The balance between photosynthesis and respiration explains the niche differentiation between Crocosphaera and Cyanothece

Graphical abstract


Introduction
Nitrogen (N) is the major limiting factor for primary productivity in the subtropical and tropical ocean gyres. In such regions, specialized prokaryotes, 'dinitrogen (N 2 ) fixers' or 'diazotrophs', are able to use N in the most abundant form on Earth and in seawater, N 2 gas. Diazotrophs utilize the nitrogenase enzyme which cleaves the strong triple bond of the N 2 molecule to form bioavailable ammonium (NH 4 + ). Thus, N 2 fixation introduces a source of new bioavailable N to surface waters and is considered the most important external source of N to the ocean, supporting ocean productivity and biological pump [2,15,16,39]. Marine autotrophic unicellular diazotrophs play a key role in biogeochemical cycles [44]. Marine autotrophic unicellular diazotrophs are phylogenetically divided into two groups. The unicellular group B (UCYN-B) is most closely related to Crocosphaera watsonii [24]. UCYN-B lives singly, colonially or in symbioses with a large chain-forming diatom (Climacodium frauenfeldianum) [4,5] containing cultivated strains (i.e. C. watsonii WH8501, WH0003, PS0609) [25,42]. The unicellular group C (UCYN-C) is the group identified by nifH sequence and is most closely related to the free-living unicellular diazotroph Cyanothece sp. strain ATCC51142 [35]. UCYN-B widely distributes in oligotrophic sub-tropical and tropical open ocean [27,40], while, the distribution of UCYN-C is reported in coastal regions [37,41].
The distribution of these diazotrophs is constrained by the growth capacity, which is supported by capacities of photosynthesis and N 2 fixation. Since nitrogenase is inactivated by the oxygen [12,14], both of these groups segregate these processes by temporal separation: restrain N 2 fixation to the night when oxygen evolution of light-dependent photosynthesis is absent [1,45]. The timing of these processes is primarily regulated by the circadian clock [6], and the nocturnal nitrogen fixation is fueled by carbon (C) accumulated during the light period [8,36]. Given the above, how is niche separation between Crocosphaera and Cyanothece related to their relationship between photosynthesis and N 2 fixation?
In this study, we performed physiological experiments to investigate the capacities of photosynthesis and N 2 fixation in Crocosphaera watsonii WH8501 and Cyanothece sp. ATCC51142 under optimum growth conditions. Further, we elaborated on a simple ecological model to show that their different niches can be sufficiently explained by the observed differences in their photosynthetic capacities and rates of C consumption.

Materials and methods
We obtained Crocosphaera watsonii WH8501 (hereafter referred to as Crocosphaera) from the culture collection of the Royal Netherlands Institute of Sea Research in Yerseke (Strain CCY 0601) and maintained it in YBC-II medium without any enriched combined nitrogen source [7] at 28°C in flat panel photobioreactors (FMT150; Photon System Instruments, Drásov, Czech Republic) [29]. Light intensity followed a sinusoidal 12:12 h light:dark cycle with a maximum irradiance of 400 lmol photons m À2 s À1 in the middle of the light period. Cyanothece sp. American Type Culture Collection (ATCC) 51142 (hereafter referred to as Cyanothece) has been recently renamed to Crocosphaera subtropica [23]. We maintained it under similar conditions as Crocosphaera, but grew it in ASP-2 media without NaNO 3 [32] and set the maximum irradiance to 130 lmol photons m À2 s À1 . The chosen light intensities represent the optimum light conditions of these two species based on the corresponding light saturation intensities for carbon incorporation, Ek C , which were 331 lmol quanta m À2 s À1 and 88 lmol quanta m À2 s À1 for Crocosphaera and Cyanothece, respectively (Table 1).
We used the optical densities of the cultures at 735 nm (OD 735 ) as an indicator of cell density. The photobioreactor monitored and averaged the OD 735 data over every 1 or 5 min intervals and we normalized them to the OD 735 values recorded at 1 h after the onset of the light phase (1L). Based on our earlier studies [20,26,34], we considered the increase in OD 735 during the light phase as a proxy for C accumulation by photosynthesis. In contrast, the decrease in OD 735 at the end of the light phase and also during the dark phase as a proxy for C consumption by respiration (see Discussion).
To determine the rates of N 2 fixation by an acetylene reduction assay ( [46]), we dispensed 5 mL of cell suspensions into HCl-rinsed glass vials (n = 3). After sealing with a septum, we injected 10 mL of acetylene gas (99.7 % [v/v]; Linde Gas) into each vial by replacing the same volumes of the headspace. The samples were incubated at 28°C in the dark for 1 h. We took subsamples from the headspace immediately after acetylene addition and also at the end of the incubation period to determine their ethylene content with a flame ionization gas chromatograph (HRGC 5300, Carlo Erba Instruments, Strumentazione, Italy). We calculated the rate of ethylene production according to Breitbarth [3] and converted it to rate of N 2 fixation using a theoretical molar ratio of acetylene reduction to cellular N 2 reduction of 4:1 [28].
Using a Clark-type electrode combined with a DW2/2 electrode chamber (Hansatech, UK), we determined O 2 evolution and respiration rates at 28°C in the presence of 5 mM sodium bicarbonate. Depending on the culture cell density, we spun down different culture volumes by 10 min of centrifugation at 7500 g and resuspended the pellet in 2.7 mL fresh medium. We calculated the rate of gross O 2 evolution as a difference of net O 2 evolution measured at a saturating irradiance of 600 lmol photons m À2 s À1 (KL1500, Schott, Mainz, Germany) and the of respiratory O 2 consumption measured in the dark right after the light exposure.
To determine electron transport rates, we took aliquots (2 mL) from the photobioreactor after 1, 3, 5, 7, 9 and 11 h into the light period (hereafter referred to as (1L, 3L, 5L 7L, 9L and 11L) and transferred them into the measuring FastAct head of the benchtop FastOcean fast repetition rate FRR fluorometer (Chelsea Technologies Group, West Molesey, UK). We obtained photosynthesisirradiance (P-E) curves by exposing the cells to increasing irradiances from 0 to 1495 lmol photons m À2 s À1 with 11 steps. Absolute electron transport rates (ETR) normalized to Chl a concentration (lmol electrons (lg Chl a) À1 h À1 ) were calculated according to the ''absorption" method of [30] as: where F o , and F m , are the minimum and maximum Chl a fluorescence in the dark, F 0 and F 0 m are the steady-state and maximal Chl a fluorescence measured at given light intensity, E is the intensity of the irradiance (in lmol photons m À2 s À1 ), K A is the instrumentspecific calibration factor (11800 m À1 ), 3600 is factor to convert seconds to hours, and [Chl a] is the Chl a concentration (in mg/ m 3 ). Then, we estimated the maximum electron transport rates (ETR max ) by fitting the data to the model of [11].
To determine carbon incorporation of 14 C-carbon, we collected 1 mL subsamples which were inoculated with 14 C-labelled sodium bicarbonate (MP Biochemicals, CA, USA) at a concentration of 0.5 lCi/mL [17], and placed into a photosynthetron at 28°C for 30 min at irradiances of 0 to 1528 lmol photons À2 s À1 . Carbon uptake was then terminated by addition of 50 lL formaldehyde. Subsequently, we acidified samples with 250 lL of 3 N HCl and placed them onto an orbital shaker overnight to purge off unincorporated label. For counting radioactive decay, we added 5 mL of Ecolite liquid scintillation cocktail (MP Biochemicals, CA, USA) to each sample and placed them in a scintillation counter (PerkinElmer, MA, USA) for counting. The resultant rates of carbon incorporation were normalized to Chl a concentration to obtain the assimilation number, P m We then calculated the electron demand for carbon assimilation U, either as the ratio of ETR max and P m B under saturating irradiance where a ETR is the Chl a normalized absorption coefficient for electron transport rate (a ETR , lmol e (lg Chl a) À1 h À1 (lmol quanta m À2 s À1 ) À1 ) and a C is the Chl a normalized absorption coefficient for carbon incorporation (a C , lmol C (lg Chl For analysis of membrane proteins and their complexes, we isolated cyanobacterial membranes by breaking the cells with glass beads followed by differential centrifugation [21]. Afterwards, we solubilized the membranes in 1 % n-dodecyl maltoside (DM) and solubilized proteins and we separated their complexes by clear native polyacrylamide gel electrophoresis (CN PAGE). For assessment of the standard D1 (sD1) and rogue D1 (rD1) protein content, the membranes were analyzed in denaturing 12-20 % linear gradient polyacrylamide gel containing 7 M urea. We stained the gel with SYPRO Orange and transferred the separated proteins onto a polyvinylidene difluoride (PVDF) membrane. We incubated this membrane with specific primary antibodies against sD1 or rD1, then with a secondary antibody-horseradish peroxidase conjugate (Sigma, St. Louis, MO, USA) and specific signal of each protein was developed in the presence of chemiluminiscent substrate Immobilon Crescendo (Merck, USA).

Growth model
We modelled the growth rates of the studied microorganisms according to Eq. (2) (CFM-CC: Cell Flux Model of Crocosphaera and Cyanothece). It expresses growth rate l i : (unit: d À1 , where i indicates microorganism, i.e., either Crocosphaera (Cro) or Cyanothece (Cya)). For microorganism i, we assumed that growth rate increases with both the maximum photosynthesis rate (P i Max , unit: d À1 ) and the nutrient repletion factor (f N , dimensionless unit) in the medium. The term P i Max f N thus represents the rate of photosynthesis, which agrees with the general observation that nutrient repletion positively affects the rate of photosynthesis. The f N value represents the level of environmental nutrient repletion taking values from 0 (deplete environment) to 1 (replete environment). The equation also consists of a constant respiration rate (m i , unit: d À1 ), which decreases the overall growth rate. Based on our experimental results (Fig. 2), we assigned higher photosynthesis P i Max and respiration m i rates to Cyanothece than to Crocosphaera. The list of parameters and parameter values are given in Tables S1 and S2, respectively.

Results and discussion
In Crocosphaera, the normalized OD 735 values increased up to $2.3 fold of the initial value at 9L, declining thereafter continuously until the end of the dark period to a final value of $1.4 fold of the initial value (Fig. 1A). In Cyanothece on the other hand, the corresponding normalized OD 735 values increased only up to $1.6 fold of initial value by 9L peaking at 12L and decreased to $1.1 fold of the initial value during the dark period (Fig. 1B).
In both species, the maximal gross O 2 evolution capacities already started to increase in the middle of the dark phase and reached maxima of 363 nmol O 2 (lg Chl a) À1 h À1 at L6 and 580 nmol O 2 (lg Chl a) À1 h À1 at L8 in Crocosphaera and Cyanothece, respectively (Fig. 2). In both species, O 2 evolution declined during the early dark phase, but started to increase again in the second half of this period. Cyanothece respired actively both in the dark and light phases and its total diurnal respiration was about 2.1-fold higher (1.1 lmol O 2 (lg Chl a) À1 light phase À1 ) than that of Crocosphara (0.5 lmol O 2 (lg Chl a) À1 light phase À1 ) and was 1.2-fold higher overall (2.0 lmol O 2 (lg Chl a) À1 day À1 in Cyanothece compared to 1.7 lmol O 2 (lg Chl a) À1 day À1 in Crocosphaera) ( Table 1).
Crocosphaera started N 2 fixation gradually so that a marked increase was not measurable until 6 h into the dark (6D) and peaked at 10D with 31.7 ± 0.5 nmol N 2 (lg Chl a) À1 h À1 (Fig. 2A). On the other hand, Cyanothece started to fix N 2 from the beginning of the dark phase and peaked at 6D (43.0 ± 4.8 nmol N 2 (lg Chl a) À1 h À1 ) (Fig. 2B). Over the course of an entire day, Crocosphaera fixed 77.1 nmol N 2 (lg Chl a) À1 , which corresponds to 82 % of the amount fixed by Cyanothece (93.7 nmol N 2 (lg Chl a) À1 ).
The average chlorophyll-normalized maximal carbon incorporation rates P m B during the light period were 0.237 ± 0.037 lmol C (lg Chl a) À1 h À1 and 0.175 ± 0.036 lmol C (lg Chl a) À1 h À1 for Crocosphaera and Cyanothece, respectively, while the corresponding mean PSII-mediated ETR max were 1.260 ± 0.245 lmol e -(lg Chl a) À1 h À1 and 2.290 ± 0.199 lmol e -(lg Chl a) À1 h À1 , respectively (Fig. 3). Thus, the electron demand for carbon incorporation (U max = ETR max / P m B ) calculated from these values was 2.5 times higher in Cyanothece (13.6 ± 3.1 e C À1 ) compared to Crocosphaera (5.3 ± 1.2 e C À1 ), suggesting that the coupling between photosynthetically generated electrons and C fixation at saturating irradiances in Crocosphaera is much tighter and more effective compared to Cyanothece (Fig. 3). However, the electron requirement for carbon incorporation under non-saturating light intensities (U Lim ) were comparable for both Crocosphaera (3.7 ± 1.3 e C À1 ) and Cyanothece (3.7 ± 0.9 e C À1 ). The optical densities recorded by the photobioreactor revealed that population dynamics are highly reproducible during consecutive diel cycles (Fig. 1). OD 735 is a measure of light scattering due to particulate material. It increases when either cell density increases and/or increases in sizes increase, or cells accumulates intracellular storage products such as granules. The non-linear relationships between OD 735 and cell abundance have been shown previously, both in Crocosphaera and Cyanothece [26,34]. The daily changes in cell size as the cells undergo division are also small (<10 %) in both species. Therefore, the diurnal increase in OD 735 reflects mostly biomass production, specifically the increase in cellular C content as a result of photosynthesis, whereas the nocturnal decrease in OD 735 is caused by the consumption of cellular C as a consequence of respiration [26,34]. Since ODs can be influenced by various factors, absolute OD 735 of two different species may not reflect the same C content. Therefore, to eliminate these differences, we normalized OD 735 measurements to the corresponding values recorded at 1L. We also recognize that there are other intracellular storage compounds such as cyanophycin, a major nitrogen storage compound, and/or phosphorus granules. However, their content is generally much smaller than that of storage carbohydrates [19,20,31] and can be neglected. Under exponential growth conditions, we observed much more dynamic diel changes in OD 735 in Crocosphaera ($3 lm) compared to Cyanothece ($3 lm). High variability in OD 735 in Crocosphaera may reflect the larger cellular size and/or more peripheral allocation of carbohydrate within these cells [19]. Therefore, we define the rate of changes in OD 735 as a proxy of C incorporation and consumption; photosynthesis and respiration.
Our experiments were designed to compare the photosynthetic and N 2 fixation capacity of the two studied strains under optimal growth conditions, i.e. during exponential growth, without any nutrient or light limitation. Under such conditions, the diel pattern in photosynthetic activities and N 2 fixation in Crocosphaera and Cyanothece showed similarities, but also pronounced differences. The latter of which were as follows: firstly, Crocosphaera maintained high O 2 evolving capacity even at the very end of the light phase, whereas Cyanothece reduced O 2 evolving capacity in parallel with the decreasing light intensity after 8L (Fig. 2). Secondly, Crocosphaera lost the capacity for photosynthetic O 2 evolution when actively fixed N 2 , whereas Cyanothece still retained its capacity to evolve O 2 in the middle of the dark phase (Fig. 2). This uncoupling of the photosynthetic capacity from N 2 fixation in the dark in Crocosphaera is well documented [26,33] and can be explained by inactivation of PSII complexes, their monomerization and disassembly, most probably related to the decreased protein synthesis [26]. The nocturnal decline of PSII activity has also been reported for Cyanothece [38]. However, analysis of the membrane protein complexes using CN PAGE (Fig. 4) revealed that the monomerization and disassembly was not as significant in that microorganism as what was observed for Crocosphaera [26]. As observed earlier [38], non-functional rD1 protein was accumulated specifically during dark phase, similar to that observed in Crocosphaera [26].
Finally, the timing of N 2 fixation also differed in the two cyanobacteria. Cyanothece started to fix N 2 shortly after the start of the dark period, whereas Crocosphaera started to fix N 2 only after 4 h of darkness. We assume that the swift shift from photosynthesis to N 2 fixation in Cyanothece is supported by the observed rela-   tively high respiration rates, which was comparable in the dark and light phases (Fig. 2B, Table 1). In contrast, the respiration rates in Crocosphaera during the light phase was less than half of its dark rates (Table 1). Besides, the rate of respiration compared to gross O 2 evolution during the light phase was higher in Cyanothece (36 %), compared to that of Crocosphaera (25 %). These results suggest that a balance between C incorporation and consumption may be the main reason of the observed smaller diel OD dynamics in Cyanothece (Fig. 1). Interestingly, the duration of active N 2 fixation was about 6 h in both species, and active N 2 fixation stopped about 2 h before the start of the daylight phase.
The lower ETR max coupled with higher P m B in Crocosphaera suggests that this species is more efficient in C incorporation under saturated irradiance compared to Cyanothece, as shown by low U max . However, under light-limiting conditions, the electron requirement was identical for both strains (Fig. 3), which suggests that Crocosphaera that contains less thylakoid membranes as well as chlorophyll per cell captures electrons using larger lightharvesting antenna compared to Cyanothece. On the other hand, assuming that 8 electrons are necessary to fix one N 2 [13] in both Crocosphaera and Cyanothece, ETR at given irradiance suggested that 9.1 %, and 9.3 % of the transported electrons are devoted to and PSI (1), trimeric and monomeric PSI complexes, respectively; PSII (2) and PSII (1), dimeric and monomeric core PSII complexes, respectively; RC47, PSII complex lacking CP43, u.CP43 and u.CP47, unassembled CP43 and CP47; F.P., free pigments. 5 lg of Chl were loaded for each sample. (C) Membranes were analyzed by denaturing SDS-PAGE, gel was electroblotted to PVDF membrane and the membrane was probed with antibodies specific for the standard D1 (sD1) and rD1. 2 lg of Chl were loaded for each sample. N 2 fixation, respectively. Thus, the efficiency of electron utilization in nitrogen fixation is comparable in these species.

Simulated competition between Crocosphaera and
Cyanothece in a simple ecosystem model The above described results showed that Cyanothece exhibits higher photosynthesis as well as dark respiration rates when compared with Crocosphaera. The question arises whether these differences in C-related metabolisms could explain their ecological success in different niches. To address this question, we developed a simple metabolic model for phytoplankton (CFM-CC (Fig. 5A), see also Materials and Methods). The model is based on a simple equation but has an aspect of a coarse-grained model [18], resolving key metabolic pathways including C fixation, respiration and growth. The residual C after C fixation and respiration is converted to biomass of new cells via cell division (growth). In general, in an interspecies competition, when two species use the same resources, the faster growing strain outcompetes the other one.
We calculated growth rates of Crocosphaera and Cyanothece based on the CFM-CC model and plotted as a function of nutrient repletion factor in Fig. 5B. Importantly, this factor indicates how  nutrient limitation influences the rate of C fixation. The higher this factor the smaller the nutrient limitation. For example, at nutrient repletion factor of 0.3 and 0.7, 30 % and 70 % of maximum photosynthesis occurs due to relatively high and low nutrient limitations, respectively. Under nutrient-replete conditions as are frequently found in its coastal environment (characterized with high nutrient repletion factor), Cyanothece has higher growth rates indicating an ecological advantage under such conditions, which allow the organism to take advantage of their inherently high rate of photosynthesis and an overall higher maximum photosynthesis rate (Table 1). In contrast, under nutrient limiting conditions, Crocosphaera predominates as indicated by higher growth rates under low nutrient repletion factors therefore enabling it to occupy the niche of low-nutrient open ocean waters due to its low respiration rates coupled with a highly efficient metabolism. These conclusions are summarized in Fig. 6. Despite the simple parameterizations, our results clearly show a niche separation of these microorganisms, underpinning the significance of differences in C metabolisms in shaping of ecological niches.
In this study, we focused on the growth rate, but factors affecting the mortality rate may affect the organismal competition. For example, if the grazing patterns for these organisms are different, it may affect the competition. However, such differentiated grazing for these organisms have not been reported under the same condition. If the size of these organisms is the same, under the same environmental conditions, grazing rates for these organisms are likely to be similar. Given the identical grazing/mortality rates, the competition is mainly governed by the growth rate. At the same time, uncertainties in grazing rates are large; the grazing rates vary from nearly zero to as high as 0.7 d À1 for Crocosphaera [10,43], whereas grazing rates for Cyanothece seem to be slightly more stable (0.18-0.58 d À1 ) [9]. Pinning down the effect of grazing will ultimately require additional experiments for these two organisms and their potential grazers under a set of identical growth conditions.

Conclusions
Overall, our integrated study of laboratory scale measurements showed that highly reproducible diel changes in OD 735 is a proxy for population metabolic dynamics. The observed diel changes were much more dynamic in Crocosphaera compared to Cyanothece. This dynamic change is possibly a consequence of the strict temporal segregation of photosynthesis and respiration in Crocosphaera. A simple ecosystem model with two competing species suggested that differences in C incorporation and consumption may lead to different niche acquisition: High C fixing capability enables Cyanothece to grow actively in coastal waters, and low C loss enables Crocosphaera to survive in oligotrophic water.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.