was used to develop a defined artificial seawater medium for

Ca. P. ubique that consisted of inorganic salts, defined types and amounts of organic carbon, reduced sulfur and vitamins. Subsequent in vitro experimentation was used to show that Ca. P. ubique requires simultaneous additions of glycine and methionine to meet cellular requirements for glycine and sulfur, respectively. A new requirement for pyruvate was identified and linked to the production of alanine. We found that pyruvate could be replaced by additions of glucose or oxaloacetate and that glycine could be replaced with serine or glycine betaine. Interestingly, glycolate partially fulfilled the glycine requirement of Ca. P. ubique, likely because of dual use as a carbon and glycine source. Once these major organic nutrient requirements were established, the defined medium was used to identify specific trace requirements of Ca. P. ubique for vitamins or vitamin precursors. Previously, analysis of the Ca. P. ubique genome did not identify complete biosynthetic pathways for thiamine (vitamin B1), pantothenate (vitamin B5), pyridoxine (vitamin B6), biotin (vitamin B7) or cobalamin (vitamin B12), suggesting that these vitamins cannot be synthesized de novo. We describe requirements for the thiamine-pyrimidine precursor 4-amino-5-hydroxymethyl-2-methylpyrimidine (HMP) and vitamin B7 as determined from laboratory cultures. We determined in situ (the

Sargasso Sea) concentrations of HMP by liquid chromatography coupled tandem mass spectrometry. Our measurements show that in situ HMP concentrations are ample for the native SAR11 population. Genomic and experimental evidence that vitamins B5 and B6 are synthesized de novo is presented and genomic evidence that vitamin B12 is not required by Ca. P. ubique is discussed. A second Pelagibacter isolate, Pelagibacter sp. str. HTCC7211 was also grown on a defined medium. Pelagibacter sp. str. HTCC7211 encodes a suite of genes, absent in Ca. P. ubique, dedicated to the acquisition and storage of inorganic phosphate and the utilization of organic phosphorus. On a defined medium, the growth of both Ca. P. ubique and Pelagibacter sp. str. HTCC7211 was limited by excluding inorganic phosphate (Pi) and we show that Ca. P. ubique has an apparent

-1 requirement for 10.2 attomoles Pi cell and Pelagibacter sp. str. HTCC7211 has an

-1 apparent requirement for 45.7 attomoles Pi cell . Discrete phosphorus utilization physiotypes were evident, whereby Pelagibacter sp. str. HTCC7211, but not Ca. P. ubique, utilized assorted Pi-esters and phosphonates in place of Pi to meet its cellular P- requirement. When grown on methylphosphonic acid (Mpn), the apparent P-requirement decreased in Pelagibacter sp. str. HTCC7211 to 11.4 attomoles cell-1. We investigated the underlying transcriptional dynamics that give rise to the observed physiotypes using

DNA microarrays. Ca. P. ubique responded rapidly to P-limitation by upregulating a high-affinity transport system operon (pstSCAB-phoUB) and stress response genes suggestive of a stringent response. Conversely, Pelagibacter sp. str. HTCC7211 responded slowly to the onset of P-limitation; no transcripts were differentially regulated for 68 hours, after which, phosphonate transport and utilization genes showed the greatest increases in expression. Collectively, these experiments exemplify the utility of growing

Ca. P. ubique (and other Pelagibacter isolates) on a defined medium and have opened the door for additional experiments that were previously impossible to conduct on natural seawater. This work also represents an important step towards developing Ca. P. ubique into a model system that can be used to understand the cellular adaptations that make oligotrophy a successful life strategy in the sea.

Genome-Enabled Investigation of the Minimal Growth Requirements and Phosphate Metabolism for Pelagibacter Marine Bacteria

by Paul J. Carini

A DISSERTATION

submitted to

Oregon State University

in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Presented May 23, 2013 Commencement June 2014

Doctor of Philosophy dissertation of Paul J. Carini presented on May 23, 2013.

APPROVED:

Major Professor, representing Microbiology

Chair of the Department of Microbiology

Dean of the Graduate School

I understand that my dissertation will become part of the permanent collection of Oregon State University libraries. My signature below authorizes release of my dissertation to any reader upon request.

Paul J. Carini, Author

ACKNOWLEDGEMENTS

I express my sincerest appreciation and love to my wife Kiri for her unwavering love and support (and tolerance) that has been the foundation for my successful completion of this degree. To my daughter Lucia (arguably my greatest experiment to date), whose smiles and hugs reverse even the most discouraging of days, thank you for your unconditional love. I thank Stephen Giovannoni for the opportunity to excel under his tutelage; as an advisor, he didn’t push me to become a better scientist, instead he challenged me, which is exactly what I needed to grow into the scientist I leave as today

(although I didn’t always realize it). Thank you to Emily O. Campbell for her belief in me as a mentor and hard work necessary to complete the experiments described in Chapters

3-5. H. James Tripp has been instrumental to my knowledge of the methodologies of metabolic reconstruction and cultivation of SAR11 cells. Through the years, our professional relationship has grown into a close friendship that I value greatly. Thank you to my family in Wisconsin who has endured (or reveled in?) my six years away from home and to my in-laws in Oregon who have given me a new sense of home. A special thank you to Carmen for introducing me to Kiri and Mariah for being there to keep Kiri and Lucia company when I was busy doing experiments.

CONTRIBUTION OF AUTHORS

Dr. Laura Steindler and Dr. Sara Beszteri performed experiments related to the cell division patterns linked to alanine use in chapter 2. Emily O. Campbell performed some of the growth experiments in chapter 3, chapter 4 and chapter 5. Jeff Morré and Dr.

Samuel Bennett assisted with the extraction and quantification of vitamins from seawater described in chapter 3. Dr. Sergio Sañudo-Wilhelmy performed vitamin measurements important to the results obtained in chapters 3 and chapter 5. Dr. J. Cameron Thrash constructed the phylogenetic trees in chapter 3. Dr. Taghd Begley synthesized the 4- amino-5-hydroxymethyl-2-methylpyrimidine and 4-amino-5-aminomethyl-2- methylpyrimidine for experiments in chapter 3. Dr. Angelicque E. White assisted with methane measurements described in chapter 4. Dr. Stephen J. Giovannoni and the author conceptualized all experiments.

TABLE OF CONTENTS

Page Chapter 1 General Introduction ...... 1 The foundations of microbial ecology ...... 2 Marine microbiology ...... 2 The ‘great plate count anomaly’ ...... 4 Discovery and cultivation of SAR11 ...... 5 The SAR11 genome sequence and metabolic reconstruction ...... 8 Unraveling the metabolism of Ca. P. ubique ...... 8 Genome streamlining ...... 13 Growth on a defined medium ...... 15 Chapter 2 Nutrient Requirements for the Growth of the Extreme Oligotroph ‘Candidatus Pelagibacter ubique’ HTCC1062 on a Defined Medium ...... 19 Abstract ...... 20 Introduction ...... 21 Methods...... 24 Results ...... 27 Discussion of results ...... 29 Chapter 3 ‘Candidatus Pelagibacter ubique’ HTCC1062 is Dependent on 4-Amino-5- Hydroxymethyl-2-Methylpyrimidine, An Abundant Thiamine Precursor In The Sea ..... 46 Abstract ...... 47 Introduction ...... 48 Results & Discussion ...... 49 Methods...... 58 Chapter 4 Distinct Phosphate-Acquisition Physiotypes and Adaptations for Extreme Marine Oligotrophy in SAR11 Bacteria ...... 66 Abstract ...... 67 Introduction ...... 68 Methods...... 71 Results ...... 75 Discussion ...... 78

Chapter 5 Analysis of the Vitamin B5, B6, B7 And B12 Requirements of ‘Candidatus Pelagibacter ubique’ HTCC1062: Evidence for B7 Auxotrophy...... 97 Abstract ...... 98 Introduction ...... 99 Methods...... 104 Results ...... 106 Conclusions ...... 109 Chapter 6 Conclusions ...... 121 Bibliography ...... 123 TABLE OF CONTENTS (Continued)

Page Appendices ...... 139 Appendix 1 ...... 140 Appendix 2 ...... 145 Appendix 3 ...... 164

LIST OF FIGURES

Figure Page

Figure 2-1: Simplified illustration of central metabolism in Ca. P. ubique...... 41 Figure 2-2: Growth of Ca. P. ubique in AMS1 with organic carbon additions...... 42 Figure 2-3: Maximum cell yields of Ca. P. ubique in response to pyruvate and glycine additions…………………………………………………………………………….43 Figure 2-4: DNA content and morphology of SYBR Green I-stained stationary phase cells from pyruvate-deplete and replete batch cultures...... 44 Figure 2-5: Glycolate assimilation gene organizations in E. coli and Ca. P. ubique...... 45 Figure 3-1: Comparative genomics of thiamine biosynthesis...... 63 Figure 3-2: Growth of Ca. P. ubique under thiamine-limiting conditions and responses to thiamine pyrimidines...... 64 Figure 3-3: Depth distribution of dissolved 4-amino-5-hydroxymethyl-2- methylpyrimidine (HMP) and thiamine in the Sargasso Sea...... 65 Figure 4-1: Linear dose responses of Ca. P. ubique and Pelagibacter sp. str. HTCC7211 to P-sources...... 93 Figure 4-2: P-source utilization by Ca. P. ubique and Pelagibacter sp. str. HTCC7211. 94 Figure 4-3: Comparative genomics of gene expression during Pi-starvation in Ca. P. ubique and Pelagibacter sp. str. HTCC7211...... 95 Figure 4-4: Methane (CH4) production by Pelagibacter sp. str. HTCC7211 utilizing Mpn as a sole P-source...... 96 Figure 5-1: Simplified illustration of B-vitamin-requiring reactions in Ca. P. ubique. .. 113 Figure 5-2: Comparative genomics of vitamin B5 biosynthesis in Ca. P. ubique...... 114 Figure 5-3: Comparative genomics of vitamin B6 biosynthesis in Ca. P. ubique...... 115 Figure 5-4: Comparative genomics of vitamin B7 biosynthesis in Ca. P. ubique...... 116 Figure 5-5: Routes of canonical vitamin B12 biosynthesis…………………………...…117 Figure 5-6: Ca. P. ubique is bradytrophic for vitamin B5 (pantothenate) in the presence of vitamin B7 (biotin)...... 118 Figure 5-7: Vitamin B7-scrubbed medium allows for demonstration of vitamin B7- requirement in Ca. P. ubique...... 119

LIST OF TABLES

Table Page

Table 2-1: Constituents of the Artificial Medium for SAR11 (AMS1) ...... 38 Table 2-2: Potential sources of pyruvate for Ca. P. ubique when grown in AMS1 with 25 µM glycine and 10 µM methionine ...... 39 Table 2-3: Potential sources of glycine for Ca. P. ubique when grown in AMS1 with 50 µM pyruvate and 10 µM methionine ...... 40 Table 5-1: Common vitamin B12-requiring enzymes are absent in Ca. P. ubique...... 120

LIST OF APPENDIX FIGURES

Figure Page

Figure A1- 1: Transfer of Ca. P. ubique batch cultures from natural seawater to AMS1 ...... 140 Figure A1- 2: Flow cytometry profiles and corresponding microscopic images at different pyruvate:glycine molar ratios...... 141 Figure A1- 3: DNA fluorescence proﬁles and maximum cell densities in response to alanine additions...... 142 Figure A1- 4: Effect of vitamin additions on Ca. P. ubique growth...... 143 Figure A2- 1: Simplified illustration of Ca. P. ubique’s central metabolism, highlighting thiamine-requiring biosynthetic reactions…………………………………………145 Figure A2- 2: Reproduction of Figure 3-2a in main text with microarray sample points labeled...... 146 Figure A2- 3: Growth responses of Ca. P. ubique to potential thiamine precursors...... 146 Figure A2- 4: Maximum-likelihood phylogenetic reconstruction of the SAR11_0811 amino acid sequence...... 147 Figure A2- 5: Illustration of predicted thiamine regulatory elements or thiamine pyrimidine salvage genes genetically associated with the putative HMP transport protein coding sequence...... 149 Figure A2- 6: Profiles of in situ temperature and chlorophyll a fluorescence at Hydrostation S in the Sargasso Sea during water collection for vitamin analysis...156 Figure A3- 1: Growth curves of Pelagibacter sp. str. HTCC7211 on alternate phosphorus sources showing diauxic growth pattern…………………………………………..164 Figure A3- 2: Growth curves for linear dose responses of Pelagibacter sp. str. HTCC7211 grown on Pi...... 165 Figure A3- 3: Growth curves for linear dose responses of Ca. P. ubique grown on Pi...166 Figure A3- 4: Growth curves for linear dose responses of Pelagibacter sp. str. HTCC7211 grown on Mpn...... 167 Figure A3- 5: Simplified illustration of phospholipid biosynthetic routes and predicted reactions involved in lipid remodeling in Pelagibacter sp. str. HTCC7211...... 169 Figure A3- 6: Scanning transmission electron microscopy images of Pelagibacter sp. str. HTCC7211 cells grown under different P-regimes...... 170

LIST OF APPENDIX TABLES

Table Page

Table A2- 1: Gene transcripts more abundant in thiamine-limited conditions ...... 157 Table A2- 2: Gene transcripts more abundant in thiamine-replete conditions ...... 160 Table A3- 1: Specific growth rates of Ca. P. ubique & Pelagibacter sp. str. HTCC7211 in AMS1 with different amounts of Pi or Mpn……………………………………….171 Table A3- 2: Genes differentially regulated in P-limited Ca. P. ubique 4 hours after resuspension...... 172 Table A3- 3: Genes differentially regulated in P-limited Ca. P. ubique 20 hours after resuspension...... 173 Table A3- 4: Genes differentially regulated in P-limited Ca. P. ubique 38 hours after resuspension...... 175 Table A3- 5: Genes differentially regulated in P-limited Pelagibacter sp. str. HTCC7211 68 hours after re-suspension...... 177 Table A3- 6: Genes differentially regulated in P-limited Pelagibacter sp. str. HTCC7211 96 hours after re-suspension...... 178

DEDICATION

During my time in Oregon I have come to realize that my love for nature and my interest in how the world works is the result of years of listening to my grandfather

Joseph Horbas. Although not formally educated in science, he is a Naturalist, in the truest sense of the word. His love for nature and penchant for observation has resonated with me as long as I can remember. And, on days where I’m working outside or walking with

Lucia, I think of the times I spent working in his yard fondly, or find myself passing on this love of nature to the next generation of Naturalists. This dissertation work is dedicated to him.

Genome-Enabled Investigation of the Minimal Growth Requirements and Phosphate Metabolism for Pelagibacter Marine Bacteria

Chapter 1 General Introduction

For over a century, the field of microbial ecology has been predicated on the ability of scientists to cultivate and study environmentally relevant microbes under controlled laboratory conditions. The goal of the research contained herein is to develop isolates of the abundant and ubiquitous SAR11 clade of α-proteobacteria into model systems that can be used to understand the metabolism of oligotrophs and the biogeochemical processes they mediate in situ. Studies such as these have traditionally been at the forefront of microbial ecology. Though, with the 20th century advances in molecular biology and biochemistry - primarily nucleic acid sequencing and the expression of foreign genes in heterologous hosts - cultivation of environmentally important organisms has been minimized, effectively having been replaced by metabolic models of uncultivated microbes deduced from DNA sequence information. However, sequence data is merely descriptive, requiring experimental evidence to contextualize it in terms of organismal and ecosystem function. Therefore, as is presented here, the ideas and hypotheses constructed from genomes and metagenomic analyses should be addressed by experimentation for a truer understanding of microbial metabolism and the roles of microbes in ecosystem dynamics.

The foundations of microbial ecology

The understanding that bacteria act as the functional agents of geochemical change was largely derived from the experiments of Sergio Winogradsky and Martinus

Beijerinck in the late 19th century. Winogradsky discovered that key steps in the sulfur and nitrogen cycles were mediated by specific groups of bacteria. Through isolation and study of pure bacterial cultures, Winogradsky proposed the idea of chemolithotrophy – a form of metabolism that derives energy from inorganic compounds to fuel CO2 fixation.

Winogradsky’s contemporary, Beijerinck, also understood that specific assemblages of bacteria in the natural environment exploited distinct geochemical niches. Through his own work designing and testing different microbial growth media, Beijerinck developed the concept of the enrichment culture. He reasoned that by altering the nutritional composition of growth media, one could select for (that is, enrich for) specific nutrient transformations or consortia of bacteria that share similar growth requirements. By studying both pure and enriched cultures under laboratory conditions, Winogradsky and

Beijerinck pioneered what are now foundational ideas and methodology in microbial ecology. By linking laboratory results to environmental observations, they proposed that microbes had an active and important role in environmental nutrient cycling.

Marine microbiology

The prevailing view in the late 19th century was that although microbes were abundant in soils, they could not survive in the harsh conditions of the sea. High salt concentrations, extremely low amounts of organic carbon and other nutrients, together with large doses of ultraviolet light from the sun, were thought to render the open ocean

3 devoid of microbial life. Reports from Bernard Fischer in the late 1800’s changed this perspective (reviewed in (1)). Fischer experimentally developed growth media for marine microbes and used it to cultivate, characterize and study environmental distributions of bacteria in the sea. Fischer determined that an ideal medium for the growth of marine bacteria should contain 1.0% peptone, 0.5% fish extract, and 10% gelatin (or 2% agar) in a seawater base. Using this medium, he isolated and characterized a number of marine organisms, including the first species of luminescent bacteria (Photobacterium indicum).

Fischer also identified ecological relationships related to bacterial distributions. For example, he observed that nearly all samples of seawater contained cultivatable bacteria and that seawater contained an average of 1,083 bacteria per milliliter. In 1894 he noted that there were more bacteria in coastal waters than in the open ocean and that bacteria were more abundant at depth (200-400 meters) than at the surface.

In 1941 Claude ZoBell identified a reliable growth medium that produced higher cell counts per unit of seawater than the media described by Fischer (2). Known as

ZoBell medium 2216 (and currently known commercially as ‘Marine Agar 2216’), it consisted of Bacto™ peptone (5.0 g L-1), ferric phosphate (0.1 g L-1) and Bacto™ agar (15 g L-1), dissolved in aged or “rotted” seawater. Variations on ZoBell medium 2216 included the addition specific carbon sources or inorganic nutrients such as glucose (for example (3)) to further enrich for specific bacterial groups or activities. However, ZoBell himself understood the limitations of his medium – specifically, that no single growth medium could meet the growth requirements of all cells in a seawater sample (1, 2).

The ‘great plate count anomaly’

The culture-based methods used to enumerate marine bacteria as exemplified by

Fischer, ZoBell and others, routinely produced lower cell counts per unit volume than direct counts of planktonic cells by microscopy. For example, Waksman (4) counted

1000-fold more bacteria per milliliter of seawater under the microscope than were identified when using agar plate-based quantification methods. Jannasch and Jones (5) found that direct counts of seawater were 13 to 9,700-fold higher than those obtained by cultural methods. Similar observations were made in other aquatic ecosystems (6). The incongruence between direct counts and plate counts was presented by Staley and

Konopka as “the great plate count anomaly” (7), and was believed to result from one of two paradigms: 1) the cells observed by microscopy were dead or otherwise nonviable; or

2) the growth medium used to cultivate bacteria did not support their growth requirements. ZoBell succinctly summarized the topic in his 1946 monograph (1): “At best, direct counts give data which only supplement and aid in the interpretation of results obtained by cultural procedures.”

Departing from approaches that used high-nutrient growth media, Button developed a new methodology called “dilution culturing” for use in marine ecosystems

(8, 9). Button hypothesized that abundant marine bacteria are probably in a constant state of starvation due to the paucity of nutrients in their oligotrophic environments (8). This paradoxical phenomenon, known as ‘oligotrophy’ (reviewed in (10)), describes the metabolism of cells that grow optimally in nutrient-deplete conditions. Button reasoned that sterilized seawater (usually autoclaved) could supply sufficient native dissolved

5 organic matter to support the growth of oligotrophs. He calculated that his method of cultivation should favor numerically dominant organisms (which he hypothesized were oligotrophic) over cells that persist in natural waters at lower densities but tend to grow well on high-nutrient media. This method of dilution culturing was successfully applied by Button and colleagues to isolate the marine oligotrophs Sphingopyxis alaskensis

RB2256 and Cycloclasticus oligotrophus (9, 11).

Discovery and cultivation of SAR11

In the late 1970’s and 1980’s a trio of discoveries were made that would forever revolutionize microbial ecology – i) the development of more reliable and convenient

DNA sequencing (‘Sanger sequencing’ (12)); ii) the polymerase chain reaction (PCR)

(13); iii) and a microbial classification system based on rRNA gene sequences (14). With these new molecular tools and insights into microbial evolution, researchers began investigating microbial diversity in the environment not with a microscope or agar plates, but with PCR primers and sequence analysis tools (15). Stephen Giovannoni used this type of cultivation-independent approach to assay the microbial diversity of the Sargasso

Sea, an extremely oligotrophic ocean gyre in the North Atlantic Ocean (16). Giovannoni used PCR to amplify 16S ribosomal RNA genes from DNA extracted from Sargasso Sea bacterioplankton. The resulting PCR products were cloned, sequenced and used as a basis for molecular phylogeny. Molecular phylogeny revealed that many of the sequences belonged to a deeply-rooted group of uncultivated α-proteobacteria (16). A DNA sequence, representative of the abundant group of sequences was named ‘SAR11’ (short for clone number 11 identified from the SARgasso sea). Hybridization of SAR11

6 sequence probes to total rRNA collected from bacterioplankton obtained at different locations showed that the SAR11 rRNA was widely distributed and comprised a relatively large amount of total rRNA in the locations assayed (16). Continued study of

SAR11 distributions, using fluorescent in situ hybridization (FISH) of SAR11 probes to filtered bacterioplankton cells, showed that cells that hybridized to the SAR11 probes were extremely small, planktonic cells that can account for up to 50% of all cells near the ocean’s surface and up to 25% of all planktonic cells at depth (17). Today it is understood that the SAR11 cells are widely distributed in marine waters (17-19) and some freshwater ecosystems (20).

For nearly a decade after the discovery of its DNA in the environment, researchers struggled to cultivate SAR11 in the laboratory on traditional marine growth media (for example, ZoBell media 2216). Prior to the use of sterile seawater by Button, most growth media contained large amounts of enzymatically digested protein, fish extract, or other suitable constituents, in hopes to boost the number and diversity of cultivatable isolates. In contrast to this, Button hypothesized that unamended (or minimally amended) seawater could supply carbon (in the form of complex and undefined, but importantly, native, dissolved organic matter) and other nutrients in sufficient quantities to favor the growth of numerically abundant oligotrophs. Button’s dilution culture technique was improved by Giovannoni and colleagues by miniaturizing cultivation chambers and developing high-throughput arrays for filtering, enumerating and identifying cells at lower densities (21). The speed and efficiency of enumeration was later enhanced by implementation of a 96-well microtiter plate-based flow cytometer,

7 which allowed for the rapid and accurate determination of cell counts with a limit of detection of about 1000 cells ml-1 (22). This approach, one of high-throughput cultivation by dilution, would be used repeatedly to successfully cultivate cells that were routinely observed by the microscopists, but hypothesized to be dead or dormant by ZoBell and others.

In 2002, Giovannoni and colleagues isolated the first SAR11-clade bacterium from the environment (the northeast Pacific Ocean off of the coast of Oregon, USA) using dilution culturing into autoclaved natural seawater (23). The name proposed for the first SAR11 isolate was ‘Candidatus Pelagibacter ubique’ strain HTCC1062 (Ca. P. ubique, herein). Although cultivated, Ca. P. ubique was given (and still retains)

‘Candidatus’ status because growth experiments required for classification by the

Bacteriological Code (24) could not be performed for technical reasons. Analyses of Ca.

P. ubique’s growth kinetics and metabolism in natural seawater showed the compounding factors that had contributed to previous cultivation difficulties. First, growth was completely inhibited in the presence of 0.001% (w/v) protease peptone – a complex mixture of nutrients in most conventional marine media preparations (23). Additionally, the maximum cell yield was under 106 cells ml-1, well below densities for optical density measurements (a technique typically used for assessment of broth cultures) and colony formation (for visualization on agar plates). Low specific growth rates of 0.4 - 0.58 d-1 may have also contributed to cultivation difficulties. Thereafter, Ca. P. ubique was routinely propagated on a base medium of autoclaved natural seawater, usually obtained from the northeast Pacific Ocean.

The SAR11 genome sequence and metabolic reconstruction

Since its initial discovery, SAR11 cells have been identified worldwide and the clade is generally inferred to have a large role in the trophic interactions related to the global carbon cycle because of their abundance and ubiquity. However, identifying the specific metabolic nuances of SAR11 cells proved to be a challenge, owing to their inherent growth characteristics. The genome of Ca. P. ubique was sequenced in 2005 with the goal of identifying clues from metabolic reconstruction that could be used to unravel the growth requirements of these cells (25). The closed genome was just over 1.3

Mbp in length and is one of the smallest genomes of any free-living organism. Metabolic reconstruction showed that Ca. P. ubique likely encodes complete metabolic pathways for all 20 amino acids and all other cellular constituents (except for five B-vitamins, discussed in subsequent chapters) (25). Cells appeared to be non-glycolytic (although some, but not all genes encoding proteins used in the Entner-Doudoroff glycolytic pathway were identified), aerobic heterotrophs that made efficient use of dissolved organic matter in the ocean. Although metabolic reconstruction did not identify obvious reasons for the inability to grow Ca. P. ubique on defined medium, clever analysis and design of laboratory experiments were successful in uncovering nutrient requirements for increased cell yields on natural seawater.

Unraveling the metabolism of Ca. P. ubique

Using the reconstructed metabolism of Ca. P. ubique as a guide, Tripp and

Schwalbach conducted a series of growth experiments to test hypotheses generated from genomic information, and ultimately identified three classes of nutrients that were

9 required for growth: reduced organosulfur compounds, serine (or glycine), and low molecular weight organic acids. The first anomaly identified in the Ca. P. ubique genome was that the genes for assimilatory sulfate reduction (cysDNCHIJ) were absent (26). This was unusual because sulfate is not only extremely abundant in the ocean, but at the time, only one other aerobic heterotrophic bacterium (Idiomarina iloihiensis) lacked assimilatory sulfate reduction genes (26). Tripp hypothesized that because Ca. P. ubique was deficient in assimilatory sulfate reduction, an already reduced source of sulfur was required for growth. In the sea, methionine and dimethylsulfoniopropionate (DMSP) are relatively abundant reduced sulfur compounds (27), and may act as sources of reduced sulfur for Pelagibacter cells. Previously, Malmstrom showed that SAR11 cells compete effectively for DMSP and free amino acids, providing support for the hypothesis that

DMSP is metabolized by Ca. P. ubique (28). To test this hypothesis in vitro, Pelagibacter was grown in a natural seawater-based medium with amendments of inorganic nutrients and mixed carbon, but no reduced sulfur additions. Ca. P. ubique cell densities responded linearly to DMSP additions and the sulfur from 35S-DMSP was incorporated into cellular biomass, showing that it is required and used for growth (26). DMSP could be replaced by methionine, implying that either of these compounds can serve as a reduced sulfur source for Ca. P. ubique. Other SAR11-clade genomes lack the genes encoding proteins that are necessary for sulfate reduction (29), suggesting that the reduced sulfur requirement of Ca. P. ubique is common among wild SAR11-clade cells.

Additional metabolic reconstruction revealed an unusual gene complement and chromosomal organization of genes encoding for enzymes involved in the synthesis and

10 regulation of serine, glycine and threonine. Typically, serine is derived from the gluconeogenic precursor 3-phosphoglycerate (3PG) through the action of the SerA (D-3- phosphoglycerate dehydrogenase), SerB (3-phosphoserine phosphatase), and SerC (3- phosphoserine aminotransferase) proteins. Glycine is derived directly from serine by action of the GlyA (serine hydroxymethyl-transferase) protein that removes a methyl group from serine and transfers it to tetrahydrofolate (THF), forming 5,10- methylenetetrahydrofolate (CH3-THF), glycine and water. Ca. P. ubique lacks the serB and serC genes encoding proteins required for serine biosynthesis. However, a putative secondary route to glycine and serine synthesis was identified in Ca. P. ubique via a threonine aldolase. In Saccharomyces cerevisiae, the threonine aldolase cleaves threonine to form glycine (30). Therefore, because threonine synthesis is complete, a threonine aldolase is present and the GlyA reaction is reversible (that is, when there is sufficient

CH3-THF and glycine present, serine can be formed), the initial metabolic reconstruction of Ca. P. ubique’s glycine and serine synthesis appeared to be complete.

However, because the threonine aldolase is a low-specificity enzyme in E. coli

(31), Tripp hypothesized that Ca. P. ubique was ‘effectively auxotrophic’ for glycine and serine synthesis. To test this, Ca. P. ubique was grown in a natural seawater medium with amendments of inorganic nutrients, mixed carbon and a source of reduced sulfur. Ca. P. ubique cell yields responded linearly to additions of glycine, suggesting that glycine is a required nutrient (32). Serine, but not threonine, replaced glycine in this natural seawater medium, providing evidence that the threonine aldolase is not active in Ca. P. ubique, or its activity was too low to detect under the experimental conditions (32).

In addition to effective serine/glycine auxotrophy, unusual features pertaining to the regulation of glycine metabolism were identified. Similar to other organisms, Ca. P. ubique contains a glycine-activated riboswitch preceding the coding sequence of the glycine cleavage system T-protein (gcvT). The GcvT protein is part of the glycine cleavage complex that catalyzes the cleavage of glycine to form ammonia and a methyl group that is transferred to the methyl pool (via THF), where it can be oxidized for energy (reviewed in (33)). Riboswitches are mRNA leader sequences that fold to form a

2° structure capable of binding specific effector molecules that affect transcription and/or translation of the downstream mRNA (reviewed in (34)). When ample glycine is present, the gcvT transcript containing the riboswitch binds glycine, allowing for translation of the

GcvT protein. Conversely, when the glycine concentration is low, the switch is configured to prevent translation of gcvT. The system acts to regulate the degradation of excess glycine and to use the products of glycine degradation as a source of ammonia and energy.

A second glycine riboswitch was predicted in the Ca. P. ubique genome in an uncommon position preceding the coding sequence of the glcB gene, encoding the malate synthase-G (32). In E. coli, glcB is in an operon with genes encoding the glycolate oxidase (glcDEFGB) which, when expressed, allows for the oxidation of glycolate to glyoxylate (via GlcDEF) (35) and the subsequent condensation of glyoxylate with acetyl-

CoA (via GlcB) to form the tricarboxylic acid (TCA) cycle intermediate malate (36).

Through this mechanism, carbon from glycolate can be shunted into the TCA cycle where it is converted to biomass or burned for energy. However, in the Ca. P. ubique

12 genome, glcB is not co-localized with the glcDEF genes (32). Instead, it appeared that

GlcB was involved in completing the glyoxylate bypass in Ca. P. ubique (along with the isocitrate lyase) and that the glyoxylate bypass of the TCA cycle was regulated by glycine (32). Tripp theorized that the unusual arrangement of riboswitches – one preceding glcB and one preceding gcvT - aided Ca. P. ubique cells in balancing its energy and carbon requirements (32).

The primary reconstruction of Ca. P. ubique’s carbohydrate metabolism identified unusual features for an aerobic heterotroph. The hallmark pathway of glycolysis – the

Embden–Meyerhof–Parnas (EMP) pathway of glycolysis - was missing, and instead, an unusual variant of the Entner–Doudoroff (ED) glycolytic pathway was proposed to be present. Schwalbach explored the unusual nature of carbohydrate utilization with two

SAR11 isolates – the Pacific Ocean isolate Ca. P. ubique and Pelagibacter sp. str.

HTCC7211, isolated from the oligotrophic Sargasso Sea in the Atlantic Ocean. Genomic comparison of the two strains showed that Pelagibacter sp. str. HTCC7211 lacked the genes predicted to be involved in the ED-variant glycolytic pathway. Schwalbach hypothesized that because of these genomic differences; Ca. P. ubique but not

Pelagibacter sp. str. HTCC7211 would metabolize glucose. Radiotracer experiments supported this by showing that only Ca. P. ubique assimilated and oxidized glucose to

CO2 (37). Through elegant use of DNA microarrays and growth experiments, Schwalbach showed that when grown on glucose, the predicted ED-variant genes were significantly upregulated in Ca. P. ubique, relative to cultures grown with pyruvate as a primary carbon source. Perhaps more importantly, Schwalbach noted that neither isolate

13 metabolized a range of hexose or pentose sugars (except for glucose in Ca. P. ubique) that are generally thought of as ‘high-value’ carbon sources in nature. Instead, both Ca. P. ubique and Pelagibacter sp. str. HTCC7211 utilized a range of low molecular weight organic acids, including pyruvate, lactate, taurine, acetate and oxaloacetate (37). Taken together, these findings showed that low molecular weight organic acids are important carbon sources for SAR11, and that not all SAR11 isolates are able to utilize glucose.

Further support for these findings comes from comparative genomics studies of SAR11 genomes, where poor conservation of genes used for carbohydrate metabolism was observed across seven SAR11 genomes, suggesting that variation in carbohydrate metabolism is not unusual within the SAR11 clade (29).

Genome streamlining

The studies by Tripp and Schwalbach identified three distinct classes of compounds that Ca. P. ubique depends on for cellular metabolism: reduced sulfur compounds, glycine (or serine), and low molecular weight organic acids (and sometimes glucose). Only when constituents from all three classes were present did they observe increases in maximal cell densities (10-fold above those observed in unamended natural seawater) (26, 32, 37). The apparent co-requirement for such distinct compounds is unusual compared to the physiology of marine bacteria cultivated on standard growth media. The uncommon co-requirements have been described within the context of adaptive selection for gene loss in the Pelagibacter lineage by a process termed ‘genome streamlining.’ Reduction of genome size is not uncommon within small bacterial populations that experience a high degree of genetic drift such as obligate symbionts

(reviewed in (38)). In cases of obligate symbiosis, the effect of relaxed selection to maintain gene function leads to the accumulation of mutations in non-essential genes.

Because the effective population of symbionts is small, these mutations are more likely to become fixed within the population by genetic drift. Over time, the combination of mutation, deletion and genetic drift leads to genome degradation and a decrease in size.

In contrast to the small population sizes of obligate symbionts, the Pelagibacter population is massive, and as such, genetic drift is not expected to have had such a distinct effect. Instead, it is proposed that selection is responsible for the reduction of

Pelagibacter genomes. This leads to a logical question: what trait is selection acting upon to keep Pelagibacter genomes small? One hypothesis is that by reducing the size of the genome, Pelagibacter saves on the cost of maintaining DNA and expressing genes. That is, the cost of maintaining certain genes (or the pathways their products are involved in) has become too metabolically expensive from either an energetic point of view or from the availability of the raw materials (for example, core nutrients such as N, P and C) (25,

39, 40).

Morris has put forward an explanation regarding a putative selective pressure that would result in adaptive gene loss and termed it the ‘black queen hypothesis’ (41). The hypothesis, named by analogy to the queen of spades in the card game of Hearts, states that within a given mixed-member community, a metabolically expensive function (e.g. - sulfate reduction) can be lost from a portion of the community (Pelagibacter cells, for example) if the remaining members of the community can compensate for that lost function (provide a reduced sulfur source for Pelagibacter). In the context of the

15 parenthetical example above, phytoplankton reduce sulfur and make methionine and

DMSP as compatible solutes. Presumably the amount of these compounds that leak from cells or are released upon cell death is sufficient to support the SAR11 populations observed in nature. Thus, if maintenance of the sulfate reduction genes (or running the sulfate reduction pathway) was metabolically expensive to SAR11, loss of the genes where environmental reduced sulfur is ample may have conferred a competitive advantage.

Growth on a defined medium

Although headway was made by the identification of the three major ‘food groups’ that Ca. P. ubique needs to grow, further advances were necessary in order to develop Ca. P. ubique into a system for understanding marine oligotrophy and the trophic role of SAR11 cells in the environment. Though an abundance of data has been collected for numerous microbes on ‘complex’ (multiple constituents) and ‘undefined’ (unknown constituents) growth media, the compositions of those media were relatively stable, and the types of microbes grown in such media were robust enough to withstand the small amount of nutritional variability that may have been encompassed between different media preparations. This has not been the case for Ca. P. ubique and other SAR11 isolates. Immediately after cultivation on natural seawater, ‘batch effects’ were observed with SAR11 cells grown on seawater collected at different times whereby different batches of seawater supported varying final cell yields (23). The specific reasons for the batch effects are unknown, but one presumes that it is the result of native nutrients (or inhibitory factors) being present in one water sample but absent (or found in different

16 amounts) in another. The observation that different batches of seawater have varying effects on growth of marine microbes had been made previously (42). The lack of consistency among batches of seawater early on during Ca. P. ubique growth studies made the reproduction of experiments difficult because each batch of seawater was finite and had its own ‘character’ that tended to supersede the effects of nutrient addition experiments. Moreover, it was unknown if natural seawater provided a required, but unidentified, trace nutrient for Ca. P. ubique growth. Without a defined medium, the requirement for an unknown nutrient could not be ruled out.

One of the archetypal methodologies for understanding microbial physiology is to exclude a nutrient of interest and study the growth effect of removal on a particular organism. Hence, another problem with utilizing natural seawater, beyond that its composition is unknown, is that specific nutrients cannot be excluded from seawater to study the effect on cellular metabolism. Currently, removal of a single compound from seawater (for nutrient exclusion experiments), without altering the rest of the makeup, is not possible. And finally, the issue of seawater collection is extremely problematic. It is rather impractical to collect seawater in sufficient volumes to do routine lab work unless the sea is directly accessible.

Minimal defined media have been instrumental tools at the disposal of microbiologists studying a variety of systems, including enterobacteria and Gram positive bacilli (43, 44). To understand Ca. P. ubique’s role in environmental nutrient cycling and to determine a minimal set of requirements for Ca. P. ubique growth, experiments were designed to identify constituents for growth on a defined minimal medium (see Chapter

2). The defined medium was used to explore the trace requirements for five B-vitamins

(B1 (Chapter 3), B5, B6, B7 and B12 (Chapter 5)) for which complete canonical biosynthetic pathways were not identified in the Ca. P. ubique genome. By doing such experiments, we have determined a minimal set of carbon and sulfur sources that Ca. P. ubique requires for growth and show that at least one other Pelagibacter isolate

(Pelagibacter sp. str. HTCC7211) utilizes the same nutrients. We have uncovered a vitamin B7 (biotin) requirement and an unusual requirement for the vitamin B1 (thiamine) precursor 4-amino-5-hydroxymethyl-2-methylpyrimidine (HMP). We did not find physiological support for a vitamin B5 (pantothenate), B6 (pyridoxine) or B12 (cobalamin) requirement, but instead offer alternative evidence for why Ca. P. ubique appears to be prototrophic (or in the case of pantothenate, bradytrophic) for these vitamins.

In addition to studying the nutrient requirements of Pelagibacter cells, specific adaptations that Ca. P. ubique employs to effectively compete for trace levels of Pi are described (Chapter 4). These adaptations include the use of alternate phosphorus compounds and the probable rearrangement of lipid polar head groups to reduce phospholipid (and hence cellular Pi) content. Moreover, we verify that the phosphonate degradation pathway (C-P lyase) in Pelagibacter sp. str. HTCC7211 confers the ability to utilize phosphonates and phosphite, and show that when methylphosphonate (Mpn) is used for phosphorus, methane is stoichiometrically released. The release of methane from the aerobic sea is well described and we propose that in certain regions of the ocean,

Pelagibacter sp. str. HTCC7211-like SAR11 cells may be responsible for the production of much of the aerobically produced methane.

These studies represent important first steps in the development of Ca. P. ubique

(and other SAR11 isolates) into a model system for understanding the role of the SAR11 clade as an agent of biogeochemical change in the environment. Likewise, because Ca. P. ubique has an unusual physiology, encoded by divergent metabolic pathways with unfamiliar genetic arrangements and regulatory schemes, it provides researchers with an opportunity to study the underpinnings of oligotrophy in this abundant clade. Oligotrophy is characterized, somewhat loosely, as the paradoxical requirement for low concentrations of nutrients for optimal growth. We show herein that Ca. P. ubique clearly conforms to this conceptual definition, but that it can be manipulated in vitro to grow at high nutrient concentrations. Model systems are essential to push our knowledge forward about new and unusual types of metabolism and physiology, such as oligotrophy. Today, the same cultures Winogradsky and Beijerinck isolated and domesticated for laboratory use have been developed into model systems from which environmental, biochemical, physiological and geochemical discoveries are rooted. Finally, the strategies described here are meant to outline how to use DNA sequence information, from both genomes and from the environment, to construct testable hypotheses about microbial metabolism and understand how microbes affect nutrient cycles in the ocean.

Chapter 2 Nutrient Requirements for the Growth of the Extreme Oligotroph ‘Candidatus Pelagibacter ubique’ HTCC1062 on a Defined Medium

Paul Carini1, Laura Steindler1,2, Sara Beszteri1,3 and Stephen J. Giovannoni1

1Department of Microbiology Oregon State University Corvallis, OR 97331

2Current address: Department of Marine Biology, The Leon H Charney School of Marine Science, University of Haifa, Haifa, Israel

3Current address: Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany

Published: The ISME journal. doi:10.1038/ismej.2012.122

Abstract

Chemoheterotrophic marine bacteria of the SAR11 clade are Earth’s most abundant organisms. Following the first cultivation of a SAR11 bacterium, ‘Candidatus

Pelagibacter ubique’ strain HTCC1062 (Ca. P. ubique) in 2002, unusual nutritional requirements were identified for reduced sulfur compounds and glycine or serine. These requirements were linked to genome streamlining resulting from selection for efficient resource utilization in nutrient-limited ocean habitats. Here we report the first successful cultivation of Ca. P. ubique on a defined artificial seawater medium (AMS1), and an additional requirement for pyruvate or pyruvate precursors. Optimal growth was observed with the collective addition of inorganic macro- and micronutrients, vitamins, methionine, glycine and pyruvate. Methionine served as the sole sulfur source but methionine and glycine were not sufficient to support growth. Optimal cell yields were obtained when the stoichiometry between glycine and pyruvate was 1:4, and incomplete cell division was observed in cultures starved for pyruvate. Glucose and oxaloacetate could fully replace pyruvate, but not acetate, taurine or a variety of tricarboxylic acid cycle intermediates. Moreover, both glycine betaine and serine could substitute for glycine. Interestingly, glycolate partially restored growth in the absence of glycine. We propose that this is the result of the use of glycolate, a product of phytoplankton metabolism, as both a carbon source for respiration and as a precursor to glycine. These findings are important because they provide support for the hypothesis that some microorganisms are challenging to cultivate because of unusual nutrient requirements caused by streamlining selection and gene loss. Our findings also illustrate unusual metabolic

21 rearrangements that adapt these cells to extreme oligotrophy, and underscore the challenge of reconstructing metabolism from genome sequences in organisms that have non-canonical metabolic pathways.

Introduction

Laboratory studies of ecologically important organisms are important for understanding the principles and mechanisms that govern ecosystem behavior (45). For example, several studies have reported high connectedness in marine plankton communities (46, 47). ‘Connectance’ is a general measure of the degree to which populations within a community display correlated behavior, and, in terrestrial ecosystems, has been linked to ecosystem stability (48). The SAR11 clade of α- proteobacteria are the most abundant heterotrophs in marine euphotic zones worldwide

(17) and have been shown to contribute significantly to the overall connectedness of marine microbial plankton communities (47). Specific mechanisms that might explain

SAR11 connectedness are slowly emerging from studies of genomes and the metabolism of cells in culture (26, 32, 37, 49).

‘Candidatus Pelagibacter ubique’ HTCC1062 (Ca. P. ubique), a member of the

SAR11 clade, has a small genome and displays characteristic signatures of streamlining selection (25). According to the genome streamlining theory, when effective population sizes are large, extreme selection for the efficient use of resources in nutrient-poor environments can result in genome reduction (25, 50, 51). The metabolic consequences of genome reduction in Ca. P. ubique have been the subject of several reports (26, 32,

37).

Metabolic reconstruction from the Ca. P. ubique genome highlighted the conspicuous absence of genes common to aerobic chemoorganoheterotrophs (Figure 2-1).

Subsequent batch culture experiments, using a natural seawater medium, showed that Ca.

P. ubique required an unusual combination of nutrients for growth (26, 32, 37).

Specifically, Ca. P. ubique lacks genes necessary for assimilatory sulfate reduction, and as a result, requires reduced sulfur compounds such as methionine or 3- dimethylsulphoniopropionate for growth (26). Similarly, the unusual absence of common genes for serine and glycine biosynthesis resulted in a conditional requirement for either of these amino acids (32). Concomitantly, an uncommon arrangement of two glycine- activated riboswitches was reported; one located upstream of the gene encoding for the glycine cleavage system T-protein (gcvT) and the other upstream of the malate synthase gene (glcB) (32). The first of these is a common riboswitch involved in glycine cleavage

+ to CO2 and NH4 , but the second is a rare riboswitch configuration that was implicated in glyoxylate metabolism (32). Although Ca. P. ubique is missing the complete Embden–

Meyerhof–Parnas glycolytic pathway (genes encoding pyruvate kinase and phosphofructokinase are absent, Figure 2-1), a putative operon encoding genes involved in a predicted variant of the Entner–Doudoroff glycolytic pathway is present, as is a complete gluconeogenic pathway (Figure 2-1; see (37)). Experiments showed that low molecular weight organic acids were important carbon sources for multiple SAR11 isolates and that Ca. P. ubique, but not all SAR11 isolates, oxidized glucose (37).

Moreover, genes coding for proteins involved in carbohydrate metabolism are not consistently conserved across the SAR11 clade, suggesting that the metabolism of sugars

23 and other carbohydrates may be involved in niche partitioning or ecotype variation (29).

Previously, Ca. P. ubique was grown exclusively in natural seawater-based media.

However, seawater contains an undefined mixture of naturally occurring organic compounds that vary in composition and quantity between seasons, locations and depths.

The consequences of this nutrient variability on Ca. P. ubique’s physiology was initially observed in the form of ‘batch effects’, in which seawater collected at different times yielded different cell densities or growth rates independent of nutrient additions (23). The three main limitations of conducting experiments with Ca. P. ubique cells on natural seawater-based media are: (i) the precise composition and concentrations of organic matter in seawater is unknown and variable; (ii) the finite supply of any single seawater batch limits comparisons between experiments conducted with different batches; and (iii) native concentrations of nutrients (for example, nitrogen, sulfur and carbon) cannot be excluded.

We developed a defined medium to use as a tool for studying the nutrient requirements and metabolism of Ca. P. ubique. Ultimately, without a defined medium, it was impossible to rule out that unknown compounds in natural organic matter were essential for growth of Ca. P. ubique. A simple approach to solve this problem is the addition of a diverse mixture of compounds, under the assumption that the organism will use only what it needs. This approach is commonly implemented by adding mixtures of nutrients such as yeast extract or digests of protein to growth media. However, complex, high-nutrient mixtures inhibit the growth of many oligotrophic chemoheterotrophs, making this strategy impractical (23, 26, 32). Building on previous metabolic models for

Ca. P. ubique, we hypothesized that the minimum nutrient requirements for growth on a defined medium would include pyruvate in addition to a source of reduced sulfur and glycine. We report here the propagation of Ca. P. ubique to high cell density on a defined artificial seawater medium containing inorganic micro- and macronutrients, vitamins, methionine (for reduced sulfur), glycine and pyruvate. Pyruvate, or its precursors, was identified as the last essential macronutrient needed by Ca. P. ubique for growth. We also demonstrate that several environmentally relevant compounds can be used by Ca. P. ubique in place of pyruvate and glycine, and describe unusual rearrangements of central metabolic pathways that confer these properties.

Methods

Organism source

Ca. P. ubique HTCC1062 was originally isolated from the northeast Pacific

Ocean as described elsewhere (23). Frozen glycerol stocks from the original isolation of

Ca. P. ubique were used as the source inoculum for all experiments.

Media preparation

Artificial seawater medium AMS1 (Table 2-1) was derived from the artificial seawater medium AMP1 (52). Organic buffers (for example, HEPES and EDTA) were excluded to avoid potential toxic effects (53) and the possibility that they might be used as substrates for growth (54). After autoclaving, AMS1 was sparged with 0.1 µm-filtered

CO2 for 5 h followed by sparging with air for 10 h to establish a bicarbonate-based buffer system (21). All vitamins and organics were added after autoclaving and sparging. The

25 pH of the resulting AMS1 typically ranged from 7.5 to 7.7.

Cultivation details

All cultures were grown in acid-washed and autoclaved polycarbonate flasks at

20°C with shaking at 60 r.p.m. under a 12/12-h light (140-180 µmol photons m-2 s-1)/dark cycle.

Measurement of growth

Cells were stained with SYBR Green I (Molecular Probes, Inc., Eugene, OR,

USA) and counted with a Guava Technologies flow cytometer (Millipore, Billerica, MA,

USA) at 48- to 72-h intervals as described elsewhere (22, 26).

Acclimation to growth on AMS1

Natural seawater collected from the Newport Hydroline station NH-05 (latitude:

44.651, longitude: -124.181) from a depth of 10 m in June 2008 was amended with glycine (1 µM), methionine (1 µM), pyruvate (50 µM), FeCl3 (1 µM) and vitamins and inoculated with Ca. P. ubique from glycerol stocks. When exponentially growing cells in this amended natural seawater medium exceeded a density of 2.0 × 106 cells ml-1, they were diluted (1:100) with fresh AMS1 supplemented with glycine (1 µM) (as a glycine/serine source), methionine (as a sulfur source) (1µM), pyruvate (as a carbon source) (50 µM), FeCl3 (1 µM) and vitamins. We observed no lag phase in cells that were transferred into amended AMS1 (Supplementary Figure A1-1 in appendix 1). All cultures described herein are derived from this lineage and have been maintained exclusively on

AMS1 for >15 consecutive batch culture transfers (approximately 150 generations).

Pyruvate substitution experiments

Potential pyruvate precursors (Table 2-2) were selected because they were either

(i) ‘common’ sole carbon sources for chemoorganoheterotrophs or (ii) present in seawater as products of phytoplankton metabolism. Each substitute was tested at a concentration of 50 µM in AMS1 amended with 25 µM glycine, 10 µM methionine, 1µM

FeCl3 and vitamins. The positive control was amended with 50 µM pyruvate and the same concentrations of glycine, methionine, FeCl3 and vitamins. The negative control contained no pyruvate, but was otherwise identical.

Glycine substitution experiments

Potential glycine precursors (Table 2-3) were selected because they were either (i) metabolic precursors of glycine in other organisms or (ii) predicted to be a precursor based on metabolic reconstruction in Ca. P. ubique. Each substitute was tested at a concentration of 25 µM in AMS1 with pyruvate (50 µM), methionine (10 µM), FeCl3

(1µM) and vitamins. The positive control was amended with 25 µM glycine and the negative control contained no glycine.

Cell division experiments

Growth media consisted of AMS1 amended with glycine (1 µM), methionine

(1 µM), FeCl3 (1 µM), vitamins and either 0.5 µM pyruvate (deplete conditions) or 50

µM pyruvate (replete conditions). SYBR Green I-stained cultures that exhibited relative

DNA fluorescence values of 300–325 and 475–500, were independently filtered on to 0.2

µm black polycarbonate filters, and imaged using a Leica DMRB epifluorescence

27 microscope (Wetzlar, Germany) equipped with filter sets appropriate for SYBR Green I

(excitation: 450–490 nm; emission: 580 nm). Images were captured with a Hamamatsu

ORCA-ER CCD digital camera (Hamamatsu City, Japan) and Scanalytics IPLab v3.5.5 scientific imaging software (Fairfax, VA, USA).

Chemicals

All inorganic salts were obtained from Sigma-Aldrich Co. (St Louis, MO, USA) and were of the highest available quality (typically labeled ‘ultrapure’). All other compounds were obtained from Sigma-Aldrich Co. or other commercial vendors and were of reagent grade quality.

Results

When grown on AMS1 with methionine, glycine and pyruvate, Ca. P. ubique’s maximum specific growth rate was 0.41 ± 0.01 day-1 (mean ± s.d., n=3), and batch cultures reached maximum cell densities of 9.18 ± 0.02 × 107 cells ml-1 (mean ± s.d., n=3) (Figure 2-2). Additions of methionine or carbon (as pyruvate and glycine) alone did not result in increased cell densities beyond those of the negative control (Figure 2-2).

Cell densities responded linearly to both pyruvate (R2=0.998) and glycine (R2=0.995) additions in AMS1 when other nutrients were in excess (Figure 2-3). The maximum cell density increased by 2.6 × 106 cells ml-1 µM-1 pyruvate (Figure 2-3a) and

1.0 ×107 cells ml-1 µM-1 glycine (Figure 2-3b).

Ca. P. ubique utilized oxaloacetate and glucose in place of pyruvate on AMS1

(Table 2-2). As observed previously (37), Ca. P. ubique’s specific growth rate was slower

28 with glucose as a sole pyruvate source. Addition of taurine or lactate resulted in cell densities in excess of fivefold greater than the negative control, but did not achieve the cell densities of the pyruvate treatment. Notably, in the absence of pyruvate, additions of alanine or glycine did not improve growth yield.

Glycine betaine and serine were able to fully replace glycine in AMS1 and glycolate partially substituted for glycine (Table 2-3). Ca. P. ubique grew slower but to slightly higher cell densities when glycine betaine was the sole glycine source. Glycolate led to cell density increases fourfold greater than those of the negative control. The addition of pyruvate, without glycine, did not result in higher cell densities.

While developing the AMS1 medium, unusual cell division patterns were observed under pyruvate-deplete conditions. When SYBR Green was used to stain DNA in early stationary-phase cells, fluorescence from pyruvate-limited cells was about two- fold higher than from cells raised in a pyruvate-replete medium (Figures 2-4a and b).

Microscopic images from these cultures showed that the increase in DNA fluorescence was caused by elongated cells containing two nucleoids (‘doublets’) (Figures 2-4c and d).

We observed this unusual cell division pattern when cells entered pyruvate-limited stationary-phase across a range of pyruvate:glycine ratios (Supplementary Figure A1-2 in appendix 1). This phenomenon was previously observed when Ca. P. ubique cells were grown in a natural seawater medium without added pyruvate (unpublished data).

Experiments conducted in natural seawater found alanine induced the division of cell doublets (unpublished data). The effect of different alanine concentrations on Ca. P. ubique cell morphology in stationary-phase cultures grown in AMS1 was tested using

29 relative DNA fluorescence as a proxy for the formation of cell doublets (Supplementary

Figure A1-3 in appendix 1). DNA fluorescence profiles and cell counts suggested that alanine induced cell division in pyruvate-limited cultures.

Discussion of results

Similar to the goals of Neidhardt et al. (43), who developed a defined medium for the growth of enterobacteria, one of our objectives was to prepare a ‘physiologically optimal and experimentally useful’ medium in which Ca. P. ubique, and other SAR11 isolates, could be propagated reproducibly. This artificial medium for the growth of Ca.

P. ubique contains only methionine, glycine, pyruvate, vitamins and inorganic salts. The specific growth rates observed on AMS1 (0.41 ± 0.01 day-1) were comparable to those previously published for Ca. P. ubique grown in seawater batch cultures (23, 26, 37) and to those from microcosm experiments with natural assemblages of plankton (55, 56).

Results observed with cells growing in AMS1 support the previous conclusions that Ca. P. ubique requires a reduced sulfur source, glycine and an organic acid for growth. Our results also confirm that methionine meets Ca. P. ubique’s requirement for exogenous reduced sulfur compounds (26) but cannot substitute for glycine or pyruvate as a sole carbon source. The linear responses to both pyruvate and glycine additions, when other constituents were in excess, indicated that both pyruvate and glycine were necessary for optimal growth and used in the molar ratio of 4.0:1 (pyruvate:glycine). In addition, we showed that alanine is required for septation, the final step in cell division, and that in the absence of pyruvate, Ca. P. ubique cells were not able to synthesize alanine by other routes, thus implicating pyruvate as a primary precursor for alanine

30 biosynthesis. In Escherichia coli, the rate of cell wall biosynthesis increases during septation in order to accommodate the synthesis of the new daughter cell poles (57).

Because alanine is a major component of cell walls in Gram-negative bacteria (58), we propose that in pyruvate-limiting conditions, Ca. P. ubique is unable to synthesize sufficient alanine to complete septation, resulting in the formation of cell doublets.

Our observations show that pyruvate or its precursors are required to synthesize alanine, but that alanine cannot replace pyruvate (Table 2-2). In Ca. P. ubique, L-alanine is putatively formed by an alanine dehydrogenase (encoded by ald, SAR11_0809—

Figure 2-1). Ald catalyzes the formation of alanine from pyruvate and ammonia via the oxidation of reduced nicotinamide adenine dinucleotide, and has been shown to be involved in both the catabolism and anabolism of alanine in Pseudomonas, Rhodobacter and Sphingopyxis species (59-61). Although Ald catalyzes a reversible reaction in some organisms, our data suggest that in Ca. P. ubique, Ald does not catalyze the formation of pyruvate from alanine under the conditions tested.

The observation that both glucose and oxaloacetate could replace pyruvate suggests that pyruvate is a metabolic intermediate formed during the catabolism of these compounds. Both compounds were previously identified as carbon sources for Ca. P. ubique in natural seawater (37). The use of oxaloacetate in place of pyruvate is consistent with the presence of a predicted malic enzyme gene (maeB, SAR11_0375) that may be involved in the decarboxylation of oxaloacetate to form pyruvate. Our results are also consistent with Schwalbach’s proposed glycolytic pathway in Ca. P. ubique that predicted pyruvate was an end product of glucose metabolism (37). Previously, in natural

31 seawater-based media, taurine, acetate and lactate additions resulted in cell density increases similar to those observed with pyruvate additions (37). However, in our experiments on artificial seawater media, maximum cell densities decreased in this order: pyruvate >> taurine > lactate >> acetate (Table 2-2). Taurine is putatively metabolized to acetyl-CoA in Ca. P. ubique, suggesting that taurine catabolism supplies two-carbon units for biosynthesis. In our experiments, acetate, a direct precursor to acetyl-CoA, could not replace pyruvate (Table 2-2). This contradicts the prediction of a metabolic pathway for acetate assimilation as a sole carbon source via the glyoxylate bypass in Ca. P. ubique

(Figure 2-1). We postulate that the glyoxylate bypass has instead been recruited to function in glycine metabolism, as described below.

We show for the first time that glycine betaine, in addition to serine, can meet Ca.

P. ubique’s glycine requirement. Previously, Tripp et al. (32) identified serine as the only compound able to replace glycine in natural seawater-based batch cultures of Ca. P. ubique. Consistent with this finding, we found that serine was an effective replacement for glycine on AMS1 (Table 2-3). Glycine betaine is an important osmolyte that can be degraded by marine bacteria (49, 62, 63). Ca. P. ubique’s genome encodes a full suite of genes predicted to be involved in the acquisition and stepwise oxidation of glycine betaine to form glycine (32, 49). Experiments also showed that Ca. P. ubique can oxidize methyl groups derived from glycine betaine to CO2 (49). Although glycine betaine fully replaced glycine, we observed a reduction in growth rate when glycine betaine was supplied as a sole glycine source. Tripp et al. (32, 49) previously reported that large amounts of glycine betaine had unpredictable or deleterious effects on the growth of Ca.

P. ubique in natural seawater. We propose that growth rate reduction may be in part due to toxicity resulting from the production of formaldehyde (by the predicted dimethylglycine dehydrogenase, EC:1.5.99.2, SAR11_1253) and hydrogen peroxide (by the predicted sarcosine oxidase, EC:1.5.3.1; for genes, see (49)) during glycine betaine catabolism.

We also report the new finding that Ca. P. ubique partially utilized glycolate to meet its glycine requirement. Glycolate is commonly formed by phytoplankton as a result of photorespiration (64-68). In E. coli, carbon from glycolate is assimilated into biomass after it is oxidized to glyoxylate by the glycolate oxidase (glcDEF) (35, 36). Glyoxylate is then condensed with acetyl-CoA by the malate synthase (GlcB) to form the tricarboxylic acid cycle intermediate malate (35). In Ca. P. ubique, genes encoding the glycolate oxidase are found in a putative operon with a pyridoxal-phosphate-dependent aminotransferase, annotated as an aspartate aminotransferase (aspC), but separate from the malate synthase gene (Figure 2-5). Previously, Tripp et al. (32) had proposed that this aminotransferase produces glycine from glyoxylate, but, working with natural seawater media, were unable to demonstrate replacement of glycine by glycolate to validate this hypothesis.

Results presented here help resolve an enigmatic arrangement of glycine-activated riboswitches associated with genes of the glyoxylate cycle and glycine catabolism. Ca. P. ubique has genes for the proteins required to channel glycolate into the tricarboxylic acid cycle where it can be assimilated into biomass or oxidized (glcDEF and glcB; (32)). In

Ca. P. ubique, a glycine-activated riboswitch is located in a very unusual position

33 upstream of the glcB coding sequence, and a second glycine-activated riboswitch is located in a common arrangement, upstream of gcvT, where it functions to regulate

+ glycine cleavage to CO2 and NH4 . It is now apparent that this unusual configuration of two riboswitches results in glycine concentrations regulating the fate of glyoxylate, as illustrated in Figure 2-5. We postulate that when intracellular glycine concentrations fall too low, the switch on glcB closes, shunting glycolate through glyoxylate to form glycine.

When intracellular glycine concentrations are ample, the glyoxylate bypass opens to channel glycolate-derived carbon into the tricarboxylic acid cycle. We attribute the observation that glycolate did not support as high a cell yield as glycine (Table 2-3) to its dual role in these pathways. This model also helps explain why Ca. P. ubique does not respond vigorously to acetate addition; the glyoxylate bypass, which in most cells is used for acetate assimilation, has been recruited to functions that are adaptive to the ocean environment.

One of the mysteries yet to be fully explained is the growth of Ca. P. ubique cells to 106 cells ml-1 in the absence any additional carbon compounds except vitamins. We determined the maximum amount of methionine, glycine and pyruvate carried over with the source inoculum to be 500, 500 and 156 pM, respectively. Based on the calculated per-cell glycine and pyruvate requirement (regression lines in Figure 2-3), and the previously published sulfur requirement (26), we conclude that the carryover of these nutrients does not sufficiently explain the growth of the negative controls observed in

Figures 2-2 and 2-3. One alternative explanation for this growth is that Ca. P. ubique is able to utilize the carbon and sulfur originating from one or more of the vitamins, vitamin

34 degradation products or traces of contaminating carbon in the vitamin stocks. We tested this and found that Ca. P. ubique responded in a dose-dependent manner to increasing amounts of freshly prepared vitamins (Supplementary Figure A1-4 and discussion in appendix 1). It is not clear whether Ca. P. ubique can utilize the vitamins themselves for growth or if the vitamins enable more efficient use of other traces of contaminating carbon in AMS1. Regardless of the source of the contaminating nutrients, we show that maximum cell densities are two orders of magnitude greater when methionine, glycine and pyruvate are added to the medium.

Metabolic reconstruction has shown Ca. P. ubique’s unusual requirement for a balanced supply of organic matter is a consequence of streamlining selection for minimal genome size and metabolic simplicity (Figures 2-1 and 2-5; (26, 32, 37)). We speculate that the unusual arrangement of central metabolism in Ca. P. ubique is adaptive not only because it is small and simple, and therefore requires fewer nutrients to replicate, but also because it is suited to planktonic environments where the flux of labile dissolved organic matter is low but continuous most of the time. In addition to the potential supply of pyruvate from glycolysis in some strains, pyruvate and its precursors oxaloacetate and glyoxylate are common metabolic intermediates in other organisms and also are formed by the photooxidation of dissolved organic matter (69-73). Taurine, a compound produced by both phytoplankton and animals (74), also substituted for pyruvate, but at greatly reduced efficiency. Previously it was shown that the Ca. P. ubique glycine requirement could be met by glycine or serine, which are found in seawater at nanomolar concentrations (32). We also found that the common osmolyte glycine betaine and the

35 photorespiration product glycolate could serve as precursors for glycine biosynthesis.

Glycolate is produced by oxygenic phototrophs when they become carbon limited - including both eukaryotic phytoplankton (64, 66, 67) and marine cyanobacteria (68) - the dominant marine phototrophs in temperate and tropical oceans. Tripp showed previously that either methionine or 3-dimethylsulphoniopropionate, a phytoplankton osmolyte, could meet the Ca. P. ubique requirement for organosulfur (26, 75). Therefore, the results of this study suggest that Ca. P. ubique has evolved to efficiently use a combination of ubiquitous, low molecular weight metabolites produced by phytoplankton, or resulting from the photooxidation of dissolved organic carbon, which may partially explain its high abundance in the euphotic zone.

A number of definitions have been proposed for the term ‘oligotroph’ based on optimal and inhibitory nutrient concentrations (reviewed in (10)). These definitions do not easily fit the metabolic behaviors being observed in the experiments presented here.

The observation that alanine, derived from pyruvate or its precursors, is required for cell division (Figure 2-4; Supplementary Figures A1-2 and A1-3 in appendix 1), but that large amounts of alanine adversely affect growth (Table 2-2), is an example of a deleterious imbalance caused by the presentation of a metabolic substrate to cells at either unnatural concentrations or in unnatural combinations. Unusual nutrient requirements, together with cryptic patterns of nutrient substitution and inhibition, present the experimentalist with a complex problem. This problem can be solved by the stepwise process of metabolic reconstruction coupled with experimentation if a defined medium is available.

Although this approach is time consuming (especially with slow-growing cells, such as

Ca. P. ubique), it has been successfully applied in other systems (76). The capacity of Ca.

P. ubique for growth with the ambient concentrations of organic matter found in autoclaved seawater (23), and in mineral salts without added carbon (Figure 2-2 and

Supplementary Figure A1-4 in appendix 1) are ample evidence of the adaptation of these cells to growth at low nutrient concentrations. More importantly, members of the SAR11 clade are the most successful chemoorganoheterotrophs in oligotrophic ocean systems.

Therefore, although Pelagibacter defies conventional definitions of the term ‘oligotroph’ because it can tolerate some compounds at relatively high concentrations, it clearly conforms to the concept of oligotrophy.

Metabolic reconstruction in silico results in metabolic models that are subject to uncertainties that can only be resolved by studying the physiology and metabolism of cells in vitro. The defined medium presented here for Ca. P. ubique, provides a platform from which such controlled experiments can occur. Ca. P. ubique and other cells with streamlined metabolism are particularly challenging to model because of uncertainties arising from the loss of genes that function in canonical pathways (Figure 2-1). For instance, the requirements for glycine and alanine were not evident from the initial genome analysis of Ca. P. ubique (25). In addition, the hypothesis that pyruvate could serve as the sole carbon precursor for both biosynthesis and energy production was not predicted by previous metabolic models. Uncommon gene arrangements also complicate the interpretation of metabolic models in Ca. P. ubique (Figure 2-5). In E. coli, the genes coding for proteins required for the assimilation of carbon from glycolate are found at a single locus and are transcribed to a polycistronic message. In Ca. P. ubique, the same

37 genes are physically separated from one another—but their functions are linked through the metabolic intermediate glycine (Figure 2-5). Experiments with cultures growing on a defined medium are an important step in the refinement of metabolic models such as the one presented in Figures 2-1 and 2-5.

The original success of dilution to extinction approaches was founded on the principle that if cells could be detected and cultured at low concentrations, on low nutrient media, their physiological responses to specific nutrient additions could be studied in subsequent experiments. The results presented in Tables 2-2 and 2-3, in light of the previously reported metabolic abilities of Ca. P. ubique (32, 37), highlight one of the challenges associated with cultivating specialist oligotrophic organisms. Our results suggest that perhaps the largest obstacle to overcome in the cultivation of marine oligotrophs is the specific, unusual and often combinatorial or conditional nutrient requirements. There is growing evidence that a number of abundant microbial plankton species that are not yet cultivated, or are cultivated but difficult to propagate, are similar to Ca. P. ubique in that they are abundant, small cells that contain relatively small genomes (<2.0 Mbp) (39, 77, 78). The success of cultivating Ca. P. ubique on artificial media relied on a targeted, minimal combination of nutrients deduced from metabolic reconstruction from a sequenced genome. If, as is likely, such requirements are common among organisms with streamlined genomes, it is hoped that this approach may serve as a blueprint for the cultivation of other important taxa that are difficult to grow in the laboratory.

Table 2-1: Constituents of the Artificial Medium for SAR11 (AMS1)1 Compound Final Concentration Base Salts: NaCl 481 mM MgCl2•6H2O 27 mM CaCl2•2H2O 10 mM KCl 9 mM NaHCO3 6 mM MgSO4•7H2O 2.8 mM Macronutrients: (NH4)2SO4 400 µM NaH2PO4 (pH 7.5) 50 µM Trace Metals: FeCl3•6H2O 117 nM MnCl2•4H2O 9 nM ZnSO4•7H2O 800 pM CoCl2•6H2O 500 pM Na2MoO4•2H2O 300 pM Na2SeO3 1 nM NiCl2•6H2O 1 nM Vitamins: B1 6 µM B3 800 nM B5 425 nM B6 500 nM B7 4 nM B9 4 nM B12 700 pM myo-inositol 6 µM 4-aminobenzoic acid 60 nM 1The AMS1 medium does not include nutrients to meet the requirements for reduced sulfur (see (26)), glycine (see text), or pyruvate (see text).

Table 2-2: Potential sources of pyruvate for Ca. P. ubique when grown in AMS1 with 25 µM glycine and 10 µM methionine maximum growth potential pyruvate source density1 rate2 glucose 12.1±3.83 0.27±0.02 pyruvate (positive control) 8.91±0.74 0.33±0.01 oxaloacetate 6.92±0.04 0.33±0.04 taurine 0.46±0.14 0.21±0.02 lactate 0.11±0.09 0.18±0.04 ribose 0.05±0.01 0.23±0.01 malate 0.03±0.01 0.14±0.04 citrate 0.03±0.00 0.12±0.01 acetate 0.02±0.01 0.10±0.05 no pyruvate (negative control) 0.02±0.01 0.13±0.02 succinate 0.02±0.00 0.11±0.01 alanine 0.01±0.00 0.01±0.01 glycine 0.01±0.00 0.06±0.01 1Presented as mean ± s.d. × 107 cells ml-1, n=3 2Presented as mean ± s.d. cells d-1, n=3

Table 2-3: Potential sources of glycine for Ca. P. ubique when grown in AMS1 with 50 µM pyruvate and 10 µM methionine maximum potential glycine source growth rate2 density1 glycine betaine 13.1±0.34 0.18±0.01 glycine (positive control) 10.2±0.28 0.38±0.01 serine 8.34±0.15 0.40±0.01 glycolate 1.88±0.13 0.35±0.01 acetate 0.66±0.10 0.30±0.01 malate 0.51±0.03 0.32±0.03 glyoxylate 0.50±0.01 0.38±0.01 succinate 0.46±0.06 0.31±0.01 pyruvate 0.44±0.05 0.35±0.01 citrate 0.44±0.01 0.29±0.01 no glycine (negative control) 0.43±0.02 0.34±0.00 2-oxoglutarate 0.39±0.05 0.33±0.01 1Presented as mean ± s.d. × 107 cells ml-1, n=3 2Presented as mean ± s.d. cells d-1, n=3

α-D-glucose β-D-glucose gluconate

6-phospho- α-D-glucose-6P β-D-glucose-6P ribulose-5P gluconate β-D-fructose-6P

β-D-fructose-1,6P2 2,4-keto-3-deoxy- 6-phosphogluconate serine 3-HP G3-P glycine 2- PEP SO4

ald APS pyruvate alanine O-acetyl- NH3 PAPS CO2 homoserine

CO2 2- H2S SO3 acetyl-CoA homocysteine

CO2 methionine oxaloacetate citrate glcB malate glyoxylate isocitrate

CO2 succinate 2-oxoglutarate

CO 2

Figure 2-1: Simplified illustration of central metabolism in Ca. P. ubique. Black lines: Reactions predicted to occur in Ca. P. ubique based on genome content. Red lines: Reactions predicted to occur in E. coli, but missing from Ca. P. ubique. Blue lines: Putative glucose oxidation pathway (see ref. 37). 3-HP: 3-hydroxypyruvate, G3-P: glyceraldehyde-3-phosphate, APS: adenosine 5'-phosphosulfate, PAPS: 3'- phosphoadenylyl sulfate, PEP: phosphoenolpyruvate. Gene names in green are discussed further in this manuscript. glcB: malate synthase; ald: alanine dehydrogenase. Bolded compounds were previously identified as growth substrates for Ca. P. ubique.

8 methionine + glycine + pyruvate 10 glycine and pyruvate no added nutrients methionine 107 ) -1

106

105 cell density (cells ml

104

051015202530 time (days)

Figure 2-2: Growth of Ca. P. ubique in AMS1 with organic carbon additions. Black: Cells grown without additions of glycine, methionine or pyruvate. Red: Cells amended with methionine (10 µM), glycine (50 µM), and pyruvate (50 µM). Blue: Cells amended with methionine (10 µM) only. Green: Cells amended with glycine (50 µM) and pyruvate (50 µM) only. Points are the average density of triplicate cultures. Error bars indicate ± 1.0 s.d. (n=3). When error bars are not visible, they are smaller than the size of the symbols.

Figure 2-3: Maximum cell yields of Ca. P.

) 1.6 6 6 -1 A y=2.6×10 x + 1.9×10 ubique in response to pyruvate and glycine R2=0.998 additions. A) Pyruvate titration in AMS1 supplemented with glycine (50 µM) and cells ml cells

7 1.2 methionine (10 µM). Using the formula of the regression line, we calculated the maximum cell density achievable from pyruvate 0.8 carryover with source inoculum (156 pM) to be 400 cells ml-1 (filled star). B) Glycine titration into AMS1 supplemented with 0.4 pyruvate (50 µM) and methionine (10 µM). Using the formula of the regression line, we maximum cell density ( × 10 calculated the maximum cell density 0 achievable from glycine carryover with source 012345 inoculum (500 pM) to be 5,000 cells ml-1 pyruvate added (μM) (filled star). Points are the average maximum cell densities of triplicate batch cultures. Error bars indicate ± 1.0 s.d. (n=3). When error bars B y=1.0×107x + 2.9×106 are not visible, they are smaller than the size ) 2 -1 5.0 R =0.995 of the symbols.

cells ml cells 4.0 7 3.0

2.0

1.0 maximum cell density ( × 10 0 012345 glycine added (μM)

300-325 event-1

A C counts 475-500 event-1

B D

relative DNA fluorescence intensity

Figure 2-4: DNA content and morphology of SYBR Green I-stained stationary phase cells from pyruvate-deplete and replete batch cultures. Red dashed line in A and B represents the minimum threshold of fluorescence detection. Black dashed lines in A and B represent relative DNA fluorescence values of 300-325 per event and 475-500 per event, as indicated with black arrows. A) Relative DNA fluorescence of cells from pyruvate-replete (50 µM) stationary phase cultures, and B) pyruvate-deplete (0.5 µM) stationary phase cultures. C) Fluorescent microscopy image of cells from (A). Arrowheads point to single cells. D) Fluorescent microscopy image of cells from (B). Arrowheads point to cell doublets.

Escherichia coli K-12 MG1655 glcC glcD glcE glcF G glcB yghK

‘Candidatus Pelagibacter ubique’ HTCC1062 aspC glcD glcE glcF ldh 0274 ~220 kbp accAglcB recA

acetyl-CoA

O2 H2O2 glycolate glycolate glyoxylate out in GlcB malate amino donor TCA cycle (amino acid)

2-oxo-acid glycine-activated translation of glcB glycineout glycinein glcB mRNA serine NH3+ CO2 protein

Figure 2-5: Glycolate assimilation gene organizations in E. coli and Ca. P. ubique. Reaction arrows are are colored by the genes predicted to catalyze the reaction. Green stem-loop images represent the glycine-activated riboswitch (32). Gene annotations are as described in NCBI, however we predict the reaction catalyzed by AspC to be as described in the figure. glcC: DNA-binding transcriptional dual regulator, glycolate- binding; glcD: glycolate oxidase subunit, FAD-linked; glcE: glycolate oxidase, FAD- binding subunit; glcF: glycolate oxidase, iron-sulfur subunit; G: glcG; Putative glc operon gene, function unknown; glcB: malate synthase G; yghK: glycolate transporter; aspC: probable aspartate transaminase; 0275: SAR11_0275; probable 2-hydroxyacid dehydrogenase; 0274: SAR11_0274; major facilitator superfamily transporter, possible sugar-phosphate transporter; accA: acetyl-CoA carboxylase; recA: recombinase A.

Chapter 3 ‘Candidatus Pelagibacter ubique’ HTCC1062 is Dependent on 4-Amino-5- Hydroxymethyl-2-Methylpyrimidine, An Abundant Thiamine Precursor In The Sea

Paul Carini1, Emily O. Campbell1, Jeff Morré5, Sergio Sañudo-Wilhelmy3, Samuel Bennett2, J. Cameron Thrash1, Taghd Begley4 and Stephen Giovannoni1

1Department of Microbiology, Oregon State University Corvallis, OR 97331

2Department of Environmental & Molecular Toxicology, Oregon State University, Corvallis, OR 97331

3Department of Biological Sciences, Marine Environmental Biology, and Earth Science, University of Southern California, Los Angeles, CA 90089

4Department of Chemistry, Texas A&M University, College Station, Texas 77843

5Department of Chemistry, Oregon State University, Corvallis, OR 97331

Abstract

The essential B-vitamin thiamine (B1) has complex distribution patterns in the sea and is often low or undetectable in marine surface waters. Despite this, analysis of sequenced genomes belonging to surface-dwelling organisms in the SAR11 clade of marine heterotrophic bacteria revealed incomplete thiamine synthesis pathways. A representative SAR11 isolate, ‘Candidatus Pelagibacter ubique’ HTCC1062, lacks thiC, encoding the 4-amino-5-hydroxymethyl-2-methylpyrimidine (HMP) synthase, required for de novo synthesis of thiamine’s pyrimidine moiety, but retains the remainder of the thiamine biosynthetic pathway. Here we show that Ca. P. ubique is dependent on exogenous HMP for growth. Exclusion of thiamine from Ca. P. ubique cultures growing on a defined medium amended with excess carbon and sulfur resulted in lower cell densities than cultures amended with thiamine. Cell growth was restored with picomolar additions of HMP. We report the vertical distributions (0 – 300 meters) of dissolved

HMP and thiamine in seawater collected from the Sargasso Sea. HMP was detected in all samples and ranged from 1-36 pM with maximum concentrations coinciding with the deep chlorophyll maximum. Thiamine ranged from undetectable to 22.3 pM with maximum concentrations at the surface. We provide evidence that marine

Prochlorococcus and Synechococcus bacteria are a probable source of environmental

HMP, and that HMP production may have diel periodicity. These findings are evidence of dynamic thiamine and thiamine precursor cycling the marine euphotic zone that may have implications in community energy and carbon metabolism over short time scales.

Introduction

Thiamine (vitamin B1) is a sulfur-containing coenzyme used by proteins that catalyze crucial manipulations of carbon in all living systems. Specifically, thiamine is required for enzymes of the TCA cycle, the non-oxidative portion of the pentose- phosphate pathway, light-independent photosynthetic reactions and for the biosynthesis of branched-chain amino acids, isoprenoids and some vitamins (79). The pathways and enzymes utilized for de novo thiamine synthesis and salvage have been the topic of extensive research in bacteria and eukaryotic yeasts (80), and in all organisms capable of de novo biosynthesis, the formation of thiamine monophosphate (ThP) is the result of the enzyme-catalyzed linkage of two separately synthesized moieties: 4-amino-5- hydroxymethyl-2-methylpyrimidine diphosphate (HMP-PP) and 4-methyl-5-(2- phosphoethyl)-thiazole (THZ-P). Phosphorylation of ThP yields the active thiamine coenzyme, thiamine diphosphate (ThPP) (Fig. 3-1; reviewed in (80)).

Research of oceanic vitamin distributions is experiencing a renaissance, largely due to more sensitive analytical techniques and a greater overall understanding of the factors influencing productivity. There is new evidence that oceans have complex distribution patterns of vitamins (81, 82) and that these cofactors may act synergistically with other vitamins or inorganic nutrients to increase bacterial and phytoplankton productivity (83). Direct measurements of thiamine in the sea revealed that surface-water standing stocks were low except in regions of upwelling (81, 82). To contextualize such vitamin abundance patterns and relate them to community structure and ecosystem function, an understanding of the community members that produce and consume

49 vitamins is vital (84). Previously, eukaryotic phytoplankton have been shown to both produce (85) and require (86) thiamine. Although the synthesis and utilization of other B- vitamins proceeds through complex symbiotic relationships (87), very little is known regarding thiamine metabolism of marine bacteria and archaea.

The most abundant heterotrophic bacterioplankton in the marine euphotic zone belong to the SAR11 clade of bacteria (17). Both in situ and in vitro studies have shown that SAR11 cells actively contribute to the cycling of carbon and sulfur in the ocean (26,

28, 37, 88). The first cultivated SAR11 bacterium, ‘Candidatus Pelagibacter ubique’

HTCC1062 (Ca. P. ubique), contains one of the smallest genomes found in free living organisms and shows substantial evidence for genome streamlining (25). Genome streamlining has been proposed as an explanation for the unusual combination of amino acids, reduced organosulfur compounds and organic acids required for the growth of Ca.

P. ubique (26, 32, 37, 89). Although the macronutrient requirements of Ca. P. ubique have been identified (89), the trace requirements for vitamins have not been investigated.

Herein we investigate the metabolism of thiamin by Ca. P. ubique and identify 4-amino-

5-hydroxymethyl-2-methylpyrimidine (HMP) as a required trace nutrient.

Results & Discussion

Comparative genomics of thiamine biosynthesis and regulation

Previous metabolic reconstructions using the Ca. P. ubique genome did not identify a complete biosynthetic pathway for thiamine (25), yet multiple essential genes encoding proteins that require thiamine were identified (Figure A2-1 in appendix 2),

50 suggesting thiamine is necessary for normal metabolism. Homologs for the HMP synthase (thiC) and tyrosine lyase (thiH), encoding essential proteins for de novo thiamine biosynthesis in Escherichia coli, were not identified in Ca. P. ubique (Figure 3-

1a). Genes encoding proteins that may function to supply HMP independently of ThiC were also not identified in the Ca. P. ubique genome. For example, NMT1 homologs, which catalyze the formation of HMP from vitamin B6 and histidine in Saccharomyces cerevisiae (90), were not identified. Moreover, genes commonly ascribed to thiamine salvage (thiM, thiK, tenA, tenI) (80, 91) and transport (thiBPQ) (92) were absent from the

Ca. P. ubique genome.

Genes encoding homologs for the remaining proteins required for thiamine biosynthesis were identified in the Ca. P. ubique genome (Figure 3-1), including: (i) the hydroxy-(phospho)methylpyrimidine kinase (thiD); (ii) thiazole sulfur transfer protein

(thiS); (iii) thiazole biosynthesis protein (thiG); (iv) two copies of the thiamine monophosphate synthase (thiE); and (v) thiamine monophosphate kinase (thiL). The protein sequence of E. coli’s ThiS-adenylyltransferase (encoded by thiF; b3992) is homologous to the translated product of SAR11_0403 (blastp E-value: 1 × 10-54, 86/237 identities), which is annotated as moeB and predicted to encode a protein involved in molybdopterin biosynthesis. During THZ-P synthesis in E. coli, sulfur is mobilized from cysteine to the thiazole ring by the IscS protein (encoded by iscS; b2530) (93). The best

Ca. P. ubique homolog to the E. coli IscS protein is the protein product of SAR11_0742, annotated as a selenocysteine lyase (blastp E-value: 3 × 10-28; 99/384 identities). In B. subtilis, a glycine oxidase (encoded by thiO; BSU11670) forms dehydroglycine instead

51 of the ThiH protein used by E. coli (94). The best Ca. P. ubique protein hit to the B. subtilis ThiO protein is encoded by SAR11_1221, and annotated as a sarcosine dehydrogenase (blastp E-value: 1 × 10-26 100/378 identities). However, it is possible that the formation of dehydroglycine is not enzyme-catalyzed. Experiments in vitro showed that ammonia and glyoxylate are in equilibrium with dehydroglycine in aqueous solution and can serve as thiazole precursors (95). Glyoxylate was previously proposed to have a central role in glycine synthesis in Ca. P. ubique (89), thus intracellular dehydroglycine may arise from the non-enzymatic condensation of ammonia and glyoxylate. In E. coli, the thiamine biosynthesis genes (thiCEFSGH and thiMD) are transcribed to polycistronic messages, however, in Ca. P. ubique, only two thiamine biosynthesis genes were found adjacent to one another (thiS and thiG), suggestive of co-transcription (Figure 3-1a).

The transcription of thiC and other thiamine biosynthetic and transport genes is negatively regulated by intracellular ThPP concentrations through the action of mRNA leader sequences called riboswitches. When ThPP is bound to ThPP-binding riboswitches transcription and translation of the downstream coding sequence is repressed (96).

Genomic searches for ThPP-binding riboswitch motifs have been used to identify genes putatively involved in thiamine biosynthesis and transport (97). In Ca. P. ubique, a single predicted ThPP-activated RNA riboswitch was identified upstream of a coding sequence annotated as a sodium:solute symporter family ‘proline uptake protein’ (SAR11_0811 –

Figure 3-1a) (98).

Physiology of thiamine deprivation & substitution

When Ca. P. ubique was grown in an artificial seawater medium with carbon, reduced sulfur, but without thiamine, maximum cell densities of 3.09 ± 1.6 × 107 cells ml-

1 (mean ± s.d., n=6) (Figure 3-2a,b) were attained. To determine if there were traces of thiamine in our base medium that may account for growth to 107 cells ml-1 in the absence of thiamine, we measured the thiamine concentration in uninoculated medium and found there to be 26 pM present (for method, see (81)). Efforts to remove this contaminating thiamine were unsuccessful, which is consistent with previous reports describing the difficulty of removing thiamine from growth media (99). We suspect the trace thiamine

(or other thiamine precursors) is from our reagent of the lowest purity (oxaloacetate –

98%) and is responsible for the growth of Ca. P. ubique in the absence of added thiamine.

When Ca. P. ubique was grown in the aforementioned medium amended with 1 µM thiamine, maximum cell yields increased to 3.11 ± 0.04 × 108 cells ml-1 (mean ± s.d., n=6) (Figure 3-2a,b).

To support the interpretation that the reduction of cell yield in thiamine deplete conditions (relative to thiamine replete conditions) was the result of thiamine limitation and to determine which Ca. P. ubique genes were putatively involved in thiamine acquisition and metabolism, we conducted DNA microarray experiments to compare gene transcript abundances from cells grown with and without added thiamine (for sample points see Figure A2-2 in appendix 2). As Ca. P. ubique cells approached thiamine limitation, the putatively ThPP-regulated ‘proline uptake protein’ (SAR11_0811

- Figure 3-1a) was significantly upregulated (3-fold, q-value of ≤0.01). In thiamine-

53 limited conditions, Ca. P. ubique upregulated 52 genes (Table A2-1 in appendix 2) and downregulated 71 genes (Table A2-2 in appendix 2) relative to thiamine replete treatments. The only gene upregulated in thiamine-limited conditions that is predicted to encode a thiamine biosynthetic enzyme was the hydroxy-(phospho)methylpyrimidine kinase thiD (2.1-fold, q-value <0.01; Table A2-1 in appendix 2).

We investigated the effective substitution range of HMP, thiamine and the thiamine degradation product 4-amino-5-aminomethyl-2-methylpyrimidine (AMP) across eight orders of magnitude in Ca. P. ubique cultures growing in an artificial medium

(Figure 3-2b). Growth of Ca. P. ubique was stimulated at HMP concentrations of 10-

100 pM and reached maximal cell densities at concentrations ≥1 nM (Figure 3-2b).

Thiamine and AMP were ineffective at restoring thiamine-limited cell yields at picomolar concentrations; each compound supported maximal cell yields when supplied at or above a final concentration of 1.0 µM (Figure 3-2b). No other potential thiamine precursors, including 4-methyl-5-thiazoleethanol (THZ) (100), histidine + vitamin B6 (101) or pantothenate (102) restored growth in the absence of thiamine (Figure A2-3 in appendix

2).

Molecular phylogeny and comparative genomics of SAR11_0811

We propose that the SAR11_0811 gene annotated as a ‘putative proline transporter’ encodes a transporter for HMP and that the predicted ThPP riboswitch preceding its coding sequence is a functional ThPP-dependent regulatory element. The increased expression of the SAR11_0811 gene upon thiamine-depletion is consistent with the expected expression pattern for genes regulated by ThPP-binding riboswitches (96).

Maximum-likelihood phylogenetic analysis of proteins homologous to SAR11_0811 shows that it is closely related to protein sequences encoded by a diverse group of organisms, including archaea, Gram-positive bacteria and other proteobacteria (Figure

A2-4 in appendix 2). In all but one instance, the putative HMP transporter is directly preceded by probable ThPP-riboswitches or is in a predicted operon with thiamine pyrimidine salvage genes (Figure A2-5 in appendix 2). In nine of ten sequenced

Pelagibacter genomes, SAR11_0811-like genes were identified (Figures A2-4 & A2-5 in appendix 2). No SAR11_0811-like gene sequence was identified Pelagibacter sp. str.

HIMB59, however, a putative thiamine transporter (thiBPQ) (92) was identified, suggesting thiamine or ThPP can be transported directly.

Measurements of HMP and thiamine in the environment

Previously it was suggested that thiamine pyrimidines may be relatively stable in seawater (103), yet no in situ measurements of thiamine pyrimidines have been made. To close this knowledge gap, we extracted and concentrated HMP and thiamine from

Sargasso Sea seawater collected on September 19th (20:00 local time) and 20th (08:00 local time), 2012, to a solid-phase and quantified them using high-performance liquid chromatography-coupled tandem mass spectrometry (using method of (81)). We found detectable HMP in all samples, ranging from 1.3-35.7 pM (Figure 3-3). The maximum concentration of HMP was from samples collected at 08:00 (Figure 3-3) and coincided with the deep chlorophyll maximum (Figure A2-6 in appendix 2). A decrease in the HMP concentration was evident at every depth except 0 meters in samples collected at 20:00 relative to samples collected at 08:00 (Figure 3-3). Thiamine concentrations ranged from

55 undetectable (detection limit: 0.81 pM (81)) to 23 pM and was present in samples from 0 to 160 m but not detected in samples from 250 and 300 m (Figure 3-3). Thiamine concentrations were generally highest at the surface and decreased with depth and did not vary greatly with collection time (Figure 3-3).

We show that Ca. P. ubique requires thiamine or a thiamine pyrimidine for growth and that HMP most effectively satisfies that growth requirement (Figure 3-2b).

Although thiamine and AMP can restore thiamine-limited growth, the amount required to do so is four orders of magnitude higher than the amount of HMP required for the same cell yield. No known direct measurements of thiamine in the sea have exceeded 1.0 nM

(81, 82), showing that environmental concentrations of thiamine are insufficient to elicit growth responses in Ca. P. ubique. Instead, we show that the in situ concentration (10-

40 pM, Figure 3-3) of HMP is within the concentration range to which Ca. P. ubique responds (10-100 pM, Figure 3-2b). There is additional evidence that natural microbial communities are thiamine-stressed in the Sargasso Sea. For example, metaproteomic analysis of the Sargasso Sea (104) detected peptides that mapped to ThPP-riboswitch- regulated genes, including the Ca. P. ubique sodium:solute symporter family protein we describe here (Figures A2-4 & A2-5 in appendix 2), and peptides mapping to the ThiC protein of Synechococcus and Prochlorococcus. Expression of these ThPP-riboswitch- regulated proteins would only be expected in cells experiencing thiamine stress (96).

A specific requirement for HMP has been noted in Plasmodium falciparum (105) and Listeria monocytogenes (106) but is unknown in marine organisms. The thiC gene is absent from all sequenced Pelagibacter genomes, which may be generalized of all

56 members of the SAR11 clade (29). The loss of thiC within the Pelagibacteraceae is consistent with the predicted consequences of genome streamlining as outlined by the

‘Black Queen Hypothesis’ (BQH) proposed by Morris (41). The BQH states that in some bacterial lineages, selection for the loss of metabolically expensive functions may be favored if the lost function can be compensated for by other members of the community.

Gene deletions at potentially useful but non-essential loci can be fixed by genetic drift in small bacterial populations. However, the population of SAR11 is massive and therefore the loss of thiC is unlikely to be the result of neutral processes such as drift. This implies that HMP production is costly for Pelagibacter to synthesize de novo and that a growth advantage is gained by outsourcing its production to other plankton.

Oxygenic photoautotrophic bacteria are probably contributors to HMP levels in the sea and the production of HMP may be periodic. Diel transcription of thiC was observed in both laboratory cultures of Prochlorococcus MED4 (along with diel translation of ThiC peptides) (107) and in environmental transcripts mapping to

Synechococcus (108). In both reports, maximum thiC transcript levels were observed in the mid-late afternoon shortly after periods of highest light intensity. Consistent with the hypothesis that photoautotrophs are the primary HMP producers, the highest HMP concentrations (Figure 3-3) coincided with the deep chlorophyll maximum (Figure A2-6 in appendix 2). Although additional experiments are needed to address the causes of the observed thiC expression patterns, an attractive hypothesis is that intracellular thiamine is less stable in illuminated conditions because of light-induced decomposition. Thiamine is light sensitive and UV-B induces cleavage of thiamine (resulting in AMP production)

(109, 110). UV-B radiation can pass though seawater to depths approaching 40 m in the open ocean (111) and is an important factor in the decomposition of some types of dissolved organic matter (71). Thus, following periods of high illumination, decomposition of intracellular thiamine may prevent ThPP-repression of thiC (and ThiC), perhaps resulting in HMP overproduction. The use of HMP (or AMP) instead of thiamine may be an adaptation common other organisms in the euphotic zone as a way to increase fitness in illuminated environments.

Because thiamine plays an important role in metabolic pathways pertaining to energy production and carbon metabolism, its presence or absence may have unique ramifications in how cells organize carbon and energy fluxes over short time scales. For example, when thiamine is absent, the TCA cycle cannot function due to the ThPP- requirement of the 2-oxoglutarate dehydrogenase. However, most Pelagibacter genomes encode a complete glyoxylate bypass (29, 32) that may circumvent the low flux of carbon through the TCA cycle in times of thiamine stress while enabling the formation of

TCA-cycle-derived amino acids. Pelagibacter cells can also generate energy from the thiamine-independent oxidation of one-carbon compounds, which may be an important source of energy when thiamine levels are low (49). Recently, synchronization of gene transcripts related to ribosome biogenesis and oxidative phosphorylation was observed between wild populations of SAR11, SAR86 and marine group II Euryarcheota (108). In our studies, gene transcripts that showed the largest expression increases when amended with thiamine (compared to thiamine-limited cultures) (Table A2-2 in appendix 2) were identified by Ottesen as positively correlated with ribosomal protein transcripts (108).

Although more research is necessary to elucidate the specific factors influencing the transcription patterns observed in situ and in vitro, secreted or leaked compounds essential for energy metabolism (such as thiamine, HMP and possibly other B-vitamins) may be cues for SAR11 and other taxa to adopt a specific metabolic state.

The depth-distribution measurements of HMP offer clues to the cycling of thiamine, and perhaps other B-vitamins in the euphotic zone of marine ecosystems and raise interesting questions regarding the biologically active form of individual vitamins in seawater. It is known that other B-vitamins are light sensitive or unstable in seawater

(103, 112) or play roles in light-mediated regulatory schemes (113). Furthermore, there are a number of B-vitamins for which no environmental data exists (e.g. – niacin (B3), pantothenate (B5) and folate (B9)). Given that B-vitamins are often required for essential cellular functions, understanding their environmental distributions may help elucidate succession patterns and the observed community connectivity in the environment (46,

47), making B-vitamin and precursor studies an exciting and fruitful avenue of future research.

Methods

Organism source

Ca. P. ubique HTCC1062 was revived from 10% glycerol stocks and propagated in artificial medium for SAR11 (AMS1) (89) amended with oxaloacetate (1 mM), glycine

(50 µM), methionine (50 µM) and FeCl3 (1 µM). Thiamine or precursors were added as indicated in figure legends and text.

Chemicals

All AMS1 constituents, reagents and vitamins were of the highest available quality (labeled ‘ultrapure’ when possible). Nutrient and vitamin stocks were prepared in combusted glassware (450°C for 4 h) with nanopure water, 0.1 µm-filter-sterilized and frozen in amber tubes immediately after preparation. Reasonable precautions were taken to limit the number of freeze-thaw cycles and light exposure of each reagent. HMP was synthesized as described in (114). AMP was synthesized as described in (115).

Cultivation details

All cultures were grown in acid-washed and autoclaved polycarbonate flasks and incubated at 20°C with shaking at 60 RPM in the dark. Cells for counts were stained with

SYBR green I and counted with a Guava Technologies flow cytometer at 48-72 h intervals as described elsewhere (26).

Vitamin-limitation procedure

To dilute out the vitamins routinely added to SAR11 cultures, Ca. P. ubique was transferred to AMS1 amended with organic macronutrients, but without B-vitamins, a minimum of 10 times (>1000 generations). After three passages (>30 generations) truncated cell growth was consistently observed (104 cells ml-1 to 107 cells ml-1 - for example, see Figure 3-2a).

Thiamine replacement experiments

Ca. P. ubique was grown in AMS1 with oxaloacetate (1 mM), glycine (50 µM), methionine (50 µM) and FeCl3 (1 µM). The treatments for the effective range of

60 substitution experiments were: no thiamine addition (negative control), thiamine-HCl added (1 pM to 10 µM), 4-amino-5-hydroxymethyl-2-methylpyrimidine added (1 pM to

10 µM), or 4-amino-5-aminomethyl-2-methylpyrimidine added (1 pM to 10 µM).

Experiments testing other potential thiamine precursors were conducted in AMS1 with the same organic amendments, but with additions of 1.0 µM of pantothenate, 1.0 µM 4- methyl-5-thiazoleethanol (THZ) or 1.0 µM each of histidine and B6 in place of thiamine.

Microarray analysis

Biological replicates (n=3) of Ca. P. ubique were grown in AMS1 with amendments of oxaloacetate (1 mM), glycine (50 µM), methionine (50 µM) and FeCl3

(1 µM). The thiamine-replete treatment was amended with 1.0 µM thiamin-HCl; the thiamine-deplete treatment was not amended with thiamine. Each growth condition sampled at points indicated in Figure A2-2 in appendix 2. At each sample point, cells were harvested by centrifugation at 20,000 RPM for 1.0 h at 4°C. Transcription profiles of cell pellets were immediately stabilized with RNAprotect Bacteria reagent (Qiagen –

Valencia, CA). RNA was extracted using an RNeasy Mini kit (Qiagen – Valencia, CA) and amplified using the MessageAmp-II Bacteria RNA amplification kit (Ambion –

Carlsbad, CA) per the manufacturer’s instructions. Array chip hybridization to custom

‘Candidatus Pelagibacter ubique’ Affymetrix GeneChip arrays that contained probes for

Pelagibacterales strains: HTCC1002, HTCC1062 and HTCC7211 (Pubiquea520471f) was performed as described elsewhere (37). A Bayesian statistical analysis was conducted using Cyber-T (116). The estimate of variance was calculated in Cyber-T by using window sizes of 101 and a confidence value of 10. A t-test was performed on log-

61 transformed expression values by using the Bayesian variance estimate. The program

QVALUE, was used to obtain a q-value, which accounts for multiple t-tests performed

(117). We defined genes as differentially expressed if both the q-value was ≤0.05 and the gene was differentially regulated by ≥2.0-fold.

Seawater collection for depth distributions

Seawater for vitamin analysis was collected from Hydrostation S (32°10'N,

64°30'W) from casts at 20:00 (local time) on 19 September 2012, and 08:00 (local time) on 20 September 2012. At time of collection, samples were filtered through nanopure water-rinsed 0.2 µm supor filters into acid-washed amber polypropylene bottles and frozen immediately. Temperature and fluorescence CTD data from these casts are show in Figure A2-6 in appendix 2.

Extraction of HMP from seawater

HMP was extracted from seawater to a reverse-phase C18 silica bead solid phase

(Agilent HF-Bondesil) as described in (81). In short, 3.0 g of the C18 beads were measured into chromatography columns. 100% methanol was added to the beads to form a slurry. Methanol was drained and the beads were conditioned with 20 mL HPLC-grade water. The seawater samples were thawed and allowed to equilibrate to room temperature before their pH was adjusted to 6.5 with 1 N HCl. The pH 6.5 sample (300 mL) was applied to the conditioned column at a flow rate of 1-2 mL min-1. The effluent was collected in a clean amber polypropylene bottle. The pH of the sample effluent was re- adjusted to 2.0 with 1 N HCl. The column was re-conditioned with 20 mL pH 2.0 (1 N

HCl) HPLC-grade water. After adjusting pH to 2.0, samples were applied to columns again. The column was washed with 20-30 mL pH 2.0 water to remove residual salt.

Organics were eluted from the C18 beads with 10 mL (1 mL 10 times) 100% MeOH.

Methanol was evaporated to dryness in a heated Savant Speedvac vacuum concentrator and frozen until LCMS analysis. For quantification purposes, aged seawater collected from Hydrostation S in 2009 was spiked with known amounts of HMP and thiamine ranging from 0 pM to 100 pM and extracted alongside samples.

High pressure liquid chromatography tandem mass spectrometry of HMP and thiamine

Dried samples were reconstituted in 125 µL HPLC-grade water. Samples were centrifuged to pellet insoluble matter and the supernatant was transferred to a sampling vial. 5 µL of each sample was injected into an Applied Biosystems MDS Sciex 4000 Q

TRAP mass spectrometer coupled to a Shimadzu HPLC system. An Agilent Zorbax SB-

Aq (2.1 × 100 mm, 3.5-micron) HPLC column was used for separation over a 10-minute gradient flow was used with mobile phases of pH 4 (with formic acid) methanol (MeOH) and pH 4 (with formic acid) 5 mM ammonium formate (AmF). The flow rate was 0.4 mL min-1 and a gradient starting at 98% AmF: 2% MeOH for 1 minute changing to 75%

AmF: 25% MeOH over 3 minutes, 50% AmF: 50% MeOH over 0.2 minutes, and finally to 10% AmF:90% MeOH over 0.8 minutes. The mass spectrometer was run in ‘Multiple

Reaction Monitoring’ (MRM) mode. The m/z of the HMP parent ion was 140.2, and ion transitions of 81.1 and 54.1 were used for quantification and qualification, respectively.

Peaks were analyzed using the Analyst® software package v 1.5.2 (AB SCIEX; Concord,

ON, Canada). Thiamine was detected as described in (81).

A E. coli K-12 MG1655 thiL 1.1 Mbp thiM thiD 2 Mbp thiC thiE thiF S thiG thiH Ca. P. ubique HTCC1062 thiE 39 kbp moeB 178 kbpthiE2 20 kbp thiD 120 kbp thiG S 15 kbp csdB

76 kbp 0811 226 kbp thiL

N B TenA H N N H2N 2 N N H O PO O 2 3 ThiC ThiD NH2 N HO N -3 N O6P2O N AMP HO OH NH2 NH AIR HMP(-P) 2 degradation HMP-PP -2 OPO3 -3 S. cerevisiae OP2O6 Purine histidine vitamin B ThiL 6 ThiE + s nucleotide NMT1 N N + s N N biosynthesis N NH2 N NH2 ThPP S ThiM S ThP N N -2 Active Cofactor OH OPO3

THZ THZ-P -2 OPO3 ThiG HO Dxs COOH CO-AMP COSH OH G3-P O [ThiS] MoeB [ThiS] CsdB [ThiS] dDXP pyr dh-gly

NH2 ThiH ThiO HS O tyr gly glyoxylate + NH3 OH cys E. coli B. subtilis

Figure 3-1: Comparative genomics of thiamine biosynthesis. A) Comparative genomics of genome content in Escherichia coli K12 MG1655 and ‘Ca. Pelagibacter ubique’ HTCC1062. Dashed black gene outlines: genes present in E. coli but absent in Ca. P. ubique. Colored gene outlines: shared colors represent gene homologs, and match biosynthetic steps depicted in (B). Red stem-loop structures represent predicted or characterized ThPP-binding RNA riboswitches. B) Simplified illustration of the reconstructed thiamine synthesis pathway in Ca. P. ubique. Black dashed lines: steps that are missing in Ca. P. ubique. For detailed biochemistry of thiamine biosynthesis, see (80). Abbreviations: AIR - aminoimidazole ribotide; HMP(-P) - 4-amino-5- hydroxymethyl-2-methylpyrimidine (-phosphate); HMP-PP - 4-amino-5-hydroxymethyl- 2-methylpyrimidine diphosphate; AMP - 4-amino-5-aminomethyl-2-methylpyrimidine; ThP – thiamine monophosphate; ThPP – thiamine diphosphate; THZ - 4-methyl-5- thiazoleethanol; THZ-P - 4-methyl-5-(2-phosphoethyl)-thiazole; cys – cysteine; gly – glycine; tyr – tyrosine; dDXP - 1-deoxy-D-xylulose 5-phosphate; dh-gly - dehydroglycine; pyr – pyruvate; G3-P – glyceraldehyde 3-phosphate. TenA is described in (91).

A B 4.5

) AMP 8 4.0 Thiamine 108 HMP

× 10 3.5 no vitamins -1 ) -1 3.0 10ml 7 2.5 2.0 106 1.5

density (cells 1.0 thiamine (1.0 μM) 105 no thiamine 0.5 maximum density ml (cells 0 0 5 10 15 20 25 30 35 none 1.0 10 100 1.0 10 100 1.0 10 pM pM pM nM nM nM μM μM time (days)

Figure 3-2: Growth of Ca. P. ubique under thiamine-limiting conditions and responses to thiamine pyrimidines. a) Growth curve of Ca. P. ubique on artificial seawater with methionine (50 µM), glycine (50 µM) and oxaloacetate (1 mM) with 1 µM thiamine (red) or without thiamine (black). Points are the average densities of biological replicates ± s.d. (n=3). b) Maximum densities of Ca. P. ubique cultures grown in artificial medium (as described for (a)) with thiamine, HMP or AMP at culture concentrations spanning eight orders of magnitude. Bar heights are the average densities of biological replicates ± s.d. (n=3).

concentration (pM) 0 5 10 15 20 25 30 35 40 0

100

150 depth (meters)

200 [HMP] - 20:00 cast [HMP] - 08:00 cast [Thi] - 20:00 cast 250 [Thi] - 08:00 cast

300

Figure 3-3: Depth distribution of dissolved 4-amino-5-hydroxymethyl-2- methylpyrimidine (HMP) and thiamine in the Sargasso Sea. Filled markers were collected at 20:00 (local time) on 19 September 2012. Open markers were collected at 08:00 (local time) on 20 September 2012. Squares are thiamine concentrations. Circles are HMP concentrations. HMP points are the average of replicate analyses for each sample ± 1.0 s.d. (n=3). There was no replication for the thiamine measurements. Thiamine was not detected in samples collected from 200 m at 20:00 or at 250 m and 300 m at either time. Temperature and fluorescence CTD data for each profile is shown in Figure A2-6 in appendix 2.

Chapter 4 Distinct Phosphate-Acquisition Physiotypes and Adaptations for Extreme Marine Oligotrophy in SAR11 Bacteria

Paul Carini1, Emily O. Campbell1, Angelicque E. White2 and Stephen Giovannoni1

1Department of Microbiology, Oregon State University Corvallis, OR 97331 2College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331

Abstract

Collectively, heterotrophs belonging to the SAR11 clade of oligotrophic marine

α-proteobacteria comprise one of the largest pools of biomass in the Earth’s oceans.

Despite the ubiquity of SAR11 cells, very little is understood about how their physiology and transcription patterns respond to nutrient perturbation and how those responses may affect geochemical cycles. Here we demonstrate evidence for distinct phosphorus (P) acquisition strategies by SAR11 marine bacteria that manifest as different P-utilization physiotypes. The SAR11 isolate ‘Candidatus Pelagibacter ubique’ HTCC1062 (Ca. P. ubique) responded to inorganic phosphate (Pi) depletion by rapidly upregulating genes encoding a high affinity Pi transport system (pstSCAB-phoUB). The remaining transcriptional response indicated that a stringent-like stress response was induced, as exemplified by increased abundance of stress response gene transcripts and simultaneous reduction in transcripts for ribosomal proteins, RNA polymerase subunits and cell division proteins. Of ten alternative organic and reduced P sources, only Pi effectively satisfied the P requirement of Ca. P. ubique. Conversely, Pi-depletion caused a second

Pelagibacter isolate, Pelagibacter sp. str. HTCC7211, to upregulate a suite of genes that encode a probable phosphate-ester/phosphite/ phosphonate/Pi transport system

(phnCDEE2), multiple C-P lyase complex genes that confer the ability to use phosphonates as sole Pi sources and four genes probably involved in the synthesis of P- free lipids. Experiments with cells in culture confirmed that Pelagibacter sp. str.

HTCC7211 used organic and reduced forms of P, including phosphonates, in place of Pi.

When grown on methylphosphonic acid (Mpn) as a sole P-source, the apparent cellular

68 requirement for P decreased. Moreover, cells grown on Mpn stoichiometrically released methane, implicating Pelagibacter as a source of ocean methane, and a potentially key player in the ocean methane paradox. These results show that Pelagibacterales bacteria have adopted diverse strategies to compete for low levels of Pi in the sea.

Introduction

Phosphorus (P) in its +5 valence state (phosphate; Pi) is an essential component of all living cells. It is a constituent of nucleic acids, phospholipids, phosphorylated proteins and is intimately involved in energy metabolism and some transport functions (via ATP).

Most bacteria readily assimilate Pi to meet their P-requirement, however in oligotrophic ocean gyres, Pi concentrations are extremely low (for example, 0.2-1.0 nM in the

Sargasso Sea (118)) and the availability of Pi can limit bacterial and primary production

(118-124). Several studies of P metabolism in primary producers that inhabit Pi-deplete waters have uncovered microbial adaptations that enhance the competitiveness of these lineages in low-Pi environments (125-131). However, the strategies used by major chemoheterotrophic bacteria that cohabitate these nutrient poor environments are relatively unexplored.

The SAR11 clade of oligotrophic α-proteobacteria are the numerically dominant chemoheterotrophic cells in marine euphotic zones worldwide (17). Ca. P. ubique, the first cultivated SAR11 isolate, is a small cell (volume of 0.01 µm3) and contains a streamlined genome (1.3 Mbp) (23). Both in situ (28, 88, 132, 133) and in vitro (26, 32,

37, 49) work shows that the Pelagibacter population is active and metabolizes specific forms of dissolved organic matter (DOM), including reduced sulfur compounds, amino

69 acids, one-carbon compounds and organic acids. In artificial media, Ca. P. ubique requires an unusual combination of amino and organic acids, reduced sulfur and vitamins for growth (89).

Microbes inhabiting Pi-deplete environments have evolved diverse strategies to effectively compete for trace amounts of Pi. For example, some bacteria, including marine phytoplankton, actively remodel membrane lipid structure to reduce phospholipid content, resulting in a decreased cellular P-requirement (134-137). P-starvation in many bacteria, including Escherichia coli, Sinorhizobium meliloti, and Prochlorococcus sp.

MED4, results in an increased cellular affinity for Pi, due to induction of high-affinity Pi transport systems (128, 138, 139). Furthermore, Pi stored as endogenous polyphosphate

(140) or phospholipids (137) during periods when Pi is plentiful may be metabolized when exogenous Pi is low. In low-Pi environments P is also obtained by degrading P- esters with broad-substrate range alkaline-phosphatases (APases; secreted or cytoplasmic

(141)), or phosphonates with a C-P lyase (126). Recently, Prochlorococcus was shown to utilize phosphite, a reduced P source for which environmental distributions are unknown

(142). Both metagenomic and metaproteomic approaches show that gene suites conferring both Pi and organic-P acquisition, P-storage (as polyphosphate) and metabolism are abundant and expressed in ocean environments that are low in Pi (104,

143-145), suggesting that the maintenance of complex P-regulons in low-Pi environments may allow microbes to maintain competitiveness and occupy specific P-acquisition niches.

The genome sequences of two organisms in the Ia-subclade of the

Pelagibacteraceae - Ca. P. ubique and Pelagibacter sp. str. HTCC7211 - show variation in the gene complements associated with the acquisition, storage and metabolism of Pi

(29), implying that the two bacteria may respond differently to Pi-limitation. Ca. P. ubique contains genes encoding a high affinity transport system and associated regulatory genes (pstSCAB-phoUB) (genomic context is depicted in Figure 4-3). Pelagibacter sp. str. HTCC7211 encodes, in addition to a high affinity transport system (pstSCAB- phoUB), genes involved in phosphonate acquisition (phnCDEE2), phosphonate metabolism (phnFGHIJKLNM and phnX), polyphosphate metabolism (ppx and ppk) and other genes putatively involved in Pi metabolism (genomic context is depicted in Figure

4-3). The P-acquisition and metabolism genes in Pelagibacter sp. str. HTCC7211 are found clustered in the genome.

Although assorted ‘-omic’ investigations have highlighted the genomic potential of Pelagibacter cells through in silico analyses, little is known about the specific types and amounts of P Pelagibacter use or cellular adaptations that may benefit these cells in low Pi ecosystems. To close this gap in knowledge, we undertook experiments designed to determine i) the amount of P required for Pelagibacter growth; ii) alternate forms of P that Pelagibacter cells can use and iii) the transcriptional responses to Pi-starvation. Two

Pelagibacter strains isolated from geochemically distinct ecosystems were used to accomplish these goals – Ca. P. ubique (isolated from the NE Pacific Ocean off the coast of Oregon USA (23)) and Pelagibacter sp. str. HTCC7211 (isolated from the North

Atlantic Ocean in the Sargasso Sea (22)). Experiments with both strains were conducted

71 on a chemically defined, low-Pi artificial medium for SAR11 (AMS1) amended with appropriate amino and organic acids and vitamins (89).

Methods

Organism source

Ca. P. ubique and Pelagibacter sp. str. HTCC7211 were revived from 10% glycerol stocks and propagated in artificial medium for SAR11 (AMS1) (89) amended with pyruvate (100 µM), glycine (5 µM), methionine (5 µM), FeCl3 (1 µM), and vitamins

(146) unless indicated otherwise.

Cultivation details

All cultures except those describing methane production (below) were grown in acid-washed and autoclaved polycarbonate flasks. Cultures were incubated at 20°C with shaking at 60 RPM under a 12-h/12-h light (140–180 µmole photons m-2 s-1)/dark cycle.

Cells for cell counts were stained with SYBR green I and counted with a Guava

Technologies flow cytometer at 48-72 h intervals as described elsewhere (26).

Growth conditions for phosphate replacement experiments

Pelagibacter isolates were grown in a base medium of AMS1 with no added Pi, amended with pyruvate (100 µM), glycine (5 µM), methionine (5 µM), FeCl3 (1 µM) and vitamins. The treatments were: no Pi addition (negative control), 1.0 µM Pi (as NaH2PO4

- positive control), 1.0 µM of deoxycytidine triphosphate (dCTP), 1.0 µM of deoxyguanine triphosphate (dGTP), 1.0 µM of deoxyadenine triphosphate (dATP),

1.0 µM of deoxythymidine triphosphate (dTTP), 1.0 µM phosphite (as NaPHO35H2O),

1.0 µM glucose 6-phosphate (G6P), 1.0 µM ribose 5-phosphate (R5P), 1.0 µM methylphosphonic acid (Mpn), 1.0 µM 2-aminoethylphosphonic acid (2-AEP) or 1.0 µM phosphoserine (P-ser). Flasks were incubated and counted as described in ‘Cultivation details.’

Microarray growth and sampling conditions

Biological replicates (n=3) of Ca. P. ubique or Pelagibacter sp. str. HTCC7211 were grown in AMS1 amended with pyruvate (1 mM), glycine (50 µM), methionine

(50 µM), FeCl3 (1 µM), Pi (10 µM) and vitamins. In mid- to late-logarithmic growth, cells were harvested by centrifugation (10,000 RPM for 1.0 h at 20°C). Cell pellets were evenly split and washed twice with growth media either amended with 100 µM Pi (replete conditions) or not amended with Pi (deplete conditions). After washing, pellets were resuspended in Pi-deplete or Pi-replete growth media. Samples were collected from cell suspensions at t=0 h, 4 h, 20 h, and 38 h after resuspension for Ca. P. ubique, and t=0 h,

20 h, 38 h, 68 h, and 96 h after resuspension for Pelagibacter sp. str. HTCC7211.

Sampling consisted of centrifugation at 20,000 RPM for 1.0 h at 4°C followed by resuspension of cell pellets in RNAprotect Bacteria reagent (Qiagen – Valencia, CA).

Preparation and analysis of microarrays

RNA was extracted using an RNeasy Mini kit (Qiagen – Valencia, CA) and amplified using the MessageAmp-II Bacteria RNA amplification kit (Ambion – Carlsbad,

CA) using the manufacturer’s instructions. Array chip hybridization to custom

‘Candidatus Pelagibacter ubique’ Affymetrix GeneChip arrays that contained probes for

QVALUE, was used to obtain a q-value, which accounts for multiple t-tests performed

(117). A gene was defined as differentially expressed if both the q-value was ≤0.05 and the gene was differentially regulated by ≥2.0-fold.

Methane production from Mpn

To test for methane (CH4) production from Mpn, Pelagibacter sp. str. HTCC7211 cells previously grown in media with excess Mpn as the sole P source were used to inoculate 60 mL sterile, bovine serum albumin (BSA)-coated serum bottles containing 55 mL growth media (leaving a 5 mL headspace) amended with 10 µM Mpn as the sole P source. The negative control consisted of Pelagibacter sp. str. HTCC7211 cells previously grown in media with excess Pi as the sole P source as inoculum for 60 mL sterile, BSA-coated serum bottles containing 55 mL growth media amended with 10 µM

Pi. Treatment and control bottles were capped with Viton septa, crimped with aluminum seals and incubated horizontally as described in ‘Cultivation details.’ CH4 in the headspace was measured by gas chromatography in a Shimadzu GC-8A gas chromatograph (GC) equipped with a column packed with Porapak N (80/100 mesh size) fitted with a flame ionization detector (FID). In brief, 100 µM samples of headspace from each bottle were injected into the GC at a flow rate of 25 mL min-1. Peaks were integrated

74 on a Chromatopac data processor (Shimadzu C-R8A). Quantification was accomplished by calibrating peak areas to a standard curve derived by injecting known amounts of CH4 into sealed serum bottles containing growth medium. Standards were equilibrated in a shaking water bath for 24 h at 25ºC. Single bottles were sacrificed from each treatment for cell counts over the duration of the experiment. Theoretical CH4 release from Mpn- grown cells was calculated assuming an average Mpn cell-1 usage of 11.4 amol cell-1

(calculated from regression line in Figure 4-1).

Transmission Electron Microscopy

To determine if the ppk gene conferred the ability to produce polyphosphate, we qualitatively determined whether Pelagibacter sp. str. HTCC7211 contained electron- dense polyphosphate-like granules under Pi-replete, Pi-starved and Mpn-replete conditions using STEM. Pelagibacter sp. str. HTCC7211 cells were grown in a base medium consisting of AMS1 with pyruvate (1 mM), glycine (50 µM), methionine (50

µM), FeCl3 (1 µM) and vitamins with either 10 µM Pi (Pi-replete) or 10 µM Mpn (Mpn- replete). Cells that were Pi-starved were harvested from the Pi-replete treatment and resuspended in Pi-free medium for 96 hours to starve them. Cells were harvested for microscopy by centrifugation and pellets were fixed in 2% glutaraldehyde and 1% paraformaldehyde in 0.1 M sodium cacodylate buffer. Fixed cells were stained with 2% osmium tetraoxide, rinsed and dehydrated (from 30% to 100% acetone). Dehydrated cells were infiltrated with Spurr’s resin and polymerized (65°C overnight). Cells embedded in resin were thin sectioned (50-70 nm) and stained with 4% Pb2+ (as lead citrate) for 2 min, then switched to 2% uranyl acetate for 30 min before a final 4% Pb2+ stain for 4 min.

Stained samples were examined by Scanning Transmission Electron Microscopy (STEM) using a Titan G2 80-200 microscope (FEI; Eindhoven, Netherlands) at 80 kV, spot size 3.

Results

Basic physiology

The maximum specific growth rate of Ca. P. ubique and Pelagibacter sp. str.

HTCC7211 on AMS1 was 0.59 ± 0.02 day-1 (average ± s.d. n=24) and 0.35 ± 0.02 day-1

(average ± s.d. n=24), respectively, regardless of media Pi concentration (Table A3-1 in appendix 3). Both Ca. P. ubique and Pelagibacter sp. str. HTCC7211 responded linearly to Pi additions when grown on AMS1 with other nutrients in excess (Figure 4-1). The apparent Pi requirement was calculated from Pi dose-response linear regressions and

-1 -1 found to be 10.2 attomoles Pi cell for Ca. P. ubique and 45.7 attomoles Pi cell for

Pelagibacter sp. str. HTCC7211 (Figure 4-1).

Alternate P-source utilization in Ca. P. ubique

Growth media without added Pi (negative control) supported maximal Ca. P.

6 -1 ubique cell densities of 1.7 × 10 cells ml (Figure 4-2). Growth medium with 1.0 µM Pi

7 -1 supported densities of 3.3 × 10 cells ml (Figure 4-2). Of nine alternate P-sources, dCTP, phosphite, dATP, dGTP, dTTP and R5P supported minimal to moderate cell density increases to 3.5-8.5 × 106 cells ml-1 (Figure 4-2). 2-AEP, Mpn, G5P and P-ser did not alleviate Pi limitation.

Alternate P-source utilization in Pelagibacter sp. str. HTCC7211

Growth media without added Pi (negative control) supported maximal

Pelagibacter sp. str. HTCC7211 cell densities of 7.8 × 105 cells ml-1 (Figure 4-2). Growth

7 -1 media with 1.0 µM Pi (positive control) supported densities of 3.8 × 10 cells ml (Figure

4-2). Pelagibacter sp. str. HTCC7211 grew to 4.0 × 107 cells ml-1 with 1.0 µM P-ser as a

P source (Figure 4-2). 1.0 µM additions of 2-AEP, Mpn, R5P, G6P or phosphite supported densities above 1.0 × 107 cells ml-1 (Figure 4-2). Diauxic growth was observed when Pelagibacter sp. str. HTCC7211 cells were grown with phosphite, G6P, R5P, Mpn,

2-AEP, or P-ser as sole P-sources (Figure A3-1 in appendix 3). Minimal to moderate cell density increases were observed for dCTP, dGTP, dATP and dTTP.

Transcriptional responses to Pi-starvation in Ca. P. ubique

To better understand the underlying genetic basis for the observed P-utilization patterns, we used DNA microarrays to monitor gene expression changes during the onset of Pi-limitation in cell-suspension time-course experiments. When stressed for Pi, Ca. P. ubique induced transcription of the high-affinity transport system (pstSCAB-phoUB)

(Figure 4-3; Tables A3-2 - A3-4 in appendix 3). Genes encoding proteins involved in the

SOS-response (umuD, recA, and lexA), heat shock (dnaK, dnaJ, ipbA), protein maturation and folding (groEL and trxB), protein recycling (hflCK, ftsH and hslUV), and membrane integrity (tolRAB) were upregulated during extended Pi starvation (select genes in Figure

4-3; Tables A3-3 & A3-4 in appendix 3). Genes encoding proteins involved in N- metabolism (glnA, gltBD), ammonia transport (amt), regulation (a CheY-like response regulator) and other hypothetical proteins were also upregulated in the absence of Pi

(Tables A3-3 & A3-4 in appendix 3). We observed a concomitant reduction in transcripts encoding for ribosomal proteins (rpsBEHNPK and rplBCFNP), translation machinery

(fusA and trmD), RNA polymerase subunits (rpoA and rpoC), cell division-related proteins (mraZ and ftsZ) and type IV pilus assembly (pilB and pilQ) (select genes in

Figure 4-3; Tables A3-3 & A3-4 in appendix 3). Also downregulated were genes encoding a Fur/Zur-like Fe2+/Zn2+-transcriptional regulator, periplasmic binding proteins for probable amino acid ABC transporters (ydhW and livJ) and proteins involved in the glyoxylate cycle (aceA) and methionine biosynthesis (metY).

Transcriptional responses to Pi-stress in Pelagibacter sp. str. HTCC7211

A transcriptional response to Pi-limitation in Pelagibacter sp. str. HTCC7211 was not observed 0 h, 20 h, or 38 h after resuspension. At 68 h, a probable phosphate-ester/ phosphite/phosphonate/Pi ABC transporter (phnCDEE2), multiple components of the C-P lyase (phnGHIJKLM), the high-affinity Pi transport system (pstSCABphoU), two genes annotated as putative hemolysins, a probable glycosyl transferase (rfaG-like), a calcineurin-like metallophosphoesterase (plsC) and an HD-domain-hydrolase were significantly upregulated along with genes of unknown function (Figure 4-3; Table A3-5 in appendix 3). After 96 h of Pi starvation, a probable operon consisting of a glutaredoxin-arsenate reductase (arsC) and putative glpF-family transporter (aqpS) were also significantly elevated (Figure 4-3; Table A3-6 in appendix 3), as were genes found elsewhere in the genome, including a putative alginate lyase and a cytoplasmic hydroxyglutathionine hydrolase (gloB) (Table A3-6 in appendix 3).

Methane release and change in apparent cellular P-requirement in Pelagibacter sp. str.

HTCC7211 when grown on Mpn

When grown with Mpn as the sole P source, Pelagibacter sp. str. HTCC7211 produced methane concomitantly with cell growth (Figure 4-4). Methane was not produced when cells were grown with Pi additions (Figure 4-4). Pelagibacter sp. str.

HTCC7211 responded linearly to Mpn additions in the absence of Pi (Figure 4-1) and growth rate was not dependent on the amount of Mpn in the media (Table A3-1 in appendix 3). Using the slope of the regression line in Figure 4-1, the apparent amount of

Mpn required per cell was calculated to be 11.4 attomoles Mpn cell-1 (34.3 attomoles

-1 cell less P than when grown on Pi (Figure 4-1)).

Polyphosphate production in Pelagibacter sp. str. HTCC7211

Electron-dense granules were observed in Pelagibacter sp. str. HTCC7211 cells that were grown on Mpn as a sole P-source. Similar granules were not observed in cells grown with Pi as the sole Pi-source, nor were they present in Pi-starved cells (Figure A3-6 in appendix 3).

Discussion

One of the objectives of this study was to determine the amount of P Pelagibacter cells need for growth. We found that when grown on AMS1 with excess organic

-1 amendments, Ca. P. ubique required 10.2 amol Pi cell (calculated from regression line in

Figure 4-1). 81% of the Pi-requirement could be accounted for by estimating the P contained in various cellular constituents: DNA (one chromosome cell-1) = 4.3 amol;

79 rRNA (200 ribosomes cell-1) = 1.5 amol; and phospholipids (134) 2.5 amol cell-1. The remaining 1.9 amol of P are conceivably contained in nucleotide pools (for nucleic acid synthesis and energy), mRNA transcripts, phosphorylated proteins and other biosynthetic intermediates. In contrast to Ca. P. ubique, we observed a larger apparent Pi-requirement

(that is, a lower cell yield per nmol Pi) for Pelagibacter sp. str. HTCC7211 when grown

-1 on Pi (45.7 amol cell ; calculated from the regression line in Figure 4-1).

To test whether the relatively high amount of Pi required by Pelagibacter sp. str.

HTCC7211 was due to polyphosphate accumulation, we assayed for polyphosphate using a non-quantitative scanning transmission electron microscopy (STEM) assay (similar to

(147)). Polyphosphate readily stains with Pb2+ giving rise to electron-dense granules when imaged by STEM (148). Pelagibacter sp. str. HTCC7211 cells grown in Pi-replete conditions or cells starved for Pi did not contain electron-dense granules, suggesting that long polyphosphate molecules are not synthesized in either condition (Figure A3-6 in appendix 3). Although polyphosphate accumulates in E. coli grown with excess Pi under some growth conditions, and when Pi-starved (149-151), it does not accumulate in E. coli cells grown in amino acid and Pi-replete conditions (149). In E. coli, polyphosphate accumulation is dependent on the PhoB response regulator which is expressed during Pi- stress (149). Pelagibacter sp. str. HTCC7211 did not differentially regulate phoB under

Pi-limiting conditions (Tables A3-5 & A3-6 in appendix 3), hinting that phoB transcription and polyphosphate accumulation might be controlled by alternate factors.

In addition to determining the cellular requirement for Pi, we sought to gain an understanding of the transcriptional responses to Pi-limitation in each organism. The pho-

80 regulon typically includes genes encoding a high affinity Pi transport system

(pstSCABphoU) and, as many organisms do (125, 127, 152, 153), Ca. P. ubique quickly upregulated the high affinity Pi-transport system when P-limited (Figure 4-3; Tables A3-2

& A3-4 in appendix 3). The remaining transcriptional responses in Ca. P. ubique (of both genes upregulated and downregulated) were reminiscent of transcription profiles observed in E. coli cells in which the “stringent response” was induced (154, 155). The stringent response is a bacterial stress response that results from the intracellular accumulation of the guanosine 3’,5’-bispyrophosphate (ppGpp) nucleotide (reviewed in

(156)). ppGpp is produced by Pi-starved E. coli cells (157) and is involved in regulating transcription of the pho regulon regulator, PhoB (149, 158). When the stringent response is artificially induced in E. coli, in addition to increased transcription of pstS and phoB, induction of multiple SOS-response genes is evident, including lexA, recA and umuD

(154), similar to what we observe here for Ca. P. ubique (Figure 4-3; Tables A3-3 & A3-

4 in appendix 3). The accumulation of ppGpp also represses the synthesis of tRNAs and rRNAs, and results in reduced transcription of ribosomal protein genes (reviewed in

(159)). Reduction of ribosomal protein transcripts was also observed in Pi-stressed Ca. P. ubique cultures (Figure 4-3; Tables A3-3 & A3-4 in appendix 3). Activation of E. coli’s stringent response represses some temperature shock proteins and chaperones, while inducing others (154, 160). Similar regulatory patterns were previously described for Pi- limited marine Prochlorococcus (125) and Synechococcus (127) bacteria, in which ribosomal protein and chaperone transcripts were differentially regulated under Pi-deplete conditions. Although the specific chaperone and heat shock proteins that are differentially

81 transcribed in Ca. P. ubique are not identical to those in other organisms, a conserved theme of stress-induced proteome re-configuration is evident, perhaps to prepare the cell for prolonged nutrient deprivation.

A stringent-like transcriptional response was conspicuously absent in

Pelagibacter sp. str. HTCC7211. Instead, an extended pho regulon was induced, akin to the canonical pho- regulon of E. coli, including genes for alternate P-source transport

(phnCDEE2 and pstSCABphoU), utilization (phnGHIJKLM) and other functions that are presumably advantageous in Pi-limited conditions. Similar to pho-regulon expression patterns of the phnCDE operon in E. coli (161), Pelagibacter sp. str. HTCC7211 upregulated the phnCDEE2 operon to a greater extent (20.3-fold mean across 4 genes) than the pstSCABphoU operon (3.2-fold mean across 5 genes) when Pi -stressed (Figure

4-3, Tables A3-5 & A3-6 in appendix 3).

Although it is unclear if the stringent-like response observed in Ca. P. ubique is a bona fide stringent response (i.e.- intracellular ppGpp accumulates), the absence of a stringent response-like gene expression pattern in Pelagibacter sp. str. HTCC7211 was unexpected given that nearly all of the differentially regulated genes in Ca. P. ubique are present in similar genomic context in Pelagibacter sp. str. HTCC7211. In E. coli, the stringent response alarmone, ppGpp, is kept in dynamic balance through synthesis and degradation by the paralogous proteins RelA and SpoT, respectively (159, 162).

However, most bacteria only contain SpoT (including Pelagibacter) (163), which catalyzes both the synthesis and degradation of ppGpp (159). The absence of a stringent- like response and phoB expression (which requires ppGpp (158)) in Pelagibacter sp. str.

HTCC7211 suggests that intracellular ppGpp is not accumulating. SpoT hydrolysis of ppGpp (to pp + Gpp) occurs in a HD-hydrolase domain of the SpoT protein (159, 164).

Interestingly, a gene predicted to encode a HD-hydrolase was significantly upregulated under Pi-limitation in Pelagibacter sp. str. HTCC7211 (Figure 4-3, Tables A3-5 & A3-6 in appendix 3). We propose that Pelagibacter sp. str. HTCC7211 may have recruited this

HD-hydrolase to its pho-regulon to combat elevated ppGpp levels that would normally trigger the stringent response, thus enabling cell proliferation under Pi-deficient conditions (provided an alternate P source is available).

Consistent with the induction of the phn genes in Pelagibacter sp. str.

HTCC7211, inorganic phosphite and organophosphorus sources were utilized in place of

Pi (Figure 4-2). In E. coli, genetic evidence indicated that the phnCDE operon encodes for a promiscuous ABC transporter that is able to transport phosphate-esters, phosphite, Pi and phosphonates and that the phnGHIJKLMN gene products confer the ability to metabolize phosphonates and phosphite for use as P-sources (165, 166). The phn operons are part of the pho-regulon in E. coli and are induced during Pi-stress (152). When

Pelagibacter sp. str. HTCC7211 is starved for Pi, multiple phn genes are induced (Figure

4-3, Tables A3-5 & A3-6 in appendix 3) and this induction is presumably responsible for the observed diauxic nature of growth when switching to growth on phosphonates, phosphite, or phosphate-esters as primary P-sources (Figure A3-1 in appendix 3).

Although Pelagibacter sp. str. HTCC7211 utilized the phosphate-esters G6P, R5P and P- ser for growth, the mechanism of utilization is unclear and might involve a novel phosphatase. Pelagibacter sp. str. HTCC7211 does not encode for known APases (phoA,

83 phoD and phoX homologs are absent) typically associated with phosphate-ester metabolism (167). There are two differentially regulated genes that might exhibit phosphatase activity and may explain the use of organic phosphates. The HD-hydrolase previously described in the context of ppGpp metabolism may also have phosphatase activity on phosphate esters (164). A second gene, divergently transcribed downstream of phnE2 and annotated as a ‘hypothetical protein’ (PB7211_828; Figure 4-3, Tables A3-5

& A3-6), was upregulated 12 to 15-fold in Pi-deplete conditions. PB7211_828 is unique only to Pelagibacter sp. str. HTCC7211 (a so-called ORFan (168)), and given its genomic position adjacent to an organophosphate transporter (phnCDEE2) and strong upregulation under Pi-deplete conditions, may be involved in phosphate-ester metabolism.

The utilization of Mpn for P by Pelagibacter sp. str. HTCC7211 may have an effect on atmospheric concentrations of the potent greenhouse gas methane (CH4). The open ocean’s surface is supersaturated with CH4 relative to the atmosphere in a phenomenon coined the ‘oceanic methane paradox.’ Aerobic metabolism of Mpn for use as a P-source results in the stoichiometric production of methane in Pseudomonas testosteroni (169) and has been linked to the phnCDE transport system and C-P lyase activity in E. coli (165). CH4 is released from Mpn-amended surface seawater samples containing native bacteria, showing that utilization of Mpn for P is a plausible source of the observed CH4 supersaturation (170). Previously, Trichodesmium erythraeum IMS101 was shown to utilize phosphonates and produce CH4 from Mpn (126, 171). We show that

Pelagibacter sp. str. HTCC7211 used Mpn in place of Pi (Figures 4-1 & 4-2) and when

84 doing so, released CH4 stoichiometrically (Figure 4-4). Published metagenomic analyses of multiple oceanic environments show that the Pelagibacter-like phn gene clusters are not universally distributed, but instead are more common in Pi-deplete waters (143, 144).

Recently, a novel Mpn biosynthetic pathway was discovered and characterized in

Nitrosopumilus maritimus, a member of the abundant Marine Group I ammonia- oxidizing archaea (172). Marine group I archaea are ubiquitous, have a large global population size and in marine ecosystems, can account for up to 30% of all cells below the ocean’s thermocline (18, 173), suggesting that archaeal-derived Mpn may be an important phosphorus source in the sea. Given the abundance of Pelagibacter in Pi and nutrient deplete waters such as the Sargasso Sea, Pelagibacter sp. str. HTCC7211-like cells may play a central role in the metabolism of archaeal-derived Mpn and the release of aerobically produced CH4 from Pi-deplete ocean systems.

In addition to releasing CH4, the physiology of Pelagibacter sp. str. HTCC7211 changed when it was grown on Mpn in ways that implied Mpn was utilized more efficiently than Pi. For example, the apparent cellular P-demand for Pelagibacter sp. str.

-1 -1 HTCC7211 decreased from 45.7 amol cell (when grown on Pi) to 11.4 amol cell when cells were grown on Mpn (calculated from the regression lines in Figure 4-1). This implies that Pelagibacter sp. str. HTCC7211 may decrease their cellular P-requirement when growing on Mpn. Some P-starved bacteria reduce their cellular P-requirement in low- Pi environments by replacing the P-containing polar head groups on lipids with non-

P containing moieties such as sulphoquinovosyldiacylglycerol (SQDG) (134), glycolipids

(174) or amino acid lipids (175).

Previous metabolic reconstructions did not find support for the hypothesis that

Pelagibacter sp. str. HTCC7211 can modify their lipid polar head groups; canonical genes encoding proteins used in the synthesis of P-free betaine lipids, SQDG-lipids or glycolipids were absent. However, a four-gene cluster containing two genes annotated as putative hemolysins, a probable glycosyltransferase and a calcineurin-like metallophosphoesterase, were differentially regulated under Pi-stress in Pelagibacter sp. str. HTCC7211 and might be involved in lipid polar head group remodeling (Figure 4-3;

Tables A3-5 & A3-6 in appendix 3). The first step of lipid remodeling in the α- proteobacteria Sinorhizobium meliloti, is to cleave phosphatidylcholine and phosphatidylethanolamine from phospholipids with phospholipase C (PlcP) to form diacylglycerol plus phosphocholine or phosphoethanolamine, respectively (137). A search of the Pelagibacter sp. str. HTCC7211 genome for homologs to S. meliloti’s PlcP, revealed that the calcineurin-like metallophosphoesterase gene that was upregulated

-87 under Pi-stress had a high percent identity (118/251 identities; blastp e-value: 2 × 10 ), suggestive of similar function. Immediately downstream of the putative plcP gene in a probable operon, is a group 1 glycosyltransferase (COG438; rfaG). Glycosyltransferases transfer saccharide groups from nucleotide sugars to a variety of acceptor molecules. For example, during the synthesis of SQDG-lipids, a glycosyltransferase is responsible for the transfer of a modified saccharide (sulphoquinovose) to diacylglycerol (176). Given the genomic context of the Pelagibacter sp. str. HTCC7211 glycosyltransferase adjacent to the probable phospholipase C gene, we predict that it may be involved in the glycosylation of the diacylglycerol resulting from PlcP-mediated cleavage of

86 phospholipids. A putative alginate lyase is also upregulated during Pi stress (Table A3-6 in appendix 3), hinting that the saccharide used by this glycosyltransferase may be an alginate derivative. Two divergently transcribed genes, immediately downstream of the glycosyltransferase, annotated as putative hemolysins, were also upregulated during Pi stress (Figure 4-3). These genes belong to the same cluster of orthologous groups (COG) as homologs to OlsB from S. meliloti (COG3176), which catalyze the first step in the synthesis of Pi-free ornithine lipids in Pi-stressed cultures (175, 177). We propose that this four-gene cluster acts to remodel the lipid composition of Pelagibacter sp. str.

HTCC7211 cells during Pi-stress to replace phosphorus-containing polar head groups with saccharides and/or ornithine (or other amino acids) in order to reduce the cellular P- requirement (Figure A3-5 in appendix 3). Previously, it was estimated that Pelagibacter

-1 cells grown in Pi-replete conditions contain 2.5 amoles P cell in their lipids (134), thus replacement of P-lipids with non-P lipids may save up to 25% of their total P-requirement

(assuming 10-11 amol P cell-1).

Polyphosphate is a linear polymer of Pi-residues that has been associated with many cellular activities including P-storage, stationary-phase survival, pathogenicity and

DNA uptake (reviewed in (151)). The precise physiological role of polyphosphate in marine bacterioplankton is unclear, however metagenomic analysis of polyphosphate genes in the GOS database suggests that polyphosphate may be important to microbial survival in low-P environments (143, 145). When Pelagibacter sp. str. HTCC7211 cells previously grown with Mpn as a sole P-source were transferred into P-free medium, cell yields were 30 times higher than cells previously grown on Pi, suggestive of potential P

87 stores in the Mpn-grown cells. Indeed, electron-dense granules characteristic of polyphosphate were observed in Mpn-grown cells (Figure A3-6 in appendix 3) (147).

Assuming insignificant external sources or carryover of P, the internal stores of

-1 phosphate in the cells used to inoculate the flasks was calculated to be ~660 amol P cell .

Although the polyphosphate synthesis and utilization genes (ppx and ppk, respectively) were not differentially expressed under Pi-limiting conditions, we did not explore gene regulation when growing on organic P-sources. Synthesis of polyphosphate from organic

P-sources has not been intensely investigated, although growth experiments in the diazotroph Trichodesmium IMS101 suggested that P from P-esters could be used to form polyphosphate (178).

We undertook these experiments to gain an understanding of how the genetic differences related to P metabolism are manifested physiologically in two similar organisms and how Pelagibacter may participate in oceanic P-cycling. Our results show that despite the high degree of phylogenetic similarity between Ca. P. ubique and

Pelagibacter sp. str. HTCC7211 by both 16S ribosomal sequence comparisons (22) and whole genome analysis (29, 179), non-conserved strategies resulting in distinct P- utilization physiotypes were evident. Ca. P. ubique exhibits a physiotype that is characterized by primary reliance on Pi for P nutrition. This interpretation is highlighted by the observation that Ca. P. ubique does not use organic P-esters or phosphonates effectively (Figure 4-2). Further support for this interpretation can be linked to gene expression changes under Pi-limiting conditions that show a rapid and strong upregulation of the high-affinity Pi transport system followed by expression patterns

88 indicative of a stringent-like stress response, suggesting that Pi-depletion is recognized as a global stress inducer (Figure 4-3, Tables A3-2 – A3-4 in appendix 3). In contrast to this,

Pelagibacter sp. str. HTCC7211’s physiotype is characterized by the effective and broad use of organic phosphorus sources, including phosphonates, when Pi is limiting (Figures

4-1 & 4-2). Gene expression differences support this interpretation; during Pi-starvation, genes encoding proteins conferring the use of alternate P-sources showed the greatest degree of induction (Figure 4-3, Tables A3-5 & A3-6 in appendix 3). Differential regulation of genes involved in the stringent response in Pelagibacter sp. str. HTCC7211 was absent.

The different physiotypes we describe probably reflect cellular adaptations pertinent to the environments from which each organism was isolated. For example, Ca.

P. ubique was isolated from the northeast Pacific Ocean off the coast of Oregon, USA

(23) where Pi concentrations are usually sufficient to support the native plankton populations and rarely drop below 100 nM. The high-affinity Pi transporter encoded by

Ca. P. ubique may be transiently advantageous in situations where modulation of Pi- uptake increases competitiveness, possibly due to Pi-depletion by other plankton.

However, genome sequences of multiple Pelagibacter isolates from the northeast Pacific

Ocean show that the high affinity Pi transport system is not required for environmental persistence or for growth in vitro (23). Although we are able to grow one such isolate in

AMS1, (‘Ca. Pelagibacter ubique’ str. HTCC1002) the organic growth requirements have not been optimized to the extent required to practically conduct the in vitro experiments described here. Interestingly, all of the Pelagibacter genomes lacking a pstSCAB-phoUB

89 operon contain homologs to the Ca. P. ubique gene SAR11_1181 (Figure 4-3).

Metagenomic analyses show that Pelagibacter genes homologous to SAR11_1181 are enriched in the Pacific Ocean compared to the Atlantic Ocean (143). Analysis of the

SAR11_1181 protein sequence with the RaptorX protein structure prediction web server

(180) revealed a domain with distant similarity (2.70 × 10-6 across amino acid residues

21-273) to the PHO4-family of Pi-cation symporters in Neurospora crassa (181). We observed significant upregulation of SAR11_1181 in Pi-deplete conditions, which is also suggestive of a role in Pi-acquisition (Figure 4-3; Tables A3-3 & A3-4 in appendix 3).

We suspect that SAR11_1181 encodes a Pi-symporter that may co-transport cations with

Pi, similar to the PitA transporter in E. coli (182) and may explain the absence of identifiable Pi transporters in multiple Pelagibacter genomes (29).

Organisms that thrive in the Sargasso Sea should be adapted to low Pi-levels and high concentrations of dissolved organic-P (DOP) and arsenate. The results presented here suggest that Pelagibacter sp. str. HTCC7211 is adapted to life in the Sargasso Sea.

Pelagibacter sp. str. HTCC7211 upregulated genes required for DOP utilization (Figure

4-3, Tables A3-5 & A3-6 in appendix 3) and effectively used DOP in place of Pi (Figure

4-2). DOP is more abundant than Pi in many oligotrophic environments, including the

Sargasso Sea (122, 183-185), thus our in vitro experimental results showing efficient use of DOP are compatible with the relative availability of Pi and DOP in Pelagibacter sp. str. HTCC7211’s native habitat. In the surface waters of the Sargasso Sea, arsenic concentrations often exceed those of Pi (186). Arsenate is a Pi-analog and commonly enters cells through Pi-transport systems. During Pi-limitation, we observed upregulation

90 of a two-gene operon consisting of genes encoding an arsenate reductase (ArsC) and a glpF-like aquaglyceroporin (AqpS). In S. meliloti, ArsC and AqpS were shown to facilitate arsenate reduction to arsenite (ArsC) and passive efflux from the cell (AqpS) as a mechanism of arsenate detoxification (187). We suspect that these genes function similarly in Pelagibacter sp. str. HTCC7211 and are induced to combat elevated levels of arsenate transported into the cell by the promiscuous phnCDEE2 transporter in low Pi environments.

Previous metaproteomic analysis of peptides isolated from native Sargasso Sea bacterioplankton identified a number of peptides that mapped to genes that were differentially expressed in our studies, implying that our in vitro experimental results have relevance in situ. For example, peptide spectra from the Sargasso Sea mapped to translations of Pelagibacter sp. str. HTCC7211’s pstS, pstA, pstB, phoU, phnC, phnD, phnE, PB7211_828 and PB7211_757 genes and other genes that were differentially expressed in Ca. P. ubique (104). Although the composition of DOP in the environment has not been fully chemically defined, we show that Pelagibacter sp. str. HTCC7211 efficiently used a broad range of organic-P compounds for P-nutrition. Previously, it was suggested that the range of phosphonate compounds that can be utilized by E. coli was limited by the specificity of transport by the phnCDE transporter not the activity C-P lyase (188, 189). If true, P-utilization niche specificity in Pelagibacter, and other organisms encoding the phnCDE transporter, might vary depending on the evolution of transport specificity of the phnCDE transporter.

The term ‘ecotype’ is commonly used to describe groups of closely related bacteria within which ecologically distinct populations are observed. Separate ecotypes are assumed to differ in traits conferring niche specificity. Previously, the presence of P- utilization gene suites was described to be incongruent with phylogenetic placement in

Prochlorococcus ecotypes (125, 143). Physiological assessments of P-utilization patterns in Prochlorococcus isolates also showed that different P-utilization physiotypes were evident (128, 190).

The Ia-subclade of the Pelagibacteraceae has been described as an ecotype based on 16S rRNA sequence analysis and tRFLP patterns that clearly show reproducible and coherent spatiotemporal distributions in the Sargasso Sea (191, 192). The data we present here show that within the SAR11 Ia-subclade ecotype there is substantial physiological and transcriptional variation related to P-utilization. Although we describe two different physiotypes, it is not clear how many different P-utilization physiotypes exist in the environment and whether the diversification of P-utilization physiotypes is adaptive or neutral. Metagenomic analyses of the Sargasso Sea show higher abundances of pstSCAB- phoU genes relative to C-P lyase and phnCDEE2 genes, suggesting that not all

Pelagibacter cells in the Sargasso Sea contain the extended pho-regulon that we describe here and implying that different P-utilization physiotypes co-exist in nature (143, 144).

Analysis of multiple Pelagibacter genomes suggests that there are additional P-utilization physiotypes that may vary in the scope of phosphonate utilization (29). Thus, further characterization of Pelagibacter P-utilization physiotypes may unveil interesting and

92 geochemically relevant findings pertinent to the global P cycle and help to further refine the concept of an ecotype.

Pelagibacter sp. str. HTCC7211 + Mpn 12

) Pelagibacter sp. str. HTCC7211 + Pi -1 Ca. P. ubique HTCC1062 + Pi 2 10 y=87,542x + 1,980,400; R =0.995 y=21,900x + 7,883; R2=0.991

cells ml 2

6 y=97,669x + 365,626; R =0.999 8

2 maximum cell density (× 10

0 0 10 20 30 40 50 60 70 80 90 100 concentration of P-source (nmol L-1)

Figure 4-1: Linear dose responses of Ca. P. ubique and Pelagibacter sp. str. HTCC7211 to P-sources. Each point is the mean of the maximum density achieved for biological replicates ± 1.0 s.d. (n=3). Linear regression through P-source additions of 1.0 nM, 3.3 nM, 10 nM, 33.3 nM and 100 nM are shown. Abbreviations: Mpn – methylphosphonic acid; Pi – inorganic phosphate. Full growth curves are presented in Figures A3-2 - A3-4 in appendix 3.

) Ca. P. ubique HTCC1062 -1 Pelagibacter sp. str HTCC7211 4 cells ml 7 3

max cell density (× 10 0 TP dTTP A Mpn d dGTP dCTP 2-AEP (neg ctr) P-serine (pos ctr) i Phosphite P i -P Glucose-5PRibose-5P +

Figure 4-2: P-source utilization by Ca. P. ubique and Pelagibacter sp. str. HTCC7211. Bar height is the mean of the maximum density achieved for biological replicates ± 1.0 s.d. (n=3). denotes diauxic growth pattern (diauxy shown in Figure A3-1 in appendix 3). Growth medium was not amended with Pi in the negative control. All other treatments had 1.0 µM of potential P source added.

A Ca. Pelagibacter ubique HTCC1062 ftsZ mraZ ipbA trmD rpsP umuD recA lexA 0964 0965 rpoA rpsK rpsS rplB fusA 1164 4 h 20 h 38 h 4 h 20 h 38 h -4.0 -1.5 1.5 >30.0 phoB phoU pstB pstA pstC pstS phoR 1181 amount regulated (fold) high-a!nty phosphate transport B Pelagibacter sp. str. HTCC7211

lipid modi"cation high-a!nty Pi transport poly-P metabolism LCB5 plsC rfaG hemolysins phoR 78 pstS pstC pstA pstB phoU phoB sixA ppk ppx NUDIX 913 72 h 96 h

72 h 96 h

phnX aqpS arsC HD-hyd araJ 828 phnE2 phnE phnD phnC phnF G phnH phnI phnJ phnK phnL phnN 764 phnM 1178 organic-P transport C-P lyase

Figure 4-3: Comparative genomics of gene expression during Pi-starvation in Ca. P. ubique and Pelagibacter sp. str. HTCC7211. Time elapsed since resuspension is listed to the left of each gene row. A) Select genes differentially regulated in Ca. P. ubique. Gene names, annotations and actual expression values are listed in Tables A3-2 – A3-4 in appendix 3. B) Select genes differentially regulated in Pelagibacter sp. str. HTCC7211. Gene names, annotations and actual expression values are listed in Tables A3-5 & A3-6 in appendix 3. Vertical black lines represent a lack of genomic synteny. Genes shaded grey are not differentially expressed.

1.6 A + Pi 1.4 ) + Mpn -1 Predicted CH 1.2 4

1.0

0.8 evolved ( μ mol L evolved 0.64

0.4 total CH total 0.2

0 B - P 8.0 i

) - Mpn 7

- No Pi × 10

-1 6.0

4.0

2.0 cell density ml cell (cells

0 0 5 10 15 20 25 30 35 40 time (days)

Figure 4-4: Methane (CH4) production by Pelagibacter sp. str. HTCC7211 utilizing Mpn as a sole P-source. A) CH4 production by Pelagibacter sp. str. HTCC7211 cultures grown with Mpn (open triangles) or Pi (open circles). Open circles and triangles are the mean amount of CH4 measured from biological replicates ± 1.0 s.d. (n=3). Theoretical CH4 production based on cell densities in (B) assuming 11.4 amol Mpn cell-1 (inverted filled triangles and dashed line). B) Cell densities from single bottles sacrificed for cell density measurements for Pelagibacter sp. str. HTCC7211 cultures grown with no Pi (×), Mpn (filled triangles) or Pi (filled circles).

Chapter 5 Analysis of the Vitamin B5, B6, B7 And B12 Requirements of ‘Candidatus Pelagibacter ubique’ HTCC1062: Evidence for B7 Auxotrophy.

Paul Carini1, Emily O. Campbell1, Sergio Sañudo-Wilhelmy2 and Stephen Giovannoni1

1Department of Microbiology, Oregon State University Corvallis, OR 97331

2Department of Biological Sciences, Marine Environmental Biology, and Earth Science, University of Southern California, Los Angeles, CA 90089

Abstract

‘Candidatus Pelagibacter ubique’ strain HTCC1062 is a member of the globally distributed and ecologically important SAR11 clade of marine bacteria. Previously, in metabolic reconstructions of the Ca. P. ubique genome it was noted that the biosynthetic pathways for the essential B-vitamins pantothenate (B5), pyridoxine (B6), biotin (B7) and cobalamin (B12) were absent or incomplete. Despite the absence of genes encoding proteins required in these biosynthetic pathways, genes coding for proteins that utilize vitamin B5 (or its derivative, coenzyme A), B6 and B7 were identified in the Ca. P. ubique genome, suggesting that these vitamins are required for normal cellular metabolism.

Although four of the five genes encoding proteins required for vitamin B7 synthesis are absent, a probable transporter for B7 was identified, hinting that Ca. P. ubique obtains B7 from the environment. We confirmed this experimentally by limiting the growth of Ca. P. ubique by excluding vitamin B7 in a defined growth medium and showing that picomolar additions of vitamin B7 restore cell yields. Because most, but not all, enzymes required for the biosynthesis of vitamin B5 were identified, we hypothesized that Ca. P. ubique was either auxotrophic for vitamin B5 or utilized novel enzymatic reactions to complete

B5 synthesis. To test the hypothesis that Ca. P. ubique is auxotrophic for vitamin B5, we attempted to limit the growth of Ca. P. ubique cultures by excluding vitamin B5. We were unable to limit Ca. P. ubique growth with vitamin B5, however, we observed a significant growth rate reduction in the absence of vitamin B5. These results are consistent with the conclusion that Ca. P. ubique is bradytrophic for vitamin B5, suggesting synthesis is possible at the expense of more rapid growth. Re-analysis of the Ca. P. ubique genome

99 shows that an alternate vitamin B5 pathway is likely present. Recent work in Escherichia coli uncovered a novel biosynthetic route for de novo vitamin B6 synthesis from the glycolytic/gluconeogenic intermediate 3-phosphoglycerate. Re-analysis of the Ca. P. ubique genome identified this new vitamin B6 biosynthetic pathway, suggesting de novo synthesis of vitamin B6 is possible. Although only two of the 18+ genes required for vitamin B12 synthesis were identified in the Ca. P. ubique genome, no genes encoding

B12-dependent enzymes were identified, implying this vitamin is neither produced nor used by Ca. P. ubique. We propose that reactions commonly catalyzed by B12-dependent enzymes are instead performed by B12-independent analogs encoded within the Ca. P. ubique genome. These studies are an important step in understanding the unusual organization of metabolic pathways that result in the unique oligotrophic physiology of

Ca. P. ubique. Additionally, the discovery of vitamin B7 auxotrophy suggests that environmental distributions of vitamin B7 have the potential to affect SAR11 population sizes.

Introduction

Patterns of B-vitamin production and consumption in nature are poorly understood, as are the metabolic ramifications of vitamin deficiency, at both the cellular and ecosystem levels. Traditionally, measurements of vitamins in the sea are conducted using bioassay techniques that are relatively insensitive and subject to errors caused by organismal responses to vitamin analogs or degradation products, in addition to the intact

B-vitamins themselves (193-195). Sensitive and accurate direct measurements of B- vitamins extracted from seawater have been determined by high performance liquid

100 chromatography (HPLC) coupled to either a UV-vis detector (196, 197) or tandem mass spectrometer (81). These measurements showed that some B-vitamins have complex distribution patterns in the sea. An understanding of the vitamin requirements and vitamin production by microbes in the environment is necessary to contextualize the unusual distribution patterns (84).

SAR11 chemoheterotrophic marine bacteria are found worldwide and make up a large percentage of cells in marine photic zones (17), where they are important in the cycling of carbon and sulfur (26, 28, 32, 37, 88). ‘Candidatus Pelagibacter ubique’ str.

HTCC1062 (Ca. P. ubique) was the first member of the SAR11 clade to have its complete genome sequenced (25). Metabolic reconstruction of the Ca. P. ubique genome found that B-vitamin synthesis pathways for B5, B6, B7 and B12 are incomplete. Despite the incompleteness of these pathways, genes coding for proteins that require B5, B6 and

B7 were identified, suggesting that these vitamins are necessary for normal metabolism in

Ca. P. ubique (Figure 5-1). Knowledge of the vitamin requirements of Ca. P. ubique is potentially important to help understand SAR11 population sizes in nature, which may be controlled by vitamin supply, and for further refinement of current models of Ca. P. ubique metabolism.

Biosynthetic pathways of vitamin synthesis can be difficult to predict from in silico metabolic models derived from genome sequences. Most knowledge of vitamin synthesis pathways is based on data from a small number of model organisms and divergence from these canonical synthesis pathways at one or multiple reaction steps is not uncommon (198). Vitamin biosynthetic pathways can be complex (for example

101 vitamin B12; see Figure 5-5) and often utilize unusual and distinct chemistries during their synthesis (198). Enzymes involved in vitamin biosynthesis may have multiple cellular functions, including roles in other metabolic pathways, multiple roles in the synthesis of a single vitamin, or in the synthesis of other vitamins, and in the case of vitamin B1, the synthesis of itself (198).

Vitamin B5 biosynthesis

Vitamin B5 (pantothenate) is required for the synthesis of coenzyme A, an acyl carrier that, among other reactions, transfers pyruvate-derived carbon to the tricarboxylic acid cycle, where it can be oxidized for energy. Vitamin B5 is also the precursor to phosphopantetheine, a prosthetic group on acyl carrier proteins used in the synthesis of fatty acids. Vitamin B5 is formed from the linkage of pantoate and β-alanine. Pantoate is formed from two pyruvate molecules in five enzymatic steps, four of which are shared with the initial stages of valine and leucine biosynthesis. The divergence between these pathways occurs at the metabolic intermediate 2-oxioisovalerate (Figure 5-2), thus the first committed step in pantoate biosynthesis is the hydroxymethylation of 2- oxioisovalerate to dehydropantoate (199). Dehydropantoate is reduced to pantoate via

PanE in E. coli, however some organisms utilize IlvC to exclusively perform this reaction

(200, 201). In E. coli, β-alanine is formed the decarboxylation of aspartate by PanD

(202). In other organisms, β-alanine can be derived from a number of compounds including spermine (203), propionate (204) and uracil (205).

102

Vitamin B6 biosynthesis

Vitamin B6 (pyridoxine) is an important cofactor in amino acid aminotransferase reactions, but is also used by some amino acid racemases and sulfinate desulfinases.

Vitamin B6 is synthesized by joining deoxy-xylulose-5-phosphate (dXP) with 3-amino-2- oxypropyl-phosphate (Figure 5-3). dXP is derived from pyruvate and glyceraldehyde-

5’phosphate by the deoxyxylulose-phosphate synthase (Dxp). 3-amino-2-oxypropyl- phosphate is derived from O-phospho-4-hydroxythreonine (4HT-P). In the canonical vitamin B6 biosynthetic pathway in E. coli, 4HT-P is formed from D-erythrose-4- phosphate, a pentose pathway intermediate (Figure 5-3), however in some pyridoxine synthesis mutants other compounds can substitute (206-208).

Vitamin B7 biosynthesis

Vitamin B7 (biotin) is a sulfur-containing coenzyme that is universally required for the carboxylation of acetyl-CoA to malonyl-CoA in the first committed step of fatty acid biosynthesis and other (de)carboxylation reactions. Canonical microbial synthesis of vitamin B7 starts with pimeloyl-CoA and proceeds through four enzymatic reactions that require inputs of alanine, S-adenosylmethionine and ATP (reviewed in (209)) (Figure 5-

4). In bacteria, the genes involved in vitamin B7 biosynthesis are co-transcribed in an operon that is under the control of the bi-functional BirA protein. BirA activates vitamin

B7 with ATP to form biotinyl-5’-adenylate and catalyzes the addition of vitamin B7 to biotinylated proteins (e.g. - the acetyl-CoA carboxylase). When levels of biotinyl-5’- adenylate are elevated, BirA binds biotinyl-5’-adenylate and acts as a repressor of vitamin B7 synthesis (210-215).

103

Vitamin B12 biosynthesis

Vitamin B12 (cobalamin) is a soluble organic molecule containing a single cobalt ion chelated within a tetrapyrrole structure. Vitamin B12 can be a part of many cellular processes, including methionine biosynthesis, ribonucleotide reduction and some methyltransferase and isomerase (mutase) reactions (216). The de novo biosynthesis of vitamin B12 in bacteria involves ~20 enzymes to convert the heme biosynthetic intermediate uroporphyrogen –III (uro’-III) to the active vitamin (217-220) (Figure 5-5).

In short, the corrin ring is initiated by the methylation of uro’-III to precorrin-2. There are two paths by which vitamin B12 corrin rings can be biosynthesized from precorrin-2: the oxygen-dependent (aerobic) and oxygen independent (anaerobic) pathways. In the aerobic pathway, corrin rings are formed prior to the insertion of cobalt to form cob(II)yrinic acid a,c diamide (Figure 5-5). In the anaerobic pathway, cobalt is initially inserted into the precorrin-2 molecule and the corrin rings are formed subsequently to form cob(II)yrinic acid a,c diamide (Figure 5-5). After the formation of cob(II)yrinic acid a,c diamide, the aerobic and anaerobic pathways share the same biosynthetic processes to the form the active vitamin (Figure 5-5).

B-vitamin biosynthesis in Pelagibacter

Metabolic reconstruction of the Ca. P. ubique genome found that the biosynthetic machinery for de novo vitamin B5, B6, B7 and B12 synthesis was incomplete (25). We investigated these predicted B-vitamin deficiencies by using a combination of metabolic reconstruction and growth experiments on a defined artificial medium. We show that Ca.

P. ubique is auxotrophic for vitamin B7 and bradytrophic for vitamin B5. We identified a

104 probable vitamin B7 transporter and propose a complete biosynthetic pathway for B5. We identified a recently characterized (in E. coli) vitamin B6 biosynthetic pathway, and did not observe growth defects upon vitamin B6 exclusion. Finally, we discuss the lack of vitamin B12 metabolism and identify B12-independent enzymes that probably catalyze reactions that usually employ vitamin B12.

Methods

Organism source

Ca. P. ubique HTCC1062 was revived from 10% glycerol stocks and propagated in artificial medium for SAR11 (AMS1) (89) amended with oxaloacetate (1 mM), glycine

(50 µM), methionine (50 µM) and FeCl3 (1 µM). B-vitamins were added as indicated in figure legends and text.

Chemicals

Cultivation details

All cultures were grown in acid-washed and autoclaved polycarbonate flasks and incubated at 20°C with shaking at 60 RPM in the dark. Cells for counts were stained with

105

SYBR green I and counted with a Guava Technologies flow cytometer at 48-72 h intervals as described elsewhere (26).

Vitamin-limitation procedure

To dilute out the vitamins routinely added to SAR11 cultures, Ca. P. ubique was transferred to AMS1 amended with nutrients, but without B-vitamins, a minimum of 10 times (>1000 generations). After three passages (>30 generations) truncated cell growth was consistently observed (104 cells ml-1 to 107 cells ml-1. For example, see Figure 5-6).

This truncation was attributed to thiamine-limitation in chapter 3.

Trace vitamin measurements

Vitamins were measured in un-inoculated medium using the solid-phase extraction plus HPLC tandem mass spectrometry method described in (81).

Vitamin B7-scrubbing

Trace contaminants of vitamin B7 were removed from growth medium by adding streptavidin-coated magnetic beads (0.5 µL beads L-1; rinsed 3X with Tris-EDTA) (New

England BioLabs, Ipswich, MA) and incubating at 16°C for 24 hours with gentle shaking. After incubation, beads were centrifuged out of the medium and sequestered at the bottom of the centrifuge tube with a magnet while the supernatant (the scrubbed growth medium) was transferred into a new container for use in experiments.

106

Results

Metabolic reconstruction of vitamin B5 biosynthesis

Genes encoding homologs to all but one gene (PanE) used for the synthesis of pantoate in E. coli were identified in the Ca. P. ubique genome (Figure 5-2). Specifically, genes encoding homologs for the i) acetolactate synthase complex (IlvH – small subunit

& IlvB – large subunit); ii) ketol-acid reductoisomerase (IlvC); dihydroxy-acid dehydratase (IlvD) and 3-methyl-2-oxobutanoate hydroxymethyltransferase (PanB) were identified (Figure 5-2). Although homologs to PanE were not identified in Ca. P. ubique,

IlvC has been shown to have a promiscuous activity that can replace the function of PanE

(200, 201). Ca. P. ubique encodes for IlvC, therefore we propose that IlvC is functional ketopantoate reductase. In E. coli, β -alanine formation results from PanD-catalyzed decarboxylation of aspartate. Ca. P. ubique lacks PanD homologs, thus the vitamin B5 biosynthetic pathway appears to be incomplete because of a deficiency in β-alanine synthesis (Figure 5-2).

Metabolic reconstruction of vitamin B6 biosynthesis

Ca. P. ubique is missing the canonical route for 4HT-P synthesis via the pentose- phosphate pathway, intermediate D-erythrose-4-phosphate (by action of the D-erythrose-

4-phosphate dehydrogenase (Epd), erythronate-4-phosphate dehydrogenase (PdxB) and phosphoserine aminotransferase (SerC); Figure 5-3). A recently discovered (in E. coli

(206)) alternate 4HT-P synthesis pathway from glycolaldehyde and glycine appears to be present in Ca. P. ubique. Enzymes involved in the alternate pathway include the L-

107 threonine aldolase (LtaE) and homoserine kinase (ThrB) (Figure 5-3). In E. coli, glycolaldehyde formation may have multiple routes, including one starting from 3- phosphoglycerate (Figure 5-3) (206). Ca. P. ubique is missing a gene encoding YeaB, a protein to derive glycolaldehyde from 3-phosphoglycerate, but encodes probable homologs for LtaE and ThrB (Figure 5-3).

Ca. P. ubique encodes for the remaining proteins necessary to form vitamin B6 from 4HT-P. Genes encoding homologs for the 4HT-P dehydrogenase (PdxA); pyridozine-5-phosphate synthase (PdxJ) and the pyridoxamine-5-phosphate oxidase

(PdxH) were identified (Figure 5-3).

Metabolic reconstruction of vitamin B7 biosynthesis

Ca. P. ubique encodes for only one of the five genes involved in vitamin B7 biosynthesis in E. coli: the adenosylmethionine-8-amino-7-oxononanoate transaminase

(BioA). The remaining genes encoding required vitamin B7 biosynthetic proteins in E. coli – i) biotin synthase (BioB); ii) the predicted methyltransferase of biotin synthesis

(BioC); iii) dethiobiotin synthase (BioD); and iv) 8-amino-7-oxononanoate synthase

(BioF) were not identified in Ca. P. ubique (Figure 5-4). A gene encoding a probable vitamin B7 transporter (bioY, SAR11_0202) was identified in Ca. P. ubique (Figure 5-4).

Metabolic reconstruction of vitamin B12 biosynthesis

Only two of the more than 18 genes that encode proteins required for vitamin B12 synthesis were identified in the Ca. P. ubique genome - cobS and cobT, encoding cobalt chelatase subunits (Figure 5-5). No genes predicted to encode vitamin B12 transporters

108 were identified in the Ca. P. ubique genome. Further investigation of the Ca. P. ubique genome did not find any genes coding for enzymes predicted to require vitamin B12 as a cofactor (Table 5-1). In some instances, we identified vitamin B12-independent isozymes that may perform equivalent functions in vivo (Table 5-1).

Measurements of B-vitamins in growth medium

To determine if there were trace B-vitamins in our ‘vitamin-free’ growth medium we measured the B-vitamin concentrations of vitamins B6, B7 and B12 using the method described in (81) (no assay for B5 in seawater has been developed). There were contaminating amounts of vitamin B6 (30.921 ± 1.817 pM; average ± s.d., n=3) and B7

(96.535 ± 10.08 pM; average ± s.d., n=3). Vitamin B12 was not detected (limit of detection 0.18 pM (81)).

Growth responses to vitamin exclusions

When 1.0 µM each of vitamin B5 and B7 were added to the growth medium, maximal cell densities of 1.22 ± 0.51 × 107 cells ml-1 (herein presented as mean ± s.d., n=3) were attained and cells grew at a specific growth rate of 0.52 ± 0.01 day-1 (herein presented as mean ± s.d., n=3) (Figure 5-6). When 1.0 µM of vitamin B7 alone was added to the growth medium, maximal cell densities of 1.06 ± 0.52 × 107 cells ml-1 were attained and cells grew at a specific growth rate of 0.34 ± 0.03 day-1 (Figure 5-6). When

1.0 µM of vitamin B5 alone was added to the growth medium, maximal cell densities of

4.75 ± 1.25 × 106 cells ml-1 were attained and cells grew at a specific growth rate of 0.37

± 0.04 day-1 (Figure 5-6). When no B-vitamins were added to the growth medium,

109 maximal cell densities of 3.59 ± 1.15 × 106 cells ml-1 were attained and cells grew at a specific growth rate of 0.39 ± 0.01 day-1 (Figure 5-6).

Biotin limitation of Ca. P. ubique

To reliably demonstrate a vitamin B7 requirement, it was necessary to scrub contaminating vitamin B7 from the growth medium (see methods for details). When vitamin B7 was added back to the scrubbed medium, growth resumed, indicating that the scrubbing procedure was not detrimental to the growth of Ca. P. ubique (Figure 5-7a).

When vitamin B7 was excluded from vitamin B7-scrubbed medium amended with organic carbon, sulfur and thiamine, cell yields of 1.87 × 106 ± 0.07 cells ml-1 were attained.

When the media was amended with 1 µM vitamin B7, cell yields increased to

2.64 ± 0.17 × 108 cells ml-1 (mean ± s.d., n=3) (Figure 5-7b). Cell yields responded incrementally to picomolar additions of vitamin B7 to the scrubbed medium (Figure 5-

7b).

Identification of vitamin B12-requiring enzymes and B12-independent analogs

No genes encoding vitamin B12-requiring enzymes were identified in the Ca. P. ubique genome (Table 5-1). Instead, we identified genes encoding vitamin B12- independent enzymes that catalyze reactions that are typically vitamin B12-dependent

(Table 5-1).

Conclusions

The data presented here are one of the final steps in the elucidation of the organic nutrient requirements of Ca. P. ubique, and help resolve the last of the metabolic

110 anomalies first presented with the Ca. P. ubique genome publication in 2005 (25). In addition to the unusual requirement for the thiamine precursor HMP presented in chapter

3, we show that Ca. P. ubique has an absolute requirement for vitamin B7 (Figure 5-7a,b).

When vitamin B7 is present, addition of vitamin B5 increases growth rate (Figure 5-6), suggesting that Ca. P. ubique is bradytrophic for vitamin B5 in a vitamin B7-dependent manner. We have identified the probable transporter for vitamin B7 (bioY; Figure 5-4). In

Rhodobacter capsulatus, BioY was shown to be a high capacity biotin transporter (221).

Biosynthesis of vitamin B5 appears to be dependent on a source of vitamin B7.

When vitamin B5 is added together with vitamin B7, growth rates nearly double (Figure

5-6), suggesting that vitamin B7 availability may play a role in vitamin B5 synthesis. We propose that the synthesis of β-alanine is dependent on malonyl-CoA, which requires vitamin B7 for synthesis (Figures 5-1 & 5-2). MmsA is a methylmalonate-semialdehyde dehydrogenase that can act on malonyl-CoA together with NADPH and H+, to form 3- oxopropanoate (222). BioA is a β-alanine:pyruvate transaminase that transfers the amino group from alanine to a variety of acceptor molecules, including 3-oxopropanoate, forming pyruvate and β-alanine in the process (222). Because malonyl-CoA production is vitamin B7-dependent, a shortage of vitamin B7 may cause a biosynthetic bottleneck resulting in reduced growth rate as shown in Figure 5-6.

The finding that Ca. P. ubique is missing vitamin B12-based metabolism is unusual and surprising, but may confer an advantage to SAR11 cells by reducing their

Co2+-requirement. There may be intense competition for limiting amounts of the cobalt ions required for vitamin B12 biosynthesis in marine surface waters. Cobalt is rare in

111 ocean surface waters and displays an uncommon nutrient distribution, being more concentrated in the euphotic zone and less concentrated below (223). The marine cyanobacteria Prochlorococcus and Synechococcus bacillaris have an absolute requirement for cobalt atoms (224, 225), and physiological and genomic evidence suggest that their cobalt requirement may be primarily used for vitamin B12 biosynthesis

(50, 226, 227). Despite the apparent inability to produce vitamin B12, we predict that trace amounts of Co2+ are still required by Ca. P. ubique for use by proteins that require the divalent ion as a cofactor. For example the 3-dehydroquinate synthase involved in tryptophan biosynthesis requires divalent cobalt ions. We predict that retention of two cobalt chelatase subunits, encoded by cobST, serves to ensure this trace requirement for cobalt is met in Ca. P. ubique.

The initial predicted requirement for vitamin B6 was based on the best metabolic reconstruction information available at the time (ca. 2005). Since then new experimental data has emerged that suggests Ca. P. ubique encodes the machinery to synthesize vitamin B6 from glycolaldehyde and dXP (Figure 5-3)(206). We did not conduct growth experiments with vitamin B6, however vitamin B6 was excluded from all vitamin experiments without detriment. Because Ca. P. ubique is missing a homolog to YeaB, glycolaldehyde is likely synthesized from an alternate precursor. We propose that the glycolaldehyde required for vitamin B6 is formed during the synthesis of folate (as a byproduct of the dihydroneopterin aldolase). The presence of this recently described vitamin B6 synthesis pathway in Ca. P. ubique highlights the importance for revisions in metabolic models over time. As new science expands our knowledge at the intersection

112 of genetics and metabolism, insight based on metabolic reconstruction can be improved by reinterpretation.

From the limited environmental distribution data available, vitamin B7, like other

B-vitamins, has unusual distribution patterns in the sea (81, 193-195, 228, 229). The finding that Ca. P. ubique requires B7 raises interesting questions about the sources of this vitamin in the environment. The current paradigm pertaining to vitamin metabolism in the sea is that prokaryotes are able to synthesize their own B-vitamins. Our data implies that the numerically dominant heterotrophs in the euphotic zone do not fit into this theoretical framework and that Ca. P. ubique may compete for limited stocks of vitamin B7. Currently, vitamin research in marine systems is experiencing a resurgence as technological advances provide more sensitive vitamin assays and oceanographers seek a more sophisticated understanding of factors influencing microbial diversity and nutrient flux in plankton communities. The data we present suggest that B7 concentrations in the sea could impact population sizes of the important chemoheterotroph Ca. P. ubique, underscoring the potential importance of vitamin cycling for understanding microbial community dynamics.

113

B1 (thiamin) required B (pantothenate) required 5 glucose-6P

B6 (pyridoxine) required Pentose-Phosphate Pathway B (biotin) required 7 fructose-6P xylulose-5P ribulose-5P ery-4P Non-mevalonate ribose-5P isoprenoid biosynthesis PRPP histidine sedohep-P tryptophan deoxy-D-xylulose-5P GA-3P nucleotides dxs tktBCN

α-KG CO PEP pdhD 2 B6 & B1 sucAB isocitrate succinyl-CoA retinol ubiquinone pyruvate undecaprenyl-PP pdhD CO CO TCA Cycle 2 2 aceEF succinate valine ilvBH acetolactate acetyl-CoA leucine HCO - 3 accB OAA B aatA malate 5 fatty acid malonyl-CoA biosynthesis asp argD lysA asp-4-SA lys cysK glyA thrC thrC cys ser gly thr homoserine

Figure 5-1: Simplified illustration of B-vitamin-requiring reactions in Ca. P. ubique. Gene abbreviations are listed as annotated in the Ca. P. ubique genome. Abbreviations: PEP: phosphoenolpyruvate; GA-3P: glyceraldehyde-3-phosphate; OAA: oxaloacetate; α- KG: α-ketoglutarate; PRPP: phosphoribosyl pyrophosphate; asp-4-SA: aspartate-4- semialdehyde; ery-4P: erythrose 4-phosphate; sedohep-P: sedoheptulose 7-phosphate. Amino acids are listed with standard three-letter abbreviations.

114

A Escherichia coli K12 MG1655 D 1.3 panC panB 293 panE 3400 ilvH ilvI 100 ilvD 2.6 ilvC

‘Candidatus Pelagibacter ubique’ HTCC1062 ilvC ilvH ilvB 239 mmsA 395 panB 144 bioA 161 ilvD 107 panC B pyruvate O

OH O

OH O IlvBH O O IlvC IlvD PanB O PanE O PanC OH OH OH HO HO H OH OH N OH OH HO O OH HO OH O O IlvC O O 2-acetolactate 2,3-dihydroxy- 2-oxoisovalerate 2-dehydropantoate pantoate pantothenate O isovalerate OH O O O O O MmsA BioA HO S-CoA HO NH2 pyruvate HO O malonyl-CoA 3-oxopropanoate β-alanine alanine pyruvate PanD Proposed β-alanine synthesis pathway in Ca. P. ubique O NH2 O HO OH aspartate

Figure 5-2: Comparative genomics of vitamin B5 biosynthesis in Ca. P. ubique. A) Genomic orientation of vitamin B5 biosynthetic genes in E. coli and Ca. P. ubique. B) Illustration of canonical and proposed vitamin B5 biosynthesis. Genes depicted in (A) are colored by the reaction they catalyze in (B). Dashed lines are reactions predicted to be absent in Ca. P. ubique, but present in E. coli. Proposed route of β-alanine is encapsulated in dashed box. Gene abbreviations are listed as they appear in the Ca. P. ubique or E. coli genome, respectively. D: PanD. Numbers between syntenic fragments is the distance between genes in kilobasepairs.

115

H2O3PO HO OH A O B 3-phosphoglycerate E. coli K-12 MG1655 canonical H2O3PO O OH serC 1.5 Mbp pdxB 600 kbp epd O 3-phosphohydroxypyruvate alternate YeaB 900 kbp 1 Mbp 1.2 Mbp HO thrB ltaE yeaB serA OH OH O OPO H O 3 2 3-hydroxypyruvate O OH D-erythrose-4-phosphate Ca. P. ubique HTCC1062 non-enzymatic Folate biosynthesis Epd thrB 540 kbp 0689 620 kbp serA HO OH HO O OPO3H2 O OH glycolaldehyde 4-phosphoerythronate NH2 O PdxB OH O glycine HO OPO3H2 NH2 O OH HO OH O OH 2-oxo-3-hydroxy- 4-hydroxythreonine 4-phosphobutanoate

SerC NH2 HO OPO3H2 O OH O-phospho-4-hydroxythreonine C E. coli K-12 MG1655

H N OPO3H2 2 O pdxA 120 kbp pdxH 2.5 Mbp pdxJ 3-amino-2-oxopropyl phosphate OPO3H2 Ca. P. ubique HTCC1062 HO OH O pdxH 93 kbp pdxA 330 kbp pdxJ deoxy-D-xylulose 5-phosphate OH OH OPO3H2 N pyridoxine phosphate

O OH OPO3H2 N pyridoxal phosphate

Figure 5-3: Comparative genomics of vitamin B6 biosynthesis in Ca. P. ubique. A) Illustration of canonical and alternate vitamin B6 biosynthetic routes. B) Genomic orientation of phospho-4-hydroxythreonine (4HT-P) biosynthesis genes in E. coli and Ca. P. ubique. C) Genomic orientation of vitamin B6 biosynthesis genes (from 4HT-P) in E. coli and Ca. P. ubique. Genes depicted in (B&C) are colored by the reaction they catalyze in (A). Dashed lines and genes are reactions and genes predicted to be absent in Ca. P. ubique, but present in E. coli. Gene abbreviations are listed as they appear in the Ca. P. ubique or E. coli genome, respectively. 0689: SAR11_0689 gene identifier in the Ca. P. ubique genome.

116 E. coli K-12 MG1655 A bioD bioC bioF bioB bioA Ca. P. ubique HTCC1062 bioA ~720 kbp bioY

CoA + CO SA-4-me-thio-2-OB B O O 2 HO S-CoA BioF 8-amino- BioA pimeloyl-CoA 7-oxononanoate 7,8-diaminononanoate

ATP BioD ala SAM + CO2 ADP (sulfur donor) + P O + SAM i O NH BioB NH HO dethiobiotin S biotin met + 5’ dA

Figure 5-4: Comparative genomics of vitamin B7 biosynthesis in Ca. P. ubique. A) Genomic orientation of vitamin B7 biosynthetic genes in E. coli and Ca. P. ubique. B) Illustration of canonical vitamin B7 biosynthesis in Gram negative organisms. Genes depicted in (A) are colored by the reaction they catalyze in (B). Dashed lines are reactions predicted to be absent in Ca. P. ubique, but present in E. coli. Gene abbreviations are listed as they appear in the Ca. P. ubique or E. coli genome, respectively. bioY encodes a probable biotin transport protein. Abbreviations: ala: alanine; met: methionine; SAM: S-adenosyl-methionine; SA-4-me-thio-2-OB: S- adenosyl-4-methylthio-2-oxobutanoate; 5’dA: 5’-deoxyadenosine. The sulfur incorporation into dethiobiotin to form biotin is derived from an Fe -S cluster contained within BioB and is indicated by (sulfur donor).

117

O OH OH OH O HO O O Figure 5-5: Routes of 5-AL NH HN heme biosynthesis NH HN canonical vitamin B12 biosynthesis. Red gene HO OH O O abbreviations depict genes O OH HO O Uro-III absent in Ca. P. ubique. CobA CbiK Co2+ Green gene abbreviations CysG CbiX precorrin-2 depict genes present in Ca. CobI P. ubique. Abbreviations: precorrin-3a Co2+-precorrin-2 5-AL: 5-aminolevulinic CobG CbiL acid; Uro- 2+ precorrin-3b Co -precorrin-3 III:Uroporphyrinogen III; CobJ CbiH/G 2-AP: 2-aminopropanol. Aerobic pathway Co2+-precorrin-4 precorrin-4 CobM CbiF 2+ precorrin-5 Co -precorrin-5 CobF CbiD 2+ precorrin-6x Co -precorrin-6x CbiJ CobK 2+ precorrin-6y Co -precorrin-6y CbiET CobL Co2+-precorrin-8x precorrin-8x pathway Anaerobic CbiC CobH hydrogenobryinic acid cobyrinic acid CobB CbiA hydrogenobryinic CobNST cob(II)yrinic acid- acid-a,c-diamide a,c-diamide CobW 2+ Co cob(I)yrinic acid- O NH 2 NH 2 NH a,c-diamide O 2 H2N O CobO O BtuR N N N + adenosylcobyrinic acid- Co O N NH2 N N N N a,c-diamide HO OH CobQ 2-AP H N 2 O CbiP

HN O H2N O adenosylcobyrinic acid N CobD -O O P N CbiB O O adenosylcobinamide OH HO CobP O CobU adenosylcobalamin adenosyl-GDP-cobinamide CobV CobS CobC α-ribazole α-ribazole-5P

118

107 ) -1 106

105

density (cells ml No Vitamins Pantothenate 104 Biotin Pantothenate + Biotin 0 10 20 30 time (days)

Figure 5-6: Ca. P. ubique is bradytrophic for vitamin B5 (pantothenate) in the presence of vitamin B7 (biotin). Base medium was AMS1 amended with oxaloacetate (1 mM), glycine (50 µM), methionine (50 µM) and FeCl3 (1 µM). Vitamin B5 and B7 were added at a final concentration of 1 µM, where indicated. Points represent the mean of biological replicates ± s.d. (n=3).

119

A B

8 8 10 10 ) ) -1 -1 107

107 6 10 +0 biotin density (cells ml density (cells ml +5 pM biotin scrubbed no biotin +10 pM biotin scrubbed + biotin 105 un-scrubbed no biotin un-scrubbed + biotin 106 0 10 20 30 0 5 10 15 20 time (days) time (days)

Figure 5-7: Vitamin B7-scrubbed medium allows for demonstration of vitamin B7- requirement in Ca. P. ubique. Base medium was AMS1 amended with oxaloacetate (1 mM), glycine (50 µM), methionine (50 µM), FeCl3 (1 µM) and thiamine (1 µM). A) Vitamin B7 scrubbing was necessary to remove traces of vitamin B7 and not detrimental to growth. Vitamin B7 was at a final concentration of 1 µM when added. B) Titrations of vitamin B7 into vitamin B7-scrubbed medium. Points represent the mean of biological replicates ± s.d. (n=3).

120

Table 5-1: Common vitamin B12-requiring enzymes are absent in Ca. P. ubique. Ca. P. ubique vitamin B12- vitamin B12- in Ca. P. independednt requiring enzyme gene reaction catalyzed ubique? isozyme methionine metH Homocysteine + CH4 → absent MmuM, BhmT synthase Methionine methylmalonyl- mutAB Methylmalonyl-CoA → absent N/A CoA mutase Succinyl-CoA ribonucleotide nrdD ribonucleotide + §absent NrdAB, reductase thioredoxin→deoxyribo- SAR11_0723§ nucleotide + thioredoxin disulfide D-α-lysine mutase D-lysine→2,5- absent N/A diaminohexanoate methyleneglutarate 2-methyleneglutarate→2- absent N/A mutase methylene-3- methylsuccinate β-lysine mutase kamED 3,6-diaminohexanoate absent N/A →3,5-diaminohexanoate ornithine ornithine→(2R,4S)-2,4- absent N/A aminomutase diaminopentanoate isobutyryl-CoA 2-methylpropanoyl- absent N/A mutase CoA→butanoyl-CoA leucine 2,3- (2S)-alpha-leucine→ absent N/A aminomutase (3R)-beta-leucine methylaspartate mutES L-threo-3-methyl- absent N/A mutase aspartate →L-glutamate propanediol pduCDE glycerol→3- absent N/A dehydratase hydroxypropionaldehyde ethanolamine etuBC Ethanolamine→acet- absent N/A ammonia-lyase aldehyde + ammonia glycerol dhaBCE Glycerol→3- absent N/A dehydratase Hydroxypropanal + H2O N/A-Reaction not predicted to be involved in Ca. P. ubique metabolism. § Part of a B12-dependent nucleotide reductase was identified, however it appears to be truncated, missing the catalytic cysteine residues characteristic of a vitamin B12-dependent nucleotide reductase.

121

Chapter 6 Conclusions

Ca. P. ubique was originally isolated in 2002 on autoclaved seawater where natural, unknown dissolved organic compounds provided substrates for growth.

Sequencing of the Ca. P. ubique genome provided information about specific metabolic pathways that allow these cells to thrive in the ocean environment. Direct interactions between Ca. P. ubique and native DOM are challenging to observe. Thus, we developed laboratory-based methods based on interpretations genome-scale metabolic reconstructions to understand how Ca. P. ubique impacts geochemical cycles in nature.

Metabolic pathways were reconstructed to identify complete pathways and also to identify gaps that may be indicative of a nutrient requirement. Subsequent laboratory testing of these metabolic predictions has uncovered a number of unusual metabolic features in Ca. P. ubique. Although cultivation is time consuming, we show that this methodology works and have inferred minimal nutrient requirements for growth of Ca. P. ubique from its genome sequence. These minimal requirements include sources of glycine, reduced sulfur, pyruvate, vitamin B7 and 4-amino-5-hydroxymethyl-2- methylpyrimidine (HMP).

Interestingly we found that Ca. P. ubique grew best with a glycine:pyruvate molar ratio of 1:4 and when starved for pyruvate has unusual cell division patterns.

Additionally, we show the new finding that Ca. P. ubique is auxotrophic for thiamine because of the inability to synthesize HMP de novo. We then measured HMP in the

Sargasso Sea and found that there was sufficient HMP to support Pelagibacter

122 populations. Moreover, we show that within the Ia-subclade of the Pelagibacterales, distinct phosphate acquisition ecotypes are apparent and may play an important role in the aerobic production of methane in the sea.

These findings reinforce the hypothesis that Ca. P. ubique is adapted to the low but continuous flux of DOM in the ocean. The finding that Ca. P. ubique has multiple, simultaneous nutrient requirements is unusual when compared to the physiology of other cultivated marine bacteria. However, we suspect that this ‘decentralized’ mode of metabolism may be a feature other marine oligotrophs with streamlined genomes.

Things that are still unknown about the physiology of Ca. P. ubique relate to specific metabolic abilities and nuances related to the actual metabolic fluxes through Ca.

P. ubique. Uncommon gene arrangements of genes encoding proteins involved in the

TCA cycle, the glyoxylate bypass and glycine/serine metabolism suggest that the fluxes through these essential pathways are not canonical. A combination of cultivation-based experimentation and targeted metabolomic experiments may help resolve the flow of carbon and energy through these pathways, leading to further refinement of Ca. P. ubique metabolic models. Such experiments are more easily conducted and interpreted on a defined medium devoid of excess carbon (as is the case in natural seawater).

Moreover, this work has the potential to serve as a blueprint for the cultivation of other marine microbes. Currently, genomes from single cells are being sequenced (230), with the result of shedding light on the metabolic abilities of organisms that are not cultivated. We hope that the preceding chapters are helpful in guiding other researchers during their attempts to deduce the nutrient requirements of other marine organisms.

123

Bibliography

1. ZoBell CE. 1946. Marine Microbiology, a monograph on hydrobacteriology. Chronica Botanica Company, Waltham, MA. 2. ZoBell CE. 1941. Studies on marine bacteria I: the cultural requirements of heterotrophic aerobes. Journal of marine Research 4:42–75. 3. Gutierrez T, Shimmield T, Haidon C, Black K, Green DH. 2008. Emulsifying and Metal Ion Binding Activity of a Glycoprotein Exopolymer Produced by Pseudoalteromonas sp. Strain TG12. Appl Environ Microbiol 74:4867–4876. 4. Waksman SA. 1933. On the distribution of organic matter in the sea bottom and the chemical nature and origin of marine humus. Soil Science 36:125–148. 5. Jannasch HW, Jones GE. 1959. Bacterial populations in sea water as determined by different methods of enumeration. Limnology and Oceanography 4:128–139. 6. Bere R. 1933. Numbers of bacteria in inland lakes of Wisconsin as shown by the direct microscopic method. Internationale Revue der gesamten Hydrobiologie und Hydrographie 29:246–263. 7. Staley JT, Konopka A. 1985. Measurement of in situ activities of nonphotosynthetic microorganisms in aquatic and terrestrial habitats. Annu Rev Microbiol 39:321–346. 8. Button DK, Schut F, Quang P, Martin R, Robertson BR. 1993. Viability and isolation of marine bacteria by dilution culture: theory, procedures, and initial results. Appl Environ Microbiol 59:881–891. 9. Schut F, de Vries EJ, Gottschal JC, Robertson BR, Harder W, Prins RA, Button DK. 1993. Isolation of Typical Marine Bacteria by Dilution Culture: Growth, Maintenance, and Characteristics of Isolates under Laboratory Conditions. Appl Environ Microbiol 59:2150–2160. 10. Schut F, Prins RA, Gottschal JC. 1997. Oligotrophy and pelagic marine bacteria: Facts and fiction. Aquatic Microbial Ecology 12:177–202. 11. Button DK, Robertson BR, Lepp PW, Schmidt TM. 1998. A small, dilute- cytoplasm, high-affinity, novel bacterium isolated by extinction culture and having kinetic constants compatible with growth at ambient concentrations of dissolved nutrients in seawater. Appl Environ Microbiol 64:4467–4476. 12. Sanger F, Nicklen S, Coulson AR. 1977. DNA sequencing with chain- terminating inhibitors. Proceedings of the National Academy of Sciences 74:5463–5467. 13. Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R, Horn GT, Mullis KB, Erlich HA. 1988. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science (New York, NY) 239:487–491. 14. Woese CR, Fox GE. 1977. Phylogenetic structure of the prokaryotic domain: the primary kingdoms. Proceedings of the National Academy of Sciences 74:5088– 5090. 15. Pace NR, Stahl DA, Lane DJ, Olsen GJ. 1985. Analyzing natural microbial

124

populations by rRNA sequences. ASM American Society for Microbiology News 51:4–12. 16. Giovannoni SJ, Britschgi TB, Moyer CL, Field KG. 1990. Genetic diversity in Sargasso Sea bacterioplankton. Nature 345:60–63. 17. Morris RM, Rappé MS, Connon SA, Vergin KL, Siebold WA, Carlson CA, Giovannoni SJ. 2002. SAR11 clade dominates ocean surface bacterioplankton communities. Nature 420:806–810. 18. Schattenhofer M, Fuchs B, Amann R, Zubkov M, Tarran G, Pernthaler J. 2009. Latitudinal distribution of prokaryotic picoplankton populations in the Atlantic Ocean. Environ Microbiol 11:2078–2093. 19. Ghiglione J-F, Galand PE, Pommier T, Pedrós-Alio C, Maas EW, Bakker K, Bertilson S, Kirchmanj DL, Lovejoy C, Yager PL, Murray AE. 2012. Pole-to- pole biogeography of surface and deep marine bacterial communities. Proceedings of the National Academy of Sciences 109:17633–17638. 20. Salcher MM, Pernthaler J, Posch T. 2011. Seasonal bloom dynamics and ecophysiology of the freshwater sister clade of SAR11 bacteria “that rule the waves” (LD12). ISME J 5:1242–1252. 21. Connon SA, Giovannoni SJ. 2002. High-throughput methods for culturing microorganisms in very-low-nutrient media yield diverse new marine isolates. Appl Environ Microbiol 68:3878–3885. 22. Stingl U, Tripp HJ, Giovannoni SJ. 2007. Improvements of high-throughput culturing yielded novel SAR11 strains and other abundant marine bacteria from the Oregon coast and the Bermuda Atlantic Time Series study site. ISME J 1:361–371. 23. Rappé MS, Connon SA, Vergin KL, Giovannoni SJ. 2002. Cultivation of the ubiquitous SAR11 marine bacterioplankton clade. Nature 418:630–633. 24. Murray RG, Stackebrandt E. 1995. Taxonomic note: implementation of the provisional status Candidatus for incompletely described procaryotes. Int. J. Syst. Bacteriol. 45:186–187. 25. Giovannoni SJ, Tripp HJ, Givan S, Podar M, Vergin KL, Baptista D, Bibbs L, Eads J, Richardson TH, Noordewier M, Rappé MS, Short JM, Carrington JC, Mathur EJ. 2005. Genome streamlining in a cosmopolitan oceanic bacterium. Science (New York, NY) 309:1242–1245. 26. Tripp HJ, Kitner JB, Schwalbach MS, Dacey JWH, Wilhelm LJ, Giovannoni SJ. 2008. SAR11 marine bacteria require exogenous reduced sulphur for growth. Nature 452:741–744. 27. Kiene RP, Slezak D. 2006. Low dissolved DMSP concentrations in seawater revealed by small-volume gravity filtration and dialysis sampling. Limnology and Oceanography: Methods 4:80–95. 28. Malmstrom RR, Kiene RP, Cottrell MT, Kirchman DL. 2004. Contribution of SAR11 bacteria to dissolved dimethylsulfoniopropionate and amino acid uptake in the North Atlantic ocean. Appl Environ Microbiol 70:4129–4135. 29. Grote J, Thrash JC, Huggett MJ, Landry ZC, Carini P, Giovannoni SJ, Rappé MS. 2012. Streamlining and core genome conservation among highly divergent

125

members of the SAR11 clade. mBio 3:e00252–12. 30. Monschau N, Stahmann KP, Sahm H, McNeil JB, Bognar AL. 1997. Identification of Saccharomyces cerevisiae GLY1 as a threonine aldolase: a key enzyme in glycine biosynthesis. FEMS Microbiology Letters 150:55–60. 31. Liu JQ, Dairi T, Itoh N, Kataoka M, Shimizu S, Yamada H. 1998. Gene cloning, biochemical characterization and physiological role of a thermostable low‐ specificity L‐threonine aldolase from Escherichia coli. European Journal of Biochemistry 255:220–226. 32. Tripp HJ, Schwalbach MS, Meyer MM, Kitner JB, Breaker RR, Giovannoni SJ. 2009. Unique glycine-activated riboswitch linked to glycine-serine auxotrophy in SAR11. Environ Microbiol 11:230–238. 33. Kikuchi G, Motokawa Y, Yoshida T, Hiraga K. 2008. Glycine cleavage system: reaction mechanism, physiological significance, and hyperglycinemia. Proceedings of the Japan Academy, Series B 84:246–263. 34. Winkler WC, Breaker RR. 2005. Regulation of bacterial gene expression by riboswitches. Annu Rev Microbiol 59:487–517. 35. Pellicer MT, Badia J, Aguilar J, Baldoma L. 1996. glc locus of Escherichia coli: characterization of genes encoding the subunits of glycolate oxidase and the glc regulator protein. J Bacteriol 178:2051–2059. 36. Pellicer M, Fernandez C, Badia J, Aguilar J, Lin E, Baldoma L. 1999. Cross- induction of glc and ace operons of Escherichia coli attributable to pathway intersection. J Biol Chem 274:1745–1752. 37. Schwalbach MS, Tripp HJ, Steindler L, Smith DP, Giovannoni SJ. 2010. The presence of the glycolysis operon in SAR11 genomes is positively correlated with ocean productivity. Environ Microbiol 12:490–490. 38. Moran NA, McCutcheon JP, Nakabachi A. 2008. Genomics and Evolution of Heritable Bacterial Symbionts. Annu Rev Genet 42:165–190. 39. Giovannoni SJ, Hayakawa DH, Tripp HJ, Stingl U, Givan SA, Cho J-C, Oh H- M, Kitner JB, Vergin KL, Rappé MS. 2008. The small genome of an abundant coastal ocean methylotroph. Environ Microbiol 10:1771–1782. 40. Dufresne A, Garczarek L, Partensky F. 2005. Accelerated evolution associated with genome reduction in a free-living prokaryote. Genome Biol 6:R14. 41. Morris JJ, Lenski RE, Zinser ER. 2012. The Black Queen Hypothesis: Evolution of Dependencies through Adaptive Gene Loss. mBio 3:1–7. 42. Matudaira T. 1939. The physiological property of seawater considered from the effect upon the growth of diatom, with special reference to its vertical and seasonal change. Bulletin of the Japanese Society of Scientific Fisheries 8:187– 193. 43. Neidhardt FC, Bloch PL, Smith DF. 1974. Culture medium for enterobacteria. J Bacteriol 119:736–747. 44. Demain AL. 1958. Minimal media for quantitative studies with Bacillus subtilis. J Bacteriol 75:517–522. 45. Giovannoni S, Stingl U. 2007. The importance of culturing bacterioplankton in the “omics” age. Nat Rev Micro 5:820–826.

126

46. Fuhrman JA, Hewson I, Schwalbach MS, Steele JA, Brown MV, Naeem S. 2006. Annually reoccurring bacterial communities are predictable from ocean conditions. Proceedings of the National Academy of Sciences 103:13104–13109. 47. Steele JA, Countway PD, Xia L, Vigil PD, Beman JM, Kim DY, Chow C-ET, Sachdeva R, Jones AC, Schwalbach MS, Rose JM, Hewson I, Patel A, Sun F, Caron DA, Fuhrman JA. 2011. Marine bacterial, archaeal and protistan association networks reveal ecological linkages. ISME J 5:1414–1425. 48. Dunne JA, Williams RJ, Martinez ND. 2002. Network structure and biodiversity loss in food webs: robustness increases with connectance. Ecol Lett 5:558–567. 49. Sun J, Steindler L, Thrash JC, Halsey KH, Smith DP, Carter AE, Landry ZC, Giovannoni SJ. 2011. One Carbon Metabolism in SAR11 Pelagic Marine Bacteria. PLoS ONE 6:e23973. 50. Dufresne A, Salanoubat M, Partensky F, Artiguenave F, Axmann I, Barbe V, Duprat S, Galperin M, Koonin E, Le Gall F, Makarova K, Ostrowski M, Oztas S, Robert C, Rogozin I, Scanlan D, de Marsac N, Weissenbach J, Wincker P, Wolf Y, Hess W. 2003. Genome sequence of the cyanobacterium Prochlorococcus marinus SS120, a nearly minimal oxyphototrophic genome. Proceedings of the National Academy of Sciences 100:10020–10025. 51. Lynch M, Conery JS. 2003. The origins of genome complexity. Science (New York, NY) 302:1401–1404. 52. Moore L, Coe A, Zinser E, Saito M, Sullivan M, Lindell D, Frois-Moniz K, Waterbury J, Chisholm S. 2007. Culturing the marine cyanobacterium Prochlorococcus. Limnology and Oceanography: Methods 5:353–362. 53. Zigler JS, Lepe-Zuniga JL, Vistica B, Gery I. 1985. Analysis of the cytotoxic effects of light-exposed HEPES-containing culture medium. In Vitro Cell. Dev. Biol. 21:282–287. 54. Nörtemann B. 1992. Total degradation of EDTA by mixed cultures and a bacterial isolate. Appl Environ Microbiol 58:671–676. 55. Teira E, Martinez-García S, Lønborg C, Álvarez-Salgado XA. 2009. Growth rates of different phylogenetic bacterioplankton groups in a coastal upwelling system. Environmental Microbiology Reports 1:545–554. 56. Ferrera I, Gasol JM, Sebastián M, Hojerová E, Koblízek M. 2011. Comparison of Growth Rates of Aerobic Anoxygenic Phototrophic Bacteria and Other Bacterioplankton Groups in Coastal Mediterranean Waters. Appl Environ Microbiol 77:7451–7458. 57. Wientjes FB, Nanninga N. 1989. Rate and topography of peptidoglycan synthesis during cell division in Escherichia coli: concept of a leading edge. J Bacteriol 171:3412–3419. 58. Mengin-Lecreulx D, Flouret B, van Heijenoort J. 1982. Cytoplasmic steps of peptidoglycan synthesis in Escherichia coli. J Bacteriol 151:1109–1117. 59. Bellion E, Tan F. 1987. An NAD+-dependent alanine dehydrogenase from a methylotrophic bacterium. Biochem J 244:565–570. 60. Caballero F, Cardenas J, Castillo F. 1989. Purification and properties of L- alanine dehydrogenase of the phototrophic bacterium Rhodobacter capsulatus

127

E1F1. J Bacteriol 171:3205–3210. 61. Williams T, Ertan H, Ting L, Cavicchioli R. 2009. Carbon and nitrogen substrate utilization in the marine bacterium Sphingopyxis alaskensis strain RB2256. ISME J 3:1036–1052. 62. Kiene RP, Williams L. 1998. Glycine betaine uptake, retention, and degradation by microorganisms in seawater. Limnology and Oceanography 43:1592–1603. 63. Keller M, Kiene R, Matrai P, Bellows W. 1999. Production of glycine betaine and dimethylsulfoniopropionate in marine phytoplankton. I. Batch cultures. Marine Biology 135:237–248. 64. Leboulanger C, Oriol L, Jupin H, Desolas-Gros C. 1997. Diel variability of glycolate in the eastern tropical Atlantic Ocean. Deep Sea Research I 44:2131– 2139. 65. Leboulanger C, Descolas-Gros C, Jupin H. 1994. HPLC determination of glycolic acid in seawater. An estimation of phytoplankton photorespiration in the Gulf of Lions, western Mediterranean Sea. Journal of Plankton Research 16:897– 903. 66. Leboulanger C, Martin Jézéquel V, Descolas-Gros C, Sciandra A, Jupin HJ. 1998. Photorespiration in continuous culture of Dunaliella tertiolecta (Chlorophyta): relationships between serine, glycine, and extracellular glycolate. Journal of Phycology 34:651–654. 67. Parker MS, Armbrust EV, Piovia-Scott J, Keil RG. 2004. Induction of photorespiration by light in the centric diatom Thalssiosira weissflogii (Bacillariophyceae): molecular characterization and physiological consequences. Journal of Phycology 40:557–567. 68. Bertlisson S, Berglund O, Pullin M, Chisholm S. 2005. Release of dissolved organic matter by Prochlorococcus. Vie et Milieu 55:225–231. 69. Kieber D, Mopper K. 1987. Photochemical formation of glyoxylic and pyruvic acids in seawater. Marine Chemistry 21:135–149. 70. Kieber D, McDaniel J, Mopper K. 1989. Photochemical source of biological substrates in sea water: Implications for carbon cycling. Nature 341:637–639. 71. Mopper K, Zhou X, Kieber R, Kieber D, Sikorski R, Jones R. 1991. Photochemical degradation of dissolved organic carbon and its impact on the oceanic carbon cycle. Nature 353:60–62. 72. Moran M, Zepp R. 1997. Role of photoreactions in the formation of biologically labile compounds from dissolved organic matter. Limnology and Oceanography 42:1307–1316. 73. Obernosterer I, Kraay G, De Ranitz E, Herndl G. 1999. Concentrations of low molecular weight carboxylic acids and carbonyl compounds in the Aegean Sea (Eastern Mediterranean) and the turnover of pyruvate. Aquatic Microbial Ecology 20:147–156. 74. Huxtable R. 1992. Physiological actions of taurine. Physiological reviews 72:101–163. 75. Gage DA, Rhodes D, Nolte KD, Hicks WA, Leustek T, Cooper AJ, Hanson AD. 1997. A new route for synthesis of dimethylsulphoniopropionate in marine algae.

128

Nature 387:891–894. 76. Renesto P, Crapoulet N, Ogata H, La Scola B, Vestris G, Claverie J-M, Raoult D. 2003. Genome-based design of a cell-free culture medium for Tropheryma whipplei. The Lancet 362:447–449. 77. Dupont CL, Rusch DB, Yooseph S, Lombardo M-J, Richter RA, Valas R, Novotny M, Yee-Greenbaum J, Selengut JD, Haft DH. 2012. Genomic insights to SAR86, an abundant and uncultivated marine bacterial lineage. ISME J 6:1186–1199. 78. Santoro AE, Casciotti KL. 2011. Enrichment and characterization of ammonia- oxidizing archaea from the open ocean: phylogeny, physiology and stable isotope fractionation. ISME J 5:1796–1808. 79. Lengeler JW, Drews G, Schlegel HG. 1999. Biology of the Prokaryotes. Wiley- Blackwell, Malden, MA. 80. Jurgenson CT, Begley TP, Ealick SE. 2009. The Structural and Biochemical Foundations of Thiamin Biosynthesis. Annu. Rev. Biochem. 78:569–603. 81. Sañudo-Wilhelmy SA, Cutter LS, Durazo R, Smail EA, Gomez- Consarnau L, Webb EA, Prokopenko MG, Berelson WM, Karl DMD. 2012. Multiple B- vitamin depletion in large areas of the coastal ocean. Proceedings of the National Academy of Sciences 109:14041–14045. 82. Barada LP, Cutter L, Montoya JP, Webb EA, Capone DG, Sañudo-Wilhelmy SA. 2013. The distribution of thiamin and pyridoxine in the western tropical North Atlantic Amazon River plume. Front Microbiol 4:25–25. 83. Panzeca C, Tovar-Sanchez A, Agusti S, Reche I, Duarte C, Taylor G, Sañudo- Wilhelmy S. 2006. B vitamins as regulators of phytoplankton dynamics. EOS Transactions 87:593–596. 84. Giovannoni SJ. 2012. Vitamins in the sea. Proceedings of the National Academy of Sciences 109:13888–13889. 85. Carlucci A, Bowes PM. 1970. Production of vitamin B12, thiamine, and biotin by phytoplankton. Journal of Phycology 6:351–357. 86. Croft M, Warren M, Smith A. 2006. Algae need their vitamins. Eukaryotic Cell 5:1175–1183. 87. Croft MT, Lawrence AD, Raux-Deery E, Warren MJ, Smith AG. 2005. Algae acquire vitamin B12 through a symbiotic relationship with bacteria. Nature 438:90–93. 88. Alonso C, Pernthaler J. 2006. Roseobacter and SAR11 dominate microbial glucose uptake in coastal North Sea waters. Environ Microbiol 8:2022–2030. 89. Carini P, Steindler L, Beszteri S, Giovannoni SJ. 2013. Nutrient requirements for growth of the extreme oligotroph “Candidatus Pelagibacter ubique” HTCC1062 on a defined medium. ISME J 7:592–602. 90. Wightman R, Meacock PA. 2003. The THI5 gene family of Saccharomyces cerevisiae: distribution of homologues among the hemiascomycetes and functional redundancy in the aerobic biosynthesis of thiamin from pyridoxine. Microbiology 149:1447–1460. 91. Jenkins AH, Schyns G, Potot S, Sun G, Begley TP. 2007. A new thiamin salvage

129

pathway. Nat Chem Biol 3:492–497. 92. Webb E, Claas K, Downs D. 1998. thiBPQ encodes an ABC transporter required for transport of thiamine and thiamine pyrophosphate in Salmonella typhimurium. J Biol Chem 273:8946–8950. 93. Lauhon CT, Kambampati R. 2000. The iscS gene in Escherichia coli is required for the biosynthesis of 4-thiouridine, thiamin, and NAD. J Biol Chem 275:20096–20103. 94. Settembre EC, Dorrestein PC, Park J-H, Augustine AM, Begley TP, Ealick SE. 2003. Structural and mechanistic studies on ThiO, a glycine oxidase essential for thiamin biosynthesis in Bacillus subtilis. Biochemistry 42:2971–2981. 95. Park J-H, Dorrestein PC, Zhai H, Kinsland C, McLafferty FW, Begley TP. 2003. Biosynthesis of the Thiazole Moiety of Thiamin Pyrophosphate (Vitamin B1). Biochemistry 42:12430–12438. 96. Winkler W, Nahvi A, Breaker RR. 2002. Thiamine derivatives bind messenger RNAs directly to regulate bacterial gene expression. Nature 419:952–956. 97. Rodionov DA, Vitreschak AG, Mironov AA, Gelfand MS. 2002. Comparative genomics of thiamin biosynthesis in procaryotes. J Biol Chem 277:48949–48959. 98. Meyer M, Ames T, DP S, Weinberg Z, Schwalbach M, Giovannoni S, Breaker R. 2009. Identification of candidate structured RNAs in the marine organism “Candidatus Pelagibacter ubique.” BMC Genomics 10:1–16. 99. Button D. 1968. Selective Thiamine Removal From Culture Media by Ultraviolet Irradiation. Appl Microbiol 16:530–531. 100. Droop MR. 1958. Requirement for Thiamine Among Some Marine and Supra- littoral Protista. Journal of the Marine Biological Association of the United Kingdom 37:323–329. 101. Ishida S, Tazuya-Murayama K, Kijima Y, Yamada K. 2008. The direct precursor of the pyrimidine moiety of thiamin is not urocanic acid but histidine in Saccharomyces cerevisiae. J Nutr Sci Vitaminol (Tokyo) 54:7–10. 102. Downs DM. 1992. Evidence for a new, oxygen-regulated biosynthetic pathway for the pyrimidine moiety of thiamine in Salmonella typhimurium. J Bacteriol 174:1515–1521. 103. Carlucci A, Silbernagel S, McNally P. 1969. Influence of temperature and solar radiation on persistence of vitamin B12, thiamine, and biotin in seawater. Journal of Phycology 5:302–305. 104. Sowell SM, Wilhelm LJ, Norbeck AD, Lipton MS, Nicora CD, Barofsky DF, Carlson CA, Smith RD, Giovanonni SJ. 2009. Transport functions dominate the SAR11 metaproteome at low-nutrient extremes in the Sargasso Sea. ISME J 3:93–105. 105. Wrenger C, Eschbach M-L, Müller IB, Laun NP, Begley TP, Walter RD. 2006. Vitamin B1 de novo synthesis in the human malaria parasite Plasmodium falciparum depends on external provision of 4-amino-5-hydroxymethyl-2- methylpyrimidine. Biol Chem 387:41–51. 106. Schauer K, Stolz J, Scherer S, Fuchs TM. 2009. Both Thiamine Uptake and Biosynthesis of Thiamine Precursors Are Required for Intracellular Replication

130

of Listeria monocytogenes. J Bacteriol 191:2218–2227. 107. Waldbauer JR, Rodrigue S, Coleman ML, Chisholm SW. 2012. Transcriptome and Proteome Dynamics of a Light-Dark Synchronized Bacterial Cell Cycle. PLoS ONE 7:e43432. 108. Ottesen EA, Young CR, Eppley JM, Ryan JP, Chavez FP, Scholin CA, Delong EF. 2013. Pattern and synchrony of gene expression among sympatric marine microbial populations. Proceedings of the National Academy of Sciences 110:E488–97. 109. Gubler C. 1984. Handbook of vitamins : nutritional, biochemical, and clinical aspects. Marcel Dekker, Inc., New York. 110. Okumura K. 1961. Decomposition of thiamine and its derivatives by ultraviolet radiation. J Vitaminol (Kyoto) 24:158–163. 111. Tedetti M, Sempéré R. 2006. Penetration of ultraviolet radiation in the marine environment. A review. Photochem. Photobiol. 82:389–397. 112. Mopper K, Zika RG. 1987. Natural photosensitizers in sea water: riboflavin and its breakdown products, pp. 174–190. In Photochemistry of Environmental Aquatic Systems. ACS Symposium Series. 113. Ortiz-Guerrero JM, Polanco MC, Murillo FJ, Padmanabhan S, Elías-Arnanz M. 2011. Light-dependent gene regulation by a coenzyme B12-based photoreceptor. Proceedings of the National Academy of Sciences 108:7565–7570. 114. Reddick JJ, Nicewonger R, Begley TP. 2001. Mechanistic studies on thiamin phosphate synthase: evidence for a dissociative mechanism. Biochemistry 40:10095–10102. 115. Zhao L, Ma X-D, Chen F-E. 2012. Development of Two Scalable Syntheses of 4-Amino-5-aminomethyl-2-methylpyrimidine: Key Intermediate for Vitamin B1. Org. Process Res. Dev. 16:57–60. 116. Baldi P, Long AD. 2001. A Bayesian framework for the analysis of microarray expression data: regularized t -test and statistical inferences of gene changes. Bioinformatics 17:509–519. 117. Storey JD, Tibshirani R. 2003. Statistical significance for genomewide studies. Proceedings of the National Academy of Sciences 100:9440–9445. 118. Wu J, Sunda W, Boyle EA, Karl DM. 2000. Phosphate depletion in the western North Atlantic Ocean. Science (New York, NY) 289:759–762. 119. Sañudo-Wilhelmy SA, Kustka AB, Gobler CJ, Hutchins DA, Yang M, Lwiza K, Burns J, Capone DG, Raven JA, Carpenter EJ. 2001. Phosphorus limitation of nitrogen fixation by Trichodesmium in the central Atlantic Ocean. Nature 411:66–69. 120. Thingstad TF, Krom MD, Mantoura RFC, Flaten GAF, Groom S, Herut B, Kress N, Law CS, Pasternak A, Pitta P, Psarra S, Rassoulzadegan F, Tanaka T, Tselepides A, Wassmann P, Woodward EMS, Riser CW, Zodiatis G, Zohary T. 2005. Nature of phosphorus limitation in the ultraoligotrophic eastern Mediterranean. Science (New York, NY) 309:1068–1071. 121. Mather RL, Reynolds SE, Wolff GA, Williams RG, Torres-Valdes S, Woodward EMS, Landolfi A, Pan X, Sanders R, Achterberg EP. 2008. Phosphorus cycling

131

in the North and South Atlantic Ocean subtropical gyres. Nature Geosci 1:439– 443. 122. Lomas MW, Burke AL, Lomas DA, Bell DW, Shen C, Dyhrman ST, Ammerman JW. 2010. Sargasso Sea phosphorus biogeochemistry: an important role for dissolved organic phosphorus (DOP). Biogeosciences 7:695–710. 123. Benitez-Nelson C. 2000. The biogeochemical cycling of phosphorus in marine systems. Earth-Sci Rev 51:109–135. 124. Tyrrell T. 1999. The relative influences of nitrogen and phosphorus on oceanic primary production. Nature 400:525–531. 125. Martiny AC, Coleman ML, Chisholm SW. 2006. Phosphate acquisition genes in Prochlorococcus ecotypes: evidence for genome-wide adaptation. Proceedings of the National Academy of Sciences 103:12552–12557. 126. Dyhrman S, Chappell P, Haley S, Moffett J, Orchard E, Waterbury J, Webb E. 2006. Phosphonate utilization by the globally important marine diazotroph Trichodesmium. Nature 439:68–71. 127. Tetu SG, Brahamsha B, Johnson DA, Tai V, Phillippy K, Palenik B, Paulsen IT. 2009. Microarray analysis of phosphate regulation in the marine cyanobacterium Synechococcus sp. WH8102. ISME J 3:835–849. 128. Krumhardt KM, Callnan K, Roache-Johnson K, Swett T, Robinson D, Reistetter EN, Saunders JK, Rocap G, Moore LR. 2013. Effects of phosphorus starvation versus limitation on the marine cyanobacterium Prochlorococcus MED4 I: uptake physiology. Environ Microbiol. doi: 10.1111/1462-2920.12079 129. Fu F-X, Zhang Y, Bell PR, Hutchins DA. 2005. Phosphate Uptake and Growth Kinetics of Trichodesmium (Cyanobacteria) Isolates From the North Atlantic Ocean and the Great Barrier Reef, Australia. Journal of Phycology 41:62–73. 130. Grillo JF, Gibson J. 1979. Regulation of phosphate accumulation in the unicellular cyanobacterium Synechococcus. J Bacteriol 140:508–517. 131. Donald KM, Scanlan DJ, Carr NG, Mann NH, Joint I. 1997. Comparative phosphorus nutrition of the marine cyanobacterium Synechococcus WH7803 and the marine diatom Thalassiosira weissflogii. Journal of Plankton Research 19:1793–1813. 132. Malmstrom RR, Cottrell MT, Elifantz H, Kirchman DL. 2005. Biomass production and assimilation of dissolved organic matter by SAR11 bacteria in the Northwest Atlantic Ocean. Appl Environ Microbiol 71:2979–2986. 133. Elifantz H, Malmstrom RR, Cottrell MT, Kirchman DL. 2005. Assimilation of polysaccharides and glucose by major bacterial groups in the Delaware Estuary. Appl Environ Microbiol 71:7799–7805. 134. Van Mooy BAS, Fredricks HF, Pedler BE, Dyhrman ST, Karl DM, Koblížek M, Lomas MW, Mincer TJ, Moore LR, Moutin T, Rappé MS, Webb EA. 2009. Phytoplankton in the ocean use non-phosphorus lipids in response to phosphorus scarcity. Nature 458:69–72. 135. Van Mooy BAS, Rocap G, Fredricks HF, Evans CT, Devol AH. 2006. Sulfolipids dramatically decrease phosphorus demand by picocyanobacteria in oligotrophic marine environments. Proceedings of the National Academy of

132

Sciences 103:8607–8612. 136. Minnikin DE, Abdolrahimzadeh H. 1974. The replacement of phosphatidylethanolamine and acidic phospholipids by an ornithine-amide lipid and a minor phosphorus-free lipid in Pseudomonas fluorescens NCMB 129. FEBS Lett 43:257–260. 137. Zavaleta-Pastor M, Sohlenkamp C, Gao JL, Guan Z, Zaheer R, Finan TM, Raetz CRH, Lopez-Lara IM, Geiger O. 2010. Sinorhizobium meliloti phospholipase C required for lipid remodeling during phosphorus limitation. Proceedings of the National Academy of Sciences 107:302–307. 138. Willsky GR, Malamy MH. 1980. Characterization of two genetically separable inorganic phosphate transport systems in Escherichia coli. J Bacteriol 144:356– 365. 139. Voegele RT, Bardin S, Finan TMT. 1997. Characterization of the Rhizobium (Sinorhizobium) meliloti high- and low-affinity phosphate uptake systems. J Bacteriol 179:7226–7232. 140. Harold FM. 1964. Enzymic and Genetic Control of Polyphosphate Accumulation in Aerobacter aerogenes. J Gen Microbiol 35:81–90. 141. Luo H, Benner R, Long RA, Hu J. 2009. Subcellular localization of marine bacterial alkaline phosphatases. Proceedings of the National Academy of Sciences 106:21219–21223. 142. Martinez A, Osburne MS, Sharma AK, Delong EF, Chisholm SWS. 2012. Phosphite utilization by the marine picocyanobacterium Prochlorococcus MIT9301. Environ Microbiol 14:1363–1377. 143. Coleman ML, Chisholm SW. 2010. Ecosystem-specific selection pressures revealed through comparative population genomics. Proceedings of the National Academy of Sciences 107:18634–18639. 144. Martiny AC, Huang Y, Li W. 2011. Adaptation to Nutrient Availability in Marine Microorganisms by Gene Gain and Loss, pp. 269–276. In de Bruijn, FJ (ed.), Handbook of Molecular Microbial Ecology II: Metagenomics in Different Habitats. John Wiley & Sons, Inc., Hoboken, NJ, USA. 145. Temperton B, Gilbert JA, Quinn JP, Mcgrath JW. 2011. Novel analysis of oceanic surface water metagenomes suggests importance of polyphosphate metabolism in oligotrophic environments. PLoS ONE 6:e16499. 146. Davis H, Guillard R. 1958. Relative value of ten genera of micro-organisms as foods for oyster and clam larvae. USFWS Fish Bull. 58:293–304. 147. Ward SK, Heintz JA, Albrecht RM, Talaat AM. 2012. Single-cell elemental analysis of bacteria: quantitative analysis of polyphosphates in Mycobacterium tuberculosis. Frontiers in Cellular and Infection Microbiology 2:1–7. 148. Harold FM. 1966. Inorganic polyphosphates in biology: structure, metabolism, and function. Bacteriol Rev 30:772–794. 149. Rao NN, Liu S, Kornberg A. 1998. Inorganic polyphosphate in Escherichia coli: the phosphate regulon and the stringent response. J Bacteriol 180:2186–2193. 150. Rao NN, Roberts MF, Torriani A. 1985. Amount and chain length of polyphosphates in Escherichia coli depend on cell growth conditions. J Bacteriol

133

162:242–247. 151. Kornberg A, Rao NN, Ault-Riché D. 1999. Inorganic polyphosphate: a molecule of many functions. Annu. Rev. Biochem. 68:89–125. 152. Wanner BL. 1993. Gene regulation by phosphate in enteric bacteria. J Cell Biochem 51:47–54. 153. Ishige T, Krause M, Bott M, Wendisch VF, Sahm H. 2003. The Phosphate Starvation Stimulon of Corynebacterium glutamicum Determined by DNA Microarray Analyses. J Bacteriol 185:4519–4529. 154. Durfee T, Hansen AM, Zhi H, Blattner FR, Jin DJ. 2008. Transcription Profiling of the Stringent Response in Escherichia coli. J Bacteriol 190:1084–1096. 155. Traxler MF, Summers SM, Nguyen H-T, Zacharia VM, Hightower GA, Smith JT, Conway T. 2008. The global, ppGpp-mediated stringent response to amino acid starvation in Escherichia coli. Mol Microbiol 68:1128–1148. 156. Chatterji D, Kumar Ojha A. 2001. Revisiting the stringent response, ppGpp and starvation signaling. Current opinion in microbiology 4:160–165. 157. Spira B, Silberstein N, Yagil E. 1995. Guanosine 3“,5-”bispyrophosphate (ppGpp) synthesis in cells of Escherichia coli starved for Pi. J Bacteriol 177:4053–4058. 158. Spira B, Yagil E. 1998. The relation between ppGpp and the PHO regulon in Escherichia coli. Mol Genet Genomics 257:469–477. 159. Potrykus K, Cashel M. 2008. (p)ppGpp: still magical? Annu Rev Microbiol 62:35–51. 160. Grossman AD, Taylor WE, Burton ZF, Burgess RR, Gross CA. 1985. Stringent response in Escherichia coli induces expression of heat shock proteins. J Mol Biol 186:357–365. 161. Baek JH, Lee SY. 2006. Novel gene members in the Pho regulon of Escherichia coli. FEMS Microbiology Letters 264:104–109. 162. Jain V, Kumar M, Chatterji D. 2006. ppGpp: stringent response and survival. J Microbiol 44:1–10. 163. Mittenhuber G. 2001. Comparative genomics and evolution of genes encoding bacterial (p)ppGpp synthetases/hydrolases (the Rel, RelA and SpoT proteins). J. Mol. Microbiol. Biotechnol. 3:585–600. 164. Aravind L, Koonin EV. 1998. The HD domain defines a new superfamily of metal-dependent phosphohydrolases. Trends Biochem Sci 23:469–472. 165. Metcalf WW, Wanner BL. 1991. Involvement of the Escherichia coli phn (psiD) gene cluster in assimilation of phosphorus in the form of phosphonates, phosphite, Pi esters, and Pi. J Bacteriol 173:587–600. 166. Metcalf WW, Wanner BL. 1993. Mutational analysis of an Escherichia coli fourteen-gene operon for phosphonate degradation, using TnphoA' elements. J Bacteriol 175:3430–3442. 167. Luo H, Zhang H, Long RA, Benner R. 2011. Depth distributions of alkaline phosphatase and phosphonate utilization genes in the North Pacific Subtropical Gyre. Aquatic Microbial Ecology 62:61–69. 168. Daubin V, Ochman H. 2004. Bacterial Genomes as New Gene Homes: The

134

Genealogy of ORFans in E. coli. Genome Res 14:1036–1042. 169. Daughton CG, Cook AM, Alexander M. 1979. Biodegradation of phosphonate toxicants yields methane or ethane on cleavage of the C‐P bond. FEMS Microbiology Letters 5:91–93. 170. Karl DM, Beversdorf L, Björkman KM, Church MJ, Martinez A, Delong EF. 2008. Aerobic production of methane in the sea. Nature Geosci 1:473–478. 171. Beversdorf LJ, White AE, Björkman KM, Letelier RM, Karl DM. 2010. Phosphonate metabolism by Trichodesmium IMS101 and the production of greenhouse gases. Limnology and Oceanography 55:1768–1778. 172. Metcalf WW, Griffin BM, Cicchillo RM, Gao J, Janga SC, Cooke HA, Circello BT, Evans BS, Martens-Habbena W, Stahl DA, van der Donk WA. 2012. Synthesis of Methylphosphonic Acid by Marine Microbes: A Source for Methane in the Aerobic Ocean. Science (New York, NY) 337:1104–1107. 173. Karner MB, Delong EF, Karl DM. 2001. Archaeal dominance in the mesopelagic zone of the Pacific Ocean. Nature 409:507–510. 174. Minnikin DE, Abdolrahimzadeh H, Baddiley J. 1974. Replacement of acidic phosphates by acidic glycolipids in Pseudomonas diminuta. Nature 249:268–269. 175. Geiger O, González-Silva N, López-Lara IM, Sohlenkamp C. 2010. Amino acid- containing membrane lipids in bacteria. Progress in Lipid Research 49:46–60. 176. Benning C. 1998. Biosynthesis and function of the sulfolipid sulfoquinovosyl diacylglycerol. Annu. Rev. Plant Physiol. Plant Mol. Biol. 49:53–75. 177. Gao J-L, Weissenmayer B, Taylor AM, Thomas-Oates J, López-Lara IM, Geiger O. 2004. Identification of a gene required for the formation of lyso-ornithine lipid, an intermediate in the biosynthesis of ornithine-containing lipids. Mol Microbiol 53:1757–1770. 178. White AE, Karl DM, Björkman K, Beversdorf LJ, Letelier RM. 2010. Production of organic matter by Trichodesmium IMS101 as a function of phosphorus source. Limnology and Oceanography 55:1755–1767. 179. Thrash JC, Boyd A, Huggett MJ, Grote J, Carini P, Yoder RJ, Robbertse B, Spatafora JW, Rappé MS, Giovannoni SJ. 2011. Phylogenomic evidence for a common ancestor of mitochondria and the SAR11 clade. Sci Rep 1:13–13. 180. Källberg M, Wang H, Wang S, Peng J, Wang Z, Lu H, Xu J. 2012. Template- based protein structure modeling using the RaptorX web server. Nat Protoc 7:1511–1522. 181. Versaw WK, Metzenberg RL. 1995. Repressible cation-phosphate symporters in Neurospora crassa. Proceedings of the National Academy of Sciences 92:3884– 3887. 182. Beard SJ, Hashim R, Wu G, Binet MR, Hughes MN, Poole RK. 2006. Evidence for the transport of zinc (II) ions via the pit inorganic phosphate transport system in Escherichia coli. FEMS Microbiology Letters 184:231–235. 183. Ammerman JW, Hood RR, Case DA, Cotner JB. 2003. Phosphorus deficiency in the Atlantic: An emerging paradigm in oceanography. Eos, Transactions American Geophysical Union 84:165–170. 184. Cavender-Bares KK, Karl DM, Chisholm SW. 2001. Nutrient gradients in the

135

western North Atlantic Ocean: Relationship to microbial community structure and comparison to patterns in the Pacific Ocean. Deep Sea Research I 48:2373– 2395. 185. Mahaffey C, Williams RG, Wolff GA, Anderson WT. 2004. Physical supply of nitrogen to phytoplankton in the Atlantic Ocean. Global Biogeochemical Cycles 18:1–12. 186. Wurl O, Zimmer L, Cutter GA. 2013. Arsenic and phosphorus biogeochemistry in the ocean: Arsenic species as proxies for P-limitation. Limnology and Oceanography 58:729–740. 187. Yang HC, Cheng J, Finan TM, Rosen BP, Bhattacharjee H. 2005. Novel Pathway for Arsenic Detoxification in the Legume Symbiont Sinorhizobium meliloti. J Bacteriol 187:6991–6997. 188. Wanner BL. 1992. Genes for phosphonate biodegradation in Escherichia coli. SAAS Bull. Biochem. Biotechnol. 5:1–6. 189. Wanner BL, Metcalf WW. 1992. Molecular genetic studies of a 10.9-kb operon in Escherichia coli for phosphonate uptake and biodegradation. FEMS Microbiology Letters 100:133–139. 190. Moore L, Ostrowski M, Scanlan D, Feren K, Sweetsir T. 2005. Ecotypic variation in phosphorus-acquisition mechanisms within marine picocyanobacteria. Aquatic Microbial Ecology 39:257–269. 191. Vergin KL, Beszteri B, Monier A, Thrash JC, Ben Temperton, Treusch AH, Kilpert F, Worden AZ, Giovannoni SJ. 2013. High-resolution SAR11 ecotype dynamics at the Bermuda Atlantic Time-series Study site by phylogenetic placement of pyrosequences. ISME J. doi: 10.1038/ismej.2013.32 192. Carlson CA, Morris R, Parsons R, Treusch AH, Giovannoni SJ, Vergin K. 2009. Seasonal dynamics of SAR11 populations in the euphotic and mesopelagic zones of the northwestern Sargasso Sea. ISME J 3:283–295. 193. Natarajan K, Dugdale R. 1966. Bioassay and distribution of thiamine in the sea. Limnology and Oceanography 11:621–629. 194. Natarajan K. 1970. Distribution and Significance of Vitamin B12 and Thiamine in the Subarctic Pacific Ocean. Limnology and Oceanography 15:655–659. 195. Natarajan KV. 1968. Distribution of thiamine, biotin, and niacin in the sea. Appl Microbiol 16:366–369. 196. Okbamichael M, Saņudo-Wilhelmy S. 2004. A new method for the determination of vitamin B12 in seawater. Analytica Chimica Acta 517:33–38. 197. Okbamichael M, Sañudo-Wilhelmy SA. 2005. Direct determination of vitamin B1 in seawater by solid-phase extraction and high performance liquid chromatography quantification. Limnol. Oceanogr.: Methods 3:241–246. 198. Webb ME, Marquet A, Mendel RR, Rébeillé F, Smith AG. 2007. Elucidating biosynthetic pathways for vitamins and cofactors. Nat Prod Rep 24:988–1008. 199. Teller JH, Powers SG, Snell EE. 1976. Ketopantoate hydroxymethyltransferase. J Biol Chem 251:3780–3785. 200. Merkamm M, Chassagnole C, Lindley ND, Guyonvarch A. 2003. Ketopantoate reductase activity is only encoded by ilvC in Corynebacterium glutamicum. J

136

Biotechnol 104:253–260. 201. Primerano DA, Burns RO. 1983. Role of acetohydroxy acid isomeroreductase in biosynthesis of pantothenic acid in Salmonella typhimurium. J Bacteriol 153:259–269. 202. Cronan JE. 1980. β-Alanine Synthesis in Escherichia coli. J Bacteriol 141:1291– 1297. 203. Terano S, Suzuki Y. 1978. Formation Of β-Alanine From Spermine And Spermidine In Maize Shoots. Phytochemistry 17:148–149. 204. Rathinasabapathi B. 2002. Propionate, a source of β-alanine, is an inhibitor of β- alanine methylation in Limonium latifolium, Plumbaginaceae. J Plant Physiol 159:671–674. 205. Van Kuilenburg ABP, Stroomer AEM, Van Lenthe H, Abeling NGGM, Van Gennip AH. 2004. New insights in dihydropyrimidine dehydrogenase deficiency: a pivotal role for β-aminoisobutyric acid? Biochem J 379:119–124. 206. Kim J, Kershner JP, Novikov Y, Shoemaker RK, Copley SD. 2010. Three serendipitous pathways in E. coli can bypass a block in pyridoxal-5'-phosphate synthesis. Mol Syst Biol 6:1–13. 207. Tani Y, Dempsey WB. 1973. Glycolaldehyde is a precursor of pyridoxal phosphate in Escherichia coli B. J Bacteriol 116:341–345. 208. Shimizu S, Dempsey W. 1978. 3-hydroxypyruvate substitutes for pyridoxine in serC mutants of Escherichia coli K-12. J Bacteriol 134:944–949. 209. Streit WR, Entcheva P. 2003. Biotin in microbes, the genes involved in its biosynthesis, its biochemical role and perspectives for biotechnological production. Applied microbiology and biotechnology 61:21–31. 210. Bower S, Perkins J, Yocum RR, Serror P, Sorokin A, Rahaim P, Howitt CL, Prasad N, Ehrlich SD, Pero J. 1995. Cloning and characterization of the Bacillus subtilis birA gene encoding a repressor of the biotin operon. J Bacteriol 177:2572–2575. 211. Perkins JB, Bower S, Howitt CL, Yocum RR, Pero J. 1996. Identification and characterization of transcripts from the biotin biosynthetic operon of Bacillus subtilis. J Bacteriol 178:6361–6365. 212. Weaver LH, Kwon K, Beckett D, Matthews BW. 2001. Corepressor-induced organization and assembly of the biotin repressor: a model for allosteric activation of a transcriptional regulator. Proceedings of the National Academy of Sciences 98:6045–6050. 213. Wilson KP, Shewchuk LM, Brennan RG, Otsuka AJ, Matthews BW. 1992. Escherichia coli biotin holoenzyme synthetase/bio repressor crystal structure delineates the biotin- and DNA-binding domains. Proceedings of the National Academy of Sciences 89:9257–9261. 214. Cronan JE. 1989. The E. coli bio operon: transcriptional repression by an essential protein modification enzyme. Cell 58:427–429. 215. Rodionov DA, Mironov AA, Gelfand MS. 2002. Conservation of the biotin regulon and the BirA regulatory signal in Eubacteria and Archaea. Genome Res 12:1507–1516.

137

216. Banerjee R, Ragsdale SW. 2003. The many faces of vitamin B12: catalysis by cobalamin-dependent enzymes. Annu. Rev. Biochem. 72:209–247. 217. Warren M, Raux E, Schubert H, Escalante-Semerena J. 2002. The biosynthesis of adenosylcobalamin (vitamin B 12). Nat Prod Rep 19:390–412. 218. Rodionov DA, Vitreschak AG, Mironov AA, Gelfand MS. 2003. Comparative genomics of the vitamin B12 metabolism and regulation in prokaryotes. J Biol Chem 278:41148–41159. 219. Raux E, Schubert HL, Warren MJ. 2000. Biosynthesis of cobalamin (vitamin B12): a bacterial conundrum. Cell Mol Life Sci 57:1880–1893. 220. Martens J, Barg H, Warren M, Jahn D. 2002. Microbial production of vitamin B 12. Applied microbiology and biotechnology 58:275–285. 221. Hebbeln P, Rodionov DA, Alfandega A, Eitinger T. 2007. Biotin uptake in prokaryotes by solute transporters with an optional ATP-binding cassette- containing module. Proceedings of the National Academy of Sciences 104:2909– 2914. 222. Hayaishi O, Nishizuka Y, Tatibana M, Takeshita M, Kuno S. 1961. Enzymatic studies on the metabolism of β-alanine. J Biol Chem 236:781–790. 223. Vraspir JM, Butler A. 2009. Chemistry of marine ligands and siderophores. Ann Rev Mar Sci 1:43–63. 224. Sunda W, Huntsman S. 1995. Cobalt and zinc interreplacement in marine phytoplankton: Biological and geochemical implications. Limnology and Oceanography 40:1404–1417. 225. Saito M, Moffett J, Chisholm S, Waterbury J. 2002. Cobalt limitation and uptake in Prochlorococcus. Limnology and Oceanography 47:1629–1636. 226. Rocap G, Larimer FW, Lamerdin J, Malfatti S, Chain P, Ahlgren NA, Arellano A, Coleman M, Hauser L, Hess WR, Johnson ZI, Land M, Lindell D, Post AF, Regala W, Shah M, Shaw SL, Steglich C, Sullivan MB, Ting CS, Tolonen A, Webb EA, Zinser ER, Chisholm SW. 2003. Genome divergence in two Prochlorococcus ecotypes reflects oceanic niche differentiation. Nature 424:1042–1047. 227. Moore L, Chisholm S. 1999. Photophysiology of the marine cyanobacterium Prochlorococcus: ecotypic differences among cultured isolates. Limnology and Oceanography 628–638. 228. Nishijima T, Hata Y. 1985. Distribution of vitamin B12, thiamine, and biotin in Hiuchi-Nada sea. Research Reports of the Kochi University. Agricultural Science 34:57–69. 229. Ohwada K. 1972. Bioassay of biotin and its distribution in the sea. Marine Biology 14:10–17. 230. Woyke T, Tighe D, Mavromatis K, Clum A, Copeland A, Schackwitz W, Lapidus A, Wu D, McCutcheon JP, McDonald BR, Moran NA, Bristow J, Cheng J-F. 2010. One bacterial cell, one complete genome. PLoS ONE 5:e10314. 231. Penna VTC, Martins SAM, Mazzola PG. 2002. Identification of bacteria in drinking and purified water during the monitoring of a typical water purification system. BMC Public Health 2:1–11.

138

232. Robbertse B, Yoder RJ, Boyd A, Reeves J, Spatafora JW. 2011. Hal: an Automated Pipeline for Phylogenetic Analyses of Genomic Data. PLoS Curr 3:RRN1213. 233. Sharpton TJ, Jospin G, Wu D, Langille MG, Pollard KS, Eisen JA. 2012. Sifting through genomes with iterative-sequence clustering produces a large, phylogenetically diverse protein-family resource. BMC bioinformatics 13:264– 264. 234. Eddy SR. 1998. Profile hidden Markov models. Bioinformatics 14:755–763. 235. Edgar RC. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792–1797. 236. Castresana J. 2000. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol Biol Evol 17:540–552. 237. Darriba DD, Taboada GLG, Doallo RR, Posada DD. 2011. ProtTest 3: fast selection of best-fit models of protein evolution. Bioinformatics 27:1164–1165. 238. Stamatakis A. 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22:2688– 2690.

139

APPENDICES

140

The appendices herein contain documentation submitted (or are intended to be submitted) as supplemental information documents with manuscripts for publication. Appendix 1 is published as a supplement to: doi:10.1038/ismej.2012.122. Appendix 2 will accompany the submission of Chapter 3 and appendix 3 will accompany the submission of Chapter 4.

Appendix 1

107 ) -1

106

5 celldensity (cells ml 10

104 0 2 4 6 8 10 12 time (days)

Figure A1- 1: Transfer of Ca. P. ubique batch cultures from natural seawater to AMS1. ▲: Ca. P. ubique transferred from natural seawater amended with pyruvate (50 µM), glycine (1 µM), methionine (1 µM) and vitamins to identically amended natural seawater medium. ●: Ca. P. ubique transferred from natural seawater amended with pyruvate (50 µM), glycine (1 µM), methionine (1 µM) and vitamins to identically amended AMS1. Points are the mean of triplicate cell density measurements. Error bars: ± 1.0 s.d., n=3. When error bars are not visible, they are smaller than the size of the symbols.

141

† ‡

A B C D E

F G H I J

Figure A1- 2: Flow cytometry profiles and corresponding microscopic images at different pyruvate:glycine molar ratios. All cultures were grown to stationary-phase in AMS1 containing methionine (10 µM) with the following glycine and pyruvate concentrations: A) 0 µM pyruvate, 50 µM glycine; B) 0.1 µM pyruvate, 50 µM glycine; C) 0.9 µM pyruvate, 50 µM glycine; D) 5 µM pyruvate, 50 µM glycine; E) 25 µM pyruvate, 50 µM glycine; F) 50 µM pyruvate, 0 µM glycine; G) 50 µM pyruvate, 0.1 µM glycine; H) 50 µM pyruvate, 0.9 µM glycine; I) 50 µM pyruvate, 5 µM glycine; J) 50 µM pyruvate, 25 µM glycine. For each treatment, the upper panel shows relative DNA fluorescence profiles of SYBR Green I-stained cells; the lower panel shows microscopic images of SYBR green I-stained cells. In (A), †: the relative DNA fluorescence of single cells; ‡: the relative DNA fluorescence of doublets

142

time after inoculation (days) 19 2000 nM 500 nM 250 nM 125 nM 0 nM 10 ) -1

4 cells ml 6 x 10

cell density ( 0 2000 nM 500 nM 250 nM 125 nM 0 nM alanine concentration

Figure A1- 3: DNA fluorescence proﬁles and maximum cell densities in response to alanine additions. Top panel: Relative DNA fluorescence time course of batch cultures grown in AMS1 amended with limiting pyruvate (0.5 µM) glycine (1 µM), methionine (1 µM) and different L-alanine concentrations, as indicated. Bottom panel: Maximum cell densities of cultures for which relative DNA fluorescence values are shown in the top panel.

143

1.60E+06

1.40E+06 )

-1 1.20E+06

1.00E+06

8.00E+05

6.00E+05

4.00E+05 maximum density (cells ml

2.00E+05

0.00E+00 0 2 4 6 8 10 12 Dose of vitamin mix added (fold)

Figure A1- 4: Effect of vitamin additions on Ca. P. ubique growth. Ca. P. ubique grown in AMS1 without added organic carbon and different doses of freshly prepared vitamins (as listed in Table 1). Points show the maximum density achieved of single flasks (triplicates for each dose are shown).

Discussion of Figure A1-4 in appendix 1(above): Upon exclusion of the vitamin mixture and in the absence of added organic carbon, Ca. P. ubique grew to densities of ca. 2.0 ×105 cells ml-1. Higher densities, exceeding 1.2 ×106 cells ml-1, were observed, in a dose-dependent manner, when a freshly prepared vitamin solution was added (Figure

A1-4, above). We speculate that Pelagibacter makes efficient use of organic carbon that enters the cultures as reagent contaminants (for example, in the vitamin stocks), or by

144 diffusion from plastic culture ware and the air. Previously Ca. P. ubique has been shown to use a variety of volatile organic carbon compounds as sources of electrons for respiration (49). Also in support of this interpretation, we note that the growth of bacteria in water purification systems is not uncommon (231). Collectively, we conclude that unintentional trace levels of nutrients entering our experiments, coupled with the low amounts of carbon required by Ca. P. ubique, is a probable explanation for the observed growth without added carbon.

145

Appendix 2

D-glucose-6P Pentose-phosphate pathway TktB,C,N D-fructose-6P D-xylulose-5P D-ribulose-5P D-ery-4P Non-mevalonate D-ribose-5P isoprenoid bioynthesis TktB,C,N PRPP Dxs D-sedohep-P deoxy-D-xylulose-5P GA-3P nucleotides B B 1 6 α-KG PEP SucA CO2 SucB isocitrate PdhD succinyl-CoA retinol ubi pyruvate undecaprenyl-PP AceE IlvB AceF TCA cycle CO2 CO2 IlvH PdhD succinate valine acetolactate acetyl-CoA leucine OAA pantothenate malate

Figure A2- 1: Simplified illustration of Ca. P. ubique’s central metabolism, highlighting thiamine-requiring biosynthetic reactions. Red lines and enzyme abbreviations depict thiamine-dependent biosynthetic steps and enzymes, respectively. Gene abbreviations are listed as they are annotated in the Ca. P. ubique genome: Dxs: 1-D-deoxyxylulose 5- phosphate synthase; AceE: pyruvate dehydrogenase (lipoamide) E1 component; AceF: dihydrolipoamide S-acetyltransferase; PdhD: dihydrolipoyl dehydrogenase; IlvB: acetolactate synthase large subunit (ALS III); IlvH: acetolactate synthase small subunit; TktB: transketolase; TktC: transketolase-family protein; TktN: transketolase; SucA: 2- oxoglutarate dehydrogenase E1 component; SucB: 2-oxoglutarate dehydrogenase complex E2 component. Compound abbreviations: OAA: oxaloacetate; α-KG: α- ketoglutarate; PRPP: phosphoribosyl pyrophosphate; D-ery-4P: D-erythrose 4-phosphate; D-sedohep-P: D-sedoheptulose 7-phosphate; GA-3P: glyceraldehyde 3-phosphate; PEP: phosphoenolpyruvate; ubi: ubiquinone; B1: thiamine; B6: pyridoxine. TCA cycle: Tricarboxylic acid cycle.

146

1 μM thiamine no added thiamine 8 10 -thiamine +thiamineT2 T2 ) -1

+thiamineT1 107

-thiamineT1 106 cell density ml cell (cells

105

0 10 20 30 time (days) Figure A2- 2: Reproduction of Figure 3-2a in main text with microarray sample points labeled.

3.5 ) 8 × 10 -1

ml 2.5

1.5

0.5

maximum density (cells 0 no H+B6 pant THZ thi thi Figure A2- 3: Growth responses of Ca. P. ubique to potential thiamine precursors. Bar heights are the average maximum densities achieved of biological replicates ± 1.0 s.d. (n=3). Potential precursors were added at a final concentration of 1.0 µM, as described in methods. Abbreviations: H: Histidine; B6: pyridoxine (vitamin B6); pant: pantothenate; THZ: 4-methyl-5-thiazoleethanol; thi: thiamine.

147

Figure A2- 4: Maximum-likelihood phylogenetic reconstruction of the SAR11_0811 amino acid sequence. The tree is rooted at its midpoint. The node labels are bootstrap values based on 1,000 replicates. Branches are labeled with the gene locus identifier and organism name. Branches in red are further described in Figure A2-5 in appendix 2. See supplementary text in appendix 2 for details of tree construction.

148

Figure A2-4: Maximum-likelihood phylogenetic reconstruction of the SAR11_0811 amino acid sequence.

149 G G GA C T T T GA T C GA G A C T C A A CT ACT GA CA G GC AG C GGAAAG G GT T CATA G G TT CGG A A A TGA 2 6 1 8 C 3 4 5 7 9 T A G T T CTG AT 16 42 65 70 11 12 15 17 23 28 29 32 33 34 36 38 39 40 43 51 56 60 61 62 66 67 68 A T C10 T 13 A14 A 18 19 20 21 22 24 25 26 A27 A 30 C31 T 35 37 41 44 45 46 47 G48 49 50 52 53 T54 C55 T 57 A58 59 63 64 G 69 71 GGGG G G CTGAGA T AC C AAC CTG CA TAAT AGG A A AAGGGGTGTCGTAACTGACTGAGAATAAACCCTAATAACCTGTCC--GGTAAT--TCAGAAAGGGAGAACA AAGGGGTGTCGTAACTGACTGAGATTAAACCCTAATAACCTGTCC--GGTAAT--TCAGAAAGGGAGAACA* AAGGGGTGTCGTAACTGACTGAGAATAAACCCTAATAACCTGTCC--GGTAAT--TCAGAAAGGGAGAACA AAGGGGTGTCGTAACAGACTGAGATTAAACCCTAAGAACCTGTCC--GGTAAT--TCAGAAAGGGAGAATA AAGGGGTGTCGTAAAAGACTGAGATTAAACCCTTAGAACCTGTCC--GGTAAT--TCAGAAAGGGAAAAAA AGGGGGTGTCTTTACAGACTGAGATTAAACCCCAGGAACCTGGCT--GGTAAT--TCAGTAAGGGAATAAA AAGGGGTGCCTTTAATGGCTGAGAGTAAACTCTAGGAACCTGACT--GGTAAT--TCAGTGAGGGAATAGG CAGGGGTGCGAAAAC--GCTGAGATTATACCCTACGAACCTGATG--TTTAATTACAACGAAGGGAGAACC CAGGGGCGGCTTAGC---CTGAGA-TGAACCCTTTAAACCTGATCCAGATAATGCTGGCGGAGGAAGAAAA TGTGATTCCTGGGGGCGCCATTGCGATGGCTGAGACTCCGCACTTTGTGCGGTAACCCTTCGAACCTGATCCGGGTAATGCCGGCGAAGGAAGGATAAAGCCTGCTTT* AAAAATTGTAAGGGGTGTCGTAACTGACTGAGAATAAACCCTAATAACCTGTCCGGTAATTCAGAAAGGGAGAACAATGGATAAAACAT* thiD thiM thiE Putative HMP transporter tenA tenA Putative HMP transporter tenA tenA Putative HMP transporter tenA Putative HMP transporter tenA tenA Putative HMP transporter tenA tenA Putative HMP transporter

Figure A2- 5: Illustration of predicted thiamine regulatory elements or thiamine pyrimidine salvage genes genetically associated with the putative HMP transport protein coding sequence. Phylogeny is reproduced from Figure A2-4 in appendix 2, with modifications for clarity. Nucleotide sequences are located immediately upstream (5’) of the putative HMP transport protein. Sequences that are marked with (*) were predicted to contain ThPP-binding motifs by rfam (http://rfam.sanger.ac.uk/). Dashed box encapsulates sequences from 9 of 10 sequenced Pelagibacter genomes and their consensus sequence is illustrated.

Supplementary methods construction of phylogenetic trees:

Construction of phylogenetic tree in Figure A2-4 & A2-5 in appendix 2: To determine genomic context of ‘proline uptake protein’ (SAR11_0811) homologs in other organisms, SAR11_0811 was used to construct a gene tree in MicrobesOnline

(http://microbesonline.org/). The pre-computed tree for SAR11_0811 using COG591

(amino acid residues 13-446) demonstrated conserved features in the genomes of

Methylobacillus flagellatus KT, Marinobacter sp. ELB17, Clostridium sp. OhILAs,

Haloquadratum walsbyi DSM 16790, Haloarcula marismortui ATCC 43049,

Halorhabdus utahensis DSM 12940, Haloferax volcanii DS2 and Halogeometricum borinquense PR3, DSM 11551.

150

Analysis of the putative ‘proline uptake protein’ transporter SAR11_0811 using

Hal (232) revealed homologous sequences in eight other SAR11 genomes - HTCC1002,

HTCC9565, HTCC7211, HIMB5, AAA240-E13, AAA288-G21, HIMB114, and

IMCC9063. The nine SAR11 and eight additional homologs from Gram-positive bacteria, Archaea, and proteobacteria were searched against the Hidden Markov Model

(HMM) database of over 430,000 protein families, the so-called “Sifting Families”

(SFam) (233), using hmmscan on default settings as part of the hmmer 3.0 package (234):

$ hmmscan --domtblout txport.tab -o txport.hsc Sifted_Families/SFam_HMMs_all txport_starters.faa

All 17 genes used in the initial search had best hits to one of two SFam HMMs,

346434 or 31377. Inspection of the members of these SFams revealed that 346434 contained several thousand bacterial and archaeal members with various transport annotations, whereas 31377 contained 23 genes exclusively from fungal taxa annotated predominantly as hypothetical proteins. A Prokaryotic tree was constructed using BLAST to select sequences from the 346434 SFam. The fasta file for the SFam was generated using the tree_parse.py script:

$ python tree_parse.py --model 346434 | sed 's/^/lcl|/g' | blastdbcmd -entry_batch - -db img_v350_prot > 346434_v350.faa

This file was used to generate a database for BLASTP searches:

$ makeblastdb -in 346434_v350.faa -out 346434_SEQS -parse_seqids -hash_index $ blastp -query halo_SAR_starters.faa -db 346434_SEQS -out 346434_like_SAR11 - outfmt 7 -max_target_seqs 50" -r SAR11_346434_blastp

Unique hits with E-values > 1e-30 were selected for construction of the tree:

151

These were parsed and run through a bash script (protPipeline3) automating alignment with MUSCLE (235), curation with GBlocks (236), amino acid substitution modeling with ProtTest (237), and maximum-likelihood tree construction with RAxML (238):

$ cat 346434_like_SAR11.faa halo_SAR_starters.faa > transport_group6.faa $ cat transport_group6.faa | sed 's/\ .*//g' | sed 's/lcl|//g' > transport_group6_cln.faa $ bash protPipeline3 transport_group6_cln.faa

protPipeline3:

#!/bin/bash filename=$1 muscle -in $1 -out $filename.al cat $filename.al | fasta2phy | phyToFasta.pl > $filename.tmp rm -f $filename.al mv $filename.tmp $filename.al seqs=`grep -c ">" $filename.al` Gblocks $filename.al -t=p -p=n -e=.fst -b1=$(((seqs/2)+1)) -b2=$(((seqs/2)+1)) - b3=$(((seqs/2))) -b4=2 -b5=h mv $filename.al.fst $filename.gb cat $filename.gb | fasta2phy > $filename.phy java -Xmx250m -classpath /local/cluster/spatafora/ProtTest/ProtTest.jar prottest.ProtTest -i $filename.phy -o $filename.model -sort A -S 0 -t1 T -t2 T model=`grep "Best model according to" $filename.model | awk '{print$6}' | awk -F+ '{print $1'} | tr "[:lower:]" "[:upper:]"` raxmlHPC -f a -m PROTCAT$model -n $filename -N 1000 -s $filename.phy -p 1234 -x 1234

phyTofasta.pl:

#!/usr/bin/perl –w # $Header$ # $Log$ # Revision 1.1 2006/08/07 19:09:38 reevesj # converts a phylip file to fasta format. can read/write standard in/out for pipelining. #

152

# use Getopt::Std; use Bio::AlignIO; use Bio::SeqIO; my %opt; getopts('i:o:h', \%opt); checkOpts(\%opt); my ($inaln, $aln, $outobj, $seq); if(defined($opt{i})){ close(STDIN); open(STDIN, $opt{i}); } if(defined($opt{o})){ close(STDOUT); open(STDOUT, ">$opt{o}"); } $inaln = Bio::AlignIO->newFh(-format => 'phylip', -fh => \*STDIN); warn("reading from stdin\nphyToFasta.pl -h for more help\n"); $outobj = Bio::SeqIO->newFh(-format => 'fasta'); $aln = <$inaln>; foreach $seq ($aln->each_seq()){ print($outobj $seq); } exit(0); sub checkOpts{ $opt = shift(); if($$opt{h}){ opterr(); } return; } sub opterr{ #getopts('b:g:d:m:h'); print <<__HELP__; Usage: $0 [-ioh] -h Print this help message. -i Phylip format alignment file. If unspecified, read from standard in. Optional. -o File where converted fasta file will be written.

If unspecified, write to standard out. Optional.

153

__HELP__ exit(1); }

fasta2phy :

#!/usr/bin/perl -w # $Id: fasta2phy 905 2010-05-14 22:11:23Z alexeb112 $ # # Copyright 2005,2006,2007 John Reeves and Ryan Yoder # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 3 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program. If not, see . use Getopt::Std; use File::Spec; my (undef,undef, $SCRIPT_NAME ) = File::Spec->splitpath( $0 ); my %opt; getopts('i:o:h', \%opt); checkOpts(\%opt); my (%aln, @anames, $cname); if(defined($opt{i})){ close(STDIN); open(STDIN, $opt{i}); } while(){ chomp(); if($_ =~ /^\s*$/){ next; } $_ =~ s/\s+//g; if(substr($_, 0, 1) eq '>'){ $cname = substr($_, 1);

154

$aln{$cname} = ''; push(@anames, $cname); }else{ $aln{$cname} .= $_; } } my $l = length($aln{$cname}); #make sure the alignments are actually the same length. foreach $name (@anames){ if(length($aln{$name}) != $l){ die("Error: not all alignments are the same length.\n"); } } if(defined($opt{o})){ close(STDOUT); open(STDOUT, ">$opt{o}"); } print scalar(@anames). ' '. length($aln{$cname}). "\n"; $i=0; my $donames = 1; my $tmpseq=''; my $tmpname=''; while($i<$l){ if($donames){ $donames = 0; foreach $name (@anames){ $tmpseq = substr($aln{$name}, $i, 50); $tmpseq =~ s/(.{10})/$1 /g; $tmpname=$name; $tmpname =~ s/,.*$//g; #remove commas from name, some programs hate them. print substr(sprintf('%-10s', $tmpname), 0, 10). ' '; print "$tmpseq\n"; } }else{ foreach $name (@anames){ $tmpseq = substr($aln{$name}, $i, 50); $tmpseq =~ s/(.{10})/$1 /g; printf("%11s%s\n", ' ', $tmpseq); } } print "\n"; $i += 50; } exit(0);

155 sub checkOpts{ $opt = shift(); if($$opt{h}){ opterr(); } return; } sub opterr{ print <<__HELP__; fasta2phy: convert a fasta formatted file into interleaved phylip format. Does an in place conversion if -o is given and the input is the same file as -o. Usage: $SCRIPT_NAME [-ioh] Convert a fasta formatted file into interleaved phylip format. Command line options: -h Print this help message. -i Fasta format alignment file. If unspecified, read from standard in. Optional. -o File to which converted alignment will be written. If unspecified, write to standard out. Optional. __HELP__ exit(1); }

The initial tree contained duplicate sequences, which were removed via RAxML. The trimmed alignment was re-run with RAxML using the following settings:

$ raxmlHPC-PTHREADS -x 1234 -T 2 -f a -s transport_group6_cln.red.phy -n 6_red -m PROTCATCPREV -\# 1000

156

Figure A2- 6: Profiles of in situ temperature and chlorophyll a fluorescence at Hydrostation S in the Sargasso Sea during water collection for vitamin analysis. Panel A: Cruise number AE1224, cast 2 at 20:00 (local time), 19 September 2012. Panel B: Cruise number AE1224, cast 4 at 08:00 (local time), 20 September 2012.

157

Table A2- 1: Gene transcripts more abundant in thiamine-limited conditions fold ORF name Gene Annotation q-value change SAR11_0004 miaA tRNA isopentenyltransferase 2.28 ≤0.005 [EC:2.5.1.8] SAR11_0015 pepP Xaa-Pro aminopeptidase [EC:3.4.11.9] 2.11 ≤0.005 SAR11_0021 ddlB D-alanine--D-alanine ligase B 2.74 ≤0.005 [EC:6.3.2.4] SAR11_0022 murB UDP-N- 2.92 ≤0.005 acetylenolpyruvoylglucosamine reductase [EC:1.1.1.158] SAR11_0023 murC UDP-N-acetylmuramate--L-alanine 2.30 ≤0.005 ligase [EC:6.3.2.8] SAR11_0028 murE UDP-N-acetylmuramoylalanyl-D- 2.11 ≤0.005 glutamate [EC:6.3.2.10, 6.3.2.13] SAR11_0029 ftsI cell division protein FtsI 2.67 ≤0.005 [EC:2.4.1.129] SAR11_0082 probable short-chain dehydrogenase 2.33 ≤0.005 SAR11_0093 hypothetical protein 2.04 ≤0.005 SAR11_0111 mutY A/G-specific adenine glycosylase 2.65 ≤0.005 [EC:3.2.2.-] SAR11_0121 crtB phytoene synthase [EC:2.5.1.32] 2.35 ≤0.005 SAR11_0148 kdtA 3-deoxy-D-manno-octulosonic-acid 2.26 ≤0.005 transferase SAR11_0157 lspA lipoprotein signal peptidase 2.26 ≤0.005 [EC:3.4.23.36] SAR11_0159 ribF riboflavin biosynthesis protein 2.17 ≤0.005 [EC:2.7.7.2, 2.7.1.26] SAR11_0199 chvI two-component system response 2.29 ≤0.005 regulator SAR11_0200 polA DNA polymerase I [EC:2.7.7.7] 2.04 ≤0.005 SAR11_0206 speB arginase [EC:3.5.3.1] 2.01 ≤0.005 SAR11_0219 gloB hydroxyacylglutathione hydrolase 2.33 ≤0.005 cytoplasmic [EC:3.1.2.6] SAR11_0223 proB glutamate 5-kinase [EC:2.7.2.11] 2.60 ≤0.005 SAR11_0234 priA probable primosomal protein N 2.04 ≤0.005 SAR11_0251 leuD 3-isopropylmalate dehydratase 2.27 ≤0.005 [EC:4.2.1.33] SAR11_0317 rpe ribulose-phosphate 3-epimerase 2.04 ≤0.005 [EC:5.1.3.1] SAR11_0318 hypothetical protein 2.36 ≤0.005

158

Table A2-1 (continued)

SAR11_0343 hypothetical protein 2.20 ≤0.005 SAR11_0362 hypothetical protein 2.22 ≤0.005 SAR11_0365 nth probable endonuclease III 2.76 ≤0.005 [EC:4.2.99.18] SAR11_0408 mutM formamidopyrimidine-DNA 2.05 ≤0.005 glycosylase [EC:3.2.2.23] SAR11_0443 pheT phenylalanine-tRNA ligase 2.15 ≤0.005 [EC:6.1.1.20] SAR11_0457 fmt methionyl-tRNA formyltransferase 2.01 ≤0.005 [EC:2.1.2.9] SAR11_0494 aroA 3-phosphoshikimate 1- 2.05 ≤0.005 carboxyvinyltransferase [EC:2.5.1.19] SAR11_0590 5-formyltetrahydrofolate cyclo-ligase 2.15 ≤0.005 [EC:6.3.3.2] SAR11_0602 folP dihydropteroate synthase-like protein 2.48 ≤0.005 [EC:2.5.1.15] SAR11_0604 thiD phosphomethylpyrimidine kinase 2.06 ≤0.005 [EC:2.7.4.7] SAR11_0610 ispA geranyltranstransferase [EC:2.5.1.10] 2.06 ≤0.005 SAR11_0634 umuC DNA polymerase IV [EC:2.7.7.7] 2.21 ≤0.005 SAR11_0636 hypothetical protein 2.35 ≤0.005 SAR11_0638 phosphoribulokinase/uridine kinase 2.37 ≤0.005 family protein SAR11_0684 mobA molybdopterin-guanine dinucleotide 2.12 ≤0.005 biosynthesis protein A SAR11_0712 surA SurA-like protein [EC:5.2.1.8] 2.38 ≤0.005 SAR11_0818 amtB ammonium transporter 2.42 ≤0.005 SAR11_0831 pabB para-aminobenzoate synthase 2.89 ≤0.005 component I [EC:6.3.5.8] SAR11_0920 gltX glutamyl-tRNA synthetase 2.07 ≤0.005 [EC:6.1.1.17] SAR11_0931 putative transmembrane protein 2.07 ≤0.005 SAR11_0993 recG ATP-dependent DNA helicase RecG 2.41 ≤0.005 [EC:3.6.1.-] SAR11_0996 mltB membrane-bound lytic transglycolase- 2.22 ≤0.005 related protein SAR11_1010 recJ single-stranded-DNA-specific 2.01 ≤0.005 exonuclease SAR11_1019 xerD integrase/recombinase XerD-like 2.04 ≤0.005 protein

159

Table A2-1 (continued)

SAR11_1173 bhmT betaine-homocysteine S- 2.17 ≤0.005 methyltransferase [EC:2.1.1.5] SAR11_1248 winged helix DNA-binding 2.09 ≤0.005 SAR11_1330 putative lipoprotein 2.23 ≤0.005 SAR11_1331 ppiA peptidylprolyl cis-trans isomerase 2.26 ≤0.005 precursor [EC:5.2.1.8] SAR11_1342 alpha/beta hydrolase fold 2.37 ≤0.005

160

Table A2- 2: Gene transcripts less abundant in thiamine-limited conditions

fold ORF name Gene Annotation q-value change SAR11_0001 ilvC ketol-acid reductoisomerase 2.56 ≤0.005 [EC:1.1.1.86] SAR11_0007 hflC probable integral membrane 2.50 ≤0.005 proteinase SAR11_0008 hflK probable integral membrane 2.21 ≤0.005 proteinase [EC:3.4.-.-] SAR11_0060 pilQ type II secretion PilQ 2.04 ≤0.005 SAR11_0098 fbcH ubiquinol-cytochrome-c reductase 3.26 ≤0.005 [EC:1.10.2.2] SAR11_0099 petB ubiquinol-cytochrome-c reductase 2.36 ≤0.005 [EC:1.10.2.2] SAR11_0100 petA cytochrome b6-f complex iron-sulfur 2.45 ≤0.005 subunit [EC:1.10.2.2] SAR11_0117 atpE H+-transporting two-sector ATPase 2.49 ≤0.005 (subunit C) [EC:3.6.3.14] SAR11_0130 coxC cytochrome-c oxidase [EC:1.9.3.1] 2.71 ≤0.005 SAR11_0134 cox1 cytochrome-c oxidase [EC:1.9.3.1] 2.16 ≤0.005 SAR11_0135 coxB probable cytochrome c oxidase 2.08 ≤0.005 polypeptide II (cytochrome aa3 subunit 2) [EC:1.9.3.1] SAR11_0162 groE 60 kDa chaperonin 2.01 ≤0.005 SAR11_0183 xanthine/uracil/vitamin C permease 2.52 ≤0.005 family protein SAR11_0197 ahcY S-adenosyl-L-homocysteine hydrolase 2.47 ≤0.005 [EC:3.3.1.1] SAR11_0221 rpmA 50S ribosomal protein L27 2.72 ≤0.005 SAR11_0229 atpC ATP synthase epsilon chain (ATP 2.62 ≤0.005 synthase F1 sector epsilon subunit) [EC:3.6.3.14] SAR11_0230 atpD F1-ATP synthase beta chain 3.29 ≤0.005 [EC:3.6.3.14] SAR11_0232 atpA H+-transporting two-sector ATPase 2.60 ≤0.005 (F0F1-type ATP synthase) alpha chain [EC:3.6.3.14] SAR11_0348 rho transcription termination factor rho 2.01 ≤0.005 SAR11_0423 carD hypothetical transcriptional regulator 2.04 ≤0.005 SAR11_0500 hypothetical protein 2.38 ≤0.005 SAR11_0504 cycM cytochrome c 2.31 ≤0.005

161

Table A2-2 (continued)

SAR11_0510 glcB malate synthase [EC:2.3.3.9] 2.14 ≤0.005 SAR11_0630 transcriptional regulators, TraR/DksA 2.33 ≤0.005 family SAR11_0644 icdA isocitrate dehydrogenase 2.13 ≤0.005 [EC:1.1.1.42] SAR11_0708 acpP acyl carrier protein (ACP) 2.76 ≤0.005 SAR11_0717 ndk nucleoside diphosphate kinase 2.09 ≤0.005 [EC:2.7.4.6] SAR11_0766 gabD succinate-semialdehyde 2.32 ≤0.005 dehydrogenase (NAD(P)) [EC:1.2.1.16] SAR11_0801 hisD2 histidinol dehydrogenase 2.22 ≤0.005 [EC:1.1.1.23] SAR11_0808 bioA adenosylmethionine-8-amino-7- 2.04 ≤0.005 oxononanoate aminotransferase [EC:2.6.1.62] SAR11_0809 ald alanine dehydrogenase [EC:1.4.1.1] 2.53 ≤0.005 SAR11_0835 aprM adenylylsulfate reductase membrane 3.65 ≤0.005 anchor SAR11_0836 aprB adenylyl-sulfate reductase chain B 3.12 ≤0.005 [EC:1.8.99.2] SAR11_0837 aprA adenylyl-sulfate reductase chain A 3.85 ≤0.005 [EC:1.8.99.2] SAR11_0842 garD probable galactarate dehydratase 2.57 ≤0.005 [EC:4.2.1.42] SAR11_0843 D-galactarate dehydratase/altronate 2.66 ≤0.005 [EC:4.2.1.42]; K01708 galactarate dehydratase SAR11_0844 hypothetical protein 2.71 ≤0.005 SAR11_0861 short chain dehydrogenase, unknown 2.12 ≤0.005 specificity SAR11_0865 TRAP-type mannitol/chloroaromatic 3.01 ≤0.005 compound transport system SAR11_0866 mannitol transporter 2.67 ≤0.005 SAR11_0953 yhdW ABC transporter - general L-amino 2.27 ≤0.005 acid transport system substrate- binding protein SAR11_0954 yhdX ABC transporter - general L-amino 2.12 ≤0.005 acid transport system permease protein

162

Table A2-2 (continued)

SAR11_0957 yhdZ general L-amino acid transport ATP- 2.02 ≤0.005 binding protein SAR11_0979 yajC preprotein translocase 2.19 ≤0.005 SAR11_0980 secD protein-export membrane protein 2.81 ≤0.005 SecD SAR11_0987 ppiB peptidylprolyl isomerase [EC:5.2.1.8] 2.49 ≤0.005 SAR11_1033 rpmG ribosomal protein L33 2.69 ≤0.005 SAR11_1037 rpmF 50S ribosomal protein L32 3.06 ≤0.005 SAR11_1040 hppA membrane-bound proton- 4.75 ≤0.005 translocating pyrophosphatase [EC:3.6.1.1] SAR11_1043 ribH riboflavin synthase 2.02 ≤0.005 SAR11_1048 glyA glycine hydroxymethyltransferase 2.80 ≤0.005 [EC:2.1.2.1] SAR11_1093 rpoA DNA-directed RNA polymerase 2.82 ≤0.005 [EC:2.7.7.6] SAR11_1097 secY preprotein translocase SecY subunit 2.36 ≤0.005 SAR11_1103 rpsH 30S ribosomal protein S8 2.26 ≤0.005 SAR11_1119 fusA translation elongation factor EF-G 3.33 ≤0.005 [EC:3.6.5.3] SAR11_1122 rpoC DNA-directed RNA polymerase beta 2.73 ≤0.005 prime chain [EC:2.7.7.6] SAR11_1123 rpoB DNA-directed RNA polymerase 2.48 ≤0.005 [EC:2.7.7.6] SAR11_1124 rplL probable 50S ribosomal protein L31 2.28 ≤0.005 SAR11_1129 hypothetical protein 4.42 ≤0.005 SAR11_1130 tufB translation elongation factor EF-Tu 4.19 ≤0.005 [EC:3.6.5.3] SAR11_1143 grlA glutaredoxin 2.43 ≤0.005 SAR11_1150 dhs 2-dehydro-3-deoxy-phosphoheptonate 2.35 ≤0.005 aldolase [EC:2.5.1.54] SAR11_1181 hypothetical protein 2.15 ≤0.005 SAR11_1203 tctC putative tricarboxylic transport 2.01 ≤0.005 SAR11_1210 occT ABC transporter - octopine/nopaline 2.94 ≤0.005 transport system substrate-binding protein SAR11_1211 speE spermine/spermidine synthase 3.05 ≤0.005 [EC:2.5.1.16, 4.1.1.50] SAR11_1245 folE GTP cyclohydrolase I [EC:3.5.4.16] 2.38 ≤0.005

163

Table A2-2 (continued)

SAR11_1266 membrane protein of unknown 2.07 ≤0.005 function (DUF1430) SAR11_1274 cspL cold shock DNA-binding domain 3.82 ≤0.005 protein SAR11_1274 cspL cold shock DNA-binding domain 2.03 ≤0.005 protein SAR11_1336 potD spermidine/putrescine-binding 2.02 ≤0.005 periplasmic protein

164

Appendix 3

107 ) -1

106

density (cells ml 105 No PO4 dTTP PO4 dGTP PO3 dCTP vs Me-PO3 Ribose-5P 104 2-AEP Glucose-6P dATP P-serine

5 10 15 20 25 30 35 time (days)

Figure A3- 1: Growth curves of Pelagibacter sp. str. HTCC7211 on alternate phosphorus sources showing diauxic growth pattern. Growth medium was AMS1 with amendments of pyruvate (50 µM), glycine (5 µM), methionine (5 µM), FeCl3 (1 µM) and vitamins. Each alternate phosphorus compound was supplied at a final concentration of 1.0 µM. Points are the mean density of biological replicates ± 1.0 s.d. (n=3).

165

108 1 nM 3.3nM 107 10 nM 33 nM 100 nM 6 1000 nM )

-1 10 3000 nM

105

104 cell density (cells ml

103

0 5 10 15 20 25 30 time (days)

Figure A3- 2: Growth curves for linear dose responses of Pelagibacter sp. str. HTCC7211 grown on Pi. Points are the mean density of biological replicates ± 1.0 s.d. (n=3).

166

1 nM 3.3 nM 7 10 10 nM )

-1 33 nM 100 nM 1000 nM 6 10 3000 nM

105 cell density (cells ml

104

0 5 10 15 20 25 time (days)

Figure A3- 3: Growth curves for linear dose responses of Ca. P. ubique grown on Pi. Points are the mean density of biological replicates ± 1.0 s.d. (n=3).

167

1 nM 3.3 nM 107

) 10 nM -1 33.3 nM 100 nM 3000 nM 106

cell density (cells ml 105

5 10 15 20 time (days)

Figure A3- 4: Growth curves for linear dose responses of Pelagibacter sp. str. HTCC7211 grown on Mpn. Points are the mean density of biological replicates ± 1.0 s.d. (n=3).

168

Figure A3-5: Simplified illustration of phospholipid biosynthetic routes and predicted reactions involved in lipid remodeling in Pelagibacter sp. str. HTCC7211. Red lines: Canonical lipid biosynthetic routes for phosphatidylethanolamine and phosphatidylglycerol as identified in both Ca. P. ubique and Pelagibacter sp. str. HTCC7211. Green lines: Predicted reactions involved in lipid remodeling in Pelagibacter sp. str. HTCC7211. See text for detailed rationale for predicted functions.

169

gluconeogenesis HO OPO3H2 O glycerone-P

HO OPO3H2

OH Ca. P. ubique and glycerol-P O Pelagibacter sp. str. HTCC7211 R S-CoA

O Pi Pelagibacter sp. str. HTCC7211 only LCB5

R O OPO3H2

R O

O phosphatidate CTP

serine CDP-diacyl-glycerol glycerol-P

O O O

R O OPO2HO OH

R O OPO2HO PO4H2 R O NH2 R O OH

O O

CO2 Pi O O

NH2 R O OPO2HO OH R O OPO2HO

R O OH R O

O O phosphatidylethanolamine phosphatidylglycerol plcP

NH2 PO4 OH PO4 OH

R O OH

R O O HOH O H O HO HO H HO NH2 H OH H OH NH2 NDP-hexose ornithine rfaG-like hemolysins glycosyltransferase COG3176

O OH H HO NH2 O O H HO NH H HO R O OH H R O O H R O

R O O

Figure A3- 5: Simplified illustration of phospholipid biosynthetic routes and predicted reactions involved in lipid remodeling in Pelagibacter sp. str. HTCC7211.

170

+P P -starved +Mpn i i

Figure A3- 6: Scanning transmission electron microscopy images of Pelagibacter sp. str. HTCC7211 cells grown under different P-regimes. +Pi: cells grown with excess Pi (10 µM). Pi-starved: cells grown with excess Pi (10 µM), washed and resuspended in Pi- free medium for 96 hours. +Mpn: cells grown with excess Mpn (10 µM) and no Pi. Cells for the +P and +Mpn treatments were harvested for imaging in mid-logarithmic growth phase. Cells were stained with osmium and lead as described in methods. Arrowheads in “+Mpn” indicate putative polyphosphate granules.

171

Table A3- 1: Specific growth rates1 of Ca. P. ubique & Pelagibacter sp. str. HTCC7211 in AMS1 with different amounts of Pi or Mpn specific growth rate (day-1) amount added (nM) HTCC1062 HTCC7211+Pi HTCC7211+Mpn 0 0.24 ± 0.03 0.58 ± 0.01 0.41 ± 0.01 1 0.29 ± 0.05 0.59 ± 0.03 0.42 ± 0.01 3.3 0.31 ± 0.05 0.59 ± 0.02 0.45 ± 0.03 10 0.35 ± 0.04 0.58 ± 0.01 0.43 ± 0.02 33 0.41 ± 0.02 0.6 ± 0.01 0.45 ± 0.00 100 0.41 ± 0.02 0.57 ± 0.02 0.44 ± 0.02 1000 0.39 ± 0.02 0.60 ± 0.01 0.39 ± 0.07 3000 0.39 ± 0.02 0.58 ± 0.00 0.45 ± 0.02 Average (n=24) 0.35 ± 0.07 0.59 ± 0.02 0.43 ± 0.03 1presented as mean of biological replicates ± 1.0 s.d. (n=3)

172

Table A3- 2: Genes differentially regulated in P-limited Ca. P. ubique 4 hours after resuspension.

Fold q- ORF name Gene Annotation Change value SAR11_1176 pstB transport system ATP-binding protein 2.37 ≤0.005

SAR11_1177 pstA transport system permease protein 2.92 ≤0.005 SAR11_1178 pstC phosphate transport system permease 3.74 ≤0.005 protein SAR11_1179 pstS transport system substrate-binding 3.58 ≤0.005 protein

173

Table A3- 3: Genes differentially regulated in P-limited Ca. P. ubique 20 hours after resuspension.

Fold q- ORF name Gene Annotation Change value SAR11_0032 ftsZ cell division protein FtsZ -2.16 ≤0.005 SAR11_0033 mraZ cell division protein MraZ -2.05 ≤0.005 SAR11_0059 pilB type IV pilus assembly protein PilB -2.00 ≤0.005 SAR11_0060 pilQ type IV pilus assembly protein PilQ -2.01 ≤0.005 SAR11_0181 ibpA heat shock protein A 2.70 ≤0.005 SAR11_0254 trmD tRNA (guanine-N(1)-)- -2.15 ≤0.005 methyltransferase SAR11_0635 umuD DNA polymerase V subunit 2.41 ≤0.005 SAR11_0641 recA recombination protein RecA 3.43 ≤0.005 SAR11_0747 glnA glutamine synthetase I 2.29 ≤0.005 SAR11_0921 lexA LexA repressor 4.48 ≤0.005 SAR11_0953 yhdW general L-amino acid transport system -2.01 ≤0.005 substrate-binding protein SAR11_0964 transcriptional regulator, Fur family -2.80 ≤0.005 SAR11_0965 unknown protein -4.17 ≤0.005 SAR11_1093 rpoA DNA-directed RNA polymerase -2.35 ≤0.005 subunit alpha SAR11_1100 rpsE ribosomal protein S5 -2.05 ≤0.005 SAR11_1102 rplF ribosomal protein L6 -2.31 ≤0.005 SAR11_1103 rpsH 30S ribosomal protein S8 -2.16 ≤0.005 SAR11_1104 rpsN 30S ribosomal protein S14 -2.42 ≤0.005 SAR11_1107 rplN 50S ribosomal protein L14 -2.50 ≤0.005 SAR11_1110 rplP 50S ribosomal protein L16 -2.32 ≤0.005 SAR11_1114 rplB 50S ribosomal protein L2 RplB -3.10 ≤0.005 SAR11_1119 fusA translation elongation factor EF-G -2.27 ≤0.005 SAR11_1122 rpoC DNA-directed RNA polymerase beta -2.74 ≤0.005 prime chain SAR11_1164 hypothetical protein 2.25 ≤0.005 SAR11_1174 phoB regulatory protein 3.30 ≤0.005 SAR11_1175 phoU transport regulon regulator PhoU 3.84 ≤0.005 SAR11_1176 pstB transport system ATP-binding protein 9.13 ≤0.005

SAR11_1177 pstA transport system permease protein 10.58 ≤0.005 SAR11_1178 pstC phosphate transport system permease 10.68 ≤0.005 protein

174

Table A3-3 (continued)

SAR11_1179 pstS transport system substrate-binding 40.50 ≤0.005 protein SAR11_1181 hypothetical protein 2.84 ≤0.005

175

Table A3- 4: Genes differentially regulated in P-limited Ca. P. ubique 38 hours after resuspension.

Fold q- ORF name Gene Annotation Change value SAR11_0007 hflC probable integral membrane proteinase 2.87 ≤0.005 SAR11_0008 hflK probable integral membrane proteinase 2.84 ≤0.005 SAR11_0033 mraZ cell division protein MraZ -2.29 ≤0.005 SAR11_0077 trxB thioredoxin reductase 2.69 ≤0.005 SAR11_0162 groEL chaperonin GroEL 6.33 ≤0.005 SAR11_0181 ibpA heat shock protein A 6.95 ≤0.005 SAR11_0203 response regulator 4.64 ≤0.005 SAR11_0254 trmD tRNA (guanine-N(1)-)- -2.27 ≤0.005 methyltransferase SAR11_0255 rpsP ribosomal protein S16 -2.13 ≤0.005 SAR11_0266 TRAP dicarboxylate transporter - DctP 2.29 ≤0.005 subunit SAR11_0333 hslV ATP-dependent HslUV protease, 3.41 ≤0.005 peptidase subunit SAR11_0334 hslU ATP-dependent HslUV protease ATP- 2.44 ≤0.005 binding subunit SAR11_0367 molecular chaperone DnaJ 3.40 ≤0.005 SAR11_0368 dnaK molecular chaperone DnaK 3.39 ≤0.005 SAR11_0399 rbr Rubrerythrin 2.15 0.0050 SAR11_0433 gltD glutamate synthase (NADPH) beta 2.68 ≤0.005 chain SAR11_0434 gltB glutamate synthase large subunit 2.65 ≤0.005 SAR11_0520 hypothetical protein -2.89 ≤0.005 SAR11_0595 tolR TolR transport protein 2.29 ≤0.005 SAR11_0596 tolA TolA protein 2.20 ≤0.005 SAR11_0597 tolB TolB protein 2.05 ≤0.005 SAR11_0601 ftsH metalloprotease FtsH 2.77 ≤0.005 SAR11_0641 recA recombination protein RecA 2.37 0.0270 SAR11_0747 glnA glutamine synthetase I 2.51 ≤0.005 SAR11_0765 hypothetical protein 3.26 ≤0.005 SAR11_0788 Possible transmembrane protein 3.26 ≤0.005 SAR11_0818 amtB ammonium transporter 2.20 ≤0.005 SAR11_0841 yeiH putative integral membrane protein -2.11 0.0110 SAR11_0861 short chain dehydrogenase, unknown -2.00 ≤0.005 specificity SAR11_0907 rpsB ribosomal protein S2 -2.20 ≤0.005

176

Table A3-4 (continued)

SAR11_0921 lexA LexA repressor 2.92 ≤0.005 SAR11_0964 transcriptional regulator, Fur family -2.23 ≤0.005 SAR11_1019 xerD integrase/recombinase XerD-like -3.30 ≤0.005 protein SAR11_1030 metY O-acetylhomoserine (thiol)-lyase -2.04 0.0280 SAR11_1093 rpoA DNA-directed RNA polymerase -2.96 ≤0.005 subunit alpha SAR11_1094 rpsK ribosomal protein S11 -2.36 ≤0.005 SAR11_1113 rpsS ribosomal protein S19 -2.12 ≤0.005 SAR11_1114 rplB ribosomal protein L2 -2.22 0.0080 SAR11_1117 rplC ribosomal protein L3 -2.05 0.0060 SAR11_1119 fusA elongation factor EF-G -2.26 0.0080 SAR11_1151 glmS glutamine-fructose-6-phosphate -2.09 0.0060 transaminase (isomerizing) SAR11_1163 unknown 2.63 ≤0.005 SAR11_1164 hypothetical protein 2.17 ≤0.005 SAR11_1174 phoB regulatory protein 3.94 ≤0.005 SAR11_1175 phoU transport regulon regulator PhoU 5.00 ≤0.005 SAR11_1176 pstB transport system ATP-binding protein 9.17 ≤0.005

SAR11_1177 pstA transport system permease protein 11.52 ≤0.005 SAR11_1178 pstC phosphate transport system permease 16.37 ≤0.005 protein SAR11_1179 pstS transport system substrate-binding 27.92 ≤0.005 protein SAR11_1181 hypothetical protein 2.24 ≤0.005 SAR11_1240 aceA isocitrate lyase -2.00 ≤0.005 SAR11_1346 livJ branched-chain amino acid transport -2.08 0.0070 substrate-binding protein

177

Table A3- 5: Genes differentially regulated in P-limited Pelagibacter sp. str. HTCC7211 68 hours after re-suspension. ORF name Gene Annotation Fold q- Change value PB7211_757 GYD domain superfamily 2.02 ≤0.005 PB7211_1178 conserved hypothetical protein 2.23 ≤0.005 PB7211_503 phnM alkylphosphonate utilization protein 7.26 ≤0.005 PhnM PB7211_450 phnL phosphonate C-P lyase system protein 2.05 ≤0.005 PhnL N/A phnK alkylphosphonate utilization protein 2.21 ≤0.005 PhnK PB7211_943 phnJ phosphonate metabolism PhnJ 3.21 ≤0.005 PB7211_192 phnI phosphonate metabolism PhnI 4.39 ≤0.005 PB7211_797 phnH alkylphosphonate utilization protein 3.78 ≤0.005 PhnH PB7211_1322 phnG alkylphosphonate utilization protein 2.92 ≤0.005 PhnG PB7211_620 phnC phosphonate ABC transporter, ATP- 21.76 ≤0.005 binding protein PB7211_926 phnD phosphonate ABC transporter, 17.79 ≤0.005 periplasmic binding protein PB7211_232 phnE probable phosphonate ABC transporter, 9.60 ≤0.005 permease protein

PB7211_725 phnE2 phosphonate ABC transporter, 7.29 ≤0.005 permease protein PB7211_828 hypothetical protein 12.01 ≤0.005 PB7211_545 probable HD-hydrolase 2.74 ≤0.005 PB7211_679 phoU transport regulon regulator PhoU 2.14 ≤0.005 PB7211_412 pstB ABC transporter 3.22 ≤0.005 PB7211_586 pstA ABC transporter; transport system 3.46 ≤0.005 permease protein PB7211_733 pstC ABC transporter; transport system 3.47 ≤0.005 permease protein PB7211_1190 pstS transport system substrate-binding 3.74 ≤0.005 protein PB7211_635 Hemolysin (COG3176) 5.17 ≤0.005 PB7211_1302 Hemolysin (COG3176) 2.92 ≤0.005 PB7211_960 probable glycosyl transferase 2.74 ≤0.005 PB7211_983 calcineurin-like metallophosphoesterase 2.68 ≤0.005

178

Table A3- 6: Genes differentially regulated in P-limited Pelagibacter sp. str. HTCC7211 96 hours after re-suspension. ORF name Gene Annotation Fold q- Change value PB7211_613 gloB cytoplasmic hydroxyacylglutathione 2.21 ≤0.005 hydrolase PB7211_351 putative alginate lyase 3.13 ≤0.005 PB7211_503 phnM alkylphosphonate utilization protein 9.87 ≤0.005 PhnM PB7211_764 conserved hypothetical protein 2.13 ≤0.005 N/A phnN phosphonate metabolism protein phnN 2.46 ≤0.005 PB7211_450 phnL phnL; phosphonate C-P lyase system 3.29 ≤0.005 protein PhnL N/A phnK phnK; alkylphosphonate utilization 3.68 ≤0.005 protein PhnK PB7211_943 phnJ phosphonate metabolism PhnJ 5.23 ≤0.005 PB7211_192 phnI phosphonate metabolism PhnI 6.02 ≤0.005 PB7211_797 phnH alkylphosphonate utilization protein 4.77 ≤0.005 PhnH PB7211_1322 phnG alkylphosphonate utilization protein 4.13 ≤0.005 PhnG PB7211_620 phnC phosphonate ABC transporter, ATP- 30.55 ≤0.005 binding protein PB7211_926 phnD phosphonate ABC transporter, 21.11 ≤0.005 periplasmic binding protein PB7211_232 phnE probable phosphonate ABC transporter, 17.75 ≤0.005 permease protein

PB7211_725 phnE2 phosphonate ABC transporter, 11.93 ≤0.005 permease protein PB7211_828 hypothetical protein 15.89 ≤0.005 PB7211_545 probable HD-hydrolase 2.71 ≤0.005 PB7211_619 arsC probable arsenate reductase 2.30 ≤0.005 PB7211_98 glpF putative transport transmembrane 2.07 ≤0.005 protein PB7211_412 pstB ABC transporter 2.86 ≤0.005 PB7211_586 pstA ABC transporter; transport system 3.12 ≤0.005 permease protein PB7211_733 pstC ABC transporter; transport system 3.11 ≤0.005 permease protein PB7211_1190 pstS transport system substrate-binding 3.66 ≤0.005 protein

179

Table A3-6 (continued)

PB7211_635 Hemolysin (COG3176) 8.17 ≤0.005 PB7211_1302 Hemolysin (COG3176) 2.75 ≤0.005 PB7211_960 probable glycosyl transferase 4.07 ≤0.005 PB7211_983 calcineurin-like metallophosphoesterase 3.61 ≤0.005 PB7211_1423 ycfV ABC transporter ATP-binding protein 2.11 ≤0.005 PB7211_870 homoserine O-acetyltransferase 2.20 ≤0.005 PB7211_1410 conserved hypothetical protein 2.17 ≤0.005 PB7211_497 unknown major facilitator family 2.03 ≤0.005 permease