Results and discussion

Shipboard iron-light manipulation experiment

The starting seawater collected for this experiment contained high concentrations of dissolved inorganic macronutrients that were typical of surface waters in the Ross Sea during early summer: -20 iM nitrate + nitrite, -1.5 iM phosphate, and -55 iM silicic acid. Analysis of iron in 0.4-iM filtered subsamples of the starting seawa-ter using methods described by Sedwick et al. (2005) revealed an initial dissolved Fe concentration of 0.38 ± 0.03 nM (n = 2). Thus, our additions of DFOB to the low-iron treatments were nearly an order of magnitude higher than the ambient dissolved iron concentration, on a molar basis. Initial phytoplankton biomass was relatively high, with -4 ig l1 chlorophyll a. Shipboard microscopy and subsequent HPLC measurements of the molar ratios of 19'-hexa-noyloxyfucoxanthin to fucoxanthin (-3.8) and chlorophyll c3 to chlorophyll a (-0.3) in the starting seawater indicate that algal biomass was overwhelmingly dominated by Phaeocystis antarctica (DiTullio and Smith 1996; Van Leeuwe and Stefels 1998; DiTullio et al., this issue).

The results from this shipboard incubation experiment are presented in Fig. 2, which shows the concentrations of dissolved nitrate + nitrite, particulate chlorophyll a, and particulate 19'-hexanoyloxyfucoxanthin:fucoxanthin in the incubation bottles during the course of the 80-h experiment. There was a significant decrease (drawdown) in dissolved nitrate + nitrite in all treatments during the experiment (Fig. 2a), relative to the starting seawater, suggesting net growth and accumulation of biomass in all treatments. There was no statistically significant difference between nitrate + nitrite drawdown among the different iron treatments at a given light level, suggesting that algal community growth rates were not increased by iron addition, nor decreased by DFOB addition. However, after 80 h incubation there was a significantly lower (P < 0.05) concentration of nitrate + nitrite (mean = 14.6 ± 0.6 iM) in the six high-light treatments than in the six low-light treatments (mean = 16.6 ± 0.2 iM), implying higher net growth rates in the bottles that were exposed to an irradiance higher than estimated in situ levels.

The chlorophyll a data (Fig. 2b) also show no significant difference between the various iron treatments for a given light level after 80 h incubation, although significantly lower (P < 0.05) levels of chlorophyll a (mean = 6.1 ± 1.6 ig l-1) were measured in the high-light treatments relative to concentrations in the low-light incubations (mean = 9.5 ± 0.5 ig l-1). If chlorophyll a were to be used as a proxy for cell biomass, this result would run contrary to our conclusions based on nitrate drawdown. A likely explanation for this discrepancy is that chlorophyll a concentrations in the high-light treatments reflect light-driven changes in cellular chlorophyll a content, with cells producing less chlorophyll a in

Fig. 2 Concentrations of (a) dissolved nitrate + nitrite, (b) particulate chlorophyll a, and (c) ratio particulate 19'-hexanoloxyfucoxanthin/fucoxanthin (hex/fuco) in the incubation bottles versus incubation time, from the shipboard iron-light manipulation experiment

the high-light treatments relative those incubated at low (approximately in-situ) irradiance, consistent with the experimental results of Van Leeuwe and Stefels (1998). This finding suggests that chlorophyll a is not a suitable proxy for cell biomass in such bioassay experiments with P. antarctica.

The ratio of 19'-hexanoloxyfucoxanthin (hex) to fucoxanthin (fuco) in the experimental bottles (Fig. 2c) showed a considerable range after 80 h incubation, from -0.7 to -3.2, with a general decrease relative to the ratio of -3.8 in the starting seawater. In contrast to the data trends for nitrate + nitrite and chlorophyll a, there were no significant differences in the mean hex/fuco ratio of the low-light versus high-light treatments. However, there were statistically significant differences in the hex/fuco ratio between the high-iron, control-iron, and low-iron treatments after 80 h incubation (one-way ANOVA), with lower hex/fuco ratios observed in the Fe-amended bottles. These differences may reflect an increase in the relative abundance of diatoms following addition of iron (DiTullio and Smith 1996; Sed-wick et al. 2000), and/or iron-driven changes in the hex/fuco ratio of Phaeocystis cells (Van Leeuwe and Stefels 1998; DiTullio et al., this issue). However, there was no significant drawdown of silicic acid in any of the treatments (data not shown), implying little growth by diatoms. This is consistent with our visual observations, and suggests that colonial P. antarctica dominated the algal biomass that accumulated in all incubation bottles. Thus the lower hex/fuco ratios in the Fe-amended bottles most likely reflect changes in the hex/fuco ratio of Phaeocystis cells, a conclusion corroborated by the results of pigment analyses from our laboratory dose-response iron-addition experiment (see next section and DiTullio et al., this issue).

Additions of nanomolar concentrations of the iron-chelating ligand DFOB to open-ocean algal assemblages have been shown to reduce the concentration of iron available to many types of phytoplankton (Wells et al. 1994; Wells 1999; Hutchins et al. 1999; Timmermans et al. 2001). However, there is evidence that some algal species can access DFOB-bound iron (Soria-Dengg and Horstman 1995; Hutchins et al. 1999; Maldonado and Price 1999; Eldridge et al. 2004; Maldonado et al. 2005). A potentially important observation from our shipboard experiment is that DFOB, when added to incubation bottles in significant excess (-10x) over dissolved iron, did not appear to limit the growth of colonial Phaeocystis antarctica over the course of the experiment. The majority of dissolved Fe in open-ocean surface waters is thought to be complexed by an excess of uncharacterized Fe(III)-binding organic ligands that have conditional stability constants (KFeLFe3+) of around 1020-1024 M-1 (Gledhill and van den Berg 1994; Rue and Bruland 1995,

1997; van den Berg, 1995; Wu and Luther 1995; Boye et al. 2005). A significant fraction of this organically complexed iron is thought to be available to phytoplankton (Rue and Bruland 1997; Hutchins et al. 1999; Maldonado and Price 1999, 2001; Boye et al. 2005; Maldonado et al. 2005). Assuming this to be the case for surface seawater at our Ross Sea station A, then the addition of 3 nM DFOB (which has a KFeLFe3+ value similar to or higher than that of Fe-binding ligands in seawater; Rue and Bruland 1995; Maldonado et al. 2005) would have sequestered a significant fraction of the dissolved iron in the DFOB-Fe complex. Thus our experimental results imply that DFOB-bound iron, and potentially other organic-iron complexes, are, to some extent, biologically available to colonial P. antarctica. The mechanism by which Phaeocystis might access this DFOB-bound Fe remains unknown, but it may reflect the ability of hapto-phytes to produce specific Fe-binding ligands (C. Trick, personal communication, 2006).

Although the addition of DFOB in excess of dissolved iron had no clear effect on the growth rate of colonial Phaeocystis antarctica, the data shown in Fig. 2c suggest that DFOB addition did result in a higher hex/fuco ratio in P. antarctica relative to cells grown in the control treatments, presumably in response to complexation of dissolved Fe by the DFOB. This observation is consistent with experimental results reported by Van Leeuwe and Stefels (1998), and with the pigment data from our laboratory dose-response iron-addition experiment (DiTullio et al., this volume), which indicate that the hex/fuco ratio of P. antarctica cells may vary significantly in response to changes in the concentration and/or speciation of dissolved iron, even though dissolved iron concentrations are not low enough to limit algal growth rate. Van Leeuwe and Stefels (1998) suggest that P. antarctica may convert fuco to hex as a photoprotective mechanism under conditions of decreased iron availability. Thus the higher hex/fuco ratio observed in our DFOB-amended incubation bottles, relative to control samples, may reflect photoprotective adaptation to a decrease in the concentration of readily available dissolved Fe.

In summary, the results of our shipboard iron-light manipulation experiment suggest that: (1) the proximate control on the growth of the colonial P. antarctica at station A was irradiance, with higher growth rates observed in colonies exposed to irradiance higher than the range of estimated in-situ values; and (2) the ambient dissolved Fe levels of -0.38 nM were sufficient to meet the growth requirements of colonial Phae-ocystis in surface waters at this site. However, one caveat that must be considered here is the relatively short duration (-3 days) of this incubation experiment, which was imposed by our cruise schedule. Results of other shipboard incubation experiments in the Ross Sea (e.g., Martin et al. 1990; Sedwick et al. 2000; Coale et al. 2003) have shown time lags of 2 or more days between iron amendment and biological response, in terms of increases in algal biomass and drawdown in macronutrients. Thus we cannot say definitively that the growth rate of colonial P. antarctica collected at station A did not change in response to additions of Fe or DFOB.

Laboratory dose-response iron-addition experiment

As well as the results of the shipboard iron-light experiment described in the preceding section, a preliminary laboratory experiment that we performed using Phaeocystis isolated from Ross Sea station 25 provided further evidence that colonial P. antarctica are able to access iron from seawater in which DFOB is present in -10x excess of dissolved iron concentration (Garcia et al., manuscript in preparation). This finding, and evidence for the ability of phytoplankton to access Fe bound by other organic ligands, including EDTA, raise concerns regarding the use of chelating compounds, such as EDTA and DFOB, to maintain low concentrations of biologically available iron in algal culture experiments (e.g., Anderson and Morel 1982; Brand et al. 1983; Sunda et al. 1991; Sunda and Huntsman 1995, 1997). Indeed, Gerringa et al. (2000) have eloquently argued against the use of EDTA or other added ligands to control dissolved inorganic Fe(III) in such experiments.

With this issue in mind, we designed the laboratory iron-addition experiment that is described in the methods section ''Laboratory dose-response iron-addition experiment'', whereby our starting culture of colonial Phaeocystis antarctica was successively diluted with low-iron (-0.17 nM dissolved Fe) station B seawater using stringent trace-metal clean techniques. We thus produced a unialgal culture of predominantly colonial P. antarctica in low-iron (-0.2 nM) seawater containing a negligible concentration of EDTA (-0.02 nM), which was then used in a dose-response iron-addition experiment (see the methods section ''Laboratory dose-response iron-addition experiment''). To our knowledge, this represents the first successful application of an iron-clean culture experiment using colonial Phaeocystis antarctica in a land-based laboratory. A similar method has been applied to unialgal cultures of diatoms in shipboard experiments (Timmermans et al. 2001), and, more recently, in land-based laboratory experiments (Timmer-mans et al. 2004).

Visual observations indicated that Phaeocystis biomass was dominated by nearly spherical colonies in all incubation treatments during the course of our 31-day incubation experiment. Figure 3 shows the time course of dissolved nitrate + nitrite and particulate chlorophyll a concentrations in the incubation bottles. After 31 days incubation, the iron-amended bottles (except for the +1.8 nM Fe treatment) displayed a statistically significant decrease (drawdown) in dissolved nitrate + nitrite relative to the control treatments (Fig. 3a), with the +0.6 nM Fe treatments achieving the greatest mean drawdown in nitrate + nitrite (6.5 pM) relative to the starting seawater. Assuming that the drawdown in nitrate + nitrite reflects net accumulation of cellular biomass in the incubation bottles (see section ''Interpretation of experimental results''), we conclude that the growth rate of colonial P. antarctica in the starting seawater, which contained 0.22 nM dissolved Fe, was limited by iron deficiency. The chlorophyll a data (Fig. 3b) also indicate Fe-driven increases in chlorophyll biomass, with the highest mean chlorophyll a concentration (5.3 ig l-1) achieved in the +0.6 nM Fe treatment after 31 days. These chlorophyll a data are wholly consistent with the nitrate + nitrite results, implying that iron addition did not mediate large changes in cellular chlorophyll a. This result supports our assertion that changes in cellular chlorophyll a during the shipboard iron-light incubation experiment (see previous section) were largely driven by changes in irradiance.

Our experimental data indicate that the cellular biomass of colonial Phaeocystis increased significantly in response to added Fe, with the notable exception of the +1.8 nM Fe treatments, which displayed a lesser biological response than the +0.2 nM and +0.6 nM Fe treatments after 31 days incubation. This unexpected result for the +1.8 nM Fe treatment may reflect an effective solubility limit for dissolved Fe in such experiments, whereby bottles with an initial dissolved Fe concentration greater than -1-2 nM (in the absence of EDTA) experience a significant loss of biologically available Fe during the incubation period, via precipitation of iron hydroxides (Kuma et al. 1996, 1998; Nakabayashi 2002). Following this reasoning, we suggest that the final concentration of dissolved Fe in the +1.8 nM Fe treatments was perhaps much less than the initial concentration of 2.02 nM. In support of this explanation, we note that the concentration of dissolved Fe rarely exceeds -1-2 nM in open-ocean surface waters, even in the presence of much higher concentrations of particulate iron (e.g. see Sedwick and DiTullio 1997; Brown et al. 2005; Sedwick et al. 2005). This effective solubility limit for dissolved Fe in ocean surface waters is thought to be determined by the concentration of naturally occurring Fe(III)-binding organic ligands, which allow for dissolved Fe(III) concentrations that far exceed the solubility of iron hydroxides in organic-free seawater (Kuma et al. 1996, 1998; Wu et al. 2001; Liu and Millero 2002; Nakabayashi 2002).

We have estimated nitrate-specific net growth rates (iN) of Phaeocystis for each experimental treatment between days 25 and 31, using the mean values of nitrate + nitrite drawdown, relative to the starting seawater, at these timepoints. This calculation assumes that: (1) growth was exponential in all bottles between days 25 and 31, and (2) decreases in nitrate + nitrite were directly

Fig. 3 Concentrations of (a) dissolved nitrate + nitrite, (b) particulate chlorophyll a, in the incubation bottles versus incubation time, from the laboratory dose-response iron-addition experiment. Data are plotted as the mean value of duplicate treatments, with error bars showing ± one standard deviation from the mean

Fig. 3 Concentrations of (a) dissolved nitrate + nitrite, (b) particulate chlorophyll a, in the incubation bottles versus incubation time, from the laboratory dose-response iron-addition experiment. Data are plotted as the mean value of duplicate treatments, with error bars showing ± one standard deviation from the mean proportional to net increases in the standing stock of colonial Phaeocystis cells (see the section ''Interpretation of experimental results''). In Fig. 4, we plot these ,N values against the initial concentration of dissolved Fe in the incubation bottles. In other studies that have used bioassay experiments to examine the relationship between algal growth rate and dissolved Fe concentration (e.g. Coale et al. 1996, 2003; Blain et al. 2002; Timmermans et al. 2001, 2004), the experimental data have been described by a Monod (Michaelis-Menten) saturation function of the form i = imax[dFe/(K, + dFe)], where , is the net growth rate, dFe is the dissolved Fe concentration in the experimental treatments, ,max is maximum Fe-replete growth rate, and K, is the half-saturation constant for growth with respect to dissolved Fe. An important assumption of these studies is that dissolved Fe concentrations do not change significantly during the course of growout experiments; as discussed above, we suggest that this assumption is not valid for our +1.8 nM Fe treatments. If we exclude results from the +1.8 nM Fe treatments on this basis, then a Monod function can be fitted to the experimental data plotted in Fig. 4.

To do this we use the Eadie-Hofstee linear transformation of the Monod hyperbolic function, which is a relatively robust approach with regard to error-prone data such as our iN values, because it gives equal weight to all data points (Zivin and Waud 1982). The Eadie-Hofstee plot of iN versus iN/dFe yields a least-squares best fit with a slope (=-KM) of -0.45 nM, a y-intercept (=1max) of 0.33 d-1, and an r2 value of 0.85. This exercise yields a half-saturation constant for growth (KM) of 0.45 nM dissolved Fe for colonial Phaeocystis antarctica grown at an irradiance of

This K,, estimate is within the range of experimental KM estimates that Timmermans et al. (2004) report for four species of Southern Ocean diatoms grown at an irradiance of 60 iE m-2 s-1 (Km = 0.19-1.14 nM Fe), but it is more than an order of magnitude higher than the Kl values estimated for Ross Sea prymne-siophytes (presumably solitary P. antarctica) by Coale et al. (2003). It should be noted that the curve shown in Fig. 4 is constrained by only three data points, and that there are large uncertainties in the iN values calculated from the nitrate + nitrite data shown in Fig. 3a. As a result, there are large uncertainties associated with our estimates of KM and imax. Nonetheless, our experimental results clearly demonstrate that colonial P. antarctica require relatively high dissolved Fe concentrations—certainly well above 0.2 nM—to achieve maximum growth rates at the relatively low irradiance of -20 iE m-2 s-1.

0 0

Post a comment