Hydrobiologia

, Volume 623, Issue 1, pp 191–202

Effect of nitrogen forms on growth, cell composition and N2 fixation of Cylindrospermopsis raciborskii in phosphorus-limited chemostat cultures

Authors

    • Balaton Limnological Research Institute of the Hungarian Academy of Sciences
  • Hesham M. Shafik
    • Balaton Limnological Research Institute of the Hungarian Academy of Sciences
    • Botany Department, Faculty of ScienceSuez Canal University
  • Attila W. Kovács
    • Balaton Limnological Research Institute of the Hungarian Academy of Sciences
  • Sándor Herodek
    • Balaton Limnological Research Institute of the Hungarian Academy of Sciences
    • Balaton Limnological Research Institute of the Hungarian Academy of Sciences
Primary research paper

DOI: 10.1007/s10750-008-9657-9

Cite this article as:
Kenesi, G., Shafik, H.M., Kovács, A.W. et al. Hydrobiologia (2009) 623: 191. doi:10.1007/s10750-008-9657-9

Abstract

The aim of this research was to test whether NH4+ and NO3 affect the growth, P demand, cell composition and N2 fixation of Cylindrospermopsis raciborskii under P limitation. Experiments were carried out in P-limited (200 μg l−1 PO4-P) chemostat cultures of C. raciborskii using an inflowing medium containing either 4,000 μg l−1 NH4-N, 4,000 μg l−1 NO3-N or no combined N. The results showed the cellular N:P and C:P ratios of C. raciborskii decreased towards the Redfield ratio with increasing dilution rate (D) due to the alleviation of P limitation. The cellular C:N and carotenoids:chlorophyll-a ratios also decreased with D, predominantly as a result of an increase in the chlorophyll-a and N content. The NH4+ and NO3 supply reduced the P maintenance cell quota of C. raciborskii. Consequently, the biomass yield of the N2-grown culture was significantly lower. The maximum specific growth rate of N2-grown culture was also the lowest observed. It is suggested that these differences in growth parameters were caused by the P and energy requirement for heterocyte formation, nitrogenase synthesis and N2 fixation. N2 fixation was partially inhibited by NO3 and completely inhibited by NH4+. It was probably repressed through the high N content of cells at high dissolved N concentrations. These results indicate that C. raciborskii is able to grow faster and maintain a higher biomass under P limitation where a sufficient supply of NH4+ or NO3 is maintained. Information gained about the species-specific nutrient and pigment stoichiometry of C. raciborskii could help to access the degree of nutrient limitation in water bodies.

Keywords

Cylindrospermopsis raciborskiiHeterocytic cyanobacteriaPhosphorus limitationN2 fixationChemostat

Introduction

Cylindrospermopsis raciborskii is an invasive, potentially toxic, N2-fixing heterocytic cyanobacterium of subtropical–tropical origin. However, it is expanding into rivers and shallow water bodies in temperate regions (Padisák, 1997; Briand et al., 2004; Stüken et al., 2006). In Lake Balaton (the largest shallow lake in Central Europe), blooms of C. raciborskii have occurred during hot summers since the 1980s.

The factors triggering cyanobacterial bloom formation in lakes have been summarized in several articles (e.g. Blomqvist et al., 1994; Hyenstrand et al., 1998; Dokulil & Teubner, 2000). The factors responsible for the dominance of C. raciborskii have also been investigated, including its competition strategy in Lake Balaton. Istvánovics et al. (2000), for example, found the ability of C. raciborskii to store P important to its ecological success. Whilst, according to Kovács (2004), conditions of low light and high temperature in Lake Balaton favour C. raciborskii in competition with related cyanobacterial species. Its high uptake affinity for NH4+ at low concentrations also favours C. raciborskii (Padisák, 1997; Présing et al., 2001; Burford et al., 2006). Furthermore, Briand et al. (2004) suggest that the spread of C. raciborskii towards temperate latitudes may result from a high physiological tolerance, adaptation to lower temperatures and the effects of global warming. Despite several investigations, it is still difficult to predict bloom formation of C. raciborskii, largely because it is dependent on the interaction of many and often stochastic factors (Padisák & Reynolds, 1998).

In addition to physical conditions, nutrient availability and stoichiometry are the main factors controlling phytoplankton biomass and species composition in natural waters. In most freshwaters, the growth of phytoplankton is regulated by the availability of P. The phytoplankton in Lake Balaton is also believed to be P-limited (Herodek et al., 1995), although it has been suggested that N-limited periods may also occur in the lake (Présing et al., 2001). The N:P loading ratio in Lake Balaton is relatively low and has decreased over the last decade (Pomogyi, 1993; Herodek et al., 1995). This will have been advantageous for cyanobacteria and, in particular, for diazotrophic species (Schindler, 1977; Smith, 1983; Bulgakov & Levich, 1999; Nõges et al., 2007) such as C. raciborskii.

The effect of different combined N forms on the growth rate (μmax), cell composition and N2 fixation of Anabaena and Aphanizomenon species has been investigated under P limitation (Rhee & Lederman, 1983; Layzell et al., 1985; De Nobel et al., 1997a, b). However, in the case of C. raciborskii, these effects under P limitation have not been studied, despite it being widely acknowledged that responses can be different between closely related species, or even between strains of the same species (Meeks et al., 1983). The investigation of nutrient stoichiometry of C. raciborskii may help to understand its success in Lake Balaton, where this species together with Anabaena and Aphanizomenon species often composes 80–90% of total biomass (Présing et al., 1996). Furthermore, according to Geider & La Roche (2002), there is little information about the C:N:P stoichiometry of N2-fixing cyanobacteria. An improved understanding of the C:N:P stoichiometry characteristics of C. raciborskii could also be useful for assessing nutrient limitation or predicting bloom formation (Ahn et al., 2002).

The purpose of this study was thus to determine the effects of combined N forms (NH4+ and NO3) on the growth, nutrient stoichiometry and N2 fixation of C. raciborskii under P-limited conditions with a view to obtaining new insights into the factor(s) influencing its success in Lake Balaton.

Materials and methods

Chemostat experiments

C. raciborskii cultures were grown in chemostat devices previously described by Shafik et al. (2001). The strain of C. raciborskii (ACT 9502) used in the experiments was originally isolated from Lake Balaton (Hungary). The culture was non-axenic, but the biomass of contaminating bacteria was routinely monitored by fluorescence microscopic inspection after acridine orange staining and never exceeded 2% of that of C. raciborskii. Growth medium was continuously supplied by peristaltic pumps (Masterflex 7523-13 or 7523-14) into 2 l culture vessels. The cultures were continuously aerated and mixed by pumping sterile air into the chemostats and were maintained at conditions (pH 7.5–7.9, temperature 26 ± 1°C and illumination of 100 μmol photons m−2 s−1 at the wall of the vessels) previously found to be not limiting for C. raciborskii growth (Shafik et al., 2001; Briand et al., 2004). The cultures were continuously illuminated by Circline™ Cool White T9/32W/CW. The light intensity was measured in the centre of the culture vessel using a spherical (4π) quantum sensor (Walz US-SQS/L) attached to a LI-COR 1400 data logger. In the centre of the vessels, the light intensity increased with μ from 40–50 to 100 μmol photons m−2 s−1 due to decreasing self-shading, with the exception of the two densest of the NO3 steady-state cultures (D = 0.75 and 0.5) where the light intensity was lower (IC = 20 and 30 μmol photons m−2 s−1, respectively). To calculate mean light intensity experienced by the continuously stirred cells (I*), we adopted the formula of Reynolds (1997) (Eq. 1),
$$ I^{*} = \sqrt {I_{0} * I_{\text{C}} } $$
(1)
where (I0) and (IC) are the irradiance at the surface and in the centre of the vessels, respectively.

The cultures were grown in modified BG-11 medium (Shafik et al., 2001), but with some further alterations. In addition to 200 μg l−1 PO4-P, it contained either 4,000 μg l−1 NH4-N or NO3-N (N:P supply ratio of 20 by mass) or, alternatively, was free of combined N. The concentration of Fe(III)-citrate was reduced to 25% of that used by Shafik et al. (2001) to avoid the precipitation of P in the media. The steady states for cultures grown in all N forms—i.e., where dilution rate (D) = growth rate (μ)—were achieved initially at low dilution rates. The dilution rate was then gradually increased until D > μmax. Steady states were achieved at dilution rates of 0.25, 0.35, 0.50, 0.70, 0.75, 0.90 day−1 in the N2-fixing culture; at 0.25, 0.35, 0.50, 0.75, 1, 1.5 day−1 in the NH4+-culture and at 0.35, 0.50, 0.75, 1, 1.25 day−1 in NO3-culture. The growth of cultures was checked daily by measuring the dilution rate, concentrations of chlorophyll-a (Chl-a) and total carotenoids (Car), optical density at 750 nm in a 1-cm path length cuvet (OD), cellular dry weight (dw) and pH. The cultures were considered to be in a steady state, when the measured parameters varied by less than 5% for 3–5 days. During this period, the nitrogen and carbon contents of the cells [i.e., the particulate nitrogen (PN) and carbon (PC)], dissolved NH4-N, NO3-N and PO4-P concentrations, total phosphorus (TP), irradiance and N2 fixation were also measured. Particulate phosphorus (PP) was calculated as TP minus PO4-P. Twenty-millilitre samples were preserved with Lugol’s iodine for further microscopic investigation. The number of heterocytes was determined using an inverted microscope at 400× magnification (Utermöhl, 1958); a minimum of 400 heterocytes were counted per sample or, where the number of heterocytes was low, at least 100,000 μm of filament per sample was counted to obtain an accuracy of less than 10%.

Analytical methods

Pre-combusted (450°C, overnight) and pre-weighed glass fibre filters (Whatman GF/C) were used for the determination of cellular dry weight, PC and PN. Dry weight was measured by a METTLER 30 microbalance with a reading accuracy of 1 μg and a coefficient of variation less than 2%. PC and PN were measured by a stable isotope mass spectrometer (European Scientific, ANCA-IRMS with a coefficient of variation less than 2%). The concentrations of chlorophyll-a (Iwamura et al., 1970) and carotenoids (Wetzel & Likens, 1991) were determined spectrophotometrically (using Shimadzu UV-VIS 1601) after extraction in hot methanol. The concentrations of soluble NH4-N, NO3-N and PO4-P were determined in triplicate by the methods detailed previously in Présing et al. (2001) (coefficient of variation less than 5%).

Measurement of N2 fixation rate

N2 fixation experiments were performed in two parallel experiments using 40 ml screw-cap vials equipped with silicone septum. Measurements were performed on 1 day of each steady state using incubation periods of 0, 2, 4, 6 and 8 h. 15N2 gas was injected into the air space above 10–20 ml of steady-state cell suspensions, and the cultures were vertically shaken at the same irradiance and temperature used with chemostat cultures. After the incubation period had elapsed, a 1-ml sample from the gas phase was removed in triplicate and injected into 12 ml vials containing helium gas. Cells were then filtered and dried at 60°C for 12 h. The 15N/14N ratio of the gas samples was determined in triplicate and the N content and 15N/14N ratios of the cells were determined in duplicate according to the methods of Présing et al. (2001) (coefficient of variation of less than 2%).

Data analysis

A non-linear curve fitting the Droop (1973) function (GraFit 3.00 version) was used to determine the growth kinetic parameters of C. raciborskii (Eq. 2),
$$ \mu^{*} (Q) = \mu_{\max }^{*} \left( {1 - \frac{{Q_{0} }}{Q}} \right) $$
(2)
where Q is the cellular P quota, Q0 is the cellular P quota at μ = 0 (minimum or maintenance P quota), μ* is the actual growth rate and \( \mu_{ \max }^{*} \) is the apparent maximal growth rate (if Q were infinite). The relationship between the growth rate of C. raciborskii versus the nutrient and pigment content and the relationship between the Car:Chl-a and C:N ratios were examined using Spearman’s ρ correlation analysis (SPSS v.13). The relationship between OD and Chl-a, OD and dw, N content and Chl-a content were examined by linear regression (SPSS v.13).

Results

Biomass and growth kinetic parameters

The different parameters of the steady-state cultures were tested to select the best indicator of biomass. Dry weight and OD were strongly correlated (Fig. 1a); in contrast, chlorophyll-a did not show such a strong correlation with OD (Fig. 1b). Chl-a per OD (Fig. 2b) and Chl-a per dry weight (as detailed below) varied with growth and nutrient limitation. The dry weight to OD ratio decreased only at the highest μ in the combined N-supplied cultures (Fig. 2a) due to changes in the morphology of cells (e.g. gas vesicles, or long and coiled filaments). The determination of biovolume was the least accurate method and was also influenced by morphological changes. In addition, cell shrinkage during sample preservation might increase the uncertainty of biomass estimation (Hawkins et al., 2005). Thus, from the two most accurate indicators of biomass (PC and dw), dry weight was chosen because it is used more frequently in similar studies. The tendencies remained unchanged when normalizing to PC (not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig1_HTML.gif
Fig. 1

Optical density at 750 nm as a function of dry weight (a) and concentration of chlorophyll-a (b) in P-limited chemostat cultures of C. raciborskii. Parameters of the linear regressions are: (OD versus dw: y = 0.0017x + 0.0163, R2 = 0.95, P < 0.000; OD versus Chl-a: y = 0.0002x + 0.035, R2 = 0.56, P = 0.001). Dots represent the mean of three to five data in each steady state. SD < 5%

https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig2_HTML.gif
Fig. 2

Dry weight (a) and chlorophyll-a (b) normalized to OD at 750 nm as a function of growth rate in P-limited chemostat cultures of C. raciborskii. N2 (∆), NO3 (✱), NH4+ (●). Mean of three to five data in each steady state. SD < 10%

The biomass changed non-linearly with μ with a peak at low to intermediate growth rates. In comparison to the NH4+- or NO3-supplied cultures, the N2-grown culture maintained a lower biomass and filaments began to wash out from the chemostat at lower value of D than observed for cultures grown in the presence of combined N (Fig. 3a, b). The growth kinetic parameters were estimated by fitting the Droop equation to the mean cellular P quotas (μg P mg dw−1; Fig. 4). The N2-fixing cells were found to have a higher Q0 and a lower \( \mu_{ \max }^{*} \) in comparison to those measured in the NO3- and NH4+-cultures (Table 1). However, no significant differences in growth kinetic parameters (Q0 or \( \mu_{ \max }^{*} \)) were observed between the NH4+- or NO3-cultures (Table 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig3_HTML.gif
Fig. 3

Dry weight (a) and concentrations of chlorophyll-a (b) of C. raciborskii as a function of growth rate in P-limited chemostat cultures. N2 (∆), NO3 (✱), NH4+ (●). Mean of three to five data in each steady state. SD < 5%

https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig4_HTML.gif
Fig. 4

Growth rate as a function of phosphorus quotas (Q) of C. raciborskii in P-limited chemostat cultures. Points represent the mean of three to five values in each steady state N2 (∆), NO3 (✱), NH4+ (●). Error bars represent the error of means

Table 1

The growth kinetic parameters of C. raciborskii under P limitation with different N sources (the Droop model was fitted non-linearly)

N forms

\( \mu_{ \max }^{*} \) (day−1)

Q0 (μg P mg dw−1)

χ2

N2

1.17 ± 0.16

2.32 ± 0.30

0.016

NO3

1.45 ± 0.28

1.56 ± 0.32

0.041

NH4+

1.49 ± 0.18

1.61 ± 0.24

0.049

Nutrient and pigment contents

The cellular P content (P quota) of C. raciborskii increased with μ (Fig. 4). In contrast, the C content was relative constant at ca. 45% of the dry weight of cells and did not change tendentiously with respect to either growth rate or the N source (Fig. 5a). The N content of C. raciborskii was similar irrespective of the N source at low growth rates (ca. 6–7% of the dry weight). However, in the NH4+- and NO3-cultures, the N content increased up to ca. 12% and 10% N at μ > 1, respectively (Fig. 5b). Simultaneously, the concentration of dissolved N rose to over 1,000 μg l−1. In contrast, the N content of the N2-culture varied less with μ and reached a maximum of ca. 8%.
https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig5_HTML.gif
Fig. 5

Biomass specific C (a) and N (b) contents of C. raciborskii as a function of growth rate in P-limited chemostat cultures. Spearman’s ρ correlation coefficients between C content and μ is: rs = 0.39 (P = 0.117) and between N content and μ is: rs = 0.64 (P < 0.01). N2 (∆), NO3 (✱), NH4+ (●). Mean of three to five data in each steady state. SD < 10%

The carotenoids to dry weight ratio in the C. raciborskii cultures changed non-linearly with μ, and it was relatively constant regardless of the N source, except for a slight decrease observed at the highest growth rates (Fig. 6a). The Chl-a to dry weight ratio increased with μ (Fig. 6b) and resulted in the highest values at NH4+ supply at the highest growth rates. However, the Chl-a contents at lower and intermediate μ were higher in the N2-culture, when compared to the contents of NH4+- and NO3-cultures (Fig. 6b). The N and Chl-a contents were strongly correlated in the N2- and NH4+-grown steady states (Fig. 7), while in the NO3-culture the correlation was weaker and demonstrated considerable scatter. The cells generally received an irradiance (I*) of 60–100 μmol photons m−2 s−1, except in the two densest NO3-grown steady states, where values of I* for D = 0.75 at 900 Chl-μg l−l and D = 0.5 at 660 Chl-a μg l−1 were 45 and 55 μmol photons m−2 s−1, respectively.
https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig6_HTML.gif
Fig. 6

Changes in carotenoids (a) and chlorophyll-a (b) contents of C. raciborskii on dry weight basis with growth rate in P-limited chemostat cultures. Spearman’s ρ correlation coefficients between Car content and μ is: rs = −0.03 (P = 0.899) and between Chl-a content and μ is: rs = 0.86 (P ≪ 0.01). N2 (∆), NO3 (✱), NH4+ (●). Mean of three to five data in each steady state. SD < 10%

https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig7_HTML.gif
Fig. 7

The regression between N and chlorophyll-a contents of C. raciborskii in P-limited chemostat cultures. Parameters of the linear regression for are: y = 2.6159x − 9.7927, R2 = 0.91 (P < 0.01) N2; y = 0.8705x + 2.0437, R2 = 0.36 (P < 0.05) NO3; y = 1.5162x − 5.3318, R2 = 0.86 (P < 0.01) NH4+. Points represent the mean of three to five values in each steady state N2 (∆), NO3 (✱), NH4+ (●)

Rate of N2 fixation and the number of heterocytes

The rate of N2 fixation by C. raciborskii increased with μ in the N2-grown cells, reaching a maximum value of ca. 2 μg N mg dw−1 h−1 at μ ≥ 0.7 (Fig. 8a). In the NH4+- and NO3-supplied cultures, N2 fixation rates and the number and N2 fixation rate per heterocyte were significantly lower than in the N2-culture (Fig. 8a–c). In the NO3-supplied culture, N2 fixation did increase moderately with μ, from intermediate growth rates. In the presence of NH4+, N2 fixation decreased with μ and was completely inhibited at growth rates μ ≥ 1. There were almost no heterocytes in the NH4+-supplied cultures at the highest μ values, in contrast to the NO3-culture at comparable growth rates (Fig. 8b). Extremely long (>1,500 μm) and coiled filaments were observed in the NH4+-supplied culture at μ ≥ 1, with the lowest rate of N2 fixation.
https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig8_HTML.gif
Fig. 8

Biomass specific rate of N2 fixation (a) number of heterocytes (b) and specific rate of N2 fixation per heterocyte (c) in P-limited chemostat cultures of C. raciborskii. N2 (∆), NO3 (✱), NH4+ (●). Mean of three to five data in each steady state. Error bars represent standard deviation

Nutrient and pigment ratios

The carotenoids to chlorophyll-a ratio (Fig. 9) and the C:N ratio (Fig. 10a) of C. raciborskii decreased with μ irrespective of the supplied N source [Spearman’s ρ correlation coefficient between them was rs = 0.485 (P < 0.05)]. At low and intermediate growth rates, the C:N ratio was typically close to the Redfield ratio (i.e., 5.6 by mass) and no marked differences were observed between the cultures grown on different N sources (Fig. 10a). However, as μ increased, the cellular C:N ratio decreased in the NH4+- and NO3-cultures. In the N2-culture, the C:N ratio did not demonstrate a similar tendency. Hence, higher C:N ratios were observed at the highest growth rates in the N2-culture (μ ≥ 0.75) than in the NH4+- and NO3-cultures (μ ≥ 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig9_HTML.gif
Fig. 9

Changes in carotenoids to chlorophyll-a ratio (by mass) of C. raciborskii with growth rate in P-limited chemostat cultures. Spearman’s ρ correlation coefficient between Car:Chl-a ratio and μ is: rs = −0.77 (P ≪ 0.01). N2 (∆), NO3 (✱), NH4+ (●). Mean of three to five data in each steady state. SD < 5%

https://static-content.springer.com/image/art%3A10.1007%2Fs10750-008-9657-9/MediaObjects/10750_2008_9657_Fig10_HTML.gif
Fig. 10

Changes in a C:N, b C:P and c N:P ratios (by mass) of C. raciborskii with growth rate in P-limited chemostat cultures (the Redfield ratio is indicated by the dashed line). Spearman’s ρ correlation coefficient between C:N ratio and μ is: rs = 0.62 (P < 0.01), between C:P ratio and μ is: rs = 0.72 (P ≪ 0.01), between N:P ratio and μ is: rs = 0.53 (P = 0.026). N2 (∆), NO3 (✱), NH4+ (●). Mean of three to five data in each steady state. SD < 10%

In response to increasing μ, the C:P ratio of C. raciborskii decreased from 145 to 65 (by mass) in the N2-grown culture and from 170 to 75 and 155 to 55 in NO3- and NH4+-cultures, respectively (Fig. 10b). The N:P ratio of C. raciborskii decreased from 25 to 14 and 23 to 13 (by mass) in the case of NO3- and NH4+-cultures, respectively, and from approximately 20 to 9 in the case of the N2-culture. The cellular C:P ratios of cultures grown on all N forms converged towards the Redfield ratio at the highest μ values (Fig. 10b). However, the N:P ratio of the NO3- and NH4+-supplied cultures was higher than that observed in the N2-culture at the highest growth rates (Fig. 10c).

Discussion

Biomass and growth kinetic parameters

The biomass changed in accordance with the chemostat theory. However, slight decreases were observed in cell density at the lowest growth rates (i.e., μ ≤ 0.35) at all N sources. This might be due to the higher proportion of older cells and a higher rate of mortality in the cultures (Lee & Rhee, 1997). The difference between the biomass yields of the N2-fixing and combined N-supplied cultures (Fig. 3a, b) can be attributed to the inferior growth (Q0 and \( \mu_{ \max }^{*} \)) of the N2-culture under P limitation (Fig. 4, Table 1). Higher P demand and maintenance cell quota (Q0) of the N2-grown culture might be associated with the high P content of the heterocytes, as this nutrient is required for the rapid synthesis of RNA (i.e., to produce nitrogenase), for the formation of membrane-bound phospholipids and for the ATP requirement of N2 fixation (Layzell et al., 1985; De Nobel et al., 1997a). This assumption is confirmed by the number of heterocytes, which was the highest in the N2-fixing culture (Fig. 8b). Our estimates of \( \mu_{ \max }^{*} \) were certainly less satisfactory than those of Q0 (Fig. 4). Nevertheless, the 20% reduction in \( \mu_{ \max }^{*} \) of the N2-culture compared to those supplied with NH4+ and NO3 was reasonable (Table 1). Biochemical and physiological evidences have shown the energetic costs of N2 fixation may decrease μ by up to 30% (Turpin et al., 1985). Similarly, Shafik et al. (2001) found that μ of P-sufficient C. raciborskii decreased by 25% and 35% at temperatures of 20–25 and 30°C, respectively, when grown on N-free medium as compared to cultures grown on excess NH4+ supply.

The difference in the Q0 and \( \mu_{ \max }^{*} \) values are also in agreement with previous P-limited chemostat experiments, where the favourable effect of combined N forms on the growth of Anabaena and Aphanizomenon species has been demonstrated (Rhee & Lederman, 1983; Layzell et al., 1985; De Nobel et al., 1997a, b). Since energetic costs associated with reduction of NO3 (i.e. reducing power, nitrate reductase synthesis) are known to constrain the kinetics of cyanobacterial growth (Thompson et al., 1989), one might expect some difference in the growth parameters of the NO3- and NH4+-supplied cultures. However, under non-limiting light intensities, we found no significant differences in the growth kinetics of C. raciborskii when supplied it with either NO3 or NH4+ under P limitation. Sprőber et al. (2003) reported similar results for P-replete chemostat cultures of C. raciborskii.

The difference in the Q0 values between our combined N- and N2-grown cultures under P limitation indicates the availability of dissolved nitrogen forms reduces the amount of P required by C. raciborskii. This is in agreement with previous results obtained from NH4+-supplied cultures of Anabaena and Aphanizomenon (De Nobel et al., 1997a, b). However, under natural conditions, available dissolved N forms may give the possibility to non-N2-fixing species, also for growth or may even favour them over C. raciborskii; thus, the proportional contribution of C. raciborskii to total biomass could decrease in spite of its reduced P demand.

The Q0 values of the investigated C. raciborskii strain were in the lower end of the range reported for other cyanobacteria (Q0 ≫ 2–8 μg P mg dw−1; Healey, 1978; Istvánovics et al., 2000). In comparison to the results of De Nobel et al. (1997b), our strain showed a similar P requirement to Anabaena sp., while its Q0 was three times lower than that of Aphanizomenon sp. C. raciborskii is predominantly an affinity specialist as it has a high P uptake affinity and has a low Q0 value, even for a cyanobacterium. However, where transient P pulses occur, it also has a capacity for intracellular P-storage (Istvánovics et al., 2000). In P-deficient environments, small differences in P demand may be decisive in determining the outcome of inter-specific competition. For example, during summer blooms of heterocytic cyanobacteria in Lake Balaton, beside other factors P demand can be important.

The rate of N2 fixation

In the N2-culture, an increasing rate of N2 fixation was necessary to sustain growth (Fig. 8a). The alleviation of P limitation with μ could be seen as a prerequisite for enhanced N2 fixation (e.g. Mulholland & Bernhardt, 2005). In the NO3-culture, N2 fixation rate and heterocyte formation were not suppressed entirely not even at the two highest growth rates. While the N content was observed to increase at high μ, however it was not sufficient to completely suppress N2 fixation and heterocyte formation (Fig. 8a, b). The number of heterocytes actually increased with μ in the NO3-culture, probably due to increasing P availability. Hence, C. raciborskii appeared to preferentially decrease N2 fixation per heterocyte, rather than heterocyte formation, under NO3 supply (Fig. 8b, c). In the NO3culture, the lowest N2 fixation rate was measured in the two densest steady states (D = 0.5 and 0.75; Fig. 8a), which is in accordance with the enhanced repression of N2 fixation by combined N forms under suboptimal light conditions (Yoch & Gotto, 1982) and with the high ATP and reductant demand of N2 fixation. In the NH4+-supplied steady states, N2 fixation was most strongly suppressed at the highest μ values (Fig. 8a), where the highest N contents (Fig. 5b) and no heterocytes (Fig. 8b) were observed. In these cases, extremely long filaments were formed. Kapustka & Rosowski (1976) found similar results with Cylindrospermum species. This can be related to heterocyte formation which is blocked by the high cellular N content (Meeks & Elhai, 2002). The heterocytes of C. raciborskii are formed at the ends of its trichomes (Shafik et al., 2003); hence, C. raciborskii cannot fix considerable amount of N2 in extremely long filaments. Layzell et al. (1985) reported comparable patterns in the rate of N2 fixation per unit dry weight in P-limited Anabaena cultures when grown on NO3, NH4+ or N2, with N2 fixation only entirely suppressed at very high N contents. Sprőber et al. (2003), in P-replete culture of C. raciborskii, noted that N2 fixation was inhibited only 4–8 h after the addition of NH4+ or NO3 when the N content of cells increased. These findings suggest that the rate of N2 fixation in the C. raciborskii cultures might be suppressed by the high cellular N contents (Fig. 5b, 8a–c), rather than increased concentrations of dissolved N forms. According to the results of the present study, there should be a threshold value of N content (around 13%), or (C:N ratio lower than 4.1–4.4 by mass), above which the rate of N2 fixation in C. raciborskii is entirely suppressed. This value is in reasonable agreement with those published for related Nostocales [(e.g. Kulasooriya et al., 1972 (C:N 4.5 by mass); Yakunin et al., 1995 (11% N and 4.5 C:N); Layzell et al., 1985 (around 11.3% N and 4.1 C:N ratio)].

Changes in nutrient and pigment contents with growth rate

The parallel increase in N and Chl-a contents with increasing μ and nutrient supply (Fig. 5b, 6b, 7) could occur for several reasons. Rapid cell synthesis requires substantial photosynthetic machinery to produce reducing power and ATP, and the photosynthetic complexes and thylakoid membranes contain N rich proteins (Bryant, 1994; Geider & La Roche, 2002). Also, in rapidly growing cells, total protein and nucleic acid concentrations are usually high (Vargas et al., 1998; Geider & La Roche, 2002; Klausmeier et al., 2004). A strong positive correlation has been observed between N and Chl content in P-limited Anabaena cultures (Rhee & Lederman, 1983; Oh et al., 1991), as well as in our N2-fixing and NH4+-cultures (Fig. 7). In our N2-culture, more Chl-a was needed for assimilation of N (to capture light and produce ATP and reductant to fix N2) than in the energetically more preferential NH4+-supplied one as experienced by Layzell et al. (1985). The higher amount of N in the NH4+-culture could also be caused by the excess utilization of NH4+. The correlation was weak for the NO3-culture, where the two densest nitrate-supplied steady states (μ = 0.5 and 0.75 day−1) might have been light-deficient, at least certainly more than other steady states. Hence, their Chl-a:N ratio deviated upwards from the regression line. For the reduction of NO3, more reductant is needed than in the N2- or NH4+-cultures. We assume, that C. raciborskii compensated its high reductant demand for growth on NO3 by increasing its Chl-a content to sustain the same μ as the N2- or NH4+-cultures. In addition, its N content may have been reduced when grown on NO3 under low light conditions, as reviewed by Thompson et al. (1989). The high Chl-a may have been achieved by an increased number of photosynthetic units and/or by a higher PSI/PSII ratio (as PSI has two to three times higher chlorophyll content than PSII; Bryant, 1994; De Nobel et al., 1998).

Several authors define the Ik value as the beginning of light sufficiency (Briand et al., 2004). In our case the lowest I* values in these steady states were above the Ik values of C. raciborskii [Ik and Iopt at 27°C are 26 and 80 μmol photons m−2 s−1, respectively (Shafik et al., 2001; Briand et al., 2004)]. C. raciborskii has superior adaptation ability and low light requirement compared to other cyanobacteria (Briand et al., 2004). However, the suboptimal light conditions probably influenced its physiology (e.g. chlorophyll content and the rate of N2 fixation) in these two dense steady-state cultures.

The moderate decreasing tendency in chlorophyll content at the highest growth rates may have been caused by increased light intensity (Sciandra et al., 1997).

The C:N ratio decreased with μ, due to the increasing N content. However, in the N2-fixing culture, this increase was not pronounced, where a slight increase was observed in the C:N ratio at the highest μ values. Ward & Wetzel (1980) experienced similar results in Anabaena and Aphanizomenon cultures. The decrease with μ in the Car:Chl-a and C:N ratios were mainly caused by increasing contents of Chl-a and N (Fig. 5b, 6b). Heath et al. (1990) and Liu et al. (1999) also noted that the Car:Chl-a and C:N ratio can decrease as a function of growth rate and nutrient supply and, in common with the present study (Fig. 9), reported a Car:Chl-a ratio close to one when approaching the highest growth rates. The strong correlation between the particulate C:N and Car:Chl-a ratios has been also observed in natural environments (Watson & Osborne, 1979) and, as such, could provide a useful indicator of nutrient deficiency or a relative index of nutrient limitation in nutrient enrichment bioassays (Schlüter et al., 1997).

The increase in P content with μ was facilitated by the alleviation of P limitation, as cells receive more P at higher μ due to the reduced biomass and the increase in P supply. On the other hand, P demand also increases with μ because rapidly dividing cells have a high DNA content and intensive protein synthesis requires a higher rate of RNA synthesis (e.g. Geider & La Roche, 2002; Ågren, 2004). The N:P ratios of our N2-fixing and P-limited culture (Fig. 10c) were probably close to the “structural N:P ratio” of C. raciborskii, when neither excess N nor excess P is stored. These N:P ratios were lower, when comparing them to the N:P ratios in our NH4+- and NO3-cultures, especially at higher μ. This disparity could have been caused by luxury uptake of N in the NH4+- and NO3-cultures due to the increase in the N supply at the diluted biomass. Alternatively, the N2-culture may have been slightly N limited due to its insufficient ability to adjust the rate of N2 fixation to growth rate (Fig. 8a), as hypothesized by De Nobel et al. (1997b). In P-replete, N2-fixing, chemostat cultures of C. raciborskii (Shafik et al., 2001), probably luxury uptake of P occurred, because (whilst the N content was similar) the N:P ratios were lower (ca. 6 by mass) than those in our comparable P-limited N2-fixing culture. The N:P ratio of our N2-grown culture converged towards the P-replete N:P ratio measured by Shafik et al. (2001) with increasing μ due to the alleviation P limitation.

Nutrient and pigment ratios in natural water bodies are frequently used for assessing nutrient limitation (e.g. Kim et al., 2007; Ahn et al., 2002). However, the interspecific difference in the optimal nutrient ratios and the ability of species to optimize these ratios to nutrient supply and growth requirements may hinder this approach. There is little information about the C:N:P stoichiometry of N2-fixing cyanobacteria, and it had not previously been investigated in P-limited C. raciborskii cultures. These P-limited ratios described in this study must be close to the stoichiometric ratios characterizing C. raciborskii in the P-deficient Lake Balaton. Affinity specialists generally have higher N:P ratios due to more N rich photosynthetic and transport proteins (Geider & La Roche, 2002). It would seem that this strain of C. raciborskii, in similarly to Anabaena sp., is more affinity specialist than Aphanizomenon sp. (De Nobel et al., 1997a, b).

Conclusion

Under P limitation, dissolved N forms (NH4+, NO3) inhibited the rate of N2 fixation and heterocyte formation, influenced growth, physiology, nutrient stoichiometry and reduced the P demand of this strain of C. raciborskii. N2 fixation was partially inhibited by NO3, and completely by NH4+, but only at very high NH4+ concentrations which do not naturally occur in the pelagic region of Lake Balaton, where the entire suppression of N2 fixation has never been experienced (Présing et al., 2001). This C. raciborskii strain is able to grow faster and maintain a higher biomass under P limitation in the presence of NH4+ or NO3. However, decreasing the N load of Lake Balaton would not be desirable since it would favour nuisance N2-fixing species. The suggested restoration measure for improving the water quality of the lake is to reduce the external P load and not reduce the N load.

Acknowledgements

This study was financially supported by a grant from the Balaton Project of the Office of the Prime Minister of Hungary (MEH) and from the Hungarian Program for Research and Development (NKFP) under contract 3B/022/2004 (BALÖKO). The authors also wish to acknowledge and thank Terézia Horváth, Erika Kozma for their assistance. We wish to thank Peter Hunter for correcting the manuscript and Lajos Vörös for his help with the microscopic techniques.

Copyright information

© Springer Science+Business Media B.V. 2008