Nutrient induced fluorescence transients (NIFTs) provide a rapid measure of P and C (co-)limitation in a green alga

Abstract

Nutrient Induced Fluorescence Transients (NIFTs) have been shown to be a possible way of testing for the limiting nutrient in algal populations. In this study we tested the hypothesis that NIFTs can be used to detect a (co-)limitation for inorganic phosphorus (Pi) and CO₂ in the green alga Chlamydomonas acidophila and that the magnitude of the NIFTs can be related to cellular P:C ratios. We show a co-limitation response for Pi and CO₂ via traditional nutrient enrichment experiments in natural phytoplankton populations dominated by C. acidophila. We measured NIFT responses after a Pi- or a CO₂-spike in C. acidophila batch cultures at various stages of Pi and inorganic C limitation. Significant NIFTs were observed in response to spikes in both nutrients. The NIFT response to a Pi-spike showed a strong negative correlation with cellular P:C ratio that was pronounced below 3 mmol P: mol C (equivalent to 0.2 pg P cell^–1). Both cellular P and C content influenced the extent of the Pi-NIFT response. The NIFT response to a CO₂-spike correlated to low CO₂ culturing conditions and also had a negative correlation with cellular P content. A secondary response within the Pi-NIFT response was related to the CO₂ concentration and potentially reflected co-limitation. In conclusion, NIFTs provided a quick and reliable method to detect the growth-limiting nutrient in an extremophile green alga, under Pi-, CO_2- and Pi/CO₂ (co-)limited growth conditions.

Key words:

• acidophile
• Chlamydomonas
• CO₂ concentrating mechanism
• CO₂ limitation
• extremophile
• nutrient limitation
• photosynthesis response
• phytoplankton
• stoichiometry

INTRODUCTION

Primary productivity in aquatic ecosystems is frequently limited by nutrient availability. Although it was thought for many years that only one factor can be growth-limiting at any one time, co-limitation by multiple nutrients might be a common phenomenon in natural phytoplankton populations (Elser et al., 2007; North et al., 2007; Moore et al., 2007; Moore et al., 2008; Saito et al., 2014) and has also been observed in single species cultures of a green alga (Spijkerman et al., 2011) and a haptophyte (Buitenhuis et al., 2003). Nutrient enrichment experiments can detect such limitations via full factorial design additions of the nutrients and monitoring the change in biomass production (Moore et al., 2008; Harpole et al., 2011) but do not provide exclusive results on the way the co-limitation acts upon the species or phytoplankton community. Alternatively, models based on Monod kinetics can distinguish between different types of co-limitation (Saito et al., 2008; Buitenhuis & Geider, 2010) but these experiments are very time consuming. Three different forms of co-limitation have been identified, based on nutrient enrichment experiments: simultaneous, independent and serial co-limitation (Harpole et al., 2011). In simultaneous and independent co-limitation the community biomass increases only with the addition of all co-limiting nutrients. In the serial co-limitation response, the synergistic effect of simultaneous nutrient additions only occurs when the primary limitation is surpassed. A serial co-limitation response for Pi and CO₂ was found in nutrient enrichment experiments with cultures of the green alga Chlamydomonas acidophila Negoro (Spijkerman, 2010), where Pi addition resulted in a large response, CO₂ addition gave no response, but the combined addition of Pi and CO₂ resulted in a synergistic effect. From a biochemical point of view this can be considered a co-limitation with different nutrients limiting different processes (Wolf-Gladrow & Riebesell, 1997; Moore et al., 2008), one limiting nutrient biologically substituting with another, either directly within the same macromolecule or indirectly by substituting one macromolecule for another, or when the ability to take up low concentrations may depend on the availability of another nutrient (Moore et al., 2013).

Chlamydomonas acidophila is an extremophilic alga which experiences relatively low Pi and CO₂ concentrations in its natural environment and is potentially (co-)limited in its growth by these two nutrients (Tittel et al., 2005; Spijkerman, 2008a). Pi is limiting because the bio-available Pi concentrations are decreased by the typically high total iron concentrations (0.1–10 mM; (Spijkerman, 2008a) and inorganic carbon (Ci) concentrations are low because at low pH (2––3.5), > 99% of Ci is CO₂ and the total Ci concentration at air equilibrium is only ~14 μM (Stumm & Morgan, 1970). Recent work on Ci acquisition in C. acidophila revealed the presence of a CO₂ concentrating mechanism (CCM), which was down-regulated under high CO₂ and/or Pi-limiting conditions (Spijkerman, 2005; Spijkerman, 2011). Previous results have suggested that co-limiting conditions for Pi and CO₂ resulted in a trade-off in the resources allocated to nutrient transporters on the cytoplasmic membrane, depending on the most limiting nutrient (Spijkerman et al., 2011).

When nutrient-limited plants or algae are exposed to a pulse of the limiting nutrient, a change in the chlorophyll a (chl a) fluorescence is often observed (Shelly et al., 2010). This has been termed a nutrient-induced fluorescence transient, or NIFT (Wood & Oliver, 1995), and has been suggested as a fast and non-invasive monitoring tool for nutrient limitation (Beardall et al., 2001) because the chl a fluorescence response is generally not observed if a non-limiting nutrient or distilled water is added to the cells. NIFTs have an advantage over standard enrichment experiments because they give an indication of the immediate nutrient status rather than just indicating potential limiting factors (Beardall et al., 2001). The NIFT response to a transient pulse or ‘spike’ in nutrient begins almost immediately and the full response is observed over a few minutes. With this in mind we hypothesized that the serial co-limitation response previously seen in C. acidophila using nutrient enrichment assays could be further unravelled using this type of NIFT approach. Indeed, recent work has demonstrated the use of NIFTs in (co-)limited seagrass (den Haan et al., 2013).

To this end, we tested for a (co-)limitation in C. acidophila in its natural environment by conventional bio-assays. We applied NIFTs to batch cultures to test for a limitation in C. acidophila by Pi and CO₂, to determine if this limitation can be correlated to the cellular P:C ratio and to ascertain if we can detect a CO₂ (co-)limitation in this green alga via a NIFT response.

MATERIALS AND METHODS

Field sampling

Primary productivity in the acid lakes 107 (pH 2.4) and 117 (pH 3.4; both 51º29′ N 13º38′ E), situated in the east part of Germany, has been much studied (Beulker et al., 2003; Kamjunke et al., 2005) and compared with lakes of similar nutrient status but neutral pH (Nixdorf et al., 2003). A mixed water sample from the epilimnion above the deepest point of Lake 107 was taken on 27 June 2010 and from Lake 117 on 24 July 2010. Directly after sampling, water was filtered through a 200 µm zooplankton net. This separation did not eliminate the important predators in these acidic lakes because they consist of small-sized organisms such as rotifers, heliozoans and the chrysophyte Ochromonas sp. (Kamjunke et al., 2004). Waters from these lakes were used in a nutrient enrichment experiment as described below. In both lakes C. acidophila is the most important primary producer, whereas Ochromonas sp. often builds up a higher biomass but is predominantly phagotrophic (Kamjunke et al., 2004; Spijkerman, 2008a).

Culturing

Batch cultures of C. acidophila (CCAP 11/137; Gerloff-Elias et al. 2005) were grown at 20 ± 1 °C in modified Woods Hole medium at pH 2.65 as described in Spijkerman (2010). A multivariate approach with two different CO₂ and Pi concentrations was applied in duplicate. Cultures were either non-aerated (low CO₂, −C) or gently aerated with 5% CO₂ in air (high CO₂, +C) and were inoculated from a full strength medium (50 µM P) into a medium containing either a Pi concentration of 0 µM P (Pi-depleted, −P) or 50 µM P (Pi-replete, +P). Cultures started with an optical density (OD; 1 cm pathway at 750 nm) of 0.01 and growth in low CO₂ cultures was followed over a 12 d period, though high CO₂ cultures were terminated after 8 d because cell densities were very high in the Pi-replete cultures (i.e. 4.6 × 10⁶ cells ml^–1) and possibly experienced light limitation. We decided to follow growth by OD as this was the most rapid and reliable method and includes changes in both cell number and cell volume. Light intensity at the surface of the Erlenmeyer flasks (Biospherical Instruments Inc. Model QSL-2101, San Diego, California) was 130–140 µmol m^–2 s^–1 and was supplied continuously. Every second day, samples were taken for OD, NIFT response, cellular carbon and phosphorus content, and cell density. We checked for a normal distribution of the data, homogeneity of variance and performed statistical tests in SPSS 20.0 (SPSS for Windows, IBM, Chicago, Illinois). All data passed the prerequisites for a parametric test, and therefore differences between groups were examined with ANOVA, paired measurements with a paired t-test and Pearson correlation.

Nutrient enrichment bioassay

To test for the growth limiting factor in the phytoplankton, lake water was enriched with: (a) 5 µM Pi (as K₂HPO₄), (b) CO₂ provided via aeration with 4.5% CO₂ in normal air (v/v), and (c) a combination of both Pi and CO₂ and incubated at 18 ± 1ºC in a constant temperature room with 16:8 h light:dark cycle. Biomass production was monitored after 8 d via chl a fluorescence (recorded on a TD-700, Turner designs, Bremerhafen, Germany). A control was enriched with de-ionized water and treated similarly. Nutrient enrichments were performed in triplicate.

NIFT measurements

NIFTs were carried out on the batch cultures of C. acidophila using a PhytoPAM fluorometer (Heinz Walz GmbH, Effeltrich, Germany) by recording the chl a fluorescence (F_t) every 3 s without the application of a saturating pulse (Phytowin_v1.47). The actinic light source in the measuring cuvette was set to 120 µmol photons m^–2 s^–1, approximately matching the light intensity at which the algae were grown. Gain settings on the PAM varied between 8 and 18 during the experiment, as it needed to be adjusted for undiluted culture material. We choose to use undiluted culture material so as not to impose stress on the cells, although this could have affected the response to a nutrient spike as the concentration added per cell differed. Chl a fluorescence was first recorded for at least 1 min to obtain a stable value, after which the response to a spike of Pi, CO₂, Ci or distilled water was subsequently monitored for several minutes (Fig. 1). All three additions were performed on sub-samples of each batch culture, but the order was changed between replicates and measuring days. For Pi-addition, unless otherwise stated, we used a final standard concentration of 10 µM KH₂PO₄ as suggested previously (Holland et al., 2004; Roberts et al., 2008), which was prepared in acidified water (pH 2.65, not buffered). To test the effect of CO₂ we added NaHCO₃ to the same acidified water (pH 2.65) and spiked the subculture to a final concentration of 100 µM CO₂ after exactly 60 s of reaction time. This time was calculated as sufficient to give 100% conversion of HCO₃^– to CO₂ within 20 s at this low pH (Brinkman et al., 1934). Finally, direct HCO₃^– pulses were given at the end of the run (final concentration 650 µM) to test for a direct response to Ci (Fig. 1).

Fig. 1. Example of changes in chlorophyll a fluorescence (F_t) after addition of several nutrients or acidified water.

At the end of the growth experiment (on d 8 for –P+CO₂, and d 12 for –P–CO₂), NIFT responses were recorded on Pi-depleted cells by the addition of eight different Pi concentrations, ranging from 0.25–7.5 µM P (as KH₂PO₄). A fresh sample was taken for every measurement to avoid effects from prior additions. An intermediate recovery, or possibly a second NIFT occurred within the NIFT response to a Pi-spike. This intermediate recovery occurred without any external input. After Pi-addition, F_t decreased rapidly, increased again (to a ‘bump’), remained more or less constant or decreased (Fig. 2), and then increased again to regain the starting value of F_t. We calculated the extent of this ‘bump’ (F_t increase) as a percentage of the (initial) NIFT response (F_t decrease) for each Pi concentration.

Fig. 2. Example of secondary effect (‘bump’) in a chlorophyll a fluorescence (F_t) transient typically occurring after applying a moderate Pi-spike.

Chemical analyses and algal density

Cellular P:C ratios were determined by measuring the particulate P and C in the cultures. The particulate P concentration was determined on filtered culture suspension (0.45 μm polysulfon filter, Pall Corporation, Port Washington, New York) after heating in an autoclave for 20 min with 20 mM K₂S₂O₈ and 60 mM H₂SO₄. Measurements were performed spectrophotometrically using molybdate and ascorbic acid and comparing absorbance with a standard curve generated from similarly treated solutions of KH₂PO₄ (Murphy & Riley, 1962). For particulate C analysis a minimum of 10 µg C culture suspension was filtered on pre-combusted GF/F filters (4 h at 450ºC; Whatman), dried at 50ºC for 2 d and subsequently measured in a carbon analyser (EuroVector CHNS-O Elementaranalysator, Wegberg, Germany). Infrared gas chromatography was calibrated against standards in every run.

The OD of cultures was measured, immediately after sampling, on a Cary UV-Vis spectrophotometer (50 Bio, Varian). Cell numbers were determined after finishing the experiment, in Lugol-fixed samples, using an automatic cell counter (CASY 1, Model TT, Schärfe, Reutlingen, Germany).

RESULTS

In the nutrient enrichment bioassays in lake water, biomass production, relative to the control after 8 d, was not enhanced by the addition of CO₂, but was significantly enhanced when Pi was added (Fig. 3, ANOVA, F_2,23 = 91.2, P < 0.001). The phytoplankton of Lake 117 also showed a synergistic enhancement from the addition of CO₂ plus Pi in comparison with Pi alone (Tukey, P < 0.001), although this was not significant for Lake 107 samples (P = 0.74; Fig. 3). Microscopic examination revealed that C. acidophila was the dominant alga in Lake 117 phytoplankton before and after the enrichment, whereas in Lake 107, Ochromonas was present in high densities and likely grazed the potential growth of C. acidophila (biomass densities not shown, but see Spijkerman, 2008a for some examples).

Fig. 3. Relative increase in biomass to the control (untreated water from each lake) in enrichment experiments with water from Lake 107 and Lake 117 after an 8 day incubation. The straight line at 1 illustrates no response in relation to control. The biomass in untreated water increased by a factor of 2.6 and 2.7 in Lake 107 and Lake 117, respectively. Mean ± SD of triplicates (ANOVA, F_2,23 = 91.2, P < 0.001).

In the batch culture experiments performed to test the NIFTs, growth of C. acidophila was highest over the first 4 d (Fig. 4), with the +P+CO₂ treatments growing fastest (0.94 ± 0.02 d^–1) followed by –P+CO₂ (0.65 ± 0.02 d^–1), and then the −CO₂ treatments (both ~0.50 d^–1). The optical density at d 8 was likewise highest in the +P+CO₂ treatments (1.05 ± 0.05, equivalent to 4.6 × 10⁶ cells ml^–1), with the other three treatments being about one-sixth as dense, with little difference between them (OD ~0.15, ~1 × 10⁶ cells ml^–1). By d 14 there was a clear distinction between the cell density in the +P−CO₂ and the −P−CO₂ treatments, with the former approximately 60% higher than the latter (Fig. 4).

There were clear differences between the cellular P:C ratio in batch-grown algae over time when comparing cells grown in Pi-replete or Pi-depleted medium (Fig. 5). Independent of the presence of CO₂-aeration, the P:C ratio rapidly decreased in cells grown in Pi-depleted medium to 0.57 ± 0.11 mmol P: mol C (mean ± SD of values obtained on the last 2 measuring days, i.e. d 6 and 8 for +CO₂ and d 10 and 12 for –CO₂ cultures). In contrast, P:C ratio increased in the cells grown in Pi-replete medium, but after 4 d rapidly decreased in the high CO₂ cultures because Pi became depleted in the medium. The highest P:C ratio was measured in cells grown in Pi-replete medium at low CO₂ on d 8, reaching 22.1 ± 0.5 mmol P: mol C (mean ± SD; Fig. 5) after which the medium Pi concentrations were depleted. P:C ratio values can thus vary at least 40-fold in C. acidophila.

To determine if NIFTs matched responses seen in the nutrient enrichment bioassay in lake water and semi-continuous cultures of C. acidophila (Spijkerman, 2010), batch cultures were tested for their NIFT response throughout the growth period. Two days after inoculation in Pi-depleted medium, the cells showed a NIFT in response to the addition of Pi, with an approximately 20% reduction in chl a fluorescence (Fig. 6). This NIFT response coincided with a 2-fold decrease in P:C ratio (Fig. 5). The extent of the response was the same in high and low CO₂-grown cells and steadily increased with increasing culturing time and, concomitantly, with decreasing P:C ratio. The maximum effect ranged between 30 and 50% of the initial chl a fluorescence at the end of the growth experiment. The order in which the nutrient spikes were supplied (i.e. whether the Pi-spike came before or after the CO₂-spike) had no effect on most NIFT responses (paired t-test, n ≥ 4, P > 0.08), except for the CO₂-spike in the Pi-depleted plus CO₂ cultures (paired t-test, n = 4, t = 10.4, P < 0.05). In these cultures, if Pi was added first, the NIFT for CO₂ was lower than when CO₂ was added first. Although the Pi concentration in the medium of ‘Pi-replete’ plus high CO₂ cultures had decreased after 4 d and P:C ratio decreased to close to that in Pi-depleted cells, cells did not respond to a Pi-spike over the 8 d of incubation. This possibly resulted from the cells entering the stationary phase and becoming light limited.

Furthermore, in the low CO₂-grown batch cultures a reproducible NIFT in response to the addition of CO₂ was also observed, as compared with the lack of response in high CO₂-grown cells (ANOVA, F_4,26 = 13.5, P < 0.001), although it was much less pronounced than the response of Pi-depleted cells to Pi and never exceeded 10% of F_t (Fig. 7). The CO₂-NIFT was the same in both the Pi-replete and Pi-depleted cells after 2 and 4 d of batch growth. Against expectations, the CO₂-NIFT response after d 2 remained the same or slightly increased in the Pi-depleted cells (−P−CO₂), whereas it decreased in the Pi-replete cultures (+P−CO₂) to levels measured in the control (H₂O spikes). This deviation in NIFT response between the +P and −P low CO₂ cultures occurred when the +P−CO₂ cultures reached their maximum P:C ratio (at d 6 and 8; Fig. 5). Although P:C ratio decreased slightly in the +P–CO₂ cultures after d 8, it appeared that a high P:C ratio coincided with a diminished NIFT response to CO₂ (Fig. 7). The response to CO₂ addition in the high CO₂, Pi-depleted cells was the same as that in the control. The –P–CO₂ was the only treatment whose CO₂-NIFTs were significantly different to the response to acidified water (Tukey, P < 0.001), whereas +P–CO₂ only differed at the 10% level (Tukey, P = 0.052). In addition, we observed a ‘reverse NIFT‘ (an increase in Ft) in response to a CO₂-spike in the +P+CO₂ cultures, but only at d 4 (Fig. 7). This effect coincided with a high cellular P:C ratio at that time (Fig. 5), which together with the decreasing NIFT in the +P–CO₂ cultures suggests that a high P:C ratio might result in an increase in F_t after a CO₂ spike (compare Fig. 5 with 7).

We hypothesized that the NIFT response was directly related to P:C ratio, and this was confirmed when we analysed both the response to a Pi-spike in relation to P:C ratio for Pi-depleted cells (Fig. 8, Pearson correlation, r = −0.65, n = 20, P < 0.005) and the response to a CO₂-spike in relation to P:C ratio for the low CO₂-grown cultures (Fig. 9, Pearson correlation, r = –0.43, n = 24, P < 0.05). While care has to be taken in interpreting these correlations in the fast changes in P:C ratio occurring in batch cultures, we found no relationship between NIFTs and the cellular P content (focusing on Pi-depleted cells that contained < 0.2 pg P cell^–1; Pearson correlation r = −0.40, n = 20, P = 0.08; Fig. 10) or to their cellular C content (Pearson correlation r = 0.42, n = 20, P = 0.07; Fig. 11) in response to a Pi-spike (see Supplementary Figs 1 & 2 for the total data set). These results suggest that, unexpectedly, both the cellular P and C content equally contributed to the NIFT response to Pi.

We came to a different conclusion after comparing the response to a CO₂-spike with the cellular contents of the low CO₂-grown cells (Figs 12, 13); we found no correlation with the cellular C content (Pearson correlation r = 0.10, n = 24, P = 0.65; Fig. 12) but a negative correlation with the P content (Pearson correlation r = −0.52, n = 24, P < 0.05; Fig. 13). These results suggest that the cellular C content had no influence on the NIFT response to CO₂, coinciding with the absence of response to CO₂ enrichment in the bioassays, but that a higher cellular P content decreased the NIFT.

The addition of a range of different concentrations of Pi to Pi-depleted batch-grown cells resulted in a concentration-dependent response that followed Michaelis–Menten kinetics (Supplementary Fig. 3 & Supplementary Table 1). An intermediate recovery in chl a fluorescence occurred within the NIFT response to a Pi-addition, which we denoted as a ‘bump’ (Fig. 2). This bump was an intrinsic physiological process as it occurred without the further addition of nutrients. The extent of this bump, expressed as a percentage of the total, initial NIFT response, varied with the concentration of the Pi-spike (Fig. 14). In this case it is important to notice that we kept the volume of the spike constant (10 µl), and thereby the applied CO₂ concentration was also constant. When higher Pi concentrations were applied (> 2.5 µM P in low CO₂ and > 0.75 µM P in high CO₂ cultures), then the second decrease in F_t resulted in a lower F_t than before the bump. The extent of the bump as a proportion of the NIFT response directly following Pi-addition first increased with increasing Pi concentration up to a concentration of 0.5 µM Pi and then decreased with higher concentrations (Fig. 14). The extent of the bump was 2-fold higher in low CO₂ cells (maximum 53%) than in high CO₂ cells (maximum of 22%), a result that could not have been influenced by the different density of the culture, as high CO₂, Pi-depleted cultures were only 2.2% more dense than low CO₂, Pi-depleted cultures as determined by OD at the time of measurement.

Figs 4, 5. Changes in the optical density (OD; Fig. 4) and cellular P:C ratio (Fig. 5) of Chlamydomonas acidophila grown in batch cultures over time when cultured in four different combinations of Pi and CO₂ concentrations. High CO₂, Pi-replete cultures ran out of P on day 4 and low CO₂, Pi-replete cultures on day 8. All Pi was determined to be inside the cells on day 2 in the Pi-starved cultures. Mean ± SD of duplicates.

Figs 6, 7. Extent of NIFT response to a Pi-spike (Fig. 6) and CO₂-spike (Fig. 7) of C. acidophila grown in batch cultures over time. Mean ± SD of duplicates. In addition the response to acidified water is shown in both graphs.

Figs 8, 9. Extent of NIFT response to a Pi-spike (10 µM) as a function of P:C ratio in Pi-depleted batch cultures (Fig. 8) and extent of NIFT response to a CO₂-spike (100 µM) as a function of P:C ratio in low CO₂ batch cultures (Fig. 9) of Chlamydomonas acidophila. Lines represent significant correlations (Pearson correlation; r = −0.65, P < 0.005: Fig. 8 and r = −0.43, P < 0.05: Fig. 9).

Figs 10, 11. Extent of NIFT response to a Pi-spike (10 µM) as a function of cellular P content (Fig. 10) and cellular C content (Fig. 11) in Pi-depleted batch cultures of Chlamydomonas acidophila.

Figs 12, 13. Extent of NIFT response to a CO₂-spike (100 µM) over cellular C content (Fig. 12) and cellular P content (Fig. 13) in low CO₂ batch cultures of Chlamydomonas acidophila. In Fig. 13, line represents a significant (Pearson correlation; r = −0.52, P < 0.05) correlation.

Fig. 14. Extent of the bump in relation to Pi-spike as measured in Pi-depleted cells of Chlamydomonas acidophila cultured under two different CO₂ concentrations. Values are the mean and SD of duplicates.

DISCUSSION

The bioassays indicate that growth of the phytoplankton of Lake 117 was co-limited for Pi and CO₂, similar to that observed for semi-continuous cultures of C. acidophila (Spijkerman, 2010). The phytoplankton of Lake 117 was dominated by C. acidophila (> 60%), whereas that of Lake 107 was not (< 20%). The phytoplankton of Lake 107 did not show the synergistic response to CO₂ and Pi, which can be attributed to the low density of C. acidophila in relation to its predator, Ochromonas sp. (Tittel et al., 2003; Spijkerman, 2008a). We detected a synergistic response in C. acidophila, which is consistent with a serial co-limitation (Harpole et al., 2011) because no effect was observed after a single CO₂ addition and a significant primary Pi-limitation was present. Since addition of CO₂ alone made no difference to the biomass after 8 days, cells were unlikely to be CO₂-limited though there were clear interactions between CO₂ and Pi. Bioassays have been criticized in the past for indicating which nutrients may become limiting rather than those that are limiting in situ at the time of sampling (Holland et al., 2004). Thus, we applied the potentially more immediate technique of NIFTs and analysed the extent and shape of the response to Pi and CO₂ in the extremophile C. acidophila.

NIFT-type responses have been observed in several algal species and phytoplankton communities and in response to several different nutrient limitations. For example, a NIFT response to nitrogen (N) has been observed in the green algae Selenastrum minutum (Birch et al., 1986; Turpin & Weger, 1988; Holmes et al., 1989) and Dunaliella tertiolecta (Turpin, 1983; Young & Beardall, 2003) and in the cyanobacterium Microcystis aeruginosa (Wood & Oliver, 1995). Responses to Pi-additions have been recorded from Dunaliella tertiolecta (Roberts et al., 2008) and Selenastrum minutum (Gauthier & Turpin, 1997), and even in spinach chloroplasts (Cerovic et al., 1991). There have been a limited number of studies showing a NIFT response to Ci additions in Dunaliella tertiolecta (Ihnken et al., 2014), Chlamydomonas reinhardtii (Lucker & Kramer, 2013) and the cyanobacteria Nostoc sp. (Qiu & Liu, 2004) and Synechococcus UTEX 625 (Miller et al., 1991). In the latter study, quenching of chl a fluorescence was directly correlated with active CO₂ uptake and the size of the internal inorganic carbon pool (Miller et al., 1991). There is also some evidence that NIFTs may occur with silicate (Lippemeier et al., 2001). However, none of these previous NIFT studies considered co-limitation effects.

Effect of P

We showed pronounced NIFT responses after a Pi-addition to Pi-depleted batch-grown cells (those with P:C ratio < 3 mmol P: mol C). After two days of Pi-starvation, the NIFT was already significant, as was the decrease in P:C ratio, and with increasing Pi-starvation the effect increased up to 40% of the F_t at the start of the perturbation. Such responses fall well within those observed in non-acidophilic algal species, such as a maximum quenching of 45% in the green alga Dunaliella tertiolecta (Roberts et al., 2008) and 35% in Selenastrum minutum (Gauthier & Turpin, 1997). The NIFT response thus is also a valuable tool to detect Pi-limitation in the extremophile green alga C. acidophila. NIFTs are easy to measure, quick and non-invasive, and provide a similar response to that obtained by the more labour-intensive bioassays and/or measurements of photosynthetic or dark respiration rates. The latter, for example, has been shown for Pi-limited Selenastrum minutum, which increased its dark respiration rate 2–2.5-fold after a Pi-addition (Gauthier & Turpin, 1994) and exhibited a decrease in net O₂ evolution by 26% (Gauthier & Turpin, 1997). Pi-limitation has also been detected using delayed chl a fluorescence in Scenedesmus obtusiusculus (Mellvig & Tillberg, 1986). In delayed chl a fluorescence (emitted light after a period of illumination) the effect of a nutrient limitation does not primarily affect the extent of the chl a fluorescence response, but rather the timing of the delayed fluorescence peak (Berden-Zrimec et al., 2011), as shown in Dunaliella tertiolecta in response to N-limitation (Berden-Zrimec et al., 2008).

After two days of growth, Pi-replete cells were growing at high rates (0.94 ± 0.02 d^–1 and 0.64 ± 0.03 d^–1 for high and low CO₂ cultures, respectively) that lie well within the range reported previously (Spijkerman, 2005, 2008b; Tittel et al., 2005). At that time, cells growing in Pi-depleted medium grew at a similarly high rate (0.86 ± 0.07 d^–1 and 0.60 ± 0.05 d^–1 for high and low CO₂ cultures, respectively), but already showed a pronounced NIFT response (Fig. 6). Our observations thus indicate the NIFTs can provide early indications of the onset of limitation when other measurements show no effect. For instance, our data can be contrasted with the absence of photosynthetic suppression in response to additions of the limiting nutrient in ammonium limited Dunaliella tertiolecta when the growth rate exceeded c. 80% of its maximum (Turpin, 1983). In C. acidophila the response to the Pi-spike was related to the decrease in its P:C ratio in that the response appeared to be switched on when the P:C ratio fell below 3 mmol P: mol C. Against expectations based on a clear positive relation between maximum photosynthetic rates and P:C ratio (Spijkerman, 2010), cellular P and C alone did not contribute to the extent of a Pi-NIFT.

‘The bump’

An interesting phenomenon during the NIFT response to Pi-addition was observed in C. acidophila that has not been previously recorded in other algal species using a similar set-up (Fig 2): an intermediate recovery, or possibly second NIFT occurred within the NIFT response to Pi-addition. In other green microalgae, F_t steadily increases after reaching the minimum F_t (Holland et al., 2004; Roberts et al., 2008). However in C. acidophila a second decrease after a short recovery (‘the bump’) was always observed, although it was hardly visible when higher (i.e. 10 µM Pi) Pi-concentrations were applied. The extent of the bump was related to the magnitude of the Pi-spike, with an optimum at 0.5 µM Pi. In addition, the extent of the bump was also related to the CO₂ concentration in the medium, as it was two-fold higher in low CO₂ cells (maximum 50%) than in high CO₂ cells (maximum of 22%).

Possibly, the ‘bump’ is related to similar observations made in isolated chloroplasts from spinach (Cerovic et al., 1991). Cerovic and co-authors showed a similar chl a fluorescence rise as we have shown here, which they suggested to be characteristic of photosynthetic induction in leaves. Cerovic et al. (1991) stated that the ‘bump’ was a consequence of an inadequate rate of recycling of Pi by the cytosol. After the first decrease in F_t, the quenching relaxes as the rate of Pi uptake declines (Roberts et al., 2008) and the rate of electron transport increases. However, at low Pi concentrations the maximal rate of Pi recycling can only be attained transiently, resulting in a second transient rise in chl a fluorescence. Consequently, the quenching would thus be related to a temporary decrease in cellular ATP content associated with ATP-demanding Pi-uptake processes. Fast changes (occurring within 5 seconds) in cellular ATP pool sizes and adenylate energy charges (lasting for a few minutes) have been nicely illustrated in Pi-limited Selenastrum minutum in response to a Pi-spike (Gauthier & Turpin, 1994). Furthermore in transients following ammonium additions to nitrogen limited algae, a marked drop in endogenous ATP has been shown (cited in Turpin, 1983), though Turpin points to a range of acclimation processes that might contribute to observed changes in chl a fluorescence. Although the precise physiology behind NIFTs is still poorly understood (Petrou et al., 2008), it is believed that the transient change in chl a fluorescence is, in part at least, attributable to a reallocation of energy from photosynthesis to nutrient uptake (Holland et al., 2004). For Pi there is a suggestion that if cytosolic Pi concentrations increase during uptake, then rapid ATP export from the chloroplast may deplete stromal metabolite pools, decreasing regeneration of Calvin cycle intermediates and, subsequently, Calvin cycle activity (Heineke et al., 1989; Gauthier & Turpin, 1997).

There was a clear relationship between the concentration of Pi and the extent of the bump in C. acidophila, and the bump was higher in low CO₂ cells, which suggests inorganic carbon transport might also be involved. Similar to what has been observed in isolated chloroplasts (Cerovic et al., 1991), the bump probably relates to transporter processes on the chloroplast envelope and the described interaction between Pi supply and CCM activity (Beardall & Giordano, 2002; Raven & Beardall, 2014). Inorganic carbon transport across the chloroplast membrane towards RUBISCO is thought to be an active process in C. acidophila (Spijkerman, 2011) and could in some way thus interact with the P-translocator (Gauthier & Turpin, 1997). The concentration at which the bump is at a maximum is, however, the same in high and low CO₂ acclimated cells, suggesting that the underlying mechanism might not depend on the activation of a CCM (Spijkerman, 2005) but rather on the Pi concentration in the cytosol. Possibly, the latter also regulates the extent of the CCM in C. acidophila (Spijkerman, 2011). This requires further investigation.

Effect of C

The NIFT response to a CO₂-addition in CO₂-limited algae was less pronounced but also significant. CO₂ is a potentially growth-limiting nutrient as C. acidophila is an extremophilic algae living in very low pH environments (Tittel et al., 2005). At a pH of 2.6, as used for culturing in this study, CO₂ is the predominant form of Ci (> 99%; Stumm & Morgan, 1970) and thus we assume that CO₂ was the sole Ci source used by intact algae. The NIFT response to CO₂ (and also to HCO₃^–) was clearly related to CO₂ limitation because it was present in low CO₂ grown cells but absent in high CO₂ cells. Interestingly there appeared to be a negative relationship between cellular P:C ratio and the response to CO₂ in these low CO₂ grown cells. In addition, we found an increase in F_t after a CO₂ perturbation in cells with a high P:C ratio, as was found for perturbation by NO₃^– in N-limited algae (Young & Beardall, 2003). The response to CO₂ spikes in the +P–CO₂ cultures thus suggests the presence of compensatory dynamics that dampened the sole effect of CO₂ in this treatment. This conclusion was confirmed by the observation in the –P+CO₂ cultures that a first Pi-spike (resulting in an increased cellular P content) decreased the response to a CO₂ spike. Our observations contrast with those made in the seagrass L. variegata where the order of spike addition did not matter (den Haan et al., 2013). As with the addition of Pi to Pi-starved green algae, CO₂-uptake in CO₂-limited C. acidophila might be an ATP demanding process resulting in a state transition from state 1 to 2 (Gauthier & Turpin, 1997; Petrou et al., 2008; Roberts et al., 2008) and/or an enhancement of cyclic electron transport as shown in Chlamydomonas reinhardtii (Lucker & Kramer, 2013). The fast NIFT responses recorded here (the F_t responses to Pi and CO₂-spikes were approximately four times faster (100 s to minimum F_t compared with 400–500 s) than those of Dunaliella tertiolecta; Petrou et al., 2008), suggest that energy quenching and/or cyclic electron transport is of higher relative importance than state transitions in C. acidophila. Possibly, and especially if high concentrations of P are accumulated in the cell, demands on linear electron transport might dominate Ci-assimilation and result in an increase in F_t after a CO₂-spike. An increase in F_t has been shown after a NO₃^–-spike to N-limiting Dunaliella tertiolecta, and was supposedly related to enhanced linear electron transport (Young & Beardall, 2003).

In the cyanobacterium Synechococcus UTEX 625, a large NIFT response to Ci addition was observed that was related to active CO₂ uptake (Miller et al., 1991) and Na⁺-dependent HCO₃^– uptake (Crotty et al., 1994). A large NIFT response was also detected in the cyanobacterium Nostoc sp. (Qiu & Liu, 2004). In this Nostoc sp. the quenching of chl a fluorescence (both the rate and the extent) after a KHCO₃ spike was related to the concentration of the spike and also related to the net photosynthetic activity (Qiu & Liu, 2004). These studies concluded that the rate of chl a fluorescence quenching reflects active Ci uptake, whereas the extent of the chl a fluorescence quenching reflects the internal Ci pool (Miller et al., 1991). The small extent of the response in C. acidophila coincides with the results of a recent study showing that the extent of a CCM (measured as the accumulation of Ci internally, relative to external concentrations) never exceeded an accumulation factor of 10 (Spijkerman, 2011; Spijkerman et al., 2014), although CO₂ uptake was an active process that was > 95% inhibited by incubation in the dark, or in the light but in the presence of DCMU (see supplementary material in Spijkerman, 2011). In contrast to expectations, changes in the cellular C content could not explain the extent of the CO₂-NIFTs, whereas cellular P had a negative effect on this response.

Effect of co-limitation

Antagonistic responses might result from NIFTs performed on co-limited algae, but in C. acidophila the response to Pi addition in cells limited in Pi and CO₂ (Spijkerman, 2010) appeared similar to the response in other green algae that were only Pi-limited (Roberts et al., 2008). We thought that an enhanced NIFT response might be observed in co-limited algae when both nutrients are added in concert, or close together. When studying the growth response of C. acidophila to the addition of the limiting nutrient, providing CO₂ alone had no effect and only the combination of Pi and CO₂ resulted in a higher growth response than when only Pi was added (Spijkerman, 2010). We therefore expected the NIFT response to a CO₂ spike to be higher when provided after the Pi-spike than vice versa. However, in putatively co-limited cultures (Pi-depleted, low CO₂) the NIFT response to a CO₂-spike was the same when added before or after the Pi-spike, and only in Pi-depleted, high CO₂ cultures was an effect of the CO₂-spike observed; thus rejecting our hypothesis.

Another response that might reveal co-limitation is the bump within the NIFT response to Pi-addition, which may be related to the CO₂-concentration in the culture. The bump, potentially a second NIFT, might be a response to a physiological acclimation to CO₂ in the culture because the bump was higher in low CO₂ than in high CO₂ cells. The transient response within the NIFT to Pi thus suggests a dependence in CO₂-limitation on Pi-limitation, which was confirmed by the correlation between the NIFT response to CO₂-addition and cellular P status. More experiments will be necessary to fully reveal the effect of co-limitation on the NIFT response, possibly including delayed chl a fluorescence recordings and quenching analyses in cells grown in well-defined continuous culture conditions.

NIFTs are a quick and reliable method to detect the growth limiting nutrient or nutrients, as seen in in CO₂ and Pi/CO₂ (co-)limited algae as well as in N/P (co-)limited seagrass (den Haan et al., 2013). Intermediate concentrations of a Pi-spike might provide information about additional effects of CO₂ in co-limited algae after further testing on CO₂-(co-)limited phytoplankton (Jansson et al., 2012). Although the synergistic growth response to Pi and CO₂ in C. acidophila suggests a serial co-limitation, the NIFT response reveals that the time scale of this response lies within a minute. We show that NIFTs after a spike of the limiting nutrient did not solely depend on the cellular content of that particular nutrient but were a response to a complex interaction between different nutrients and that the ability to take up low CO₂ concentrations may depend on the availability of Pi. The possibility that high CO₂ simply down-regulated CCM activity, freeing up additional energy for Ci fixation and Pi acquisition, needs to be tested but since CO₂ addition alone (Fig 4) did not support a large increase in biomass, we think this is unlikely.

Thus in summary, we show with traditional bioassays that growth of the phytoplankton of Lake 117, which was dominated by C. acidophila (> 60%), was co-limited for Pi and CO₂. In batch cultures we showed pronounced NIFT responses after Pi-addition when cells had a cellular P:C ratio < 3 mmol P: mol C. Also in response to CO₂ a NIFT response was recorded, that could not be explained by the cellular C content but was inversely related to the cellular P content. The additional observation that medium CO₂ concentration affected the shape of the response to a Pi-addition in a second transient, suggests that the cellular P content mainly controls a co-limitation for Pi and CO₂. However, care must be taken when interpreting the correlations between cellular contents and the NIFT response in the fast changing environment of batch cultures.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author(s).

SUPPLEMENTARY INFORMATION

The following supplementary material is accessible via the supplementary Content tab on the artivle’s online page at http://dx.doi.org/10.1080/09670262.2015.1095355

Supplementary Table 1. Pi-uptake kinetics of Chlamydomonas acidophila grown in Pi-limiting semi-continuous cultures (³³P uptake) or Pi-limited batch cultures (fluorescence; this study) at 2 different CO₂ concentrations

Supplementary Figs 1-2. Extent of NIFT response to a Pi-spike (Fig. 1: 10 μM) and CO₂-spike (Fig. 2: 100 μM) plotted versus the cellular P content (Fig. 1) and cellular C content (Fig. 2) of Chlamydomonas acidophila cultured under 4 different combinations of Pi and CO₂ conditions.

Supplementary Fig. 3 The rate of change in fluorescence (ΔF% min^-1) of Chlamydomonas acidophila grown in Pi-limited batch cultures and 2 different CO₂ concentrations in relation to the concentration of the Pi-spike.

ACKNOWLEDGEMENTS

ES & SL thank Sabrina Ryl and Barbara Schmitz for practical assistance. Two anonymous reviewers provided helpful comments on the manuscript.

Additional information

Funding

ES & SL acknowledge support from the German Research Foundation, DFG [SP695/4-2 & SP695/5-1]. Work in JB’s laboratory on inorganic carbon uptake and interactions with nutrient availability was supported by the Australian Research Council.

Notes on contributors

Daryl Holland

E. Spijkerman, S. Stojkovic, D. Holland, J. Beardall: designing experiments; E. Spijkerman, S. Stojkovic, D. Holland: culture experiments; E. Spijkerman, S.C. Lachmann, J. Beardall: data analysis and interpretation; E. Spijkerman, S. Stojkovic, D. Holland, S.C. Lachmann, J. Beardall: writing manuscript.