Various factors affecting N2 fixation of a cultured strain of Trichodesmium sp. (GBRTRLI101) from the Great Barrier Reef Lagoon were investigated. The diurnal pattern of N2 fixation demonstrated that it was primarily light-induced although fixation continued to occur for at least 1 h in the dark in samples that had been actively fixing N2. N2 fixation was dependent on the light intensity and stimulated more by white light when compared with blue, green, yellow and red light whereas rates of N2 fixation decreased most under red light. Inorganic phosphorous concentrations in the lower range of treatments up to 1.2 μM significantly stimulated N2 fixation and further additions promoted little or no increase in N2 fixation. Organic phosphorous (Na-glycerophosphate) also stimulated N2 fixation rates. Added combined nitrogen (NH4+, NO3−, urea) of 10 μM did not inhibit N2 fixation in short-term studies (first generation), however it was depressed in the long-term studies (fifth generation).
Nitrogen (N) is generally regarded as the most common nutrient limiting primary production for phytoplankton in much of the ocean [1–3], although large areas of the ocean have been shown to be iron (Fe) limited . The large reserves of N in the deep ocean have been considered to be the main external source supplying the inorganic nitrogen needs of primary production in the surface ocean. However recent studies suggest that N2 fixation by Trichodesmium spp. and other diazotrophs play a major role in the supply of N to the oceans . Research on N2 fixation of Trichodesmium in the open ocean had its beginnings in 1961 when Dugdale and co-workers suggested that Trichodesmium could fix N2. To date, the bulk of research on N2 fixation in oceanic systems has focused on Trichodesmium spp. N2 fixation and its biogeochemical consequences have been studied  and quantified in the Arabian Sea , the open Atlantic , Pacific oceans , Great Barrier Reef (GBR) Lagoon  and the coastal waters of Tanzania . Trichodesmium is considered to be responsible for a substantial proportion of oceanic N2 fixation and the primary biological source of new nitrogen in the euphotic zone of the tropical oceans [13–14]. Trichodesmium provides 87.4 Tg N annually, representing about 36% of nitrogen fixed from all sources .
The diazotrophic nature of Trichodesmium renders it independent of exogenous supplies of combined nitrogen and favours its growth under low N:P ratios. Availability of other nutrients such as phosphorous, iron or other environmental factors such as salinity and light may constrain the extent of N2 fixation in a system. Understanding the effects of these factors on N2 fixation will further our ability to estimate the capacity of specific oceanic ecosystems to produce and export carbon  and predict climate change .
A recent study with cultures of Trichodesmium (strain GBRTRLI101) isolated from the GBR has shown that additions of Fe over the range 9–450 nM had a significant effect on the N2 fixation rate of Trichodesmium sp. when compared with the results from the medium containing no added Fe . N2 fixation rate increased four-fold for the 9 nM treatment and 11-fold for the 450 nM treatment whereas the cell yields only doubled for the 9 nM treatment and increased three-fold for the 450 nM treatment. These results are in general agreement with those of Rueter et al.  who observed a two-fold increase in N2 fixation rates of samples cultured in 100 nM treatment when compared with those obtained for samples cultured in 10 nM treatment. Another study with this same strain of Trichodesmium sp. has shown that active growth and N2 fixation occurred over salinities in the range 22–43 psu and the maximum growth and N2 fixation rates occurred in the range 33–37 psu .
In this study we examine the effects of some other factors on the N2 fixation rate of Trichodesmium. In particular the effects of diurnal light patterns, light intensity and quality (colour), the availability of phosphorous and different N sources (NH4+, NO3−, and urea) on N2 fixation rates are investigated. The effect of light quality (colour) was investigated because it has been observed that Trichodesmium can grow at depths up to 200 m in the open ocean . Also higher concentrations of Trichodesmium tend to occur at depths of 15–25 m  in the open ocean and at depths of 5–10 m in coastal waters (11), and hence it is hypothesised that light quality could be an important factor in controlling growth and N2 fixation of Trichodesmium.
2Materials and methods
Trichodesmium sp. (strain GBRTRLI101) used in the present study was isolated from waters near to Low Isles in the Northern GBR Lagoon. The initial culture was established in an enriched seawater medium and subsequent cultures of this strain were established on filter-sterilised (0.22-μm Millipore filter), nutrient-enriched artificial seawater (RAQA) based on Aquil medium  but without the addition of Si or inorganic nitrogen (NH4+, NO3−) and the P-PO4 concentration was reduced to 3 μM (Bell et al., in preparation). The growth experiments were carried out in 100-ml conical flasks which were stoppered with cotton wool under fluorescent light (45±2 μmol quanta m−2 s−1) on a 14-h light/10-h dark regime at 25±3°C. All glassware was soaked in 0.1 M HCl for at least 1 day, rinsed with MQ water and autoclaved at 120°C for 30 min. All growth experiments were conducted in triplicate during the exponential growth phase which normally lasts for ∼14 days in the RAQA medium.
Biomass was estimated by counting the number of filaments per ml and determining the average length per filament with an eyepiece micrometer. The average cells per filament were counted on the microscopic slide under a 400-fold magnification. At least three sub-samples were examined on the counting cell.
2.3Determination of N2 fixation
N2 fixation rates were measured by the acetylene reduction assay method as described by Capone . Seven-ml sub-samples of the cultures contained in media of different treatments were treated in 15-ml wide-mouth serum bottles, sealed by red silicone–rubber serum stoppers. One ml of acetylene was injected into the bottles and samples were then shaken gently. A gas phase sample (0.1 ml) was extracted from each bottle at zero time using a gas-tight syringe and was analysed immediately for ethylene concentration using a Photovac 10s plus portable gas chromatograph (GC) fitted with a photoionisation detector. Samples were incubated under cool fluorescent lights (45 μmol quanta m−2 s−1) for 3 h at room temperature (25°C). Gas samples were then extracted and analysed for ethylene concentration on the GC. The gas phases of several control blanks (without Trichodesmium but with acetylene) were also analysed at zero time and after the incubation. N2 fixation rates were calculated assuming an ethylene:nitrogen molar ratio of 4 . In all cases N2 fixation rates were normalised to the numbers of cells.
Four experiments were set up to investigate the effect of light on N2 fixation. In Experiment 1, N2 fixation rates were measured over a complete light:dark cycle from the late daily light period (16.00) through the following dark and light periods respectively (18:00 to 4:00 and 4:00 to 18:00) and for the initial part of the next dark period. In Experiment 2, the effect of light intensity was investigated by incubating cultures under five external photon flux densities: 10, 25, 45, 75, 160 μmol quanta m−2 s−1. In Experiment 3, incubation bottles, which had been incubated in the light for 3 h, were placed in the dark until N2 fixation was not detected and then re-exposed to the light. N2 fixation rates were determined in the dark on the dark-incubated samples and compared with the results from control samples, which were incubated under a constant light intensity. In Experiment 4, the effect of light quality (colour) was determined by wrapping coloured cellophane paper around the culture vessels. The external photon flux densities for all cultures were the same (40 μmol quanta m−2 s−1); this was achieved by wrapping the light sensor in the respective cellophane papers and locating positions for the culture vessels at the desired photon flux densities. The transmission spectra of coloured cellophane papers are given in Fig. 1.
Media containing various amounts of inorganic phosphorous (P-PO4) (added as KH2PO4) were prepared. The concentrations (0.1, 0.5, 1.2, 3.5, 6.5, 14.5, 20.8, 35.2 μM) of P-PO4 in the media were determined prior to use with the method described by Murphy and Riley . The stock culture for all experiments was prepared by growing cells in a low-inorganic-phosphorous medium (0.24 μM) for about 7 days. The trichomes were then transferred to media containing the various P-PO4 concentrations or 3 μM organic phosphorous (added as Na-glycerophosphate). Cultures grown in the medium enriched with Na-glycerophosphate were maintained for three generations before beginning experiments. Sub-cultures from the same primary culture were incubated with 0.5 μM, 3.5 μM P-PO4 and 3 μM Na-glycerophosphate for comparing and testing their ability to use organic phosphorous.
Cultures grown in RAQA N-free medium were transferred to RAQA medium containing 10 μM NO3−, NH4+, urea and 2 μM NH4+ to test the effect of combined nitrogen sources on the N2 fixation rate. For the 10 μM treatments short-term treatments (using the cells grown for one generation) and long-term treatments (grown for five generations) were examined. For the 2 μM NH4+ treatment, N2 fixation rates was determined using the cells grown for three generations.
One-way analysis of variance (ANOVA) was conducted in order to determine whether the results obtained from different individual treatment incubations on the effects of light quality and phosphorous were significantly different. Differences among mean values were tested for statistical significance (P<0.05) with the Tukey method of multiple comparisons .
3Results and discussion
3.1Effect of light on N2 fixation
3.1.1Diurnal variation in N2 fixation
N2 fixation of Trichodesmium GBRTRLI101 shows a characteristic diurnal pattern (Fig. 2) similar to that reported by Ohki et al.  for Trichodesmium NIBB1067. Maximum fixation rates occurred 6–8 h following the initiation of the light period and dropped off rapidly to near zero prior to the onset of the dark period. N2 fixation was not detected during the following dark period. Samples that were actively fixing N2 and were put into the dark continued to fix N2 for only a short while but the N2 fixation was restored on re-exposure of these samples to light (Fig. 3). The results are in agreement with previous studies which have shown that N2 fixation by Trichodesmium is primarily a light-driven phenomenon [6,25–27].
The ability of Trichodesmium to fix N2 in the light suggests it has developed some mechanisms of protection from photosynthetically evolved O2. Trichodesmium does not contain heterocysts which would provide for a definite spatial separation of photosynthesis and N2 fixation, but it has been shown that only 10–20% of cells along the filaments contain nitrogenase  and hence this could provide for some spatial separation of the two processes. Although they might be photosynthetically competent , these cells are suggested to contain active O2-consuming systems  such as a respiratory enzyme cytochrome oxidase, which could reduce molecular O2 generated during photosynthesis . Also, the light-dependent O2-consuming Mehler reaction in photosystem I is operative in Trichodesmium which would also reduce O2 concentrations.
Ohki et al.  concluded that the observed diurnal variation in the nitrogenase activity is principally due to the dark deactivation and light activation of the nitrogenase enzyme. However such a simple proposition would not explain why the nitrogenase activity drops off to near zero during the latter stages of the light period. Ohki et al.  recognised this and suggested that the reduction of N2 fixation in the latter part of the light period was possibly due to an accumulation of intermediates in the nitrogen metabolism. However more recent studies have shown that other factors are important in determining the observed diurnal variation in N2 fixation. In particular Berman-Frank et al.  found that both field-collected samples and laboratory-cultured samples exhibited a significant reduction in net oxygen production during the period of N2 fixation which suggests Trichodesmium also exhibits a temporal variation strategy for minimising the effects of evolved O2 on the deactivation of the nitrogenase enzyme. Their results also suggest that the deactivation of the nitrogenase and hence the reduction of N2 fixation in the latter stages of the light period is principally due to a high net production of O2 during that latter period.
As noted above N2 fixation persisted for a short time in the dark with samples that had been transferred from the light to the dark while still actively fixing N2; these samples showed a fast recovery to active N2 fixation on being re-exposed to the light (Fig. 3). This suggests that metabolites accumulated during the light period can support N2 fixation for a limited period of time in the dark. It is unclear however whether this is related to the availability of stored ATP, reductants, and fixed carbon or whether nitrogenase itself is activated during the light and deactivated in the dark [25,33].
3.1.2Dependence of N2 fixation on light intensity
N2 fixation rates of Trichodesmium increased with increased irradiances (Fig. 4) with the maximum rate occurring after about 8 h of incubation for all light intensities except the lowest light intensity of 10 μmol quanta m−2 s−1. Irradiance is required to obtain maximum rates of N2 fixation because this process involves a considerable energy commitment requiring a large expenditure of ATP and reductant . The results of Bergman et al.  suggest ATP and reductant would be formed and stored during the early phase of the light cycle and used during the mid-phase of the light cycle when the N2 fixation occurs. Also it has been demonstrated that higher light intensities increase the extent of Fe protein synthesis and the abundance of the active form of the Fe protein in nitrogenase and hence stimulate N2 fixation .
3.1.3Dependence of N2 fixation on light quality
N2 fixation was stimulated by a wide range of light quality with a general trend of stimulation being white>blue>yellow>green>red (Fig. 5). Tukey's analysis showed the N2 fixation under white light was significantly higher than under the other light qualities (P<0.05), but there was no significant difference under blue, yellow and green light (P>0.05). Also N2 fixation rate under red light was significantly lower than those under green, yellow and blue light (P<0.05). This trend is in general agreement with the observation of Kumar  for Anabaena sp. Fay  suggested that N2 fixation depends on the photosystem I where 90% of chlorophyll a (Chl a) is located . In photosystem I the quanta are mainly collected by Chl a. The lower overall rate of N2 fixation under green light could therefore be a consequence of less-efficient harvesting of green light. White light promoted the most N2 fixation as it permits maximum photosynthesis and it contains both blue and far-red wavebands which are absorbed most effectively by Chl a. It is also noteworthy that the colour of Trichodesmium is brown-red due to the pigment of phycoerythin and so Trichodesmium would probably tend to reflect red light and hence reduce its photosynthetic efficiency/N2 fixation under red light. Kamiya and Kowallik  found that blue light enhances carbohydrate degradation in green algae; this could then promote N2 fixation through increased photorespiration or through the carbon use mechanism suggested by Gallon and Stal .
3.2Effect of phosphorous on N2 fixation
Trichodesmium GBRTRLI101 responded to the addition of K2HPO4 with an increased rate of N2 fixation up to 6.5 μM with a decrease at higher concentrations (Fig. 6). A significant response to the 0.5 μM treatment in batch cultures was detected and treatments higher than about 1.2 μM tended to saturate the system. The Tukey test revealed that significant differences occurred within the treatment range 0.1–1.2 μM (P<0.05) whereas it did not exhibit significant differences in the range 1.2–14.5 μM phosphorous (P>0.05). However, significant differences were observed between 20.8, 35.2 and 6.5 μM (P<0.05). Our results are in agreement with those of Stewart and Alexander  who found for three freshwater diazotrophs that N2 fixation was linear with P-PO4 up to about 0.6 μM and saturated at ∼1.6 μM. The responses in N2 fixation between organic and inorganic phosphorous are comparable suggesting that organic phosphorous is readily used by Trichodesmium GBRTRLI101 (Fig. 7). This is in agreement with the findings of previous studies, which showed that Trichodesmium exhibits alkaline phosphatase activity and can grow in organic P-based media [43–44]. Also the results support the findings for work carried out in the open ocean. In particular it has been reported that in the central Atlantic Ocean N2 fixation was highly correlated to the phosphorous content of Trichodesmium. In surface waters of the North Pacific, excess total dissolved phosphorous (TDP), i.e. 0.22 μM, is detected, suggesting that P may not limit N2 fixation in the Pacific, but the TDP concentration in the North Atlantic is much lower (0.075 μM), suggesting that phosphorous is an important limiting factor there for N2 fixation . Wu et al.  note that 94–99% of the TDP was in fact organic phosphorus and hence the ability of Trichodesmium to utilise organic phosphate for N2 fixation (and growth) provides it with an advantage over phytoplankton without this ability. Another advantage possessed by Trichodesmium is that it can migrate vertically using its gas vacuoles to acquire phosphorous from below the mixed layer .
3.3Effect of combined nitrogen sources on N2 fixation
N2 fixation rates were not inhibited on the first-generation cells of Trichodesmium GBRTRLI101 grown in media containing combined nitrogen sources, 10 μM NO3−, NH4+ and urea, they being similar to N2 fixation rates of cells grown in nitrogen-free medium (Fig. 8). However by the fifth-generation cells grown in media containing 10 μM NO3−, NH4+ or urea all exhibited a significant reduction in N2 fixation rates relative to the control while the addition of 2 μM NH4+ did not affect the N2 fixation rates (Fig. 9).
The results presented here show that Trichodesmium can fix N2 in the presence of exogenous nitrogen substrates. This finding confirms the previous observation of Mulholland et al. [48–49] but is contrary to those of Ohki et al. , who found that N2 fixation was undetectable even after 5 h incubation in media containing 0.02 mM NH4+, 2 mM NO3− or 0.5 mM urea. This discrepancy may be due to the different species/strains or physiological states of Trichodesmium used in the respective studies. The reduced N2 fixation on the fifth generation relative to the first generation could be due to the accumulation of intracellular metabolites that inhibits N2 fixation. Also it is possible that combined nitrogen may inhibit synthesis of nitrogenase or may affect the functioning of pre-formed nitrogenase and hence affect the rates of N2 fixation .
Other workers have shown that the inhibition of N2 fixation by combined nitrogen does not always occur; it depends on the level and type of combined nitrogen supplied [49–50]. Our study confirmed this contention in that the addition of 2 μM NH4+ did not show any reduction in the N2 fixation rate. These results are consistent with the observation that natural populations of Trichodesmium continue to fix N2 even if combined N increases within the bloom [47,51].
The overall results presented here show that the diurnal pattern of N2 fixation in Trichodesmium GBRTRL101 was primarily light induced and the results are consistent with other work that has found that the N2 fixation rate is largely controlled by the net rate of O2 production during the light period. Dark fixation occurred in some experiments, but was related in part to the prior exposure to light. N2 fixation was dependent on the light intensity and was stimulated more by white light when compared with blue, green, yellow and red light, whereas rates of N2 fixation decreased most under red light. Phosphorous concentrations in the lower range of treatments up to 1.2 μM significantly stimulate N2 fixation and further additions promote little or no increase in the N2 fixation rates. Similar N2 fixation rates were obtained for cultures grown in media containing organic phosphorous (Na-glycerophosphate) and inorganic phosphorous (KH2PO4) which suggests that Trichodesmium can use organic phosphorous to meet its N requirement. Trichodesmium can fix N2 in the presence of combined N (NO3−, NH4+ or urea), but fixation is inhibited by high concentrations of combined N (10 μM) after several generations of growth.