SAR11 Cells Rely on Enzyme Multifunctionality to Transport and Metabolize a Range of Polyamine Compounds

In the ocean surface layer and cell culture, the polyamine transport protein PotD of SAR11 bacteria is often one of the most abundant proteins detected. Polyamines are organic cations produced by all living organisms and are thought to be an important component of dissolved organic matter (DOM) produced in planktonic ecosystems. We hypothesized that SAR11 cells transport and metabolize multiple polyamines and use them as sources of carbon and nitrogen. Metabolic footprinting and fingerprinting were used to measure the uptake of five polyamine compounds (putrescine, cadaverine, agmatine, norspermidine, and spermidine) in two SAR11 strains that represent the majority of SAR11 cells in the surface ocean environment, Ca. Pelagibacter st. HTCC7211 and C. P. ubique st. HTCC1062. Both strains transported all five polyamines and concentrated them to micromolar or millimolar intracellular concentrations. Both strains could use most of the polyamines to meet their nitrogen requirements, but we did not find evidence of use as carbon sources. We propose potABCD transports cadaverine, agmatine, and norspermidine, in addition to its usual substrates of spermidine and putrescine, and that spermidine synthase, speE, is reversible, catalyzing the breakdown of spermidine and norspermidine, in addition to its usual biosynthetic role. These findings provide support for the hypothesis that enzyme multifunctionality enables streamlined cells in planktonic ecosystems to increase the range of DOM compounds they oxidize. Importance Genome streamlining in SAR11 bacterioplankton has resulted in a small repertoire of genes, yet paradoxically they consume a substantial fraction of primary production in the oceans. Enzyme multifunctionality is hypothesized to be an adaptation that increases the range of organic compounds oxidized by cells in environments where selection favors genome minimization. We provide experimental support for this hypothesis by demonstrating that SAR11 cells use multiple polyamine compounds and propose that a small set of multifunctional genes catalyze this metabolism. We also report polyamine uptake rates can exceed metabolism, resulting in high intracellular concentrations of these nitrogen-rich compounds and an increase in cell size. Increases in cytoplasmic solute concentrations during transient episodes of high nutrient exposure has previously been observed in SAR11 cells and may be a feature of their strategy for maximizing the share of labile DOM acquired when in competition with other cell types.


Introduction 48
Polyamines are low molecular weight organic polycations that are ubiquitous in living 49 organisms. They play a role in stabilizing DNA, RNA, and proteins, are required for cell growth, 50 and have been implicated in biofilm formation (1)(2)(3). Polyamine compounds and concentrations 51 vary between cell types and can depend on nutrient status, temperature, and salinity (4). 52 Polyamines are found at low nanomolar concentrations in the coastal and open ocean, reaching 53 maximal concentrations of 30 nM during algal blooms, but typically are around 1 nM (5-7). 54 Polyamines from the environment are metabolized by bacteria as nitrogen and carbon sources at 55 rates similar to those of dissolved free amino acids and supply up to 14% of bacterial nitrogen 56 demand in coastal regions. (8-10). 57 Putrescine (PUT) and spermidine (SPD), the most abundant polyamines in the ocean, are 58 typically 3-5 nM in the environment, but spermine, cadaverine (CAD), and norspermidine 59 (NSD) have been detected at lower levels (6, 10, 11). Several other polyamines, such as 1,3-60 diaminopropane (DAP), agmatine (AGM), homospermidine (HSD), spermine, and larger, more 61 complex polyamines are known to be produced and/or metabolized by cells from all domains of 62 life (3, 4, 9, 12, 13). Metabolic pathways for common polyamines are shown in Figure 1. 63 SAR11 alphaproteobacteria make up the majority of bacteria in the ocean (14). SAR11 64 cells primarily utilize labile, low molecular weight molecules (15). They pack their relatively 65 large periplasmic space (16) with large numbers of ABC transporter substrate-binding proteins 66 (SBPs) (17, 18), increasing the encounter rate and binding of substrate molecules with SBPs, 67 resulting in high whole-cell uptake affinities (19,20). Recent modeling work has extended this 68 observation, suggesting this strategy may contribute to the slow growth rates of SAR11 cells 69 (21). SAR11 bacteria evolved minimal genomes in response to streamlining selection, which 70 favors efficient use of resources in nutrient-limited ecosystems (22). Enzyme multifunctionality 71 has been hypothesized to reduce gene content in streamlined cells and has been confirmed for the 72 SAR11 glycine betaine transporter (20). 73 SAR11 cells produce large numbers of PotD, the SBP involved in polyamine transport, 74 both in cultures and the environment, making it the most highly expressed transporter for N-75 related compounds by SAR11 cells (17,18,23). N-limited cultures of SAR11 strain HTCC1062, 76 a member of the cold, high-latitude Group Ia.1 ecotype, did not upregulate genes for polyamine 77 transport or metabolism, except for an enzyme implicated in PUT and CAD metabolism (24), but 78 genes involved in the metabolism and transport of other organic N sources were upregulated 79 (24). Incubation experiments with natural seawater communities provided evidence that SAR11 80 cells may sometimes respond to additions of polyamines PUT and SPD -transcripts for SAR11 81 genes involved in polyamine metabolism increased in the first hour of incubation and accounted 82 for over a quarter of all transcripts (25). In other experiments with PUT and SPD amendments to 83 seawater, it was observed that SAR11 cell abundance did not change during a 48-hour period in 84 response to PUT and SPD addition (26, 27); oligotrophs frequently decrease in relative 85 abundance in incubation experiments due to their slow growth rates, while copiotrophs increase 86 rapidly due to their fast growth rates under high nutrient conditions used in incubation 87 experiments (15). 88 In this study we used targeted metabolic footprinting and fingerprinting to examine the 89 types and amounts of polyamines taken up and metabolized by two SAR11 strains. Both strains 90 of SAR11 used in this study come from the Ia subgroup: Ca. Pelagibacter ubique HTCC1062 91 belongs to the cold, high-latitude Group Ia.1 ecotype and Ca. P. st. HTCC7211 is from the 92 equatorial, warm water Ia.3 ecotype (15). Targeted metabolic footprinting uses LC-MS/MS to 93 measure changes in the concentrations of specific metabolites dissolved in spent culture media 94 (28), while fingerprinting quantifies the concentrations of targeted metabolites within cells (29). 95 We hypothesized that SAR11 cells would use polyamines as N sources and that polyamines 96 might supply these cells with metabolic carbon via a branch of metabolism that proceeds through 97 pyruvate (25,30). 98

Results 99
Footprinting and Fingerprinting Experiments 100 We focused on five polyamine compounds (Table 1): putrescine (PUT), cadaverine 101 (CAD), agmatine (AGM), norspermidine (NSD), and spermidine (SPD). These compounds were 102 picked either for their prevalence in the environment and in bacteria cells or for their role as 103 precursors to other polyamine compounds (4, 6, 31, 32). These compounds also showed the best 104 recovery in solid-phase extraction and were amenable to simultaneous quantification by LC-105 MS/MS. We used polyamine concentrations 10-100X ambient environmental concentrations, 106 similar to what would be found in nutrient patches (33), as has been done previously (20,25). In 107 preliminary experiments, we found that high polyamine concentrations inhibited growth; 108 HTCC1062 growth was inhibited when all polyamines were added together at individual 109 concentrations above 500 nM, while HTCC7211 growth was inhibited at concentrations above 110 250 nM. We chose 500 nM for HTCC1062 and 250 nM for HTCC7211, the highest 111 concentrations that did not significantly inhibit growth, for further experiments ( Figure S1(A-112 B)). 113 The five polyamines were added to SAR11 cultures under nutrient replete conditions to 114 measure uptake and oxidation of these compounds. Cultures were grown to late exponential 115 phase before harvesting; growth rates were slightly lower with polyamines added: 0.44 compared 116 to 0.46 d -1 for HTCC1062 and 0.50 compared to 0.60 d -1 for HTCC7211 with and without 117 polyamines ( Figure S2(C-D)). For both strains, average intracellular levels of all five polyamine 118 compounds were significantly greater in the experimental treatment (polyamines added) 119 compared to the negative control (no polyamines added), except for SPD in HTCC1062, which 120 had non-significant higher levels in experimental cultures (Figure 2A, C; p-values in Table 2). 121 When the intracellular levels are converted to intracellular concentrations using a cell volume of 122 0.03 µm 3 for HTCC1062 (16) and 0.04 µm 3 for HTCC7211 (34), it is apparent that the cells are 123 concentrating all compounds into intracellular concentrations greater than their environment 124 (Table 2). The intracellular concentrations in the experimental treatment for HTCC7211 were 125 much higher than in HTCC1062, especially SPD, which was 40X higher. 126 In the extracellular fractions of both strains, there were no significant differences between 127 the experimental treatment and the no cell control (polyamines added, no cells), except for AGM 128 in HTCC1062 (p-value of 0.04, one-sided t-test) ( Figure 2B, D). In HTCC1062, all five 129 compounds were lower in concentration in the experimental treatment compared to the no cell 130 control ( Figure 2B). For HTCC7211, all five compounds were at similar concentrations between 131 the two treatments, except for AGM, which was lower in the experimental treatment (p-value of 132 0.06, one-sided t-test). 133

Flow Cytometry Experiment 134
Based on the very high intracellular polyamine concentrations measured in HTCC7211 135 cells in experimental treatments, we postulated that cells would change in size to accommodate 136 the influx of polyamines. To test this, we used flow cytometry to monitor the forward scatter 137 (FSC), a proxy for cell size, of nutrient-replete HTCC7211 cultures exposed to either no 138 polyamines (control) or 250 nM of each polyamine added at the beginning of growth (early 139 addition) or after 4 days of growth (late addition). On average, both experimental treatments had 140 higher FSC than the control ( Figure 3B). Mean FSC of the early addition cultures was 141 consistently higher than the control across all measured time points ( Figure 3B). Mean FSC of 142 late addition cultures was similar to early addition cultures 4 hours after addition of polyamines 143 to late addition cultures; it then decreased to control levels at day 7, finally increasing above the 144 control after day 11 ( Figure 3A). 145

Carbon Substitution Experiments 146
Growth experiments were used to examine whether the five polyamine compounds could 147 substitute for two unusual growth requirements of SAR11 cells: pyruvate, or related compounds 148 which lead to a branch of SAR11 metabolism that includes the biosynthesis of alanine, and 149 glycine or related compounds, required for another branch of SAR11 metabolism that includes 150 glycine synthesis. The five polyamine compounds were added together at final concentrations of 151 250 nM each as a replacement for either pyruvate or glycine, and the growth of the cultures 152 compared to negative control treatments with either no pyruvate or no glycine added. With both 153 strains, experimental treatments with added polyamines achieved higher maximum cell densities 154 and more rapid growth rates than negative controls, but the differences were not significant, 155 indicating that these compounds did not substitute for glycine or pyruvate ( 156 Table 3). 157 Only addition of SPD resulted in a higher cell density than the negative control, but the 174 difference was not significant. For HTCC7211, addition of SPD resulted in a significantly higher 175 maximum cell density than all three controls. Cultures to which CAD was added had 176 significantly higher maximum cell densities than the equimolar control and the negative control. 177 Cultures with PUT added had a significantly higher maximum cell density than only the 178 equimolar control. In both strains, the addition of AGM resulted in lower maximum cell densities 179 than any control treatments; in HTCC7211, the treatments with NSD and with all polyamines 180 combined also had lower maximum cell densities than any controls. Interestingly, diauxic growth 181 was observed in HTCC1062, with an early peak around 14 d and a larger peak later around 40 d 182 (Figure 4), which was not observed in HTCC7211. 183

Nitrogen Substitution Growth Experiments
Metabolic Pathways 184 Figure 1 shows genes for polyamine metabolism for the two SAR11 strains used in this 185 study, overlayed on common pathways for polyamine metabolism (8, 12, 30, 35-37). In both 186 strains, AGM is postulated to be converted by agmatinase (speB) to PUT, which is catabolized 187 by the transamination pathway, since neither strain encodes the final enzyme in the γ-188 glutamylation pathway, and the transamination pathway was upregulated in SAR11 cells in 189 response to PUT addition (25). CAD is likely metabolized to succinate via the lysine degradation 190 pathway. Many genes involved in polyamine metabolism are known to be multifunctional in 191 other cell types, as we also predict in SAR11 (Figure 1, Figure S2). 192 In metabolic reconstruction from genome sequences (38), we found that neither SAR11 193 strain encoded a clear pathway of SPD or NSD metabolism, although both compounds were 194 taken up from the medium and metabolized (Figure 2, Figure 4). Microbial enzymes responsible 195 for NSD metabolism have not been extensively characterized. Several Vibrio strains produce and 196 metabolize NSD via carboxynorspermidine (39), but the necessary enzymes are lacking in 197 SAR11. Neither SAR11 strain has homologs for the canonical genes responsible for SPD 198 metabolism: SPD dehydrogenase (spdH), which cleaves SPD to produce 1,3-diaminopropane 199 (DAP) and 4-aminobutanal, and SPD acetyltransferase, which converts SPD to acetylspermidine, 200 a less toxic version of SPD. We speculate that either the SPD synthase enzyme, SpeE, is bi-201 directional, producing PUT from SPD, or there is another, unknown enzyme capable of cleaving 202 SPD. SpeE is not known to be bi-directional in other bacteria (40). A possible candidate enzyme 203 for SPD metabolism was discovered during metabolic reconstruction: dys2, a putative 204 deoxyhypusine synthase (dhs) gene ( Figure S). Dys2 is highly conserved in these and other 205 SAR11 genomes between two genes involved in PUT formation, speC and speB. In prokaryotes, 206 dhs usually acts as a homospermidine (HSD) synthase, a promiscuous enzyme capable of acting 207 on multiple polyamines in addition to its native function of producing HSD from PUT, instead of 208 the function it performs in eukaryotes and archaea of post-translationally modifying elongation 209 factor 5 (EF5) while cleaving SPD (41, 42). 210 To differentiate between these two alternative routes of SPD metabolism and identify 211 pathways of NSD metabolism in SAR11, we used fingerprinting to search for possible by-212 products of SPD and NSD metabolism, including DAP and HSD. If SAR11 cells use the reverse 213 SPD synthase reaction to metabolize SPD, the products would be PUT and S-214 adenosylmethionine, while the by-products of the Dys2 enzyme, if it is a HSD synthase, would 215 primarily be HSD and DAP. We compared polyamine levels in SAR11 cells grown either 216 without any polyamines or with either 500 nM SPD or NSD ( Figure 5). As expected, in both 217 strains, the treatments with SPD or NSD added had higher levels of that compound in the 218 respective treatment (one-sided t-test, p-values of 0.02 and 0.003 for +SPD and +NSD vs. the 219 negative control, respectively; p-values were 0.06 and 0.2 for HTCC1062), indicating an uptake 220 of those two compounds by both strains. 221 In HTCC7211, the +SPD cultures also had significantly higher levels of PUT than the 222 negative control (one-sided t-test, p-value of 0.05) but only a slight, non-significant increase in 223 DAP ( Figure 5B). In the +NSD cultures, there were significantly higher levels of DAP than the 224 negative control (one-sided t-test, p-value of 0.001). There was also a significant increase in SPD 225 and HSD (one-sided t-test, p-values of 0.05 and 0.04) in +NSD cultures compared to the negative 226 control. In HTCC1062, there were no differences in the levels of any other compounds in the 227 +NSD or +SPD treatments compared to the negative control, aside from SPD and NSD 228 themselves and a non-significant increase in DAP in the +NSD treatment ( Figure 5A). The growth of SAR11 cultures was found to be inhibited by high polyamine 242 concentrations ( Figures S1, 3). Polyamines are known to be toxic to bacteria when added to 243 media at high concentrations (generally mM range), but the mechanism is not known (37, 43, 244 44). SAR11 cells often lack transcriptional regulation for carbon oxidation functions (15) and 245 previous work indicated they do not upregulate metabolic enzymes for polyamines when N 246 limited (24). The growth inhibition observed at high polyamine concentrations might be due to 247 adverse effects of the buildup of polyamine compounds inside the cells. Similar results 248 previously have been observed in cells experiencing metabolic pathway saturation (45). 249 Intracellular concentrations of polyamines were higher in HTCC7211 than in HTCC1062 and 250 HTCC7211 was also more susceptible to growth inhibition by polyamines, indicating these 251 might be linked ( Figure S1(A-B), Table 2). 252 there were few observable depletions of extracellular concentrations for any polyamines with 291 either strain because the intracellular polyamine pools were small relative to the surrounding 292 volume ( Figure 2C, D). The observed accumulation of intracellular polyamines was estimated to 293 result in pmolar drawdowns of the dissolved polyamine pool, which in most cases was less than 294 the precision of our measurements (Table S3). An exception was the accumulation of 239 295 pmoles of intracellular SPD in HTCC7211, which should have produced a measurable depletion 296 of SPD in the medium, yet no significant reduction was observed (standard deviation of 83.3 297 pmoles). This observation suggests that HTCC7211 cells used other transported polyamines to 298 synthesize SPD. To support this interpretation, HTCC7211 cells given only SPD had 10X lower 299 SPD levels than when given all five compounds (compare Figure 2C and Figure 5B). It appears 300 that transported polyamines are converted intracellularly to SPD, which then accumulates in 301 HTCC7211. 302

Uptake of Polyamine Compounds by SAR11 Cells
Our analysis indicates the PotABCD transport system in SAR11 is responsible for 303 transporting the five polyamines we tested, given the structural similarity between these 304 compounds and the absence of other candidates for polyamine transport functions. Both strains 305 of SAR11 lack homologs for CAD, NSD, or AGM transporters found in other bacteria. In Vibrio 306 cholerae, NSD is transported by a potABCD homolog (51), but CAD and AGM have not 307 previously been identified as substrates for potABCD. In E. coli, PotABCD is primarily a 308 PUT/SPD/SPM transporter (52), which could help explain the higher accumulation of these 309 compounds in SAR11 cells, even when accounting for native PUT and SPD production. This is 310 the second multifunctional transporter identified experimentally in SAR11 cells, in addition to 311 the ABC transporter for glycine betaine, which was shown to transport seven different substrates 312 (20). It is likely that other ABC transporters in SAR11 cells are also multifunctional, given the 313 use of a wide variety of amino acids and carboxylic acids by SAR11 cells (53-55). 314 Interestingly, there was a large difference between the two SAR11 strains in the amount 315 of polyamines taken up. Polyamine concentrations in HTCC7211 were 40 -500X higher than 316 HTCC1062, despite HTCC7211 being exposed to 2X lower concentration of polyamines. This 317 difference cannot be explained entirely by cell size; HTCC7211 cells are only ~1.3X larger, as 318 measured by C content, than HTCC1062 (34). The differential could be because the HTCC7211 319 transport system has a higher V max for polyamine transport than HTCC1062, due either to the 320 properties of the proteins themselves (one of the two permease proteins involved in polyamine 321 transport, PotB, is only 82% identical), differing abundance of transport proteins, or the 322 cytoarchitecture of the cells. There were no major differences between the two strains in 323 presence/absence of polyamine metabolic genes, nor in the location of those genes ( Figure S2). 324 One possible ecological explanation for the difference between these two strains, if they are 325 typical of the ecotypes they represent, is that HTCC1062, a member of the primarily coastal 326 subclade of SAR11, may have been influenced by selection that limits toxic buildups of 327 polyamines at the higher polyamine concentrations found in coastal regions. HTCC7211, a 328 member of a primarily open ocean subclade of SAR11, would rarely experience the high 329 polyamine concentrations found in coastal regions and so might not experience selection to limit 330 intracellular buildups. 331 Use of Polyamines by SAR11 Cells 332 SAR11 cells have unique growth requirements, needing a reduced sulfur source (e.g. 333 methionine or methane thiol), a glycine source, specific vitamins, and a carbon source that can 334 serve as precursor to alanine (usually pyruvate) (54). Most of the tested polyamine compounds 335 are predicted to be metabolized to succinate, a TCA cycle intermediate ( Figure 1). In previous 336 work, succinate was shown to not substitute for pyruvate in HTCC1062, in accord with our 337 experimental findings (56). This does not rule out the use of polyamines as an energy source, 338 however. Other small compounds have been found previously to be used by SAR11 cells only as 339 an energy source and not as a pyruvate substitute (20, 55). Polyamines are also required for a 340 variety of other cellular processes, and it is likely that SAR11 cells used the supplied polyamines 341 in those processes in addition to metabolizing them. 342 In both strains of SAR11 tested, several polyamine compounds (SPD, CAD, and PUT) 343 were able to be used as a N source, with SPD supporting the highest maximum cell density of 344 any of the polyamines ( Figure 4A). NSD does not appear to be a N substitute for either strain of 345 SAR11 at the NSD concentration tested, but it is transported ( Figure 5). The use of multiple 346 polyamines as a N source is consistent with previous reports that SAR11 cells use a variety of 347 organic N-containing compounds as N sources (24). 348 Interestingly, several compounds (AGM and NSD) were inhibitory to SAR11 growth 349 under N limiting conditions (Figure 4). One potential cause for the AGM inhibition is the by-350 product of AGM degradation by the agmatinase enzyme, urea. Neither SAR11 strain encodes a 351 urease (57). We speculate that the influx of AGM causes a build-up of inhibitory urea in cells, a 352 process that is known to occur in oligotrophs due to metabolic pathway saturation (15,45 Diauxic growth is generally observed when cells switch from using one source of nutrients to 358 another. However, the cultures used to start these experiments were acclimated to the same 359 medium (without N) prior to the experiment starting. Previously, diauxic growth was observed in 360 HTCC7211 grown on alternate P sources, which was attributed to the switch from using 361 inorganic P to organic P sources (58). Another unexpected observation, found in both strains and 362 across several repetitions, was that the equimolar positive control always had a lower maximum 363 cell density than the negative control with no N added. 364 Metabolic pathways for Spermidine and Norspermidine 365 SPD metabolism has been observed in marine bacteria without a spermidine 366 dehydrogenase gene (spdH) (9, 11, 25). It has been speculated that the enzyme that synthesizes 367 SPD from PUT, SpeE, is bi-directional, although this activity was not confirmed experimentally 368 (9). In HTCC7211, it appears that SPD is primarily metabolized via the reverse SPD synthase 369 reaction, not via the Dys2 enzyme, as no significant increase in HSD was detected, while an 370 increase in PUT was observed ( Figure 5B). The SpeE enzyme in SAR11 previously has been 371 found to be the result of a gene fusion event and is highly multifunctional, displaying high 372 biosynthetic activity with multiple polyamines substrates, in addition to this enzyme's commonly 373 predicted substrate, PUT (59). Our data suggest that this enzyme is not only multifunctional in 374 substrate range, but also in its ability to carry out catalytic reactions in reverse of its usual 375 biosynthetic activities. The results from HTCC1062 on SPD metabolism are not as clear, as no 376 other differences between the +SPD treatment and the negative control were observed aside from 377 an increase in SPD ( Figure 5A). 378 We propose that NSD is being metabolized in HTCC7211 by the enzyme SpeE, similar 379 to SPD, since we observed increased levels of DAP in the NSD treatment and the SpeE enzyme 380 in SAR11 is known to have a wide substrate rage ( Figure 5B). With these data, we cannot rule 381 out the Dys2 enzyme metabolizing NSD. In HTCC7211, there was also an increase in SPD and 382 HSD in the +NSD treatment ( Figure 5B). The Dys2 enzyme in SAR11 cells may be responsible 383 for producing SPD and HSD from the excess NSD and DAP, since dhs homologs in bacteria are 384 known to produce SPD from PUT and DAP, in addition to producing HSD from PUT (42, 60, 385 61). It appears that the dys2 gene in both SAR11 strains is not primarily acting as a HSD 386 synthase, since there was relatively low production of HSD under any condition, in contrast to 387 other prokaryotes with a HSD synthase gene where HSD is the sole polyamine present (60, 61). 388 We evaluated the hypothesis that SPD synthase might be catalyzing reactions that are the 389 reverse of its ordinary action of synthesizing polyamines. We explored thermodynamic models 390 that predicted the energies of the compounds in the primary reaction catalyzed by SPD synthase 391 (forming SPD from PUT; Figure S3), without considering entropy terms (Text S1, Table S4). 392 The estimated ∆E value for the total reaction was positive (14.32 kcal/mol) when water was used 393 as the proton acceptor (it is expected that ΔH° values will be quite similar to ΔE values) (Table  394 S5). More favorable acceptors (e.g., imidazole) easily yield negative ∆E values of -23.49 395 kcal/mol (Table S5). Our findings suggest that the direction of this reaction is easily tunable by 396 including proton carriers of varying strengths. For this calculation, we used standard conditions 397 and did not consider the very high intracellular concentrations of polyamines we observed 398 experimentally, which might further drive this reaction to reverse its normal biosynthetic 399 function. 400

Conclusion 401
Some properties of plankton cells that are important to understanding and modeling their 402 behavior in natural ecosystems can only be measured by experimentation. Recently, we 403 demonstrated very low whole-cell affinities and multifunctionality in the osmolyte transport 404 system of cultured SAR11 cells. We attributed these competitively advantageous cell properties 405 to synergism between kinetic features of the glycine betaine transporter ProXYZ and unusual 406 aspects of SAR11 cell architecture, notably their small size and large periplasm packed with 407 abundant substrate binding proteins (20). Here, we investigate SAR11 metabolism of 408 polyamines, which are transported into cells by the highly abundant SAR11 transporter system 409 PotABCD. We find this system is also multifunctional, and that the two SAR11 strains 410 metabolized a variety of polyamines, which served the cells as N sources. We cannot rule out 411 polyamines, which we predict to be metabolized to succinate, being used by SAR11 as carbon 412 sources, but our findings do not show that polyamines are a major source of carbon for these 413 cells. In previous work, we have shown that several C1 compounds are used by SAR11 as energy 414 sources via tetrahydrofolate-mediated oxidation but not as a source of carbon for biomass 415 production (55), which may be the case with polyamines. Our data strongly support the 416 hypothesis that SAR11 use many polyamines via a simplified system of few enzymes and a 417 single transporter. They mainly use these compounds as an N source and perhaps to supplement 418 their intracellular polyamine pool, potentially important adaptations in N-limited marine systems. 419 Polyamine transport rates exceeded metabolic rates, leading to mM intracellular 420 polyamine accumulations and an increase in cell size over a period of hours. We propose that 421 SAR11 cells use the multifunctional enzyme spermidine synthase, SpeE, in reversible reactions 422 that can both produce SPD and catabolize SPD and NSD. The findings we report indicate 423 enzyme multifunctionality expands the range of DOM compounds these cells harvest, which 424 may partially explain how these cells attain high success in competition for DOM resources. Our 425 findings also support previous observations which indicated SAR11 cells concentrate some 426 metabolites during pulses of availability, metabolizing them subsequently (47). In principle this 427 cell behavior could increase the success of SAR11 cells in competition for nutrient patches, but 428 further experimental work and modeling are needed to evaluate this hypothesis. In any case, the 429 properties of cells that we uncovered here are neither typical nor trivial -they change our 430 understanding of how competition for DOM resources has led to the emergence of specialized 431 cell types and will likely inform future experimental research and modeling aimed at 432 understanding cell evolution and the ocean carbon cycle. minimal degradation was observed, with over 60% recovery (Table S1). Because no intracellular 499 polyamine standard reference exists, reported intracellular measurements in this paper were not 500 corrected for recovery efficiency. 501

Extracellular SPE Extraction 502
Polyamines dissolved in the culture media were extracted using a solid phase extraction 503 (SPE) as described previously (64). 10 mL of the supernatant from the initial centrifugation of 504 cultures (described above) was used for each sample. Polyamines were extracted via gravity 505 alone (nominal flow rate of 0.07 mL/min) onto a 1000 mg Bond Elut-C 18 SPE column (3 mL, 506 Agilent) pre-conditioned with methanol and bicarbonate buffer at pH 12. Salts were removed 507 from the column by washing three times with 1 mL of 0.1 M borate buffer, pH 12. Polyamines 508 were eluted into cryovials with three washes of equal volumes of 1 M acetic acid and 509 acetonitrile, final volume 5 mL. Greater than 85% recovery for CAD, PUT, NSD, and SPD was 510 observed, while AGM had 48% recovery (Table S1) software packages were used for instrument operation and quantification respectively. A Phenyl-520 3, 150 x 4.6 mm 5 µm HPLC column (GL Sciences) was used for chromatographic separations, 521 using a 2.0 µm pre-filter as a guard column (Optimize Technologies). The sample rack was 522 cooled to 10˚C to prevent degradation of polyamines. The column temperature was maintained at 523 40˚C. HPLC mobile phases were MS grade water (Fisher) with 0.1% formic acid and MS grade 524 acetonitrile (Fisher) with 0.1% formic acid. A 10-minute binary gradient with a flow rate of 0.8 525 mL/min was used. The initial concentration of 3% acetonitrile ramped to 30% acetonitrile in 5 526 minutes. The column then re-equilibrated at 3% acetonitrile for 5 minutes. The ESI Source used 527 a spray voltage of 5200 V and source temperature of 600 C. Sheath gas pressure was 50 PSI and 528 auxiliary gas pressure was 40 PSI. The mass spec was run in positive ion mode. Compound-529 specific MRM parameters, column retention times, and limits of detection are presented in Table  530 1. The instrumental limits of detection (LOD; Table 1)

Computational Modeling 562
The reaction catalyzed by spermidine synthase (Figure S3) was computationally modeled 563 to determine reaction energetics. Water and histidine were used separately as bases (B:). Amine 564 nitrogens in putrescine and spermidine were fully protonated, but the nitrogens in the adenosyl 565 fragment were left unprotonated (neutral). First, conformational spaces for compounds I and III 566 were explored using Spartan'14 (67), with the MMFF force field ( Figure S4). DFT studies were 567 performed using Gaussian16 (68). The B3LYP functional (69-72) was employed using the cc-568 pVDZ basis set (73). An SCRF solvation model using water (74) was applied. All structures 569 were optimized and showed only real vibrational frequencies. SCF energies with solvation 570 correction were used as the primary measure of molecular energy. Reaction energies were 571 estimated using the minimum energy conformer for each compound. Tables 788 Table 1 Table S2 Maximum cell densities for SAR11 cultures grown with polyamine compounds as the 886 sole nitrogen source with corresponding p-values. All p-values are from a one-sided t-test to 887 assess whether the treatment had a significantly higher maximum cell density than the 888 corresponding control. When an experimental treatment's maximum cell density was lower than 889 or equal to the corresponding control, a t-test was not conducted. 890 Table S3 Intracellular