Molecular Underpinnings and Biogeochemical Consequences of Enhanced 8 Diatom Growth in a Warming Southern Ocean

bioRxiv preprint doi: https://doi.org/10.1101/2020.07.01.177865; this version posted July 2, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

4 Main Manuscript for:

7 Molecular underpinnings and biogeochemical consequences of enhanced 8 diatom growth in a warming Southern Ocean

11 Loay Jabrea, Andrew E. Allenb,c*, J. Scott P. McCaina, John P. McCrowb, Nancy Tenenbaumd, 12 Jenna L. Spackeene, Rachel E. Siplere,f, Beverley R. Greeng, Deborah A. Bronke,h, David A. 13 Hutchinsd*, Erin M. Bertranda,*#

14 a Dept. of Biology, Dalhousie University, 1355 Oxford Street, PO BOX 15000, Life Sciences 15 Center, Halifax, NS, Canada, B3H 4R2 16 b Microbial and Environmental Genomics, J. Craig Venter Institute, La Jolla, CA 92037, USA 17 c Integrative Oceanography Division, Scripps Institution of Oceanography, University of 18 California, San Diego, La Jolla, CA 92037, USA 19 d University of Southern California, 3616 Trousdale Parkway, Los Angeles, CA 90089, USA 20 e Virginia Institute of Marine Science, College of William & Mary, Gloucester Point, VA, 23062, 21 USA 22 f Memorial University of Newfoundland, 0 Marine Drive, Ocean Sciences Centre, St. John’s, NL, 23 Canada, A1A 4A8 24 g Dept. of Botany, University of British Columbia, #3200-6270 University Boulevard, Vancouver, 25 BC, Canada, V6T 1Z4 26 h Bigelow Laboratory for Ocean Sciences, 60 Bigelow Drive, East Boothbay, ME 04544, USA 27 *Corresponding Authors. 28 #Corresponding Author for Manuscript submission. 29 30 Corresponding Authors: 31 Erin M. Bertrand - [email protected] 32 Andrew R. Allen - [email protected] 33 David A. Hutchins - [email protected]

35 36 Keywords: Southern Ocean | metatranscriptomics | iron limitation | temperature | diatoms 37 38 Author Contributions: 39 E.M.B, A.E.A, D.A.H. designed the experiments; E.M.B. D.A.H., A.E.A., J.E.S., N.T., D.A.B and 40 R.E.S. conducted the experiments. L.J, J.S.P.M, J.P.M, B.R.G, and E.M.B. analyzed the data with 41 help from all authors; L.J. and E.M.B. wrote the paper with help from all authors. 42 43 This PDF file includes: 44 - Main Text 45 - Figures 1-5 46 - Main Text References 47 48

64 Abstract

65 The Southern Ocean (SO) harbours some of the most intense phytoplankton blooms on Earth. 66 Changes in temperature and iron availability are expected to alter the intensity of SO 67 phytoplankton blooms, but little is known about how environmental change will influence 68 community composition and downstream biogeochemical processes. We performed 69 experimental manipulations on surface ocean microbial communities from McMurdo Sound in 70 the Ross Sea, with and without iron addition, at -0.5 °C, 3 °C, and 6 °C. We then examined 71 nutrient uptake patterns as well as the growth and molecular responses of two dominant 72 diatoms, Fragilariopsis and Pseudo-nitzschia, to these conditions. We found that nitrate uptake 73 and primary productivity were elevated at increased temperature in the absence of iron 74 addition, and were even greater at high temperature with added iron. Pseudo-nitzschia became 75 more abundant under increased temperature without added iron, while Fragilariopsis required 76 additional iron to benefit from warming. We attribute the apparent advantage Pseudo-nitzschia 77 shows under warming to upregulation of iron-conserving photosynthetic processes, utilization 78 of iron-economic nitrogen assimilation mechanisms, and increased iron uptake and storage. 79 These data identify important molecular and physiological differences between dominant 80 diatom groups and add to the growing body of evidence for Pseudo-nitzschia’s increasingly 81 important role in warming SO ecosystems. This study also suggests that temperature-driven 82 shifts in SO phytoplankton assemblages may increase utilization of the vast pool of excess 83 nutrients in iron-limited SO surface waters, and thereby influence global nutrient distributions 84 and carbon cycle.

86 Significance Statement

87 Phytoplankton assemblages contribute to the Southern Ocean’s ability to absorb

88 atmospheric CO2, form the base of marine food webs, and shape the global distribution of 89 macronutrients. Anthropogenic climate change is altering the SO environment, yet we do not 90 fully understand how resident phytoplankton will react to this change. By comparing the 91 responses of two prominent SO diatom groups to changes in temperature and iron in a natural 92 community, we find that one group, Pseudo-nitzschia, grows better under warmer low-iron 93 conditions by managing cellular iron demand and efficiently increasing photosynthetic capacity. 94 This ability to grow and draw down nutrients in the face of warming, regardless of iron 95 availability, may have major implications for ocean ecosystems and global nutrient and carbon 96 cycles.

98 Main Text

99 The Southern Ocean (SO) occupies less than 10% of Earth’s ocean surface area but plays 100 a major role in driving global climate and biogeochemical cycles. It connects the Pacific, Atlantic 101 and Indian ocean basins, supplies nutrients to lower latitudes, and absorbs a considerable

102 portion of global heat and CO2 (Frölicher et al. 2015; Moore et al. 2018). SO ecological 103 processes are driven by seasonally productive phytoplankton assemblages that require iron and

104 suitable temperatures to grow. These phytoplankton sequester atmospheric CO2 through the 105 biological pump, sustain food webs, and affect dissolved nutrient availability globally. Climate 106 change models predict overall warming trends in the SO (Turner et al. 2005; Ainley et al. 2010; 107 Boyd et al. 2015; Moore et al. 2018; IPCC 2019) and changes in iron availability (Boyd et al. 108 2015; Hutchins and Boyd 2016; Tagliabue et al. 2017). Despite this, we still have a limited 109 understanding of how these changes will influence the cellular mechanisms that govern 110 phytoplankton dynamics and biogeochemistry in the region. 111 Low iron availability limits phytoplankton growth and is the primary cause of the high- 112 (macro)nutrient low-chlorophyll (HNLC) conditions in the SO (de Baar et al. 1990; Martin et al. 113 1990). Despite this, diatoms are highly successful under these conditions, and contribute 114 heavily to primary productivity throughout the SO (Arrigo et al. 1999; Kang et al. 2001). Low 115 iron-adapted diatoms utilize several strategies to survive chronic, episodic, or seasonal iron 116 limitation. For example, they can reduce photosynthetic iron demand by increasing the size of 117 light harvesting antennae, and their photosynthetic electron transport chains (ETC) can

118 substitute iron-containing ferredoxin with flavodoxin or cytochrome c6 with plastocyanin (Peers 119 and Price 2006; Pankowski and McMinn 2009a; Strzepek et al. 2019). Diatoms may also use 120 iron-free rhodopsin to supplement photosynthetic energy capture (Marchetti et al. 2015), have 121 transport mechanisms that can be up-regulated to maximize iron acquisition (Allen et al. 2008; 122 Morrissey et al. 2015; Kazamia et al. 2018; McQuaid et al. 2018; Coale et al. 2019), and contain 123 iron storage proteins (ferritin) to buffer the sporadic availability of this micronutrient 124 (Marchetti et al. 2009).

125 Temperature has also been shown to limit SO phytoplankton growth in the field and the 126 laboratory. Shipboard warming incubations of iron-limited SO microbial communities increased 127 nutrient drawdown and improved the growth of several diatom groups (Rose et al. 2009), with

128 similar results reported in iron-limited laboratory cultures (Zhu et al. 2016). Warming increases 129 the turnover rate of iron-containing enzymes including nitrate reductase (Gao et al. 2001; Di 130 Martino Rigano et al. 2006), and may thus cause drawdown of SO nitrate even in the absence of 131 iron addition (Hutchins and Boyd 2016; Spackeen et al. 2018). Increased temperature (up to a 132 critical threshold) improves cellular iron use efficiency (Sunda and Huntsman 2011; Jiang et al. 133 2018) and may cause a reduction in the amount of enzymes -and iron- needed to carry out 134 metabolic functions. Antarctic diatoms also exhibit reduced cell size under elevated 135 temperature, which can increase surface area to volume ratios and facilitate nutrient uptake 136 and growth (Zhu et al. 2016; Jabre and Bertrand 2020). Additionally, concurrent increases in 137 temperature and iron have a much larger effect on growth compared to the individual effects 138 of each factor (Rose et al. 2009; Zhu et al. 2016; Hutchins and Boyd 2016), with some species 139 like Fragilariopsis cylindrus requiring iron supplementation to benefit from warming (Jabre and 140 Bertrand 2020).

141 Despite shared adaptations across various taxa, diatoms are an extremely diverse group 142 of phytoplankton. Within the SO, major diatom groups have different morphological and 143 physiological traits, with unique optimal growth temperatures and different tolerances to low 144 iron (Hutchins et al. 2001; Sackett et al. 2013; Tréguer et al. 2018; Strzepek et al. 2019). The 145 molecular mechanisms that underpin these differences are still poorly understood, and are 146 overlooked by the majority of current marine ecosystem models, which consider all diatoms as 147 one group (Laufkötter et al. 2015). This oversimplification cannot capture the effects of specific 148 diatom taxa on important ecological processes, or how environmentally mediated changes in 149 diatom community composition and structure may affect nutrient and carbon biogeochemistry.

150 Here, we present an experimental manipulation study performed at the sea ice edge in 151 the Ross Sea of the Southern Ocean that investigated the growth, nutrient drawdown, and 152 molecular responses of a SO microbial community to warming and changes in iron availability. 153 We show that nutrient drawdown increases in response to temperature and iron, both 154 independently and synergistically, and that two closely related diatom groups, Fragilariopsis 155 spp. and Pseudo-nitzschia spp., respond differently to changes in iron and temperature. These 156 two taxa contribute substantially to diatom assemblages in the SO, and strongly influence the

157 ecology and biogeochemistry of the region (Kang and Fryxell 1992; Rose et al. 2009). Pseudo- 158 nitzschia cell numbers increased more than Fragilariopsis under warming alone, while the latter 159 required iron addition to benefit from increased temperature. Metatranscriptomes suggest that 160 growth of Pseudo-nitzschia under warming is likely facilitated by temperature-responsive light 161 harvesting and iron management strategies that were not utilized by Fragilariopsis. This ability 162 to grow and draw down nutrients in the face of warming, regardless of iron availability, could 163 have major implications for the ecological and biogeochemical response of the HNLC SO to a 164 changing climate.

165

166 Results and Discussion

167 We incubated SO surface water with and without iron addition at -0.5 °C, 3°C, and 6 °C, 168 and examined the microbial community after 24 hours (T1) and again after five (T5) and seven 169 days (T7) (Methods). Increasing temperature and iron separately caused a significant but small 170 increase in primary productivity measured as bicarbonate uptake (Fig. 1A; Fig. S1, two-way 171 ANOVA, temperature effect p = 1.1 x10-6, Fe effect p = 2.4 x 10-6), as well as an increase in 172 nitrate uptake rates (Fig.1A; Fig. S1, two-way ANOVA, temperature effect p = 0.001, Fe effect p 173 = 7.9 x10-5), nitrate consumption (Fig. 1B; Fig. S1, two-way ANOVA, temperature effect p = 2.9 174 x10-11, Fe effect p = 1.2 x10-11) and phosphorus consumption (Fig. 1B, Fig. S1, two-way ANOVA, 175 temperature effect p = 2.3 x10-6, Fe effect p = 1.0 x10-6). Incubations under concurrent warming 176 and iron addition showed an even larger increase in primary productivity (Fig. 1A, two-way 177 ANOVA, temperature x Fe effect p = 4.7 x10-5), nitrate uptake rates (Fig. 1A, two-way ANOVA, 178 temperature x Fe effect p = 0.007), nitrate consumption (Fig. 1B; Fig. S1, two-way ANOVA, 179 temperature x Fe effect p = 6.0 x10-9), and phosphorus consumption (Fig. 1B, Fig. S1, two-way 180 ANOVA, temperature x Fe effect p = 7.5 x 10-5). This synergistic iron-temperature effect has 181 been observed previously (Rose et al. 2009; Spackeen et al. 2018), and shows that in addition to 182 iron, temperature plays an important role in constraining SO phytoplankton growth and 183 macronutrient utilization. These data suggest that future warming conditions can strongly 184 influence SO productivity and nutrient drawdown, especially under increased iron availability.

185 Microscopy showed that Phaeocystis, dinoflagellates, ciliates, Fragilariopsis and Pseudo- 186 nitzschia were among the most abundant eukaryotic taxa at the beginning of the incubation 187 experiment (the in-situ community), and after temperature and iron incubations (Fig. 1C; Fig. 188 S2). We focused our analyses here on the responses of Fragilariopsis (mostly F. kerguelensis) 189 and Pseudo-nitzschia (mostly P. subcurvata), which play integral roles in SO ecology. Pseudo- 190 nitzschia but not Fragilariopsis cell counts increased significantly under warming without added 191 iron (Fig. 1D, two-way ANOVA, temperature effect p = 0.003 and p = 0.5 respectively). 192 Additionally, concurrent warming and iron supplementation caused an increase in Fragilariopsis 193 (Fig. 1D, two-way ANOVA, temperature x iron effect p = 0.001), and a much larger increase in 194 Pseudo-nitzschia cell counts (Fig. 1D, two-way ANOVA, temperature x iron effect p = 1.13 x 10-6. 195 This highlights the ability of Pseudo-nitzschia to take advantage of increased temperature under 196 both low and high iron availability.

197 We used metatranscriptomics to comprehensively capture gene expression patterns 198 and examine the molecular processes underpinning the distinct growth responses of 199 Fragilariopsis and Pseudo-nitzschia to shifts in iron availability and temperature (Fig. 2). 200 Metatranscriptomics allows for the interrogation of cellular pathways utilized across various 201 taxa, and can be used to infer metabolic states under different environmental conditions 202 (Marchetti et al. 2012; Bertrand et al. 2015; Cohen et al. 2018a). Here we identified transcripts 203 belonging to viruses, prokaryotes, and eukaryotes (Fig. S3); with dinoflagellates, ciliates, 204 Pseudo-nitzschia and Fragilariopsis contributing 37% of the total reads mapped to ORFs (Fig. 205 1C). The contribution of Phaeocystis to the metatranscriptome was lower than anticipated 206 based on microscopy, possibly due to difficulty in harvesting cells/colonies without rupturing 207 them (Kiene and Slezak 2006). KEGG Orthology (K.O.) term enrichment analysis on ORFs 208 belonging to Fragilariopsis and Pseudo-nitzschia showed that pathways corresponding to 209 photosynthesis, nitrogen metabolism and genetic information processing/translation were 210 highly differentially expressed in both taxa following temperature and iron increase (Fig. S4; Fig. 211 S5). We explored these pathways in more detail by grouping individual ORFs into groups of 212 similar sequences (clusters) using Markov clustering (MCL) (Enright 2002) (Methods). Several 213 clusters contained differential expression patterns that were similar in Fragilariopsis and 214 Pseudo-nitzschia under iron addition (63 clusters) and temperature increase (212 clusters).

215 However, a larger number of clusters were uniquely differentially expressed in either 216 Fragilariopsis or Pseudo-nitzschia under these conditions (Fig. 3A), suggesting that Fragilariopsis 217 and Pseudo-nitzschia utilize different molecular strategies to respond to changes in 218 temperature and iron.

219 Photosynthesis – Iron addition resulted in upregulation of iron-containing cytochrome b6f 220 complex transcripts and downregulation of flavodoxin transcripts in both Fragilariopsis and 221 Pseudo-nitzschia (Fig. 2, Fig. 3). This indicates that both diatoms in our study were experiencing 222 some degree of iron stress at the time of sampling, as iron-containing photosynthetic processes 223 including the electron transport chain can account for up to 40% of cellular iron quotas (Raven 224 et al. 1999) and are typically downregulated under low iron (LaRoche et al. 1996; Allen et al. 225 2008). Here we did not detect the expression of iron-dependent electron acceptor ferredoxin 226 (PetF) in Pseudo-nitzschia, regardless of iron status. Antarctic Pseudo-nitzschia sp. may have 227 constitutively reduced dependence on petF (Pankowski and McMinn 2009b; Moreno et al. 228 2018) to minimize photosynthetic iron demand.

229 Light Harvesting Complexes – Most members of the light-harvesting complex (LHC) 230 superfamily primarily absorb and direct light energy to photosynthetic reaction centers where 231 charge separation occurs, but some of them (Lhcx clade) are also involved in stress responses 232 and photoprotection (Zhu and Green 2010; Kirilovsky and Büchel 2019; Buck et al. 2019). 233 Expression patterns of transcripts belonging to the major Lhcf group and PSI-associated Lhcr 234 groups were markedly different between the two taxa (Fig. 4). In Fragilariopsis, there was a 235 general pattern of decreased expression with warming, and marked upregulation with iron 236 addition. In contrast, Pseudo-nitzchia showed a pattern of increasing expression with elevated 237 temperatures. Upregulation of Lhcf and Lhcr groups contributes to increased light harvesting 238 efficiency, and alleviates photosynthetic iron demand by reducing the number of iron- 239 containing photosynthetic components required to process light energy (Schuback et al. 2015). 240 Increased light harvesting cross section under warming has been detected in iron-stressed F. 241 cylindrus under laboratory conditions (Jabre and Bertrand, 2020), and these environmental data 242 indicate that Pseudo-nitzschia may be even better equipped to use this mechanism to support 243 growth under low iron in the SO. Several Lhcf and Lhcr clusters were also upregulated in both

244 Fragilariopsis and Pseudo-nitzschia after iron addition. Given that iron addition is expected to 245 reduce light harvesting cross section (Greene et al. 1992; Jabre and Bertrand 2020) it might be 246 expected that Fe addition could decrease demand for Lhcf and Lhcr expression. However, this 247 impact appears to be overridden by increased demand for LHC expression as a result of Fe- 248 induced increases in PSU abundance. Additional photophysiological measurements could 249 provide further insight into how light harvesting cross section responds to changes in LHC 250 expression under different iron and temperature conditions.

251 Notably, each transcript cluster belonging to the Lhcx clade was downregulated after 24 252 hours of incubation at all temperatures and iron conditions in both Fragilariopsis and Pseudo- 253 nitzchia (Fig. 4). Lhcxs are highly sensitive to environmental change, including fluctuations in 254 light and nutrient levels (Zhu and Green 2010; Taddei et al. 2016; Buck et al. 2019). Given these 255 expression patterns and previous observations of Lhcx regulation, the Lhcx transcripts detected 256 here may be responding to changes in light levels during sample collection. Notably, expression 257 of these Lhcx clades showed only minimal responses to iron addition.

258 Plastocyanin – Plastocyanin is a copper-containing protein that acts as an electron 259 shuttle in the ETC and plays an important role in reducing iron requirements in phytoplankton

260 by substituting for cytochrome c6 (Peers and Price 2006; Raven 2013; Wu et al. 2019). Here, 261 iron addition had no significant effect on plastocyanin transcript expression, while warming 262 caused its upregulation in Pseudo-nitzschia and downregulation in Fragilariopsis (Fig. 3). 263 Plastocyanin expression has been observed to be insensitive to iron availability in low iron 264 adapted diatoms (Marchetti et al. 2012). However, our data suggest that Pseudo-nitzschia 265 plastocyanin transcripts are phylogenetically distinct from those of Fragilariopsis, and may play 266 a temperature-responsive role to support growth under elevated temperatures (Fig. S6). 267 Plastocyanin transcript abundance in Pseudo-nitzschia was higher at 3 °C and 6 °C compared to 268 -0.5 °C at T1, showing a rapid response to warming (Fig. S6). This, in combination with the 269 strong and immediate response of LHCs may give Pseudo-nitzschia a photosynthetic advantage 270 over Fragilariopsis in warmer waters.

271 Nitrogen Metabolism – K.O. term enrichment analysis showed that transcripts encoding 272 nitrogen metabolism were enriched in Pseudo-nitzschia and depleted in Fragilariopsis after

273 temperature increase (Fig. S4; Fig. S5). Nitrogen is required for protein synthesis and would be 274 necessary to support the observed increase in Pseudo-nitzschia cell counts under warming, and 275 notably, dissolved nitrogen drawdown was enhanced due to warming alone and elevated 276 further upon iron addition (Fig. 5A, B). However, nitrogen acquisition and assimilation from 277 nitrate requires iron (Schoffman et al. 2016) and the observed Pseudo-nitzschia growth under 278 elevated temperature would necessitate iron-economic nitrogen metabolism. Our data show 279 that ammonium transporter transcripts were upregulated in Pseudo-nitzschia and not in 280 Fragilariopsis under warming (Fig. 3C; Fig. 5). Previous work with laboratory cultures suggests 281 that ammonium transporter expression is regulated by nitrogen demand and can be 282 independent of ammonium availability (Bender et al. 2014). This indicates that Pseudo-nitzschia 283 utilizes ammonium transporters in response to elevated nitrogen demand, which is likely a 284 result of increased growth in response to warming. Elevated ammonium transporter expression 285 would provide Pseudo-nitzschia with an additional nitrogen source at a lower iron cost for use 286 in amino acid synthesis and growth (Raven 1988; Schoffman et al. 2016; Smith et al. 2019). In 287 fact, ammonium drawdown at 6 °C was greater than at lower temperatures, regardless of iron 288 status (Fig. 5B), consistent with the notion that ammonium use at high temperature and low 289 iron could have contributed to the success of Pseudo-nitzschia. However, the high Pseudo- 290 nitzchia growth we observed cannot be supported by ammonium alone; ammonium 291 drawdown comprised only 25.2 ± 2.4 % of dissolved inorganic nitrogen (nitrate + ammonium) 292 drawdown in the 6° no added iron treatment and 7.9 ± 1.0 % of dissolved inorganic nitrogen 293 drawdown in the 6° with added iron treatment (Fig. 5A, B). Nitrate drawdown was responsible 294 for the majority of dissolved nitrogen uptake regardless of iron status, though significantly 295 (two-way ANOVA, p = 9.6 x10-11) more so under elevated iron. While nitrate drawdown was 296 elevated at high temperatures both with and without added iron (Fig. 5A), neither Pseudo- 297 nitzschia nor Fragilariopsis changed their expression of nitrate transporters in response to 298 elevated temperature. Both diatom groups up-regulated nitrate transporters in response to Fe 299 addition regardless of temperature (Fig. 3; Fig. 5E, F). Similarly, nitrate reductase transcript 300 expression was up-regulated in response to Fe addition but not elevated temperature (Fig. 3; 301 Fig. 5G, H). However, the observed elevation in nitrate drawdown and nitrate uptake rates with 302 warming (Fig. 1; Fig. 5), in the absence of a change in gene expression, is consistent with the

303 well-characterized temperature dependence of nitrate reductase activity (Gao et al. 2001). This 304 suggests that the enhanced nitrate assimilation under elevated temperature was indeed 305 accomplished without the increase in iron demand required to produce additional copies of 306 active nitrate reductase.

307 Genetic Information Processing / Translation – Protein synthesis is an energetically costly 308 process that is required for metabolism and growth. Increased kinetic energy due to warming 309 enhances ribosomal efficiency, and reduces the number of ribosomes required to maintain 310 protein synthesis rates (Toseland et al. 2013). Consistent with this response, we observed that 311 several transcript clusters corresponding to ribosomes were downregulated in both 312 Fragilariopsis and Pseudo-nitzschia under warming (Fig. 2, Fig. 3). Warming also caused 313 upregulation of transcripts encoding ribosomal recycling proteins only in Pseudo-nitzschia (Fig. 314 2), which can reduce the energetic costs associated with ribosomal synthesis. The reduction in 315 phosphorus-rich ribosome abundance increases N:P ratios in warmer ocean regions (Toseland 316 et al. 2013). However, bulk nutrient drawdown ratios measured here (Fig. S1) show that N:P 317 drawdown is not influenced by warming (two-way ANOVA, p = 0.07). This suggests that the 318 increasing dominance of diatoms, with their considerably lower N:P ratios than other SO 319 plankton groups (Arrigo et al. 1999; Zhu et al. 2016), is likely an overall driver of N:P and can 320 counterbalance the expected reduction in cellular phosphate demand at high temperature from 321 reduced ribosome content.

322 Cobalamin Metabolism – Exogenous vitamin B12 (cobalamin) acts as a cofactor in the 323 cobalamin-requiring methionine synthase enzyme (MetH) found in all diatoms to synthesize the 324 essential amino acid methionine and facilitate one-carbon metabolism (Bertrand and Allen 325 2012). Under low cobalamin availability, however, certain diatoms including Fragilariopsis but 326 not Pseudo-nitzschia, are capable of synthesizing methionine from homocysteine using a less 327 efficient cobalamin-independent methionine synthase (MetE) (Bertrand et al. 2012; Ellis et al. 328 2017). Diatoms also upregulate CBA1, a cobalamin acquisition related protein, under cobalamin 329 deprivation (Bertrand et al. 2012; Ellis et al. 2017). Following the initial 24-hour incubation 330 period, CBA1 and MetH were upregulated in Pseudo-nitzschia after warming at 6 °C (Fig. S7), 331 but were not significantly influenced by iron or temperature following 5-day incubations (Fig. 3;

332 S7). In contrast, MetE and CBA1 transcripts were significantly upregulated in Fragilariopsis 333 following iron additions after 5-day incubations (Fig. 3; S7), suggesting rearranged metabolism 334 to cope with cobalamin deprivation that likely emerged in response to iron addition (Bertrand 335 et al. 2015).

336 Rapid cobalamin uptake by Pseudo-nitzschia, facilitated by upregulation of CBA1 after 337 24 hours, may give it a competitive advantage when cobalamin is available. However, the ability 338 of Pseudo-nitzschia to maintain vigorous growth without significantly elevating CBA1 expression 339 after 5 days is notable, and suggests that these two diatoms may employ different strategies to 340 cope with low cobalamin availability. Fragilariopsis appears to reduce cobalamin demand 341 through the use of MetE while increasing investment in acquisition with CBA1. In contrast, 342 Pseudo-nitzschia may 1) have a MetH enzyme that is more efficient at elevated temperatures 343 compared to Fragilariopsis, 2) employ salvage and repair of degraded cobalamin complexes 344 (Cohen et al. 2018a) or 3) rely on bacteria in close physical association for cobalamin supply. 345 Our data show that the overall expression of genes encoding cobalamin salvage and 346 remodeling proteins were not influenced by iron status or temperature in Pseudo-nitzschia (Fig. 347 S7), despite previous evidence of their upregulation with iron addition in North Pacific diatoms 348 including P. granii (Cohen et al. 2017, 2018a). Further work comparing cobalamin uptake and 349 methionine synthase kinetics between Pseudo-nitzschia and Fragilariopsis, and examining their 350 relationships with cobalamin producing and consuming bacteria, could provide further insight 351 into how these taxa cope with episodic decreases in cobalamin availability.

352 Iron acquisition, trafficking, and storage – Increased iron uptake, storage, and intracellular 353 transport are strategies to improve growth under low iron availability. Zhu et al. (2016) report 354 iron uptake rates and Fe:C ratios that are higher in Pseudo-nitzschia under increased 355 temperature, and no warming effect on iron acquisition in Fragilariopsis. Our data show 356 downregulation of iron stress induced proteins (ISIP1, ISIP2A, ISIP3) in both Pseudo-nitzschia 357 and Fragilariopsis after iron addition, and ISIP2A upregulation only in Pseudo-nitzschia after 358 warming (Fig. 2; Fig. 3). ISIPs are involved in iron acquisition and storage systems in diatoms, 359 and specifically, ISIP2A contributes to non-reductive iron uptake (Allen et al. 2008; Morrissey et 360 al. 2015; Kazamia et al. 2018; McQuaid et al. 2018). No previous studies have inspected ISIP2A

361 response to temperature, but it is likely that the apparent temperature responsiveness of ISIP2A 362 contributes to the ability of Pseudo-nitzschia to acquire and utilize iron to support growth under 363 warming. Additionally, transcripts for ferritin, an iron storage protein (Marchetti et al. 2009), 364 were significantly upregulated only in Pseudo-nitzschia after iron addition (Fig. 3). Several 365 studies have reported ferritin upregulation in Pseudo-nitzschia under iron replete conditions 366 (Cohen et al. 2018b; a), and Fragilariopsis is also believed to utilize ferritin for long-term iron 367 storage (Marchetti et al. 2009). However, similar to what has been observed in previous field 368 studies (Lampe et al. 2018), Fragilariopsis ferritin was not significantly upregulated after iron 369 addition here (Fig. 3). These trends suggest that in addition to reducing iron demand, Pseudo- 370 nitzschia was better able to increase labile iron uptake and utilized stored ferritin-bound iron to 371 support growth under warming.

372 Ecological and biogeochemical relevance – Most recent model projections show increased 373 temperature trends across the SO with slight warming in areas previously thought not likely to 374 be affected by climate change (Boyd et al. 2015; Rickard and Behrens 2016; Moore et al. 2018). 375 Fragilariopsis grows in a wide range of environments throughout the SO but is most successful 376 in high-salinity, high-iron sea ice (~ 0 °C), while Pseudo-nitzschia is most successful in low- 377 salinity, low-iron meltwaters (~ 2 °C) (Petrou et al. 2011; Petrou and Ralph 2011; Sackett et al. 378 2013). The growth and gene expression patterns we observed in both diatoms may be driven by 379 niche-specific adaptations due to competition for resources. Our results are consistent with a 380 2013 SO incubation study where Pseudo-nitzchia dominated the community after warming 381 (Spackeen et al. 2018), and suggest that Antarctic Pseudo-nitzschia are better equipped to 382 dominate in a warmer ocean, and even more so if iron availability increases. A shift into a 383 Pseudo-nitzschia dominated community raises concerns for domoic acid (DA) production and its 384 toxic effects (Bates et al. 1998). A search for DA biosynthesis transcripts (Brunson et al. 2018) 385 did not yield strong evidence for expression of the DA biosynthesis gene cluster in this 386 experiment (Table S1; Table S2). However, a few studies have reported DA production in the SO 387 (Silver et al. 2010; Geuer et al. 2019), but the temporal, spatial and environmental factors that 388 elicit DA production remain unclear. Further work is required to investigate whether resident 389 Pseudo-nitzschia can produce the toxin under conditions not examined in our study, or if DA 390 producing Pseudo-nitzschia are likely to migrate into a warmer SO.

391 A change in SO phytoplankton community structure could have important 392 biogeochemical ramifications on local and global scales. Notably, warming-induced increases in 393 iron-conserving plastocyanin, LHCs and nitrogen assimilation mechanisms observed here likely 394 resulted in improved iron use efficiency (lower Fe:C uptake ratios) (Zhu et al. 2016). Elevated 395 iron use efficiency with warming appears to result in enhanced nutrient drawdown even in the 396 face of iron limitation. Such warming-enhanced SO nutrient drawdown could have profound 397 consequences for global nutrient distribution. As more nutrients are consumed by 398 phytoplankton and exported out of the surface ocean around Antarctica, less nutrients are 399 available to fuel productivity at lower latitudes (Moore et al. 2018). Here we have identified 400 molecular mechanisms that enable diatoms, particularly Pseudo-nitzschia, to accomplish 401 enhanced growth and nutrient drawdown with rising temperatures under low iron conditions. 402 These mechanisms; upregulation of iron-conserving photosynthetic processes, utilization of 403 iron-economic nitrogen assimilation mechanisms, and increased iron uptake and storage; 404 underpin what may be a pattern of increasing nutrient utilization in the Southern Ocean and 405 decreasing availability of nutrients for high latitudes.

406

407 Materials and Methods Summary

408 Seawater was collected at a 3-meter depth at the ice edge from McMurdo Sound, 409 Antarctica on January 15th, 2015 (165°24.7985’E, 77°37.1370’S) using previously described 410 trace-metal-clean techniques (Bertrand et al. 2015). Triplicate bottles of each treatment 411 (temperature: -0.5 °C ± 0.2 °C, 3 ± 0.5 °C and 6 ± 0.5°C, iron (2 nM): not added, added) were 412 incubated indoors at a constant irradiance of 65-85 uE m-2 sec-1. Metatranscriptome samples 413 were taken at the sea ice edge (IE), in the laboratory immediately following bottle incubation 414 setup (T0), on January 16th (T1), and again on January 20th (T5). Cell counts were measured on 415 T0 and January 22nd (T7). Dissolved nutrients (nitrate, phosphate, silicate) were measured on 416 T0, T1, T3, T5 and T7, and ammonium was measured on T0 and T7. Primary productivity 417 (bicarbonate uptake) and nitrate uptake rates were measured on T0, T1, T3 and T7 (Fig. S1).

418 For metatranscriptomics, the microbial community was harvested on 0.2 μm SterivexTM 419 filters, total RNA was extracted using Trizol reagent (Thermo Fisher Scientific) and ribosomal

420 RNA was removed with Ribo-Zero Magnetic kits (Illumina). The resulting rRNA was purified and 421 subjected to amplification and cDNA synthesis, using the Ovation RNA-Seq System V2 (TECAN). 422 One microgram of the resulting cDNA pool was fragmented to a mean length of 200 bp, and 423 libraries were prepared using Truseq kit (Illumina) and subjected to paired-end sequencing via 424 Illumina HiSeq. Illumina paired reads were filtered to eliminate primer sequences and quality 425 trimmed to Phred Q33, and rRNA identified and removed using riboPicker (Schmieder et al. 426 2012). Transcript contigs were assembled de novo using CLC Assembly Cell 427 (http://www.clcbio.com) and open reading frames (ORFs) were predicted using FragGeneScan 428 (Rho et al. 2010). ORFs were annotated for putative function using hidden Markov models and 429 blastp against PhyloDB (Bertrand et al. 2015), and filtered to eliminate those with low mapping 430 coverage (< 50 reads total over all samples), proteins with no blast hits and no known domains. 431 ORFs were assigned to taxonomic groups of interest (Fig. S2) based on best LPI taxonomy 432 (Podell and Gaasterland 2007; Bertrand et al. 2015). Reads per kilobase mapped (RPKM) 433 expression values for each ORF were calculated and taxon normalized using a normalization 434 factor representing the summed taxonomic group contribution to total nuclear-assigned reads 435 per sample. ORFs were clustered into orthologs and protein families using MCL (Enright 2002). 436 Group normalized cluster (average/total) RPKM expression values were calculated by pooling 437 the taxon-normalized expression values for each group within a cluster. Cluster annotations 438 were aggregated by annotation type (KEGG, KO, KOG, KOG class, Pfam, TIGRfam, EC, GO) and a 439 single annotation was chosen to represent each cluster based on the lowest Fisher's exact test 440 p-value (fisher.test in R). 441 Differential gene expression at T5 (foldchange magnitude and adjusted p-value) was 442 calculated using empirical Bayes quasi-likelihood F-tests (glmQLFTest in edgeR) with a p-value 443 cut off of 0.05 for statistical significance. Iron effect was determined by comparing all -Fe 444 treatments at T5 to all +Fe treatmenets at T5. Temperature effect was determined by 445 comparing all -0.5 °C to 3 °C treatments at T5, and -0.5 °C to 6 °C at T5. A gene was considered 446 upregulated with temperature if it was upregulated in both -0.5 °C vs 3 °C and -0.5 °C vs 6 °C 447 tests. 448 Stable isotope tracer techniques, using 15N-labeled nitrate and 13C-labeled bicarbonate 449 (Cambridge Isotope Laboratories), were used to determine uptake rates, similar to methods

450 previously described in Ross Sea nutrient utilization studies (Spackeen et al. 2018). Tracer 451 amounts of labeled substrates were added to bottles containing the initial ice edge community 452 (T0) and to subsampled volume from treatments at T1, T3 and T7. After incubations (~ 6 hours), 453 the microbial communities (>0.7 μm) were collected on GF/F filters (Whatman; combusted at 454 450°C for 2 hours), and stored frozen. Isotopic enrichments of 15N and 13C were measured on a 455 Europa 20/20 isotope ratio mass spectrometer, and absolute uptake rates (μM h-1) were 456 calculated according to Dugdale and Wilkerson 1986. Complete methods and uptake rate 457 measurements from all time points (Fig. S1) can be found in the SI. 458 459 460 461 Data Availability: The data reported in this paper have been deposited in the NCBI sequence 462 read archive (BioProject accession no. PRJNA637767; BioSample accession nos. SAMN15154229 463 - SAMN15154256). Contigs, assembled ORFs, and MCL cluster abundance and differential 464 expression analysis results are available at 465 https://datadryad.org/stash/share/o4oDU_xXkHACRHkrxsKuKlDcGXjlIpdTn9I7CzaUb1Q 466 467 468 469 470 471 472 473 474 475 476 477 478 479

480 481 482 483

484 485 486 Figure 1 – A) Absolute bicarbonate and nitrate uptake rates at T0 (prior to the incubations) and T7 for each 487 temperature and iron condition. Uptake rates were measured in duplicate from a single incubation bottle per 488 treatment. B) Dissolved nitrate and phosphate concentrations in triplicate incubation bottles at T0 and T7. 489 Phosphate concentration is scaled (16 x) for better visual comparison with nitrate. C) Triplicate measurements of 490 initial (T0) cell counts of eukaryotic taxa identified by light microscopy. Taxa are arranged from low to high 491 abundance. Inset pie chart represents percent contribution of taxa to the total reads mapped to ORFs. “Other” 492 represents all other taxonomic groups, including prokaryotes and viruses. D) Measurements of Fragilariopsis and 493 Pseudo-nitzschia cell counts at T0 and T7 in triplicate incubation bottles for each temperature and iron condition. 494

495

496

497

498 Figure 2 – Schematic representations of Pseudo-nitzschia and Fragilariopsis cells showing cellular processes, with 499 each process comprised of several protein clusters (MCL clusters). Lhcf = light harvesting complexes-f, OEC = 500 oxygen evolving complex, Cyt B6f complex = cytochrome b6f complex, PCYN = plastocyanin, Cyt c6 = cytochrome 501 c6, NRT = nitrate transporter, NR = nitrate reductase, NiT = nitrite transporter, NiR = nitrite reductase, AMT = 502 ammonium transporter, GOGAT = glutamine oxoglutarate aminotransferase cycle, ISIP = iron starvation induced 503 protein. Each row in a heatmap represents one Markov cluster (MCL), each column represents a temperature and 504 iron treatment at T5 with each block representing one biological replicate. Heatmaps were constructed using 505 taxon-normalized RPKM values Empty heatmap placeholders represent clusters found in Fragilariopsis but not 506 Pseudo-nitzschia. Arrows represent energy/electron flow in photosynthetic light reactions, and steps involved in 507 nitrogen assimilation using nitrate or ammonium.

508

509 510 511 Figure 3 – A) Venn diagram showing the number of transcript clusters that are significantly up (black) or down 512 (blue) regulated with iron addition or temperature increase in Fragilariopsis and Pseudo-nitzschia. The intersection 513 represents clusters up or down regulated in both Fragilariopsis and Pseudo-nitzschia. The number outside of the 514 circles represents the number of clusters from both taxa that were not differentially expressed. In both taxa, 515 warming caused more clusters to be differentially expressed than iron addition. B) Differential expression (DE) of 516 various clusters under increased temperature. Differential expression was calculated using the quasi likelihood test 517 (glmQLFTest) in EdgeR, and fold-change was calculated for high-temperature (6 °C) vs low temperature (-0.5 °C) 518 treatments at T5. C) Differential expression (DE) of various clusters after iron addition. Differential expression was 519 calculated using the quasi likelihood test (glmQLFTest) in EdgeR and fold-change was calculated for iron-added vs 520 no iron-added treatments at T5. In both B and C, positive and negative Log2 foldchange values represent up and 521 down regulation respectively. Filled circles are clusters with statistically significant DE (adjusted p-value <0.05). 522 19 523

524 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.01.177865; this version posted July 2, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

526 527 Figure 4 – Heatmaps of MCL clusters of light harvesting complexes Lhcf, Lhcr, Lhcx, Lhcz (see Table S3) in Pseudo- 528 nitzschia and Fragilariopsis measured after 24 hours (T1) and 5 days (T5) of incubation under the various iron and 529 temperature treatments. Heatmaps were constructed using taxon-normalized RPKM values. I.E represents ice 530 edge samples, T0 represents in-situ samples processed immediately before any incubations. Each block is one 531 biological replicate measurement. White-filled up/down pointing triangles represent transcripts that were 532 significantly (glmQLFTest-EdgeR p <0.05) up or down regulated due to iron addition at T5. Black-filled up/down 533 pointing triangles represent transcripts that were significantly (glmQLFTest-EdgeR p <0.05) up or down regulated 534 due to warming at T5. 535 536 537 538

539 540 541 Figure 5 – A, B) Total nitrate and ammonium drawdown from beginning to end of the experiment (difference in 542 concentration between T7 and T0) C-H) Taxon normalized expression values for clusters belonging to ammonium 543 transporters (AMT) nitrate transporters (NT) and nitrate reductase (NR) in both Pseudo-nitzschia and Fragilariopsis 544 prior to incubation (T0), and at T5 with and without added iron at -0.5, 3, and 6 °C. 545 546

547 Acknowledgements

548 We are grateful to Antarctic Support Contractors, especially Ned Corkran, for facilitating 549 fieldwork. We thank Jeff Hoffman, Zhi Zhu, and Quinn Roberts for assistance in the field and 550 Hong Zheng for her work in the laboratory. This study was funded by National Science 551 Foundation (NSF) Antarctic Sciences Awards 1103503 (to E.M.B.), 0732822 and 1043671 (to 552 A.E.A.), 1043748 (to D.A.H.), 1043635 (to D. A. B.), Gordon and Betty Moore Foundation Grant 553 GBMF3828 (to A.E.A.); NSF Ocean Sciences Award NSF-OCE-1136477 and NSF-OCE-1756884 (to 554 A.E.A.) and 1638804 (to D.A.H.), Nova Scotia Graduate Scholarship to L.J., NSERC CGS 555 Postgraduate scholarship and Transatlantic Ocean System Science & Technology scholarship to 556 J.S.P.M, NSERC Discovery Grant RGPIN-2015-05009 to E.M.B., Simons Foundation Grant 504183 557 to E.M.B., and Canada Research Chair support to E. M. B. 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576

577 References 578 Ainley, D., J. Russell, S. Jenouvrier, E. Woehler, P. O. Lyver, W. R. Fraser, and G. L. Kooyman. 2010.

579 Antarctic penguin response to habitat change as Earth’s troposphere reaches 2°C above

580 preindustrial levels. Ecological Monographs 80: 49–66. doi:10.1890/08-2289.1

581 Allen, A. E., J. LaRoche, U. Maheswari, and others. 2008. Whole-cell response of the pennate diatom

582 Phaeodactylum tricornutum to iron starvation. Proceedings of the National Academy of Sciences

583 105: 10438–10443. doi:10.1073/pnas.0711370105

584 Arrigo, K., D. H. Robinson, D. Worthen, R. B. Dunbar, G. R. DiTullio, M. VanWoert, and M. Lizotte. 1999.

585 Phytoplankton Community Structure and the Drawdown of Nutrients and CO2 in the Southern

586 Ocean. Science 283: 365–367. doi:10.1126/science.283.5400.365

587 de Baar, H., A. Buma, R. Nolting, G. Cadee, G. Jacques, and P. Treguer. 1990. On iron limitation of the

588 Southern Ocean: experimental observations in the Weddell and Scotia Seas. Marine Ecology

589 Progress Series 65: 105–122. doi:10.3354/meps065105

590 Bates, S. S., D. L. Garrison, and R. A. Horner. 1998. Bloom Dynamics and Physiology of Domoic-Acid-

591 Producing Pseudo-nitzschia Species. 27.

592 Bender, S. J., C. A. Durkin, C. T. Berthiaume, R. L. Morales, and E. V. Armbrust. 2014. Transcriptional

593 responses of three model diatoms to nitrate limitation of growth. Front. Mar. Sci. 1.

594 doi:10.3389/fmars.2014.00003

595 Bertrand, E. M., and A. E. Allen. 2012. Influence of vitamin B auxotrophy on nitrogen metabolism in

596 eukaryotic phytoplankton. Front. Microbio. 3. doi:10.3389/fmicb.2012.00375

597 Bertrand, E. M., A. E. Allen, C. L. Dupont, T. M. Norden-Krichmar, J. Bai, R. E. Valas, and M. A. Saito. 2012.

598 Influence of cobalamin scarcity on diatom molecular physiology and identification of a

599 cobalamin acquisition protein. Proceedings of the National Academy of Sciences 109: E1762–

600 E1771. doi:10.1073/pnas.1201731109

601 Bertrand, E. M., J. P. McCrow, A. Moustafa, and others. 2015. Phytoplankton–bacterial interactions

602 mediate micronutrient colimitation at the coastal Antarctic sea ice edge. Proceedings of the

603 National Academy of Sciences 112: 9938–9943. doi:10.1073/pnas.1501615112

604 Boyd, P. W., S. T. Lennartz, D. M. Glover, and S. C. Doney. 2015. Biological ramifications of climate-

605 change-mediated oceanic multi-stressors. Nature Clim Change 5: 71–79.

606 doi:10.1038/nclimate2441

607 Brunson, J. K., S. M. K. McKinnie, J. R. Chekan, and others. 2018. Biosynthesis of the neurotoxin domoic

608 acid in a bloom-forming diatom. Science 361: 1356–1358. doi:10.1126/science.aau0382

609 Buck, J. M., J. Sherman, C. R. Bártulos, and others. 2019. Lhcx proteins provide photoprotection via

610 thermal dissipation of absorbed light in the diatom Phaeodactylum tricornutum. Nat Commun

611 10: 4167. doi:10.1038/s41467-019-12043-6

612 Coale, T. H., M. Moosburner, A. Horák, M. Oborník, K. A. Barbeau, and A. E. Allen. 2019. Reduction-

613 dependent siderophore assimilation in a model pennate diatom. Proc Natl Acad Sci USA

614 201907234. doi:10.1073/pnas.1907234116

615 Cohen, N. R., K. A. Ellis, W. G. Burns, R. H. Lampe, N. Schuback, Z. Johnson, S. Sañudo-Wilhelmy, and A.

616 Marchetti. 2017. Iron and vitamin interactions in marine diatom isolates and natural

617 assemblages of the Northeast Pacific Ocean: Iron and vitamin interactions in diatoms. Limnol.

618 Oceanogr. 62: 2076–2096. doi:10.1002/lno.10552

619 Cohen, N. R., W. Gong, D. M. Moran, M. R. McIlvin, M. A. Saito, and A. Marchetti. 2018a. Transcriptomic

620 and proteomic responses of the oceanic diatom Pseudo-nitzschia granii to iron limitation:

621 Intracellular processes of an iron-limited diatom. Environ Microbiol 20: 3109–3126.

622 doi:10.1111/1462-2920.14386

623 Cohen, N. R., E. Mann, B. Stemple, and others. 2018b. Iron storage capacities and associated ferritin

624 gene expression among marine diatoms: Iron storage and ferritin expression in diatoms. Limnol.

625 Oceanogr. 63: 1677–1691. doi:10.1002/lno.10800

626 Di Martino Rigano, V., V. Vona, O. Lobosco, and others. 2006. Temperature dependence of nitrate

627 reductase in the psychrophilic unicellular alga Koliella antarctica and the mesophilic alga

628 Chlorella sorokiniana. Plant, Cell and Environment 29: 1400–1409. doi:10.1111/j.1365-

629 3040.2006.01523.x

630 Ellis, K. A., N. R. Cohen, C. Moreno, and A. Marchetti. 2017. Cobalamin-independent Methionine

631 Synthase Distribution and Influence on Vitamin B12 Growth Requirements in Marine Diatoms.

632 Protist 168: 32–47. doi:10.1016/j.protis.2016.10.007

633 Enright, A. J. 2002. An efficient algorithm for large-scale detection of protein families. Nucleic Acids

634 Research 30: 1575–1584. doi:10.1093/nar/30.7.1575

635 Frölicher, T. L., J. L. Sarmiento, D. J. Paynter, J. P. Dunne, J. P. Krasting, and M. Winton. 2015. Dominance

636 of the Southern Ocean in Anthropogenic Carbon and Heat Uptake in CMIP5 Models. J. Climate

637 28: 862–886. doi:10.1175/JCLI-D-14-00117.1

638 Gao, Y., G. J. Smith, and R. S. Alberte. 2001. Temperature dependence of nitrate reductase activity in

639 marine phytoplankton: biochemical analysis and ecological implications. Journal of Phycology

640 36: 304–313. doi:10.1046/j.1529-8817.2000.99195.x

641 Geuer, J. K., B. Krock, T. Leefmann, and B. P. Koch. 2019. Quantification, extractability and stability of

642 dissolved domoic acid within marine dissolved organic matter. Marine Chemistry 215: 103669.

643 doi:10.1016/j.marchem.2019.103669

644 Greene, R. M., R. J. Geider, Z. Kolber, and P. G. Falkowski. 1992. Iron-induced changes in light harvesting

645 and photochemical energy conversion processes in eukaryotic marine algae. Plant Physiol. 100:

646 565–575. doi:10.1104/pp.100.2.565

647 Hutchins, D. A., and P. W. Boyd. 2016. Marine phytoplankton and the changing ocean iron cycle. Nature

648 Climate Change 6: 1072–1079. doi:10.1038/nclimate3147

649 Hutchins, D. A., P. N. Sedwick, G. R. DiTullio, P. W. Boyd, B. Quéguiner, F. B. Griffiths, and C. Crossley.

650 2001. Control of phytoplankton growth by iron and silicic acid availability in the subantarctic

651 Southern Ocean: Experimental results from the SAZ Project. J. Geophys. Res. 106: 31559–31572.

652 doi:10.1029/2000JC000333

653 IPCC. 2019. IPCC Special Report on the Ocean and Cryosphere in a Changing Climate H.-O. Pörtner, D.C.

654 Roberts, V. Masson-Delmotte, et al. [eds.].

655 Jabre, L., and E. M. Bertrand. 2020. Interactive effects of iron and temperature on the growth of

656 Fragilariopsis cylindrus. Limnol Oceanogr lol2.10158. doi:10.1002/lol2.10158

657 Jiang, H.-B., F.-X. Fu, S. Rivero-Calle, and others. 2018. Ocean warming alleviates iron limitation of

658 marine nitrogen fixation. Nature Climate Change 8: 709–712. doi:10.1038/s41558-018-0216-8

659 Kang, S.-H., and G. A. Fryxell. 1992. Fragilariopsis cylindrus (Grunow) Krieger: the most abundant diatom

660 in water column assemblages of Antarctic marginal ice-edge zones. Polar Biology 12: 609–627.

661 Kang, S.-H., J.-S. Kang, S. Lee, K. H. Chung, D. Kim, and M. G. Park. 2001. Antarctic Phytoplankton

662 Assemblages in the Marginal Ice Zone of the Northwestern Weddell Sea. Journal of Plankton

663 Research 23: 333–352. doi:10.1093/plankt/23.4.333

664 Kazamia, E., R. Sutak, J. Paz-Yepes, and others. 2018. Endocytosis-mediated siderophore uptake as a

665 strategy for Fe acquisition in diatoms. Science Advances 4: 15.

666 Kiene, R. P., and D. Slezak. 2006. Low dissolved DMSP concentrations in seawater revealed by small-

667 volume gravity filtration and dialysis sampling: Filtration and dialysis for dissolved DMSP. Limnol.

668 Oceanogr. Methods 4: 80–95. doi:10.4319/lom.2006.4.80

669 Kirilovsky, D., and C. Büchel. 2019. Evolution and function of light-harvesting antenna in oxygenic

670 photosynthesis, p. 247–293. In Advances in Botanical Research. Elsevier.

671 Lampe, R. H., E. L. Mann, N. R. Cohen, and others. 2018. Different iron storage strategies among bloom-

672 forming diatoms. Proc Natl Acad Sci USA 115: E12275–E12284. doi:10.1073/pnas.1805243115

673 LaRoche, J., P. W. Boyd, M. McKay, and R. Geider. 1996. Flavodoxin as an in situ marker for iron stress in

674 phytoplankton. Nature 382: 802–805.

675 Laufkötter, C., M. Vogt, N. Gruber, and others. 2015. Drivers and uncertainties of future global marine

676 primary production in marine ecosystem models. Biogeosciences 12: 6955–6984.

677 doi:10.5194/bg-12-6955-2015

678 Marchetti, A., D. Catlett, B. M. Hopkinson, K. Ellis, and N. Cassar. 2015. Marine diatom proteorhodopsins

679 and their potential role in coping with low iron availability. The ISME Journal 9: 2745–2748.

680 doi:10.1038/ismej.2015.74

681 Marchetti, A., M. S. Parker, L. P. Moccia, and others. 2009. Ferritin is used for iron storage in bloom-

682 forming marine pennate diatoms. Nature 457: 467–470. doi:10.1038/nature07539

683 Marchetti, A., D. M. Schruth, C. A. Durkin, and others. 2012. Comparative metatranscriptomics identifies

684 molecular bases for the physiological responses of phytoplankton to varying iron availability.

685 Proceedings of the National Academy of Sciences 109: E317–E325.

686 doi:10.1073/pnas.1118408109

687 Martin, J. H., S. E. Fitzwater, and R. M. Gordon. 1990. Iron deficiency limits phytoplankton growth in

688 Antarctic waters. Global Biogeochemical Cycles 4: 5–12. doi:10.1029/GB004i001p00005

689 McQuaid, J. B., A. B. Kustka, M. Oborník, and others. 2018. Carbonate-sensitive phytotransferrin controls

690 high-affinity iron uptake in diatoms. Nature. doi:10.1038/nature25982

691 Moore, J. K., W. Fu, F. Primeau, and others. 2018. Sustained climate warming drives declining marine

692 biological productivity. Science 359: 1139–1143. doi:10.1126/science.aao6379

693 Moreno, C. M., Y. Lin, S. Davies, E. Monbureau, N. Cassar, and A. Marchetti. 2018. Examination of gene

694 repertoires and physiological responses to iron and light limitation in Southern Ocean diatoms.

695 Polar Biol 41: 679–696. doi:10.1007/s00300-017-2228-7

696 Morrissey, J., R. Sutak, J. Paz-Yepes, and others. 2015. A Novel Protein, Ubiquitous in Marine

697 Phytoplankton, Concentrates Iron at the Cell Surface and Facilitates Uptake. Current Biology 25:

698 364–371. doi:10.1016/j.cub.2014.12.004

699 Pankowski, A., and A. McMinn. 2009a. Iron availability regulates growth, photosynthesis, and production

700 of ferredoxin and flavodoxin in Antarctic sea ice diatoms. Aquatic Biology 4: 273–288.

701 doi:10.3354/ab00116

702 Pankowski, A., and A. McMinn. 2009b. Development of immunoassays for the iron-regulated proteins

703 Ferredoxin and Flavodoxin in polar microalgae. Journal of Phycology 45: 771–783.

704 doi:10.1111/j.1529-8817.2009.00687.x

705 Peers, G., and N. M. Price. 2006. Copper-containing plastocyanin used for electron transport by an

706 oceanic diatom. Nature 441: 341–344. doi:10.1038/nature04630

707 Petrou, K., M. A. Doblin, and P. J. Ralph. 2011. Heterogeneity in the photoprotective capacity of three

708 Antarctic diatoms during short-term changes in salinity and temperature. Mar Biol 158: 1029–

709 1041. doi:10.1007/s00227-011-1628-4

710 Petrou, K., and P. Ralph. 2011. Photosynthesis and net primary productivity in three Antarctic diatoms:

711 possible significance for their distribution in the Antarctic marine ecosystem. Mar. Ecol. Prog.

712 Ser. 437: 27–40. doi:10.3354/meps09291

713 Podell, S., and T. Gaasterland. 2007. DarkHose: a new method for genome-wide prediction of horizontal

714 gene transfer. Genome Biol 8: R16. doi:10.1186/gb-2007-8-2-r16

715 Raven, J. A. 1988. The iron and molybdenum use efficiencies of plant growth with different energy,

716 carbon and nitrogen sources. New Phytol 109: 279–287. doi:10.1111/j.1469-

717 8137.1988.tb04196.x

718 Raven, J. A. 2013. Iron acquisition and allocation in stramenopile algae. Journal of Experimental Botany

719 64: 2119–2127. doi:10.1093/jxb/ert121

720 Raven, J. A., M. C. W. Evans, and R. E. Korb. 1999. The role of trace metals in photosynthetic electron

721 transport in O2-evolving organisms. Photosynthesis Research 60: 111–149.

722 Rho, M., H. Tang, and Y. Ye. 2010. FragGeneScan: predicting genes in short and error-prone reads.

723 Nucleic Acids Research 38: e191–e191. doi:10.1093/nar/gkq747

724 Rickard, G., and E. Behrens. 2016. CMIP5 Earth System Models with biogeochemistry: a Ross Sea

725 assessment. Antarctic Science 28: 327–346. doi:10.1017/S0954102016000122

726 Rose, J. M., Y. Feng, G. R. DiTullio, and others. 2009. Synergistic effects of iron and temperature on

727 Antarctic phytoplankton and microzooplankton assemblages. Biogeosciences 6: 3131–3147.

728 Sackett, O., K. Petrou, B. Reedy, and others. 2013. Phenotypic Plasticity of Southern Ocean Diatoms: Key

729 to Success in the Sea Ice Habitat? S. Lin [ed.]. PLoS ONE 8: e81185.

730 doi:10.1371/journal.pone.0081185

731 Schmieder, R., Y. W. Lim, and R. Edwards. 2012. Identification and removal of ribosomal RNA sequences

732 from metatranscriptomes. Bioinformatics 28: 433–435. doi:10.1093/bioinformatics/btr669

733 Schoffman, H., H. Lis, Y. Shaked, and N. Keren. 2016. Iron–Nutrient Interactions within Phytoplankton.

734 Frontiers in Plant Science 7. doi:10.3389/fpls.2016.01223

735 Schuback, N., C. Schallenberg, C. Duckham, M. T. Maldonado, and P. D. Tortell. 2015. Interacting Effects

736 of Light and Iron Availability on the Coupling of Photosynthetic Electron Transport and CO2-

737 Assimilation in Marine Phytoplankton A.M. Cockshutt [ed.]. PLOS ONE 10: e0133235.

738 doi:10.1371/journal.pone.0133235

739 Silver, M. W., S. Bargu, S. L. Coale, and others. 2010. Toxic diatoms and domoic acid in natural and iron

740 enriched waters of the oceanic Pacific. Proceedings of the National Academy of Sciences 107:

741 20762–20767. doi:10.1073/pnas.1006968107

742 Smith, S. R., C. L. Dupont, J. K. McCarthy, and others. 2019. Evolution and regulation of nitrogen flux

743 through compartmentalized metabolic networks in a marine diatom. Nat Commun 10: 4552.

744 doi:10.1038/s41467-019-12407-y

745 Spackeen, J., R. Sipler, E. Bertrand, and others. 2018. Impact of temperature, CO2, and iron on nutrient

746 uptake by a late-season microbial community from the Ross Sea, Antarctica. Aquat. Microb.

747 Ecol. 82: 145–159. doi:10.3354/ame01886

748 Strzepek, R. F., P. W. Boyd, and W. G. Sunda. 2019. Photosynthetic adaptation to low iron, light, and

749 temperature in Southern Ocean phytoplankton. Proc Natl Acad Sci USA 116: 4388–4393.

750 doi:10.1073/pnas.1810886116

751 Sunda, W. G., and S. A. Huntsman. 2011. Interactive effects of light and temperature on iron limitation in

752 a marine diatom: Implications for marine productivity and carbon cycling. Limnology and

753 Oceanography 56: 1475–1488. doi:10.4319/lo.2011.56.4.1475

754 Taddei, L., G. R. Stella, A. Rogato, and others. 2016. Multisignal control of expression of the LHCX protein

755 family in the marine diatom Phaeodactylum tricornutum. EXBOTJ 67: 3939–3951.

756 doi:10.1093/jxb/erw198

757 Tagliabue, A., A. R. Bowie, P. W. Boyd, K. N. Buck, K. S. Johnson, and M. A. Saito. 2017. The integral role

758 of iron in ocean biogeochemistry. Nature 543: 51–59. doi:10.1038/nature21058

759 Toseland, A., S. J. Daines, J. R. Clark, and others. 2013. The impact of temperature on marine

760 phytoplankton resource allocation and metabolism. Nature Climate Change 3: 979–984.

761 doi:10.1038/nclimate1989

762 Tréguer, P., C. Bowler, B. Moriceau, and others. 2018. Influence of diatom diversity on the ocean

763 biological carbon pump. Nature Geosci 11: 27–37. doi:10.1038/s41561-017-0028-x

764 Turner, J., S. R. Colwell, G. J. Marshall, and others. 2005. Antarctic climate change during the last 50

765 years. International Journal of Climatology 25: 279–294. doi:10.1002/joc.1130

766 Wu, M., J. S. P. McCain, E. Rowland, R. Middag, M. Sandgren, A. E. Allen, and E. M. Bertrand. 2019.

767 Manganese and iron deficiency in Southern Ocean Phaeocystis antarctica populations revealed

768 through taxon-specific protein indicators. Nat Commun 10: 3582. doi:10.1038/s41467-019-

769 11426-z

770 Zhu, S.-H., and B. R. Green. 2010. Photoprotection in the diatom Thalassiosira pseudonana: Role of

771 LI818-like proteins in response to high light stress. Biochimica et Biophysica Acta (BBA) -

772 Bioenergetics 1797: 1449–1457. doi:10.1016/j.bbabio.2010.04.003

773 Zhu, Z., K. Xu, F. Fu, J. Spackeen, D. Bronk, and D. Hutchins. 2016. A comparative study of iron and

774 temperature interactive effects on diatoms and Phaeocystis antarctica from the Ross Sea,

775 Antarctica. Marine Ecology Progress Series 550: 39–51. doi:10.3354/meps11732

776777 778 779

780

781

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797 Supplementary Information for :

798 Molecular underpinnings and biogeochemical consequences of enhanced diatom 799 growth in a warming Southern Ocean

800

801 Loay Jabrea, Andrew E. Allenb,c*, J. Scott P. McCaina, John P. McCrowb, Nancy Tenenbaumd, 802 Jenna L. Spackeene, Rachel E. Siplere,f, Beverley R. Greeng, Deborah A. Bronke,h, David A. 803 Hutchinsd*, Erin M. Bertranda,*#

804 a Dept. of Biology, Dalhousie University, 1355 Oxford Street, PO BOX 15000, Life Sciences Center, Halifax, 805 NS, Canada, B3H 4R2 806 b Microbial and Environmental Genomics, J. Craig Venter Institute, La Jolla, CA 92037, USA 807 c Integrative Oceanography Division, Scripps Institution of Oceanography, University of California, San 808 Diego, La Jolla, CA 92037, USA 809 d University of Southern California, 3616 Trousdale Parkway, Los Angeles, CA 90089, USA 810 e Virginia Institute of Marine Science, College of William & Mary, Gloucester Point, VA, 23062, USA 811 f Memorial University of Newfoundland, 0 Marine Drive, Ocean Sciences Centre, St. John’s, NL, Canada, 812 A1A 4A8 813 g Dept. of Botany, University of British Columbia, #3200-6270 University Boulevard, Vancouver, BC, 814 Canada, V6T 1Z4 815 h Bigelow Laboratory for Ocean Sciences, 60 Bigelow Drive, East Boothbay, ME 04544, USA 816 *Corresponding Authors. 817 #Corresponding Author for Manuscript submission. 818 819 Corresponding Authors: 820 Erin M. Bertrand - [email protected] 821 Andrew R. Allen - [email protected] 822 David A. Hutchins - [email protected]

823 This PDF file includes

824 - Supplementary text 825 - Figures S1 – S8 826 - Tables S1 – S3 827 - SI References

828

829

830 Full Materials and Methods:

831 Experimental Design – The planktonic microbial community was sampled at 3 m depth via 832 diaphragm pump from the sea ice edge in McMurdo Sound, Antarctica on January 15th, 2015 833 (165°24.7985’E, 77°37.1370’S) from 11:00 - 12:15 using trace metal clean techniques 834 previously described (Bertrand et al. 2015). In-situ water temperature at the time of sampling 835 was -1 °C. The community was protected from light upon sampling using dark trash bags, stored 836 at 0 °C until 17:00 and then split into trace-metal cleaned polycarbonate bottles (two 1.1L and 837 one 2.7L per treatment), with and without iron supplementation at three different 838 temperatures. Bottles were kept at -0.5 ± 0.2 °C, 3 ± 0.5 °C or 6 ± 0.5 °C at constant 65-85 uE m- 839 2 sec-1 irradiance in indoor incubators for a total of 7 days. For iron supplmentation, 2 nM iron

-1 840 was added as Fe(NO3)3 from an ultrapure analytical standard solution, 1001 mg L in 2% nitric 841 acid. This was diluted to a working stock in pH 2.5 milli Q water with hydrochloric acid, resulting 842 in a negligible nitrate addition to iron amended bottles.

843 Metatranscriptome Sampling and Assembly – Four separate metatranscriptome samples were 844 taken from the initial community, one at the sea ice edge (IE; approximately 2L) and triplicates 845 in the laboratory during bottle incubation setup (T0, approximately 2L). Subsamples of single 846 replicates from each experimental treatment were harvested on January 16th (T1) and 847 subsamples of each replicate (n = 3 for each experimental treatment) were taken again on 848 January 20th (T5). Each was harvested onto 0.2 μm SterivexTM filters. RNA was extracted and 849 sequenced via paired end Illumina HiSeq. Total RNA was extracted using Trizol reagent (Thermo 850 Fisher Scientific). Ribosomal RNA was removed with Ribo-Zero Magnetic kits (Illumina). A mixed 851 Removal Solution was prepared from plant, bacterial, and human/mouse/rat Removal Solution 852 at a ratio of 2:1:1. The resulting rRNA subtracted RNA was purified and subjected to 853 amplification and cDNA synthesis, using the Ovation RNA-Seq System V2 (TECAN). One 854 microgram of the resulting high-quality cDNA pool was fragmented to a mean length of 200 bp,

855 and libraries were prepared using Truseq kit (Illumina) from the -repair step in the 856 manufacturers protocol and subjected to paired-end sequencing via Illumina HiSeq.

857 Illumina paired reads were filtered to eliminate primer sequences and quality trimmed 858 to Phred Q33, and rRNA identified and removed using riboPicker (Schmieder et al. 2012) 859 (average 13.5 % rRNA). Transcript contigs were assembled de novo using CLC Assembly Cell 860 (http://www.clcbio.com) and ORFs predicted using FragGeneScan (Rho et al. 2010). Reads were 861 mapped to ORFs using CLC (73% read mapping), and ORFs were annotated for putative function 862 using hidden Markov models and blast-p against PhyloDB (Bertrand et al. 2015). ORFs were 863 filtered to eliminate those with low mapping coverage (< 50 reads total over all samples), 864 proteins with no blast hits, and no known domains. The remaining set of ORFs were assigned to 865 chloroplast, mitochondrial or nuclear origin based on the best blast-p hit above e-value 1e-3 to 866 an organism with known organellular peptide sequences (nuclear by default), and used for 867 further comparative analysis.

868 Taxonomic groups of interest were defined (Fig. S2) and each ORF was assigned to a 869 group based on best LPI taxonomy (Podell and Gaasterland 2007; Bertrand et al. 2015). Reads 870 per kilobase mapped (RPKM) expression values for each ORF were calculated and taxon 871 normalized using a normalization factor representing the summed taxonomic group 872 contribution to total nuclear-assigned reads per sample. For example, the taxon-normalized 873 expression of an ORF assigned to Fragilariopsis in a particular library is given as reads mapped 874 to that ORF/ORF length/total Fragilariopsis nuclear assigned reads in that library. ORFs were 875 clustered into orthologs and protein families using MCL (Enright 2002). MCL clustering was run 876 in label mode (parameter -abc), with the default inflation setting 2.0 (parameter -I), on ratios of 877 best blast-p bitscore to self-hit bitscore using BLASTALL (e-value: 1e-3). Group normalized 878 cluster (average/total) RPKM expression values were calculated by pooling the taxon- 879 normalized expression values for each group within a cluster. These taxon-normalized RPKM 880 values were used to examine gene and cluster abundance. Cluster annotations were

881 aggregated by annotation type (Kegg, KO, KOG, KOG class, Pfam, TIGRfam, EC, GO) and a single 882 annotation chosen to represent each cluster based on the lowest Fisher's exact test p-value 883 (fisher.test in R) given the 2-way contingency table for each annotation coverage of each 884 cluster.

885 ISIP1 taxonomic re-assignment – Pseudo-nitzschia and Fragilariopsis ISIP1 genes could not be 886 differentiated using blast-p. Instead, nucleotide ISIP1 sequences were compared to a reference 887 database of ISIP1 genes from F. cylindrus, F. kergulensis, P. granii, P. heimii, P. multistriata and 888 P. fraudulenta using blast-n. ISIP1 sequences were assigned to Fragilariopsis or Pseudo-nitzschia 889 based on lowest e-value score. These sequences were then manually placed in clusters and 890 their abundance was normalized to each taxon as previously described.

891 Query for Domoic Acid Biosynthesis (DAB) Genes – BioEdit v7.2 was used to conduct a local 892 blast-p search for DAB genes in our data using Blosum62 similarity matrix. Amino acid query 893 sequences from Pseudo-nitzschia multiseries DAB-A (GenBank: AYD91073.1), DAB-B (GenBank: 894 AYD91072.1), DAB-C (GenBank: AYD91075.1) and DAB-D (GenBank: AYD91074.1) (Brunson et 895 al. 2018) were used.

896 LHC assignments – All Fragilariopsis and Pseudo-nitzschia ORFs annotated as ‘chlorophyll 897 binding’ or ‘light harvesting proteins’ were selected for further inspection. For Fragilariopsis 898 LHCs, a blast-p was performed against F. cylindrus CCMP1102 and annotations were retrieved 899 for the top blast hit for each amino acid sequence. Pseudo-nitzschia LHC sequences were 900 identified by comparison with those collected during the annotation of the Pseudo-nitzschia 901 multiseries CLN-47 genome. The protein sequences of both diatoms were assigned to the Lhcf, 902 Lhcr, Lhcx or Lhcz groups following a previous diatom LHC classification based on maximum- 903 likelihood phylogenetic trees and published as Supp. Information 11 and Supp. Fig. 20 of Mock 904 et al. (2017) and in Hippmann et al. (2017). The dominant Lhcf clade was further subdivided 905 into Lhcf_I, Lhcf_II (diatom-specific), and Lhcf_III groups. PID numbers and the clusters to which 906 they were assigned (see above) can be found in Table S3. 35

907 Plastocyanin Tree – Previously identified plastocyanin sequences were retrieved from MMETSP 908 (Fragilariopsis kerguelensis_0735, Pseudo-nitzschia heimii_1423, Coscinodiscus wailesii_1066), 909 JGI (Fragilariopsis cylindrus_272258), NCBI (Thalassiosira oceanica_EJK71623.1) and Cohen et 910 al. 2018 (Pseudo-nitzschia granii), and aligned with plastocyanin sequences from Pseudo- 911 nitzschia and Fragilariopsis in our dataset using Clustal Omega in SeaView v5.0 (Fig. S8). A 912 maximum-likelihood phylogeny was then estimated using PhyML with LG model and 100 913 bootstrap iterations in SeaView v5.0, and the resulting tree was edited using FigTree v1.4.4.

914 Nutrient Measurements – Samples from initial (T0) and subsequent time points (T1, T3, T5, and 915 T7) were collected, passed through a GF/F filter (Whatman; 0.7 μm nominal pore size; 916 combusted at 450 °C for 2 hours), and filtrate was stored frozen (-40 °C) until further analysis. A 917 Lachat QuickChem 8500 autoanalyzer was used to measure duplicate concentrations of 918 dissolved nitrate, phosphate and silicate (detection limit 0.03 μmol N L-1, 0.03 μmol P L-1 and 919 0.05 μmol Si L-1; Parsons et al. 1984). Samples for ammonium, collected on T0 and T7, were 920 measured in triplicate on a Shimadzu UV-1601 spectrophotometer using the manual phenol- 921 hypochlorite method (detection limit 0.05 μmol N L-1; Koroleff 1983).

922 Uptake Measurements – Nitrate and bicarbonate uptakes were assessed using the initial 923 community (T0) collected from the ice edge and during T1, T3, and T7 of the experiment (Fig. 1 924 and Fig. S1). Uptake rates were measured using 15N and 13C stable isotope tracer techniques,

15 15 - 13 925 and substrates used included N-labeled potassium nitrate (K NO3 ; 98%) and C-labeled

13 - 926 sodium bicarbonate (NaH CO3 ; 99%; both substrates came from Cambridge Isotope 927 Laboratories, Andover, MA). Uptake experiments at T0 were done in triplicate using 1 L 928 polyethylene terephthalate glycol-modified bottles. At T1, T3, and T7 a single replicate of each 929 treatment was subsampled, and uptake experiments were done in duplicate in 230 mL 930 polycarbonate conical bottles (all bottles were acid washed with 10% hydrochloric acid and 931 thoroughly rinsed with ultrapure water). After tracer level additions (less than 10% of 932 background concentrations) of 15N and 13C-labled substrates were made, bottles were returned

933 to their respective incubators for approximately 6 hours. Incubations were terminated by 934 filtering microbial communities (> 0.7 μm) onto combusted (450 °C for 4 hours) Whatman GF/F 935 filters. During T0 incubations, two microbial size fractions (0.7 – 5.0 μm collected on GF/F filters 936 and > 5.0 μm collected on Sterlitech silver membrane filters) were added together to represent 937 the > 0.7 μm microbial community. Filters were kept fozen (-40 °C) inside 1 mL cryo vials until 938 particulate nitrogen and carbon concentrations and isotopic enrichment of 15N and 13C were 939 measured on a Europa 20/20 isotope ratio mass spectrometer. Absolute uptake rates for 15N- 940 labeled nitrate and 13C-labeled bicarbonate were calculated according to Dugdale and 941 Wilkerson (1986) and Hama et al. (1983) respectively. Nitrate uptake rates were not corrected 942 for isotope dilution because concentrations of nitrate were greater than 5.5 μmol N L-1 at all 943 time points, and isotope dilution is generally negligible when concentrations are high (Baer et 944 al. 2017).

945 Cell Counts – Phytoplankton cell count samples from the initial (T0) and final days (T7) of the 946 experiment were preserved with 1% glutaraldehyde, stored refrigerated in the dark and later 947 enumerated in the lab on a Sedgwick Rafter counting chamber using an inverted compound 948 light microscope (Accu-Scope 3032), as in (Tatters et al. 2018). All plankton taxa were identified 949 to the lowest taxonomic level possible according to (Tomas 1997) and (Scott and Marchant 950 2005), with special attention to the diatom genera Fragilariopsis and Pseudo-nitzschia.

951 Statistical Analysis – We analyzed differential gene expression at T5 within observed taxa to 952 determine which genes are responsive to Fe and temperature treatments in each group. First, 953 we normalized reads mapped to each ORF by the abundance of nuclear reads assigned to that 954 taxonomic group in total, which controls for changes in community composition across 955 treatments. We then used a generalized linear model with one categorical explanatory variable, 956 with each category representing a unique experimental treatment. To examine the effect of Fe, 957 temperature, and their interaction on gene expression, we specified model contrasts. For Fe, 958 we tested the difference between the sum of coefficient estimates for all Fe treatments, minus

959 the sum of coefficient estimates for all -Fe treatments. We followed a similar approach for 960 temperature, where we tested the difference between the sum of coefficient estimates for one 961 temperature treatment versus another temperature treatment. For both approaches, we 962 divided these differences by the number of treatments in the sum (i.e. 3 for Fe test and 2 for 963 each temperature test). To test for statistical significance, we used empirical Bayes quasi- 964 likelihood F-tests (glmQLFTest in edgeR). Throughout, we used a p-value cut off of 0.05 for 965 statistical significance.

966 To examine the effects of iron on gene differential expression and fold-change magnitude, -Fe 967 treatments for all temperatures at T5 (-Fe -0.5 °C, -Fe 3 °C, -Fe 6 °C) were compared to +Fe 968 treatments for all temperatures at T5 (+Fe -0.5 °C, +Fe 3 °C, +Fe 6 °C). For temperature effect, - 969 0.5 °C was compared to 3 °C at all iron conditions at T5 (+Fe and -Fe at -0.5 °C vs +Fe and -Fe at 970 3 °C) and -0.5 °C was compared to 6 °C at all iron conditions at T5 (+Fe and -Fe at -0.5 °C vs +Fe 971 and -Fe at 6 °C). For a gene to be considered upregulated with temperature, it had to be 972 upregulated in both -0.5 °C vs 3 °C and -0.5 °C vs 6 °C (same rule applied for down regulation). If 973 a gene is upregulated in one temperature comparison and downregulated in the other, it was 974 not considered differentially expressed. Fold-change for temperature effect was calculated 975 from -0.5 °C (+Fe and -Fe) vs 6 °C (+Fe and -Fe).

976 K.O. term enrichment analysis was performed on significantly upregulated (enriched) and 977 downregulated (depleted) ORFs separately using KEGG enrichment functions in the GOstats R 978 package (Falcon and Gentleman 2007). A hypergeometric distribution test was used to test for 979 significant enrichment and depletion at p < 0.05. K.O. terms associated with less than 10 ORFs 980 were excluded from the results.

981 982 Figure S1 – A-D) Dissolved ammonium, nitrate and phosphate concentrations prior to 983 incubations (day 0), and after 1, 3, 5 and 7 days of incubation with and without added iron at - 984 0.5, 3, and 6 °C. Each point represents a mean value, where n = 2 on day 1, and n=3 on days 0, 985 3, 5, 7. In A-J, error bars represent ± 1 SD, and fall within the bounds of the symbol when not 986 visible. E,F) Nitrate and bicarbonate uptake rates prior to incubations (day 0), and after 1, 3, 987 and 7 days of incubation with and without added iron at -0.5, 3, and 6 °C. Each point represents 988 a mean value, where n = 2 on days 1, 3, 7 and n = 3 on day 0. G) Dissolved nitrogen (nitrate + 989 ammonium) : phosphate drawdown ratio at T7. Draw down is calculated as the difference in 990 concentration between T7 and T0. 39

991 992 Figure S2 – Triplicate cell count measurements of initial (T0) samples and after 7-day 993 incubations with and without iron addition at -0.5, 3 and 6 °C. Cells from the various taxonomic 994 groups were counted and identified using light microscopy.

995

996 997 Figure S3 – Number of significantly differentially expressed open reading frames (ORFs) 998 belonging to the 30 taxonomic groups identified in the metatranscriptome dataset. Red = 999 upregulation, blue = downregulation. Differential taxon-normalized expression was calculated 1000 using the quasi likelihood test (glmQLFTest) in EdgeR, and p-value cut-off of 0.05 was used for 1001 statistical significance. Iron-related DE represents differential expression patterns observed 1002 with and without iron addition at T5, temperature-related DE represents differential expression 1003 patterns observed due to warming at T5 (Methods).

1004

1005

1006

1007

1008 41

1009 1010 Figure S4 – KEGG Orthology (K.O.) term enrichment analysis using all Fragilariopsis ORFs that 1011 were annotated with a K.O. number. Black squares correspond to ‘C’ level annotations that 1012 were significantly (p <0.05) upregulated (Enriched) and/or downregulated (Depleted) at T5 by 1013 temperature increase or iron addition. Circles correspond to the individual ORFs used in the 1014 analysis for each annotation. Black circles represent statistically significant (p <0.05) up or down 1015 regulated ORFs (positive and negative Log2 fold-change values, respectively). Temperature fold 1016 change was calculated using -0.5 °C vs 6 °C treatments. Iron fold-change was calculated using - 1017 Fe vs +Fe treatments at all temperatures (Methods). 42

1018 1019 Figure S5 – KEGG Orthology (K.O.) term enrichment analysis using all Pseudo-nitzschia ORFs 1020 that were annotated with a K.O. number. Black squares correspond to ‘C’ level annotations that 1021 were significantly (p <0.05) upregulated (Enriched) and/or downregulated (Depleted) at T5 by 1022 temperature increase or iron addition. Circles correspond to the individual ORFs used in the 1023 analysis for each annotation. Black circles represent statistically significant (p <0.05) up or down 1024 regulated ORFs (positive and negative Log2 fold-change values, respectively). Temperature fold- 1025 change was calculated using -0.5 °C vs 6 °C treatments. Iron fold-change was calculated using - 1026 Fe vs +Fe treatments at all temperatures (Methods).

1027 1028 Figure S6 – A) Phylogenetic tree of the plastocyanin sequences (ORFs) comprising the 1029 plastocyanin MCL cluster from both Pseudo-nitzschia (red) and Fragilariopsis (black). Non- 1030 highlighted branch tips represent ORFs in our dataset, with corresponding point size 1031 representing mean taxon-normalized ORF expression. Highlighted branch tip labels represent 1032 previously identified plastocyanin sequences retrieved from the MMETSP dataset (Fragilariopsis 1033 kerguelensis_0735, Pseudo-nitzschia heimii_1423, Coscinodiscus wailesii_1066), JGI 1034 (Fragilariopsis cylindrus_272258), NCBI (Thalassiosira oceanica_EJK71623.1) and Cohen et al. 1035 2018 (Pseudo-nitzschia granii). B) Heatmaps of MCL clusters representing and plastocyanin 1036 (cluster_1820) in Pseudo-nitzschia and Fragilariopsis measured after 24 hours (T1) and 5 days 1037 (T5) of incubation under the various iron and temperature treatments. I.E represents ice edge 1038 samples, T0 represents in-situ samples before any incubations. Each block is one biological 1039 replicate measurement. Black-filled up/down pointing triangles represent transcripts that were 1040 significantly (glmQLFTest-EdgeR p <0.05) up or down regulated due to warming at T5. 1041

1042

1043 Figure S7 – Heatmaps of MCL clusters involved in B12 metabolism in Fragilariopsis and Pseudo- 1044 nitzschia measured after 24 hours or 5 days of incubation under various iron and temperature 1045 treatments. I.E represents ice edge samples, T0 represents in-situ samples processed in the 1046 laboratory before any incubations and each block is one biological replicate measurement. 1047 CBA1: cobalamin acquisition protein 1; MetH: cobalamin-requiring methionine synthase; MetE: 1048 cobalamin-independent methionine synthase; CobT, CobN: cobaltochelatase; BluB: gene 1049 involved in DMB production; CobB/CobQ: cobyrinic acid a,c-diamide synthase/ adenosylcobyric 1050 acid synthase. Open triangles represent clusters that were significantly (glmQLFTest-EdgeR p 1051 <0.05) up regulated due to iron addition at T5. Black-filled triangles represent clusters that were 1052 significantly (glmQLFTest-EdgeR p <0.05) down regulated due to warming at T5.

1053

1054 1055 Figure S8 – Alignment of Pseudo-nitzchia and Fragilariopsis plastocyanin sequences from our 1056 metatranscriptome data and previously identified plastocyanin sequences retrieved from the 1057 MMETSP dataset (Fragilariopsis kerguelensis_0735, Pseudo-nitzschia heimii_1423, 1058 Coscinodiscus wailesii_1066), JGI (Fragilariopsis cylindrus_272258), NCBI (Thalassiosira 1059 oceanica_EJK71623.1) and Cohen et al. 2018 (Pseudo-nitzschia granii). The alignment was 1060 conducted using Clustal Omega in SeaView v5.0. and was used to construct the maximum- 1061 likelihood phylogenetic tree for plastocyanin in Fig. S6.

1062

1063

1064

1065 Table S1 – Blast-p search results for sequences encoding domoic acid biosynthesis proteins 1066 (DabA, B, C, D) retrieved from GenBank, against all ORFs from this study. No matches were 1067 found for DabA and DabB encoding genes and no significant eukaryotic matches were found for 1068 DabC encoding genes. Score and e-values were calculated using the Blosum62 similarity matrix. 1069 E-value 1e-30 was used as the cut-off for DabD results. These DabD results are further explored 1070 in Table S2.

Query Gene Matched Sequence Score E- Taxa Hypothesized (ORF ID) value annotation DAB-A - - - - - AYD91073.1

DAB-B - - - - - AYD91072.1

DAB-C AYD91075.1 contig_254441_113_1024_+ 63 6e-010 Flavobacteria 2OG-Fe(II) oxygenase family contig_1174518_67_906_+ 61 6e-010 Alteromonadales 2OG-Fe(II) oxygenase family contig_732674_1_798_- 56 9e-008 Alteromonadales 2OG-Fe(II) oxygenase family contig_633551_118_1041_+ 56 9e-008 Flavobacteria 2OG-Fe(II) oxygenase family contig_627399_1_787_- 54 5e-007 Cyanobacteria 2OG-Fe(II) oxygenase family contig_1862_1_732_- 54 5e-007 Other Bacteria 2OG-Fe(II) oxygenase family contig_660976_1_1043_+ 52 2e-006 Other Stramenopiles 2OG-Fe(II) oxygenase family contig_253357_26_973_+ 52 2e-006 Flavobacteria 2OG-Fe(II) oxygenase family

DAB-D AYD91074.1 contig_596040_24_938_- 231 9e - 061 Pseudo-nitzschia Cyt P450, CYP4/CYP19/CYP26 contig_625510_955_1881_- 225 6e - 059 Pseudo-nitzschia Cyt P450, CYP4/CYP19/CYP26 contig_82244_85_1023_+ 156 5e - 038 Fragilariopsis Cyt P450, CYP4/CYP19/CYP26 contig_596040_947_1648_- 152 9e - 037 Fragilariopsis Cyt P450, CYP4X1 contig_82244_1395_1877_+ 145 7e - 035 Fragilariopsis Cyt P450 contig_680152_23_630_- 132 6e - 031 Cyanobacteria Cyt P450 contig_1109498_1_752_- 130 3e - 030 Other Diatom Cyt P450 1071

1072

1073

1074

1075

1076 47

1077 Table S2 – Reciprocal blast-p search results for ORFs with similarity to DabD-encoding genes 1078 (Table S1) against NCBI’s nr database.

ORF ID Top four blast-p hits Taxa Query E-Value % Accession Cover Identity - Cytochrome P450 F. cylindrus 99% 2e-133 64.62% OEU10111.1 - Unnamed protein P. multistriata 99% 4e-71 42.95% VEU44693.1 contig_596040_24_938_- - DabD P. multiseries 98% 4e-70 42.43% AYD91074.1 - Alkane-1-monooxygenaze F. solaris 96% 2e-61 39.86% GAX28661.1

- Cytochrome P450 F. cylindrus 94% 1e-111 56.48% OEU10111.1 contig_625510_955_1881_- - DabD P. multiseries 95% 3e-67 41.81% AYD91074.1 - Unnamed protein P. multistriata 95% 3e-63 41.14% VEU44693.1 - Alkane-1-monooxygenaze F. solaris 94% 6e-58 40.61% GAX28661.1

- Cytochrome P450 F. cylindrus 99% 0.0 96.43% OEU10111.1 contig_82244_85_1023_+ - Alkane - 1 - monooxygenaze F. solaris 88% 5e-46 35.21% GAX23170.1 - Alkane-1-monooxygenaze F. solaris 88% 2e-45 35.92% GAX28661.1 - DabD P. multiseries 93% 3e-41 32.00% AYD91074.1

- Cytochrome P450 F. cylindrus 93% 1e-114 77.06% OEU10111.1 contig_596040_947_1648_- - Alkane - 1 - monooxygenaze F. solaris 84% 4e-41 40.8% GAX28661.1 - Alkane-1-monooxygenaze F. solaris 84% 2e-40 40.5% GAX23170.1 - DabD P. multiseries 84% 8e-40 37.81% AYD91074.1

- Cytochrome P450 F. cylindrus 100% 4e-110 99.37% OEU10111.1 contig_82244_1395_1877_+ - Hypothetical protein T. oceanica 98% 3e-43 49.68% EJK45228.1 - DabD P. multiseries 99% 1e-38 47.47% AYD91074.1 - Unnamed protein P. multistriata 98% 2e-37 45.86% VEU44693.1

- TPA: cytochrome P450 Porticoccaceae 100% 5e-130 86.07% HAZ79708.1 contig_680152_23_630_- - Hypothetical protein Porticoccaceae 100% 2e-128 86.07% MAY69286.1 - Cytochrome P450 Porticoccaceae 100% 7e-102 69.65% WP_155531439.1 - Hypothetical protein SAR92 99% 1e-101 69.50% KRP17789.1

- Cytochrome P450 F. cylindrus 92% 7e-58 43.25% OEU10111.1 contig_1109498_1_752_- - DabD P. multiseries 91% 1e-32 29.34% AYD91074.1 - Unnamed protein P. multistriata 91% 6e-31 29.75% VEU44693.1 - Alkane-1-monooxygenaze F. solaris 87% 1e-24 30.38% GAX28661.1 1079

1080

1081 Table S3 – Protein ID (PID) assignments for ORFs in the various Pseudo-nitzschia spp. and 1082 Fragilariopsis spp. light harvesting complex (LHC) clusters. PIDs were assigned by performing a 1083 blast-p search against Pseudo-nitzschia multiseries (CLN-47) and Fragilariopsis cylindrus (CCMP 1084 1102). LHC groups were assigned based on previous diatom LHC classifications in Mock et al. 1085 2017 (Supp.Info.11) and Hippmann et al. 2017.

Cluster Pseudo-nitzschia PID Fragilariopsis PID LHC Group clust_1044 258347, 261276 169285, 205888 Lhcf_III

clust_1084 247179, 263274 195639, 261294 Lhcr

clust_1194 257565, 303201, 304112, 195777 Lhcf_I 306047

clust_1236 41763, 197371, 302398, 170761, 174589, 269349, Lhcf_I 310027 269868, 271557

clust_1787 238335, 257821 188478, 271659 Lhcx

clust_1938 318557 270184 Lhcr

clust_2256 66239, 238335 269313, 272562 Lhcx

clust_2549 307175 273003, 271559 Lhcr

clust_3219 301726 270606 Lhcr

clust_3282 307174 271561, 273005 Lhcf_II

clust_332 257565, 306047, 306447 267329, 271330, 271332 Lhcf_I

clust_3363 264176 260998 Lhcr

clust_3551 325841 210115, 213124 Lhcr

clust_402 191001, 300768, 303201, 267576, 267702, 267837, Lhcf_I 304112, 305325, 307376 268624, 269543, 269567

clust_646 14959, 178030, 234364, 177731, 187698, 195639, Lhcz 318210 270940

clust_712 66239, 264022, 307174 218498, 271659, 272024 Lhcx

clust_80 191001, 255698, 300768, 143190, 174589, 207327, Lhcf_I 300769, 303058, 303201, 267576, 271931, 268626 305325, 305720, 307376

1086 References

1087 Baer, S.E., R.E. Sipler, Q. Roberts, P. Yager., M. E. Frischer, and D.A. Bronk. 2017. Seasonal 1088 nitrogen uptake and regeneration in the western coastal Arctic. Limnology and 1089 Oceanography 62:2463-2479. doi:10.1002/lno.10580 1090 Bertrand, E.M., J.P. McCrow, A. Moustafa, and others. 2015. Phytoplankton–bacterial 1091 interactions mediate micronutrient colimitation at the coastal Antarctic sea ice edge. 1092 Proceedings of the National Academy of Sciences 112: 9938–9943. 1093 doi:10.1073/pnas.1501615112 1094 Brunson, J. K., S. M. K. McKinnie, J. R. Chekan, and others. 2018. Biosynthesis of the neurotoxin 1095 domoic acid in a bloom-forming diatom. Science 361: 1356–1358. 1096 doi:10.1126/science.aau0382 1097 Cohen, N. R., W. Gong, D. M. Moran, M. R. McIlvin, M. A. Saito, and A. Marchetti. 2018. 1098 Transcriptomic and proteomic responses of the oceanic diatom Pseudo-nitzschia granii 1099 to iron limitation: Intracellular processes of an iron-limited diatom. Environ Microbiol 1100 20: 3109–3126. doi:10.1111/1462-2920.14386 1101 Dugdale, R.C., and F.P. Wilkerson. 1986. The use of 15N to measure nitrogen uptake in eutrophic 1102 oceans; experimental considerations. Limnology and Oceanography 31: 673-689. doi: 1103 10.4319/lo/1986.31.4.0673 1104 Enright, A. J. 2002. An efficient algorithm for large-scale detection of protein families. Nucleic 1105 Acids Research 30: 1575–1584. doi:10.1093/nar/30.7.1575 1106 Falcon, S., and R. Gentleman. 2007. Using GOstats to test gene lists for GO term association. 1107 Bioinformatics 23: 257–258. doi:10.1093/bioinformatics/btl567

1108 Hama, T., Miyazaki, T., Ogawa, Y., Iwakuma, T., Takahashi, M., Otsuki, A., and S. Ichimura. 1983. 1109 Measurement of photosynthetic production of a marine phytoplankton population using 1110 a stable 13C isotope. Marine Biology 73: 31-36. doi:10.1007/BF00396282

1111 Hippmann, A. A., N. Schuback, K.-M. Moon, J. P. McCrow, A. E. Allen, L. J. Foster, B. R. Green, 1112 and M. T. Maldonado. 2017. Contrasting effects of copper limitation on the 1113 photosynthetic apparatus in two strains of the open ocean diatom Thalassiosira 1114 oceanica. P.A. Cobine [ed.]. PLoS ONE 12: e0181753. doi:10.1371/journal.pone.0181753

1115 Koroleff, F. 1983. Determination of nutrients. In Grasshoff, K., Ehrhardt, M., and K. Kremling 1116 (Eds.), Methods of seawater analysis (pp. 125-187). New York: Verlag Chemie.

1117 Mock, T., R. P. Otillar, J. Strauss, and others. 2017. Evolutionary genomics of the cold-adapted 1118 diatom Fragilariopsis cylindrus. Nature 541: 536–540. doi:10.1038/nature20803 1119 Parsons, T., Maita, Y., and C. Lalli. 1984. A manual of chemical and biological methods for 1120 seawater analysis. New York: Pergamon Press. 1121 Podell, S., and T. Gaasterland. 2007. DarkHose: a new method for genome-wide prediction of 1122 horizontal gene transfer. Genome Biol 8: R16. doi:10.1186/gb-2007-8-2-r16 50

1123 Rho, M., H. Tang, and Y. Ye. 2010. FragGeneScan: predicting genes in short and error-prone 1124 reads. Nucleic Acids Research 38: e191–e191. doi:10.1093/nar/gkq747

1125 Scott, F. I., and H. J. Marchant, eds. 2005. Antarctic Marine Protists, ABRS, Canberra and 1126 Australian Antarctic Division. 1127 Schmieder, R., Y. W. Lim, and R. Edwards. 2012. Identification and removal of ribosomal RNA 1128 sequences from metatranscriptomes. Bioinformatics 28: 433–435. 1129 doi:10.1093/bioinformatics/btr669 1130 Tatters, A. O., A. Schnetzer, K. Xu, and others. 2018. Interactive effects of temperature, CO2 and 1131 nitrogen source on a coastal California diatom assemblage. Journal of Plankton Research 1132 40: 151–164. doi:10.1093/plankt/fbx074 1133 Tomas, C. 1997. Identifying Marine Phytoplankton, Academic Press, San Diego. 1134

1135