Interactions between Thermal Acclimation, Growth Rate, and Phylogeny Influence Prochlorococcus Elemental Stoichiometry.

Variability in plankton elemental requirements can be important for global ocean biogeochemistry but we currently have a limited understanding of how ocean temperature influences the plankton C/N/P ratio. Multiple studies have put forward a 'translation-compensation' hypothesis to describe the positive relationship between temperature and plankton N/P or C/P as cells should have lower demand for P-rich ribosomes and associated depressed QP when growing at higher temperature. However, temperature affects many cellular processes beyond translation with unknown outcomes on cellular elemental composition. In addition, the impact of temperature on growth and elemental composition of phytoplankton is likely modulated by the life history and growth rate of the organism. To test the direct and indirect (via growth rate changes) effect of temperature, we here analyzed the elemental composition and ratios in six strains affiliated with the globally abundant marine Cyanobacteria Prochlorococcus. We found that temperature had a significant positive effect on the carbon and nitrogen cell quota, whereas no clear trend was observed for the phosphorus cell quota. The effect on N/P and C/P were marginally significantly positive across Prochlorococcus. The elemental composition and ratios of individual strains were also affected but we found complex interactions between the strain identity, temperature, and growth rate in controlling the individual elemental ratios in Prochlorococcus and no common trends emerged. Thus, the observations presented here does not support the 'translation-compensation' theory and instead suggest unique cellular elemental effects as a result of rising temperature among closely related phytoplankton lineages. Thus, the biodiversity context should be considered when predicting future elemental ratios and how cycles of carbon, nitrogen, and phosphorus may change in a future ocean.

Introduction

The cellular contents of class="Chemical">carbon (C), class="Chemical">n class="Chemical">nitrogen (N), phosphorus (P), and other elements in marine phytoplankton are emerging as important features of ocean biogeochemistry. For a long time, C/N/P was assumed static at Redfield proportions (106/16/1)[1]. However, variability in plankton elemental requirements can influence nutrient limitation patterns and stress [2,3], nitrogen fixation rates [4,5], the link between nutrient supply and C export [6], and atmospheric CO2 levels [7]. Recent work has demonstrated extensive differences in the elemental content and ratios of marine communities across regions or seasons [8-12]. However, the exact mechanisms controlling the observed regional differences are still uncertain as key environmental factors strongly co-vary in the ocean.

Multiple biological mechanisms controlling the elemental composition of marine phytoplankton have been proposed. The main suggested controls include nutrient availability, growth rate, temperature, and life history. Extensive experimental and model studies have demonstrated a strong effect of nutrient availability, whereby a low supply of class="Chemical">nitrogen or class="Chemical">n class="Chemical">phosphorus leads to a low cell quota (Q) of the corresponding element [13-16]. Another important factor is the cellular allocation towards P rich ribosomes at elevated growth rates. Coined the ‘Growth Rate Hypothesis’ [17], fast growth is hypothesized to result in high Q and corresponding low C/P and N/P ratios. However, this growth effect on stoichiometry appears to vary extensively by organism and environmental conditions [16,18,19]. Thus, the genetic and environmental contexts (and possible interactions) for changes in growth rate may be important to consider.

Temperature has also been proposed as a relevant factor for setting the elemental allocation in marine phytoplankton but we currently have limited understanding and data for the quantitative effect [20-22]. Toseland and co-workers showed that phytoplankton produce more P-rich ribosomes at lower temperature; putatively to compensate for lower translational efficiency. Hence, temperature was hypothesized to influence the elemental ratios in phytoplankton such that a future warming of the oceans would lead to increasing N/P ratios of marine communities [20]. Supported by a meta-analysis of eukaryotic phytoplankton lineages, Yvon-Durocher and co-workers detected an increase in C/P and N/P (but not C/N) for cells growing at higher temperature [22]. However, temperature affects many cellular processes beyond translation with unknown outcomes on cellular elemental composition. In addition, the impact of temperature on growth and elemental composition of phytoplankton is likely modulated by the life history of the organism. Important life history traits include the thermal growth optimum and more broadly adaptation of individual cellular processes to various temperature conditions. For example, an increase in temperature may have very different physiological effects depending on whether the rise occurs below or above the thermal growth optimum. Thus, the organismal context should be considered for understanding the influence of temperature on the elemental composition of phytoplankton.

The most abundant phytoplankton lineage in the ocean is the marine Cyanobacteria class="Species">Prochlorococcus [23]. The liclass="Chemical">neage is respoclass="Chemical">nsible for a substaclass="Chemical">ntial fractioclass="Chemical">n of oceaclass="Chemical">n primary productivity aclass="Chemical">nd thus ceclass="Chemical">ntral to oceaclass="Chemical">n biogeochemical fuclass="Chemical">nctioclass="Chemical">niclass="Chemical">ng. Most studies of phytoplaclass="Chemical">nktoclass="Chemical">n elemeclass="Chemical">ntal stoichiometry are doclass="Chemical">ne usiclass="Chemical">ng eukaryotic liclass="Chemical">neages with a large cell size that are either rare or abseclass="Chemical">nt iclass="Chemical">n the oceaclass="Chemical">n. Iclass="Chemical">n coclass="Chemical">ntrast, we curreclass="Chemical">ntly kclass="Chemical">now little about what regulates the elemeclass="Chemical">ntal compositioclass="Chemical">n of class="Chemical">n class="Species">Prochlorococcus but it appears that changes in growth rate could affect C/N/P [24]. Further, a prior study of Prochlorococcus strain MED4 found that concomitant with an increase in growth rate and cell size, C,N, and P quotas increased with temperature, maintaining the same stoichiometry [25]. The Prochlorococcus clade also harbors extensive genetic diversity including clades adapted to different ocean temperature regimes [26]. The HLII clade dominates in warm tropical waters, whereas the HLI clade is more common in higher latitude, cooler waters [27]. These distributions are consistent with the growth responses of representative strains, with the HLI strains growing faster than HLII at low temperature, and the HLII strains growing faster than HLI at high temperature. However, it is unknown how adaptations to different ocean regimes and temperature will modulate a thermal effect on the elemental composition.

Here, we investigated the sensitivity of the elemental quotas of n class="Species">Prochlorococcus to chaclass="Chemical">nges iclass="Chemical">n temperature, with the hypothesis that their N/P aclass="Chemical">nd C/P ratios are positively related to temperature. As a possible temperature effect will be modulated by chaclass="Chemical">nges iclass="Chemical">n growth rate as well as the life history (i.e., geclass="Chemical">notype) of the orgaclass="Chemical">nisms, we quaclass="Chemical">ntified the effect of temperature oclass="Chemical">n the growth rate aclass="Chemical">nd elemeclass="Chemical">ntal compositioclass="Chemical">n of three straiclass="Chemical">ns of the high-temperature-adapted HLII clade aclass="Chemical">nd three of the low-temperature-adapted HLI clade. This study coclass="Chemical">ntributes fuclass="Chemical">ndameclass="Chemical">ntal iclass="Chemical">nformatioclass="Chemical">n oclass="Chemical">n how temperature iclass="Chemical">nflueclass="Chemical">nces the elemeclass="Chemical">ntal compositioclass="Chemical">n of this key, abuclass="Chemical">ndaclass="Chemical">nt liclass="Chemical">neage aclass="Chemical">nd its coclass="Chemical">ntributioclass="Chemical">n to global biogeochemical cycles.

Materials and Methods

Strains and growth conditions

Six axenic class="Species">Prochlorococcus straiclass="Chemical">ns affiliated with the HLI aclass="Chemical">nd HLII clades were aclass="Chemical">nalyzed iclass="Chemical">n this study (Table 1). All straiclass="Chemical">ns except VOL29 were previously reclass="Chemical">ndered axeclass="Chemical">nic (Table 1), while VOL29 isolatioclass="Chemical">n aclass="Chemical">nd purificatioclass="Chemical">n is described preseclass="Chemical">ntly. VOL29 was isolated duriclass="Chemical">ng the POWOW1 cruise iclass="Chemical">n the N. Pacific Oceaclass="Chemical">n (29.6°N, 125.07°W) oclass="Chemical">n March 9th, 2012 at a depth of 3 m usiclass="Chemical">ng Iclass="Chemical">nstaclass="Chemical">nt Oceaclass="Chemical">n Sea Salt media (Spectrum Braclass="Chemical">nds, CA) media to grow uclass="Chemical">nder ambieclass="Chemical">nt coclass="Chemical">nditioclass="Chemical">ns (20–24°C, 40 μmol quaclass="Chemical">nta m−2 s−1 light) [28]. VOL29 was reclass="Chemical">ndered axeclass="Chemical">nic usiclass="Chemical">ng the established helper method oclass="Chemical">n class="Chemical">n class="Chemical">agarose plates [29]. A spontaneous streptomycin-resistant derivative was obtained, and plated for colonies on AMP-J agarose medium pre-seeded with the streptomycin-sensitive helper strain EZ55. Prochlorococcus colonies were inoculated in AMP-J liquid and then rendered axenic by the addition of streptomycin to eliminate the helper and subsequently verified as axenic [29,30].

Table 1

Overview of strains and temperature treatments.

Strain name	Derived from	Clade	Origin	T treatments (°C)	References
VOL 7	MED4	HLI	Med Sea	16, 19, 24, 26	[30,45]
VOL8	MIT9515	HLI	Eq. Pacific	19, 24, 26	[30,46]
VOL29	N/A	HLI	N. Pacific	16, 19, 24	This study
VOL 4	MIT9312	HLII	Gulf Stream	19, 24, 26	[30,47]
VOL1	MIT9215	HLII	Eq. Pacific	24, 26	[30,48]
UH18301	N/A	HLII	N. Pacific	19, 24, 26	[30]

class="Species">Prochlorococcus straiclass="Chemical">ns were cultured iclass="Chemical">n filtered (0.2 μm polyclass="Chemical">n class="Chemical">carbonate filter, pressure <10 mm Hg) artificial seawater AMP-J medium [29] (per L, 28.1 g NaCl, 6.9 g MgSO4*7H2O, 5.49 g MgCl2*6H2O, 0.67 g KCl. 1.47g CaCl2, 0.504 g NaHCO3 with 2 ml 0.5 M TAPS, pH 8.0, 1 ml 0.4 M (NH4)2SO4, 2 ml 0.025 M NaH2PO4 pH7.5, 100 μl 10,000 X Pro99 Trace Metal Mix) with 40 μmol quanta m−2 s−1 light on a 12:12 light:dark cycle using cool white fluorescent bulbs at temperatures from 16°C to 26°C (Table 1). Cultures were acclimated to the test temperature for at least three transfers (~20 generations) at high cell concentration (> 107cells ml-1), before transferring at 106 cells ml-1. The purity of strains was tested before and after strains were inoculated to culture. Prochlorococcus was inoculated into YTSS and 1/10 ProAC purity test broths in dark and monitored for visible signs of heterotrophic growth [29,30]. Samples were strictly taken during exponential growth. All collected data is listed in S1 Table.

Cell counting

Concentration of class="Species">Prochlorococcus was measured by flow cytometry usiclass="Chemical">ng a class="Chemical">n class="Species">Guava EasyCyte 8HT cytometer (Millipore, Billerica, MA) and growth rates were estimated.

Particulate organic matter

Particulate organic class="Chemical">carbon (POC), class="Chemical">n class="Chemical">nitrogen (PON) and phosphorus (POP) samples were each collected in duplicate from each of three biological replicates (6 total) by filtration of 50 ml of culture onto precombusted (5 h, 500°C) GF/F filters (Whatman, Florham Park, New Jersey) and stored at -20°C. To quantify POC and PON, filter samples were thawed and allowed to dry overnight at 65°C. Filters were then packed into a 30 mm tin capsule (CE Elantech, Lakewood, New Jersey) and analyzed for C and N content on a FlashEA 1112 nitrogen and carbon analyzer (Thermo Scientific, Waltham, Massachusetts) [31]. POC and PON concentrations were calibrated using known quantities of atropine and peach leaves in each run. The amount of POP was determined in each sample using a modified ash-hydrolysis method [15,32]. We also analyzed multiple blank controls.

Data analysis

All data was plotted using Matlab. Statistical analyses were done using linear models in R. To account for non-linear effects of T on the elemental content of n class="Species">Prochlorococcus straiclass="Chemical">ns, T was treated as a factor with four levels.

Phylogenetic analysis

class="Species">Prochlorococcus ITS class="Chemical">nucleotide sequeclass="Chemical">nces from each straiclass="Chemical">n were aligclass="Chemical">ned usiclass="Chemical">ng ClustalW [33]. Pair-wise DNA distaclass="Chemical">nce matrix (w. F84 substitutioclass="Chemical">n matrix) aclass="Chemical">nd class="Chemical">neighbor-joiclass="Chemical">niclass="Chemical">ng tree were calculated usiclass="Chemical">ng Phylip v. 3.69 [34] usiclass="Chemical">ng ITS sequeclass="Chemical">nces from class="Chemical">n class="Species">Prochlorococcus assemblies HNLC1 and HNLC2 as outgroup [35]. Next, we found the linear contribution of temperature, growth rate and strain identity on cell quotas and rates. To evaluate if the strain identity effects were phylogenetically structured, we then compared an Euclidian distance matrix of the strain identity effects to the pair-wise DNA distance matrix using a Mantel test in the R package ‘vegan’ [36].

Results

To identify the impact of temperature on the elemental composition of class="Species">Prochlorococcus, we quaclass="Chemical">ntified the class="Chemical">n class="Chemical">carbon, nitrogen, and phosphorus cell quota as well as growth rate of six axenic strains (Table 1). HLI and HLII clades, adapted to different temperatures, were represented by 3 strains each, and to facilitate comparisons between clades, the temperatures assayed were within the permissive range for growth of all strains. Median cell quotas across all strains of 0.44 fg P, 6.4 fg N, and 33 fg C were similar to previously measured levels [24,37]. Temperature had a significant linear positive effect on Q and Q across all strains but no direct effect on Q (Table 2 and Fig 1A–1C). Over the 10°C increase in temperature, Q and Q rose by 40% and 35%, respectively. We also examined the elemental ratios. C/N showed little variability and was close to Redfield proportions (medianC/N = 6.1)(Fig 1D). In contrast, C/P and N/P were above Redfield proportions (medianC/P = 174, medianN/P = 29)(Fig 1E and 1F). Both ratios showed some effect of temperature and there was a marginally significant positive linear trends across all strains (Table 2).

Table 2

Effects of temperature, growth rate and strain identity on the elemental composition of six Prochlorococcus strains.

		QP		QN		QC		C/N		C/P		N/P
Linear model		Estimate	p-value	Estimate	p-value	Estimate	p-value	Estimate	p-value	Estimate	p-value	Estimate	p-value
	Intercept	1.2	4.9x10-3	2.2	0.08	7.8	0.27	5.8	7.6x10-17	53	0.6	14	0.3
	Temperature	0.0	1.0	0.2	2.9x10-4	1.2	2.3x10-4	0.01	0.6	6.8	0.1	0.9	0.1
	Growth rate	-2.0	1.2x10-2	-1.6	0.6	-3.3	0.8	0.3	0.8	-65	0.7	-14	0.6
ANOVA		SS	p-value	SS	p-value	SS	p-value	SS	p-value	SS	p-value	SS	p-value
	Strain	1.1	0.1	22	1x10-4	1x103	6x10-6	4.3	7x10-3	2x105	2x10-5	4x103	1x10-4
	Temperature1	1.4	2.3x10-2	41	5x10-7	1x103	8x10-7	0.3	0.6	3x104	2.5x10-2	6.2x102	4.8x10-2
	Growth rate	0.6	3.0x10-2	6.5	1x10-3	2.8x102	3x10-4	0.8	5.3x10-2	67	0.9	10	0.7
	Strain:T	0.8	0.5	22	1x10-3	5.7x102	3x10-3	1.8	0.4	6x104	2.6x10-2	1x103	3.9x10-2
	Strain:Gr	0.1	0.9	0.9	0.8	28	0.8	0.3	0.9	3x103	0.9	1.x102	0.9
	T:Gr	0.6	0.2	0.8	0.6	6.5	0.9	0.7	0.3	2x104	0.13	5.8x102	5.7x10-2
	Strain:T:Gr	0.2	0.9	8.6	7.2x10-2	2.4x102	9.6x10-2	0.8	0.9	5x103	0.9	89	0.9
Phylogeny corr.		R	p-value	R	p-value	R	p-value	R	p-value	R	p-value	R	p-value
	Mantel test	0	0.48	0.19	0.23	0.19	0.23	0.06	0.24	-0.07	0.60	0.02	0.45

1 Temperature was treated as factor in ANOVA

Fig 1

Influence of temperature and growth on the elemental composition and ratios of across Prochlorococcus strains.

Factors measured are (A) phosphorus cell quota (Q), (B) nitrogen cell quota (Q), (C) carbon cell quota (Q), (D) C/N, (E) C/P, and (F) N/P. The color of each sample point indicates the observed growth rate. All ratios are molar based.

Factors measured are (A) class="Chemical">phosphorus cell quota (Q), (B) class="Chemical">n class="Chemical">nitrogen cell quota (Q), (C) carbon cell quota (Q), (D) C/N, (E) C/P, and (F) N/P. The color of each sample point indicates the observed growth rate. All ratios are molar based.

We also quantified growth rates of all the isolates to determine how changes in growth rate in conjunction with temperature affected the elemental composition of n class="Species">Prochlorococcus (Fig 2). At a light level of 40 μmol quaclass="Chemical">nta m−2 s−1, the growth raclass="Chemical">nged betweeclass="Chemical">n 0.13 d-1 aclass="Chemical">nd 0.39 d-1. Temperature affected the growth of HLI aclass="Chemical">nd HLII isolates slightly differeclass="Chemical">nt whereby several HLI isolates sustaiclass="Chemical">ned growth a lower T whereas HLII isolates were less iclass="Chemical">nhibited at high T. Relaticlass="Chemical">ng growth rate aclass="Chemical">nd elemeclass="Chemical">ntal quotas aclass="Chemical">nd ratios, we detected a class="Chemical">negative effect of growth rate oclass="Chemical">n Q, whereas the other cell quotas aclass="Chemical">nd ratios did class="Chemical">not display aclass="Chemical">ny liclass="Chemical">near treclass="Chemical">nds (Fig 1 aclass="Chemical">nd Table 2).

Fig 2

Growth rate of HLI and HLII cultures across a temperature gradient.

HLI cultures are VOL8 (MIT9515), VOL29, and VOL7 (MED4) and HLII cultures are MIT9215, MIT9312, and UH18301.

HLI cultures are VOL8 (MIT9515), VOL29, and VOL7 (class="Species">MED4) aclass="Chemical">nd HLII cultures are MIT9215, class="Chemical">n class="CellLine">MIT9312, and UH18301.

We next examined the influence of temperature on the cell quotas in the context of each strain as well as indirectly via changes in growth rate (Table 2 and Fig 3). We observed some similarities as well as difference in the response across the six strains. As seen in the aggregated response for all strains, individual strains displayed negative relationships between growth rate and Q. In addition, temperature also influenced Q on a per strain basis (Fig 3A), but there were no systematic differences between strains nor interactions between factors (Table 2). The HLI strains VOL8 and VOL29 had higher overall Q and Q and temperature plus growth rate influenced Q and Q across all strains (Fig 3B and 3C). Thus, there was evidence for direct influences of strain identity, temperature, and growth rate–as well as some interactions–in setting the overall elemental composition (Table 2).

Fig 3

Influence of temperature on cell quotas in six Prochlorococcus strains.

Cell quotas include (A) phosphorus (Q), (B) nitrogen (Q), and (C) carbon (Q). The error bars represent one standard deviation based on duplicate sampling of each strain.

Cell quotas include (A) class="Chemical">phosphorus (Q), (B) class="Chemical">n class="Chemical">nitrogen (Q), and (C) carbon (Q). The error bars represent one standard deviation based on duplicate sampling of each strain.

Temperature and growth rate also affected the elemental ratios of each strain in unique ways (Fig 4). For C/N, we observed differences in the overall level across the strains, whereby strain VOL8 showed the highest and class="CellLine">MIT9312 the lowest level (Fig 4A). The straiclass="Chemical">n specific C/N was also margiclass="Chemical">nally affected by growth rate but class="Chemical">not temperature (Table 2). The straiclass="Chemical">n specific C/P aclass="Chemical">nd N/P varied coclass="Chemical">nsiderably betweeclass="Chemical">n straiclass="Chemical">ns (Fig 4B aclass="Chemical">nd 4C) aclass="Chemical">nd iclass="Chemical">n particular, MIT9215 had coclass="Chemical">nsiderably higher ratios compared to the other straiclass="Chemical">ns. Temperature had a sigclass="Chemical">nificaclass="Chemical">nt impact oclass="Chemical">n C/P aclass="Chemical">nd N/P but the directioclass="Chemical">n varied betweeclass="Chemical">n straiclass="Chemical">ns. C/P aclass="Chemical">nd N/P iclass="Chemical">n straiclass="Chemical">ns VOL8, VOL29, aclass="Chemical">nd MIT9215 were positively affected. Iclass="Chemical">n coclass="Chemical">ntrast, VOL7 showed high variability with lower ratios at 16°C as well as 26°C aclass="Chemical">nd higher at the iclass="Chemical">ntermediate temperature, UH81301 showed class="Chemical">no respoclass="Chemical">nse, aclass="Chemical">nd a class="Chemical">negative respoclass="Chemical">nse was observed iclass="Chemical">n class="Chemical">n class="CellLine">MIT9312. Thus, there were complex interactions between the strain identity and temperature in controlling the elemental ratios in Prochlorococcus.

Fig 4

Influence of temperature on elemental ratio in six Prochlorococcus strains.

Elemental ratios include (A) C/N, (B) C/P, and (C) N/P. All ratios are molar based. The error bars represent one standard deviation based on duplicate sampling of each strain.

We next found little phylogenetic structuring of elemental changes across strains. After subtracting the overall influence of temperature and growth rates on cell quotas or elemental ratios, we then identified additional variation attributed to each strain. These ‘strain-specific’ contributions were then compared to the phylogenetic distance between each strain (Mantel test, Table 2). This comparison revealed that neither cell quotas nor ratios were phylogenetically structured.

Discussion

Multiple studies have put forward a ‘translation-compensation’ hypothesis for a positive relationship between temperature vs. N/P or C/P. Cells should have lower demand for P-rich ribosomes and associated depressed Q when growing at higher temperature [20,22]. A lower Q will cause elevated C/P and N/P and such an acclimation mechanism should further explain the high elemental ratios observed in cells growing in the hot, oligotrophic gyres [8,9]. However, we see little support for this hypothesis in class="Species">Prochlorococcus. Iclass="Chemical">nstead, the thermal effect leads to iclass="Chemical">ncreasiclass="Chemical">ng Q aclass="Chemical">nd Q, whereas Q shows little systematic chaclass="Chemical">nge. This poiclass="Chemical">nts towards other physiological acclimatioclass="Chemical">n mechaclass="Chemical">nisms as the primary drivers of elemeclass="Chemical">ntal chaclass="Chemical">nges iclass="Chemical">n class="Chemical">n class="Species">Prochlorococcus. The observed elemental changes are likely associated with a cell size increase as Q and Q increase in tandem. The underlying mechanism for this increase in Q and Q in Prochlorococcus is not known but the response was opposite to Scenedesmus and Asterionella [38]. Based on studies in heterotrophic organisms, it is likely associated with an increase in cellular macromolecules and especially protein content [39]. Such a change in cell size means that you cannot simply extrapolate from an increase or a decrease in an individual cell quota (like Q) to the stoichiometric ratio. Thus, our study adds to an emerging concept, whereby changes in cell size due to physiological responses to different environmental conditions are important for regulating the elemental composition and ratios in marine Cyanobacteria [16].

Q appears linked to changes in thermally induced growth rate but not temperature itself. This would indicate support for the growth rate hypothesis [17] but Q is actually decreasing at elevated growth rates across all strains as well as for most individual strains. As seen in other marine Cyanobacteria [16], it is clear that the growth rate hypothesis alone cannot explain differences in elemental composition across class="Species">Prochlorococcus straiclass="Chemical">ns. However, there is a lot of variatioclass="Chemical">n, which suggests iclass="Chemical">ndividual straiclass="Chemical">n Q respoclass="Chemical">nses to temperature aclass="Chemical">nd growth physiology. The compositioclass="Chemical">n of P coclass="Chemical">ntaiclass="Chemical">niclass="Chemical">ng macromolecules uclass="Chemical">nderlyiclass="Chemical">ng the overall cellular P coclass="Chemical">nteclass="Chemical">nt is poorly coclass="Chemical">nstraiclass="Chemical">ned iclass="Chemical">n mariclass="Chemical">ne Cyaclass="Chemical">nobacteria [15,16] as the sum of class="Chemical">n class="Chemical">phospholipids and nucleic acids does not get close to Q. Thus, it is currently uncertain, which biochemical mechanism will lead to the observed changes in Q.

We observe overall high C/P and N/P, whereas C/N is close to Redfield proportions in class="Species">Prochlorococcus. The cells are growiclass="Chemical">ng uclass="Chemical">nder class="Chemical">nutrieclass="Chemical">nt replete coclass="Chemical">nditioclass="Chemical">ns, which should lead to C/P aclass="Chemical">nd N/P at the lower eclass="Chemical">nd of the raclass="Chemical">nge for aclass="Chemical">n orgaclass="Chemical">nism [3,40]. Our observatioclass="Chemical">ns of above Redfield ratios iclass="Chemical">n class="Chemical">n class="Species">Prochlorococcus are consistent with past observations [8,24,41] and suggest this lineage has overall high C/P and N/P. As such, the presence of Prochlorococcus in low latitude marine communities will contribute to elevated elemental ratios independently of thermal and nutrient conditions.

We do not observe a direct phylogenetic structuring of cell quotas and ratios within class="Species">Prochlorococcus. However, we observed a sigclass="Chemical">nificaclass="Chemical">nt iclass="Chemical">nflueclass="Chemical">nce oclass="Chemical">n the elemeclass="Chemical">ntal compositioclass="Chemical">n via chaclass="Chemical">nges iclass="Chemical">n growth rate aloclass="Chemical">ng the temperature gradieclass="Chemical">nt. The temperature effect oclass="Chemical">n growth iclass="Chemical">n class="Chemical">n class="Species">Prochlorococcus strains have been shown to be strongly phylogenetically structured, whereby the HLI and HLII clades are adapted to lower and higher temperature regimes, respectively [26,27]. Thus, we see an indirect phylogenetic structuring through the effect of growth rate on the cell quotas and ratios. In addition, we see extensive strain variability in the elemental content and ratios due to thermal acclimation. Thus, the organismal context and potentially growth optimum appear important for the individual response. This is consistent with the thermal response in other phytoplankton lineages and strain specific variability in quotas and ratios of Gyrodinium species [42,43]. In an analysis across nine eukaryotic phytoplankton lineages, Yvon-Durocher and co-workers observed substantial variability in the link between thermal changes and elemental cellular composition [22]. Furthermore, this meta-analysis as well as our study found little thermal effect on C/N, suggesting C/N being fairly invariant to temperature changes.

The broader envclass="Chemical">ironmeclass="Chemical">ntal growth coclass="Chemical">nditioclass="Chemical">ns are importaclass="Chemical">nt to coclass="Chemical">nsider wheclass="Chemical">n evaluaticlass="Chemical">ng the elemeclass="Chemical">ntal outcome iclass="Chemical">n class="Chemical">n class="Species">Prochlorococcus to thermal changes. In this study, the cells were growing under nutrient replete conditions and Prochlorococcus may store large reserves of P overwhelming any contributions from ribosomal RNA. Multiple studies have shown the possibility for interactions between factors including interactions between nutrient limitation and temperature [22,38]. Scenedesmus showed stronger thermal responses under nutrient limited vs. replete conditions. Hence, future work studying the interaction between nutrient limitation and thermal conditions would enhance our understanding for how changes in ocean temperature would affect Prochlorococcus elemental stoichiometry.

Our study has implications for understanding both present day and future biogeochemical functioning. The oceans are projected to undergo substantial changes in temperature due to rising class="Chemical">CO2 iclass="Chemical">n the atmosphere. Such eclass="Chemical">nvclass="Chemical">n class="Chemical">ironmental changes will likely have a large impact on phytoplankton community structure and physiology [23,44]. This has been predicted to lead to an increase in N/P ratios in phytoplankton communities [20,22]. However, the observations presented here suggest unique cellular elemental effects as a result of rising temperature among closely related phytoplankton lineages. Thus, the biodiversity context should be considered when predicting future elemental ratios and how the link between the cycles of carbon, nitrogen, and phosphorus may change in a future ocean.

Data associated with this study.

(CSV)

Click here for additional data file.