Light-limited growth rate modulates nitrate inhibition of dinitrogen fixation in the marine unicellular cyanobacterium Crocosphaera watsonii.

Biological N2 fixation is the dominant supply of new nitrogen (N) to the oceans, but is often inhibited in the presence of fixed N sources such as nitrate (NO3-). Anthropogenic fixed N inputs to the ocean are increasing, but their effect on marine N2 fixation is uncertain. Thus, global estimates of new oceanic N depend on a fundamental understanding of factors that modulate N source preferences by N2-fixing cyanobacteria. We examined the unicellular diazotroph Crocosphaera watsonii (strain WH0003) to determine how the light-limited growth rate influences the inhibitory effects of fixed N on N2 fixation. When growth (µ) was limited by low light (µ = 0.23 d-1), short-term experiments indicated that 0.4 µM NH4+ reduced N2-fixation by ∼90% relative to controls without added NH4+. In fast-growing, high-light-acclimated cultures (µ = 0.68 d-1), 2.0 µM NH4+ was needed to achieve the same effect. In long-term exposures to NO3-, inhibition of N2 fixation also varied with growth rate. In high-light-acclimated, fast-growing cultures, NO3- did not inhibit N2-fixation rates in comparison with cultures growing on N2 alone. Instead NO3- supported even faster growth, indicating that the cellular assimilation rate of N2 alone (i.e. dinitrogen reduction) could not support the light-specific maximum growth rate of Crocosphaera. When growth was severely light-limited, NO3- did not support faster growth rates but instead inhibited N2-fixation rates by 55% relative to controls. These data rest on the basic tenet that light energy is the driver of photoautotrophic growth while various nutrient substrates serve as supports. Our findings provide a novel conceptual framework to examine interactions between N source preferences and predict degrees of inhibition of N2 fixation by fixed N sources based on the growth rate as controlled by light.

Introduction

Understanding the global N pan> class="Chemical">cycle is critical to ocean biogeochemical models, as nitrogen is arguably the single most limiting nutrient for oceanic primary production. A major current challenge is to determine how N biogeochemistry will change as we transition from the Holocene to the Anthropocene [1]. Nitrogen fixation is one of the key pathways predicted to change as the surface ocean becomes warmer and more acidified [2], [3], [4], [5], [6], [7] and as progressive anthropogenic eutrophication increases fixed N loading in many marine ecosystems [8], [9].

Modeled estimates of N inpan>put from marinpan>e biologipan> class="Chemical">cal N2 fixation are dependent on concentrations of other chemical species of fixed N such as nitrate (NO3 −) [10], [11]. This is largely because fixed N has been shown in past studies to have relatively strong “inhibitory” effects on N2-fixation by the ubiquitous oceanic diazotroph Trichodesmium [12], [13], [14], [15], most likely due to differences in the energetic costs involved in assimilating different N species such as NO3 − and N2 [16]. Several recent laboratory studies, however, have suggested that N2 fixation by unicellular diazotrophs such as Crocosphaera watsonii may not be as strongly inhibited by NO3 − as has been previously suggested for Trichodesmium [14], [15], [17].

While this major physiological difpan> class="Chemical">ference may relate to differences in N2-fixation strategies (Trichodesmium fixes N2 during the day; Crocosphaera fixes N2 during the night, similar to unicellular organismal physiology described by Berman-Frank et al. [18]), these recent findings imply that the ratios of N-assimilation kinetic parameters for different N sources (e.g. Vmax,N2:Vmax,NO3−) may be very different between Trichodesmium and Crocosphaera. In addition to these laboratory-based results, field studies indicate that N2-fixation rates by unicellular diazotrophs increase with decreasing depth and increasing light in upwelling water where NO3 − concentrations are high [19], [20]. Trichodesmium blooms are also frequently observed in upwelling regions that are known to have high NO3 − concentrations [21]. Lastly, Deutsch et al. [22] presented a model proposing that N2-fixation rates might be very high in the Peru upwelling system, based on the distribution of phosphorus, despite high concentrations of NO3 − in this region. The general picture of how fixed N sources such as NO3 − control N2 fixation is still unclear.

In the context of these repan> class="Chemical">cent laboratory, field and modeling studies, we asked how the growth rate, as controlled by light, influences preferences for nitrogen substrates (e.g. NH4 +, NO3 − and N2) to support growth of the unicellular N2 fixer Crocosphaera watsonii. Our data indicate that the N-source utilization ratio (NO3 −:N2) changes in a predictable manner as a function of cell growth. We present experiments suggesting that three key parameters are necessary to determine how fixed N controls N2-fixation rates by Crocosphaera watsonii: 1) the cellular demand for N, which is largely controlled by the growth rate, 2) the light-specific cellular-assimilation kinetics of the various forms of N (e.g. Vmax) and 3) the relative concentrations of the various forms of N. Our basic model relies on the tenet that light energy is the driver of photoautotrophic growth rates while substrates such as NO3 −, N2, PO4 3− etc. do not drive growth but serve as nutrient supports. Thus, a gradient in the light-energy supply rate creates a gradient in the demand for nitrogen to support growth and a gradient in the ratio of nutrient assimilation rates of various nutrient substrates. Our conceptual model may serve as a framework to understand how fixed N availability controls N2 fixation by oceanic diazotrophs. In light of expected future increases in anthropogenic fixed N inputs to both the coastal and open ocean [23], [24], these studies are needed to improve both physiological models and biogeochemical estimates of global biological N2 fixation and overall predictions of primary production trends over the next century [10], [25], [26].

Materials and Methods

We investigated short-term and long-term effepan> class="Chemical">cts of fixed N on N2-fixation rates by C. watsonii cultures (strain WH0003) in which growth rates were controlled by different light levels. In preparation for both short- and long-term experiments, C. watsonii was pre-acclimated to light environments by growing cultures in triplicate 1-L polycarbonate bottles at 25 and 175 µmol quanta m−2 s−1 and 28°C, on a 12∶12 hour light:dark cycle for 5 or more generations (as in other laboratory culture experiments; Berman-Frank et al. [18]) with an artificial seawater medium prepared according to the YBCII recipe of Chen et al. [27]. Trace metals (FeCl3·6H2O 4.50×10−7 M, MnCl2·4H2O 1.21×10−7 M, NaMoO4·2H2O 1.00×10−7 M, ZnSO4·7H2O 7.97×10−8 M, CoCl2·6H2O 5.03×10−8 M) and vitamins (Thiamine 2.96×10−7 M, B12 3.96×10−10 M, Biotin 2.50×10−9 M) were added with the dilution medium [28] with 4 µM phosphate added as HNa2PO4. Cultures were grown with a semi-continuous culturing method as in other studies [14], [18], [29], [30], [31], [32] by diluting cultures every 3 days. Cultures were diluted by enumerating cells and calculating a dilution factor to achieve a target culture cell density of 20×103 cells mL−1. We determined culture cell densities by agitating cultures just prior to collecting 5 ml of culture and enumerating live cells from subsamples microscopically. Although we did not continuously stir cultures, we did not observe cells or biomass sticking to the sides of the bottles. We calculated growth rates (µ) in between 3-day dilution periods with NT = N0e T, where N0 is the cell density at the beginning of a 3-day period (T) and NT is the cell density at the end of the period.

Short-term exposures

Initially, we exposed Crocosphaera to ranpan>ge of pan> class="Chemical">NH4 + concentrations for a short amount of time to gather basic information about how fixed N inhibits N2 fixation as a function of light-limited growth. We selected NH4 + because it has a high maximum uptake rate (Vmax) relative to other sources of fixed N in Trichodesmium [29]. Once we had collected data using NH4 + as an inhibitor, we repeated the short-term experimental design using NO3 − as the inhibitor. In short-term exposures, 50 mL samples were collected in 80 mL vials from each replicate culture and exposed to a range of NH4 + concentrations (0.2–2.0 µM, added as NH4Cl) and NO3 − (0.5–40 µM, added as NaNO3 −; n = 3 for each treatment concentration of NH4 + or NO3 −) just before the beginning of the dark period, approximately 3 hours before measurable ethylene concentrations accumulated. Replicates without added NH4 + or NO3 − served as controls. We estimated N2-fixation rates by injecting 4 mL acetylene into 30 mL headspace of the sample vials and measuring ethylene accumulation in 200 µl of the headspace over the 12-hour dark period with a gas chromatograph (model: GC-8A, Shimadzu Scientific Instruments, Columbia, MD, USA) [5], [6]. We used a 4∶1 ratio of N2:acetylene reduction to estimate N2-fixation rates [33]. Background ethylene concentrations in the acetylene source were small and subtracted from ethylene accumulation measurements. From each culture replicate, 100 mL were filtered onto combusted GF/F filters (500°C, 5 h), dried at 80°C, compressed into pellets and analyzed with an elemental analyzer (Costech instruments, model 4010) [5], [6]. The concentrations of particulate organic N were similar between cultures at the initiation of the short-term experiment (PNlowlight = 4.3±0.6 µmoles N L−1; PNhighlight = 5.5±0.7 µmoles N L−1).

Long-term exposures

Based on results from our initial short-term experiment with NO3 −, we depan> class="Chemical">cided to expose Crocosphaera to NO3 − for a longer time period to determine if long-term exposures elicited a different response relative to that in the short-term exposure. In long-term exposures to NO3 −, C. watsonii was pre-acclimated to experimental conditions in semi-continuous cultures using NO3 − as a fixed N source (added as 30 µM NaNO3), in parallel with control cultures growing without an added fixed N source. Particulate organic N of cultures was maintained at similar concentrations by semi-continuous dilution between the control (PNlowlight = 6.6±3.3 µmoles N L−1; PNhighlight = 7.0±0.8 µmoles N L−1) and added NO3 − treatments (PNlowlight = 6.7±0.9 µmoles N L−1; PNhighlight = 7.9±0.5 µmoles N L−1). We measured N2-fixation rates in 50 mL samples from each culture replicate with the acetylene reduction assay as described above at three experimental time points (Table 1). For estimates of NO3 − concentrations, we passed 20 mL of culture through a 0.45 µm syringe filter and NO3 − was measured by the analytical laboratory at the Marine Science Institute, University of California, Santa Barbara, CA, USA. We collected samples to measure the concentration of NO3 − from culture replicates 18 h after the last dilution of cultures (initial measurement) and either 48 h (high-light treatment) or 96 h (low-light treatment) after the initial measurement. To estimate cellular NO3 −-assimilation rates, we normalized diminishing NO3 − concentrations during this time to culture cell concentrations that were calculated at the mid-point between these two time points using the growth rate. We did not examine a long-term response to NH4 + exposure primarily because it generally represents a small portion of fixed N relative to concentrations of NO3 − in many natural oceanic waters.

Table 1

Measurements of culture cell density (cells L−1×106), dissolved nitrate+nitrite concentrations (NO3 − + NO2 −, µmol L−1) and N2-fixation rates (fmol cell−1 hr−1) at different time points (hours since culture dilution) in cultures used in the short- and long-term exposure experiments.

Light intensity				Cells	[NO3 −+NO2 −]	N2
25 µmol quanta m−2 s−1
	Short-term			38.4±4.8	0.15±0.05
	Long-term
		N2 only
			T18h	25.8±0.9	0.13±0.07	13.6±8.1
			T66h	37.9±3.1	n.d.	65.2±4.2
			T114h	64.8±2.1	0.06±0.05	55.1±1.8
		+NO3 − (30 µM)
			T18h	23.4±2.2	27.6±0.00	10.9±6.3
			T66h	36.4±2.9	25.8*±0.3	25.5±5.1
			T114h	59.7±5.9	23±0.7	23.8±1.8
175 µmol quanta m−2 s−1
	Short-term			31±3.3	0.16±0.05
	Long-term
		N2 only
			T18h	29.9±1.6	0.29±0.03	114.6±1.9
			T42h	60.4±1.3	n.d.	135.4±1.2
			T66h	117±12.1	0.05±0.01	115.4±12.5
		+NO3 − (30 µM)
			T18h	33.8±1.3	28.7±0.5	105.5±1.5
			T42h	76.5±6.5	24.8*±0.3	127.7±5.6
			T66h	191.4±13.6	16.7±1.4	103.1±1.0

*calculated NO3 − concentrations.

Error (±) represents the standard deviation on 3 culture replicates.

Error (±) represents the standard deviation on 3 n class="Chemical">culture replin class="Chemical">cates.

Results

We observed large differenpan> class="Chemical">ces in growth rates of C. watsonii between light treatments. In control cultures growing on N2 only, growth was significantly lower in low-light acclimated cultures (25 µmol quanta m−2 s−1; 0.23±0.02 d−1) relative to cultures growing under higher light (175 µmol quanta m−2 s−1, 0.68±0.03 d−1; t-test, p<0.05). The controlling effects of NH4 + and NO3 − on N2 fixation were different in short-term exposures, but varied as a function of growth rate. In addition, the effect of NO3 − on N2 fixation was similar between short and long-term exposures.

In slow-growing cultures apan> class="Chemical">cclimated to low light, short-term additions of 0.4 µM NH4 + inhibited N2-fixation rates to <10% of rates in control treatments without added NH4 + (Fig. 1a). In faster-growing cultures acclimated to 175 µmol quanta m−2 s−1, with biomass concentrations equivalent to those in low-light cultures (Table 1), short-term exposure to five times as much NH4 + (2.0 µM) was needed to achieve the same inhibitory effect on N2 fixation (Fig. 1a). The short-term inhibitory effects of NO3 − on N2 fixation also varied as a function of growth rate. In slow-growing, low-light acclimated cultures, short-term exposure to NO3 − reduced mean N2-fixation rates by ∼47–62% relative to rates in control treatments without added NO3 − (Fig. 1b). In fast-growing cultures acclimated to high light, however, short-term additions of NO3 − at any concentration up to 40 µM did not inhibit mean N2-fixation rates by more than 9%, relative to N2-fixation rates in control cultures without added NO3 − (Fig. 1b).

Figure 1

Short-term inhibitory effects of ammonium (NH4 +, 0–1.5 µmol L−1) and nitrate (NO3 −, 0–40 µmol L−1) on N2 fixation by Crocosphaera watsonii (WH0003) (percent of control with no added nitrogen).

Cultures were grown in steady state under high light (175 µmol quanta m−2 s−1, growth rate (µ) = 0.68 d−1, open symbols) and low light (25 µmol quanta m−2 s−1, µ = 0.23 d−1, closed symbols) before adding nitrogen. Error bars represent standard deviations on means from 3 culture replicates.

Cultures were grown inpan> steady state under high light (175 µmol quanta m−2 s−1, growth rate (µ) = 0.68 d−1, open symbols) and low light (25 µmol quanta m−2 s−1, µ = 0.23 d−1, closed symbols) before adding nitrogen. Error bars represent standard deviations on means from 3 culture replicates.

In high-light-acpan> class="Chemical">climated cultures, long-term exposure to 30 µM NO3 − yielded significantly higher growth rates (µ = 0.87 d−1) than those in control cultures without added NO3 − (µ = 0.68 d−1; p<0.05), indicating that growth was limited by the N2-assimilation rate (Fig. 2a). Diminishing NO3 − concentrations over time suggested that NO3 −-assimilation rates in fast-growing cultures (µ = 0.87 d−1) were 2.8 times higher than those in slow-growing cultures (µ = 0.23 d−1; Fig. 3a; p<0.05), but the contribution of NO3 − to the total daily N assimilation still varied as a function of growth rate. In high-light-acclimated cultures exposed to NO3 − (µ = 0.87 d−1), NO3 − assimilation represented 40% of the total daily N assimilation while N2 assimilation represented 60% (Fig. 2b). When combined, NO3 − and N2 assimilation yielded a higher total daily N-assimilation rate (187 fmol N cell−1 d−1) than that in the control treatment growing on N2 only (122 fmol N cell−1 d−1; p<0.05; Fig. 2b). Furthermore, N2-fixation rates in cultures with added NO3 − were not significantly different than those in control cultures without NO3 − (p<0.05; Fig. 2b).

Figure 2

Effects of long-term exposure to 30 µmol L−1 nitrate (NO3 −).

(a) cellular growth rates and (b) nitrogen-assimilation rates of Crocosphaera watsonii (WH0003) acclimated to high light (175 µmol quanta m−2 s−1) and low light (25 µmol quanta m−2 s−1). (b) N2-fixation rates (solid bars) are overlain on total N assimilation (N2 + NO3 − assimilation, hashed bars). Control cultures did not receive added NO3 −. Error bars represent standard deviations on means from 3 culture replicates.

Figure 3

Changes in nitrate (triangles) and cell concentrations (circles) in cultures of Crocosphaera watsonii (WH0003) in long-term exposure experiments.

Cultures were grown in steady state under (a) high light (175 µmol quanta m−2 s−1) and (b) low light (25 µmol quanta m−2 s−1) with added nitrate (30 µmol L−1; open symbols) or with N2 only (closed symbols). *Calculated NO3 − concentrations (see Methods section for details). Error bars represent standard deviations on means from 3 culture replicates.

Cultures were grown inpan> steady state under (a) high light (175 µmol quanpan>ta m−2 s−1) anpan>d (b) low light (25 µmol quanpan>ta m−2 s−1) with added pan> class="Chemical">nitrate (30 µmol L−1; open symbols) or with N2 only (closed symbols). *Calculated NO3 − concentrations (see Methods section for details). Error bars represent standard deviations on means from 3 culture replicates.

Under low light, long-term exposure to 30 µM NO3 − did not support faster growth rates (Fig. 2a, 3b) even though pan> class="Chemical">NO3 −-uptake supported 61% of the total daily N assimilation. Instead, N2-fixation rates were reduced by 55% relative to those in cultures without added NO3 − (p<0.05; Fig. 2a). Thus, in cultures that were grown with NO3 −, there was a clear shift in the ratio of N source utilization where growth-specific NO3 −-assimilation rates increased by 55% with decreasing light,while growth-specific N2-assimilation rates increased by 46% with increasing light (Fig. 4). In both the high- and low-light treatments with 30 µM NO3 − added, the concentration of NO3 − was high (>16 µmol NO3 − L−1) throughout the entire 66 h or 114 h sampling period (Fig. 3).

Figure 4

Growth-specific assimilation rates of nitrate (NO3 −; open bars) and dinitrogen (N2; closed bars) in cultures of C. watsonii (WH0003) with added NO3 − (30 µmol L−1).

Growth-specific NO3 − and N2-assimilation rates change inversely relative to each other as a function of light-limited growth. Error bars represent standard deviations on means from 3 culture replicates.

Growth-spen class="Chemical">cifin class="Chemical">c NO3 − and N2-assimilation rates change inversely relative to each other as a function of light-limited growth. Error bars represent standard deviations on means from 3 culture replicates.

Discussion

Our main finding is that N-sourpan> class="Chemical">ce utilization by C. watsonii varied as a function of the growth rate, which we controlled in our experiments with the supply of light energy. Thus, we interpret the variation in N-source utilization (e.g. NO3 −:N2 or NH4 +:N2) to be caused by a gradient in the demand for nitrogen as a substrate to support cell division. This N-source utilization ratio seems to change as a function of energy supply and growth rate because of differences in uptake kinetics between N sources (e.g. VmaxNO3−: VmaxN2) and energy requirements for the reduction and assimilation of each N source.

In our short-term exposure experiment with NH4 +, fast-growinpan>g pan> class="Chemical">cultures of C. watsonii (µ = 0.68 d−1) needed a much higher concentration of NH4 + (5x) to satisfy the nitrogen substrate demand relative to slow-growing cultures (µ = 0.23 d−1; Fig. 1a). An alternate way to view this relationship is that as the amount of the added NH4 + decreased, increasing amounts of N2 were fixed to satisfy the remaining nitrogen demand to support cell growth (Fig. 1). These results suggest that the magnitude of assimilation of various N substrates depends on the cellular N demand that is needed to support the light-controlled growth rate relative to the light-specific cellular assimilation rate kinetics of each N source. Thus, we propose that when the light-controlled, growth-modulated demand for N exceeds the cellular-assimilation rate of NH4 + or NO3 −, N2 fixation provides fixed N to fill the resulting N deficit.

The variable controllinpan>g efpan> class="Chemical">fects of NO3 − and NH4 + on N2 fixation suggest that there are large differences in the assimilation kinetics of these different N species (Fig. 1). Under low light, low concentrations of NO3 − and NH4 + (0.5 µM) had maximum inhibitory effects on N2 fixation, suggesting that the half-saturation constants (Ks) with respect to NO3 − and NH4 + are similar for C. watsonii. The incomplete inhibitory effect of NO3 − on N2-fixation rates even at high concentrations of NO3 − (Fig. 1b, 3b), however, suggests that the maximum NO3 −-assimilation rate (Vmax) by C. watsonii (WH0003) is low relative to that of NH4 +.

In our long-term experiment, we pre-acpan> class="Chemical">climated Crocosphaera with high NO3 − concentrations (∼15–30 µM, Fig. 3) for 5 or more generations before sampling cultures over a 48–96 h period. In these long-term exposures to NO3 −, we measured residual NO3 −-concentrations in the culture medium to estimate the cellular NO3 −-assimilation rate. The ratio of NO3 − -assimilation:N2 fixation varied as a function of energy supply and growth (Fig. 2), further supporting these variables as controls of fixed N inhibition of N2 fixation. Exposure to NO3 − did not affect N2 fixation by fast-growing cultures of C. watsonii, yet NO3 − comprised 40% of the total daily N, thereby supporting growth rates that were 27% higher than those in control cultures without added NO3 − (Fig. 2b). Thus, the growth of high-light cultures of C. watsonii, similar to Cyanothece, another marine unicellular N2 fixer [34], was clearly limited by the N2-assimilation rate, as the addition of 30 µM NO3 − supported higher growth rates (Fig. 2a).

These results indicate that growth rates of pan> class="Species">C. watsonii benefits from assimilating multiple N sources simultaneously, as individual assimilation rates of N2 or NO3 − alone cannot support maximum growth rates in high-light environments. Under low light, NO3 −-assimilation did not support faster growth as it did under high light, but instead comprised 61% of the total daily assimilated N (Fig. 2). This higher contribution of NO3 − to the total N demand inhibited N2 fixation by 55% relative to rates in control cultures without added NO3 −. Thus, we conclude that the inhibitory effect of NO3 − on N2 fixation by C. watsonii varies as a function of energy supply and growth rate.

Although we did not separate the direct efpan> class="Chemical">fect of light-energy supply and growth rate in our long-term experiment, our analyses of the short-term effects of NH4 + and NO3 − exposure on N2 fixation were done only during dark hours when Crocosphaera fixes N2. Thus, Crocosphaera offers a unique advantage in comparison with Trichodesmium (which fixes CO2 and N2 simultaneously in the light) because it is possible to separate direct effects of light-energy supply from the effects of the light-limited growth rate on N-source utilization preferences. Future experiments might consider experiments that separate these effects by modulating growth rates in other ways.

The assimilation rates of the various chemipan> class="Chemical">cal forms of N (e.g. NH4 +, NO3 −, N2) seem to be dictated in part by the energetic cost of reduction [16]. Many phytoplankton species are known to assimilate NH4 + more easily than NO3 − because of the lower energetic investment associated with assimilating NH4 + [35]. Although N-uptake kinetics have not been described for C. watsonii, Mulholland et al. [29] documented a maximum uptake rate for NH4 + by Trichodesmium that was presumably more than an order of magnitude higher than that for NO3 −. Based on the relatively weak inhibitory effect of NO3 − on N2 fixation by C. watsonii relative to that observed for NH4 + (Fig. 1, 3), we infer that the maximum assimilation rate of NO3 − by C. watsonii (Vmax) must be considerably lower than that of NH4 +.

Although NH4 + assimilation pan> class="Chemical">carries a cost associated with transport across the cell membrane, it is generally thought to be less expensive to assimilate than NO3 − and N2 [36], [37] because of the high costs associated NO3 − and N2 assimilation, which must first be reduced to NH4 + before being assimilated onto glutamic acid (ΔG  = +69 Kcal mol N−1 for NO3 − and +87 Kcal mol N−1 for N2) [16]. A lower assimilation cost for NH4 + might afford a high Vmax relative to that for more energetically expensive forms of nitrogen. Thus, the lower cost associated with NO3 − reduction to NH4 + relative to N2 reduction to NH4 + appears to benefit C. watsonii in a light-limited environment where growth is slow relative to a maximum NO3 −-assimilation rate (Fig. 4). In a high-light environment, the maximum assimilation rate of NO3 − relative to the growth rate is reduced in comparison with that in low-light cultures (Fig. 4), where N2 supports a higher portion of the daily N demand for growth. Future studies should quantify NO3 −-assimilation kinetics for N2 fixers and identify how they might change as a function of other environmental conditions.

In addition to the energetic pan> class="Chemical">costs for reducing NO3 − and N2, the difference between energetic and material investments associated with the production of assimilatory proteins such as nitrogenase and nitrate reductase may be at least partially responsible for the differential ratios of NO3 −:N2 reduction as function of growth. Tradeoffs in energetic investments for NO3 − and N2 reduction may come from balancing differential cellular nitrogen demands that are associated with variable growth rates [38] or from the supply of light. Further separating the effect of light-energy supply from the effect of growth on the ratio of fixed N:N2 utilization may lead to a better understanding of the release of fixed N by diazotrophs [5], [33], [39], [40], [41].

Contrary to finpan>dinpan>gs by Ohki et al. [12] that suggest a strong time dependenpan> class="Chemical">ce of exposure to NO3 −, NH4 + and urea in controlling inhibitory effects on N2 fixation in Trichodesmium, we documented consistent inhibitory effects of NO3 − on N2 fixation of Crocosphaera regardless of the duration of exposure. The results presented by Ohki et al. [12] are difficult to interpret in a context of supply and demand for N, however, because growth rates between treatments were not defined.

Although previous studies have not discussed inpan>hibitory efpan> class="Chemical">fects of fixed N on N2 fixation in a context of the supply rate of fixed N relative to the growth-modulated demand for N, four relatively recent studies have collectively examined inhibitory effects of fixed N on N2 fixation in batch cultures of Crocosphaera and/or Trichodesmium growing under 30–40, 80, 128 and 180 µmol quanta m−2 s−1, all at 26 or 27°C [14], [15], [17], [42]. In batch cultures, the biomass concentration of the culture is important to consider because of the accelerating effect of increasing biomass on the rate of disappearance of NO3 − or NH4 +. Interpretation of these studies in a context of the supply rate of fixed N relative to the growth-modulated demand for N is also difficult, mainly because biomass and/or growth rates between treatments were not defined during batch-mode growth.

In our experiments, we maintained constanpan>t exponential growth rates with a semi-pan> class="Chemical">continuous culturing method and we maintained equivalent biomass concentrations between treatments so that differences in NH4 + and NO3 − drawdown due to biomass differences would not affect cellular N2-fixation rates between treatments and between time points (Fig. 3; Table 1). In addition to our experiments with Crocosphaera, all of these previous studies indicate that NO3 − and/or NH4 + have controlling effects on N2 fixation by oceanic N2 fixers. Future studies that examine N-source preferences should focus on growth-modulated controls of fixed N on N2 fixation in both Trichodesmium and Crocosphaera. Although we presume that this model would be similar for Trichodesmium, there may be unforeseeable differences due to the major differences between the physiological mechanisms that these species use to separate oxygen generated by photosynthesis from the nitrogenase enzyme; Trichodesmium seems to use a spatial separation mechanism, as it fixes both inorganic carbon and N2 during the light period. In contrast, Crocosphaera uses a temporal separation mechanism, as it stores fixed carbon during the light period and respires it for energy during the night to fuel N2 fixation in the dark, similar to the unicellular strategy described by Berman-Frank et al. [18].

In the open oceanpan>, the primary limitinpan>g nutrients for growth of pan> class="Chemical">N2-fixing cyanobacteria are iron (Fe) and phosphorus (P) [43], [44]. In combination with light, Fe and P have an indirect effect on N demand through their support of cellular growth. Capone and Knapp [8] originally proposed that the N:P ratio is important in controlling N2-fixation rates, and recently Ward et al. [11] suggested that the N:Fe ratio is a dominant controlling factor of marine N2 fixation. Our basic model suggests that the ratio of N:X is important in controlling N2-fixation rates where “X” is a resource that influences growth rates (such as light, P and Fe), and thereby, the demand for N. Laboratory data support this, where high concentrations of P supported high N2-fixation rates relative to cultures with lower P concentrations, despite equivalent N:P supply ratios [15]. In a modeling study, Ward et al. [11] demonstrated that the N:P supply ratio is a secondary factor in defining boundaries of N2 fixation, while the N:Fe supply ratio is more important in an ecological context through competitive interactions with non-N2-fixing phytoplankton. Further, Garcia et al. [32] suggest that the Fe:P supply ratio may be more important in controlling N2 fixation than the absolute concentration of either of these limiting nutrients. Collectively, these studies suggest that links between C, N, P and Fe biogeochemical cycles depend on the relative supply of each of these nutrients and our study further suggests that the energy-supply rate or the growth rate modulates interactions between these nutrients.

Our study indicates that global models of marinpan>e biologipan> class="Chemical">cal N2 fixation should consider an interaction between assimilation kinetics of fixed N and a growth-modulated demand for N. Although our study did not focus on how Crocosphaera might respond in the natural environment, our data provide a framework around which future studies might structure investigations of N-source preferences by natural communities of N2 fixers. Reactive nitrogen from atmospheric sources and agricultural runoff are expected to increase in the future and the effects of increased N input to the oceans on phytoplankton communities is uncertain [23], [24], [45]. Thus, a clear understanding of how reactive nitrogen affects N2 fixation is needed to support predictions of how phytoplankton communities will change.

Two other relevant environmenpan>tal fapan> class="Chemical">ctors that will certainly influence growth of N2 fixers in the future are CO2 and temperature [4], [5], [6], [7], [30], [34]. Both of these factors are predicted to increase, and will likely influence the controlling effects of fixed N on N2 fixation through their effects on growth rates. Thus, our basic framework potentially has far-reaching implications for both current estimates of oceanic N2 fixation, and for estimates of N2-fixation rates that are likely to exist in the future surface oceans [3].