|Home | About | Journals | Submit | Contact Us | Français|
Experiments simulating the sea ice cycle were conducted by exposing microbes from Antarctic fast ice to saline and irradiance regimens associated with the freeze-thaw process. In contrast to hypersaline conditions (ice formation), the simulated release of bacteria into hyposaline seawater combined with rapid exposure to increased UV-B radiation significantly reduced metabolic activity.
Sea ice is a predominant feature of polar oceans and exerts a unique influence on Antarctic marine ecosystems (1, 6). For microbial communities, the ice matrix represents a harsh physicochemical environment, and productivity reflects a complex relationship between ice dynamics, the distribution of organic and inorganic nutrients, and also photosynthetically active radiation and UV-B radiation (2, 9, 20, 22). As such, the quantitative importance of bacterial production is difficult to assess (13), but it is likely that bacteria participate in a microbial loop within the ice, whereby the consumption of bacteria supports higher trophic levels (3, 5, 7, 8).
Although sea ice bacteria are known to exhibit high levels of metabolic activity (14), few authors have examined adaptation to the physicochemical extremes of the habitat (11, 12, 15, 16). In contrast, pulse amplitude modulation (PAM) fluorometry has provided insight into the intracellular stress response of microalgae to a range of experimental stimuli. For example, the release of microalgae into the hyposaline meltwater at the ice edge in the austral summer may cause more physiological stress than incorporation into the ice matrix during winter (18). The ability of Antarctic sea ice bacteria to cope with the transitional saline and irradiance levels associated with ice melt is not known but may determine their capacity to act as “seed populations” that initiate ice edge blooms.
In this study, sea ice bacteria and microalgae from the bottom (congelation layer) of annual fast ice were exposed to irradiance and saline regimens that are similar to the annual freeze-thaw process. A powered auger (Jiffy) and Kovaks corer were used to extract ice cores, and microbes were subsequently obtained by cutting and then melting the bottom 50 to 100 mm of each core over a period of 12 h into three times the volume of filtered seawater. Five saline treatments (8‰, 21‰, 32‰, 51‰, and 69‰) were prepared using the melted stock solution, and the metabolic response of bacteria to salinity stress was quantified using tetrazolium chloride (CTC) and calibrated against a community level estimate of activity ([3H]leucine). Visible light (photosynthetically active radiation), sample filtration, and 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) were also included in the simulation to determine whether bacterial activity is influenced by photosynthate derived from algae. The effect of UV-B (280 to 320 nm) on metabolic activity was examined by melting blocks of ice (60 by 50 by 15 mm) in a water bath with an overhead halogen array and UV tube (see http://www.victoria.ac.nz/sbs/staff/staff_academic/ryanken/ryan-ken.aspx for further details).
Estimates of bacterial production (protein biosynthesis) from the uptake of [3H]leucine varied between 8.2 and 1,048 picomoles liter−1 h−1. Production after a 48-h period of incubation was significantly higher than the time zero or postmelt estimate for each saline treatment (for 10‰ saline treatment, t = −4.54 [ANOVA], df = 2, P = 0.045; for 33‰ saline treatment, t = −4.69, df = 2, P = 0.043; for 55‰ saline treatment, t = −5.43, df = 2, P = 0.032). There was no difference between the three saline treatments at time zero (F = 2.536, df = 2, P = 0.159), but after 48 h, the uptake of [3H]leucine by bacteria incubated at 10‰ was significantly lower than the 33‰ and 55‰ saline treatments (F = 36.752, df = 2, P = <0.001) (Table (Table1)1) .
We observed a positive correlation between estimates of [3H]leucine incorporation and CTC-positive bacterial cells (Table (Table11 and Fig. Fig.1).1). The concentration of leucine incorporated after 48 h provides a useful validation for estimates of single-cell activity that are significantly greater than the 2 to 10% values that are typical for most marine systems. Additionally, the difference in leucine incorporation between the 0-h and 48-h incubation periods highlights a salinity response that is dependent on time, rather than osmotic shock following initial exposure to altered saline concentrations (Table (Table1).1). This finding supports earlier data on bacterial activity from McMurdo Sound in Antarctica whereby maximum thymidine and uridine incorporation was recorded at salinities from 20‰ to 30‰ with secondary peaks from 50‰ to 70‰ (12). Importantly, the inclusion of preliminary postmelt data in the current study provides a useful insight into the rapid response of bacteria following a short period of incubation.
The percentage of CTC-positive sea ice bacteria following exposure to light and saline regimens varing between 29 and 85%. There was no significant interaction between light and salinity (F = 1.787, P = 0.072); however, the effect of salinity on respiration in bottom-ice bacteria was highly significant (F = 9.787, P = <0.001). Metabolic activity at 8‰ saline treatment was significantly lower than all other saline treatments and significantly higher at 51‰ compared to the 21‰ and 32‰ saline concentrations (posthoc Tukey's test of all pairwise comparisons) (Fig. (Fig.11).
Metabolic activity varied between 8 and 69% for cells incubated at 32‰ saline, and a significant interaction was found between treatment (unfiltered, filtered, and DCMU) and incubation irradiance (F = 7.177, P = <0.001) (Fig. (Fig.2.).2.). In general, activity in the filtered treatment was coupled with increasing irradiance, while in the unfiltered treatment, activity initially increased and then declined at 262 μmol photons m−2 s−1. For the DCMU treatment, irradiance had no effect on intracellular activity. In comparison to the unfiltered treatment, however, filtration and the addition of DCMU both reduced single-cell activity after 48 h. Samples with DCMU were significantly different at 77.5 and 262 μmol photons m−2 s−1, while the filtered treatment differed at 0 and 77.5 μmol photons m−2 s−1 compared to the unfiltered treatment. There was no statistically significant difference between the DCMU and filtered treatments (posthoc Tukey's test of all pairwise comparisons).
The highest levels of single-cell activity were observed at saline concentrations that simulate sea ice formation (51‰ and 69‰), which is not unexpected given that strong selection processes favor psychrophilic bacteria following the initial freeze (21). In addition to maintaining growth at subzero temperatures, these species can produce salt-tolerant enzymes that may confer an adaptive advantage to bacteria facing changing salinity (15). While Nichols et al. (15) have demonstrated an upper and lower salinity threshold in isolated cultures of sea ice bacteria, the extent to which physicochemical variables stratify the in situ distribution of bacteria is not clearly understood. In contrast, ice melt and lowered ambient salinity may impose greater stress on bacterial metabolism than ice formation. A reduction in the activity of bacteria during the transition from sea ice to the hyposaline lenses (represented by the 21‰ and 8‰ saline treatments) that form at the receding ice edge could potentially limit the long-term cell survival and involvement of ice-derived bacteria in bloom events. While protein synthesis was reduced at 10‰ (Table (Table1),1), cellular respiration was maintained in approximately 40% of the cells after 48 h (Fig. (Fig.2).2). This response infers a degree of tolerance to low salinity and further stress during the subsequent mixing, and stabilization of the water column is unlikely to limit bacterial secondary production.
Ralph et al. (18) have recently shown that high irradiance (150 μmol photons m−2 s−1) significantly modifies the response of microalgae to salinity stress, and variation in the production of photosynthetic exudates could conceivably influence bacterial metabolism. The combined effect of light and salinity was insignificant in the current study for most treatments; however, we do provide evidence for a correlation between CTC-positive cells and microalgal photosynthetic activity for samples incubated at 32‰ saline. At this salinity, the percentage of metabolically active cells was significantly higher at 77.5 μmol photons m−2 s−1 compared to the dark treatment, but metabolic activity subsequently declined at 262 μmol photons m−2 s−1 (Fig. (Fig.2).2). Photoinhibition of microalgae exposed to high levels of light may have limited the production of exudates available for bacterial metabolism, and the lack of covariation between light and salinity in the other treatments may reflect the severe inhibitory photosynthetic response of microalgae to extremes in salinity (2, 18, 19, 23).
A significant reduction in single-cell activity was observed by removing the majority of microalgae (filtered treatment) as well as inhibiting photosynthesis with DCMU (Fig. (Fig.2).2). Importantly, although incubation irradiance had no effect on bacterial activity for samples spiked with DCMU, a positive response to increasing irradiance was observed in the filtered treatment. Despite the implications for trophic dynamics and production within the sea ice ecosystem, few studies have quantified the time required for bacteria to respond to algal metabolism. Grossmann and Gleitz (10) determined that photosynthesis influenced bacterial production, but only after an in situ incubation of several weeks. Our results, at least for melted ice cores, dramatically shorten this response time.
Bottom-ice bacteria responded to UV-B radiation in a dose-dependent manner. Data from 2006 and 2007 were analyzed separately but are shown together in Fig. Fig.3.3. Relative to the control, activity was reduced by 56% after a dose of 12.5 kJ m−2 over 12 h and by 94% at 25 kJ m−2. In 2006, approximately 12% of the bacteria were CTC positive, but there was no significant difference in activity after 12 h of exposure to UV-B irradiance of 1.2 or 6.2 kJ m−2 (F = 0.265, df = 1, P = 0.625). Higher UV-B treatments in 2007 (12.5 and 25 kJ m−2, respectively) resulted in a significant difference in metabolic activity between the control and cells exposed to UV-B (F = 10.204, df = 2, P = 0.005). Bacteria exposed to 25 kJ m−2 (1% CTC positive) were significantly less active than those exposed to 12.5 kJ m−2 (8% CTC positive), while activity at the highest UV-B treatment was significantly less than the control (18% CTC positive) (posthoc Tukey's test).
In contrast to observations from Antarctic coastal waters (4, 17), we observed significant metabolic inhibition in sea ice bacteria exposed to UV-B radiation. Extrapolating from the erythemal action spectrum used by Nunez et al. (17) and UV-B measurements collected in McMurdo Sound in Antarctica in 2007 (http://www.biospherical.com/nsf), 25 kJ m−2 is an ecologically relevant level of exposure for cells at the surface of the sea during the ice edge bloom. Whether ice-derived bacteria show adaptive mechanisms to long-term UV-B exposure or potentially benefit from the stress response of other taxa during bloom events in the Southern Ocean is not known; however, the initial exposure to increased UV-B clearly causes significant metabolic stress.
Our results show that metabolic activity in sea ice bacteria would be reduced during ice melt when hyposaline conditions are combined with increased UV-B exposure. In contrast, the bacteria present in the congelation layer at the time of sampling are likely to be psychrophilic and show a greater ability to acclimate to the hypersaline conditions that are typical of ice formation. Bacterial-algal linkages are clearly influenced by transitional light and saline regimens during the freeze-thaw process, but we show that this is largely dependent on the relative abundance of bacteria and microalgae at the time of sampling. Further research is needed to quantify this relationship with respect to the temporal variability and spatial characteristics of the sea ice matrix.
We acknowledge the logistical support of Antarctica New Zealand and in particular S. Gordon, Project Manager of the Latitudinal Gradient Project (LGP).
A. Martin was supported by a Victoria University of Wellington Postgraduate Scholarship for Ph.D. Study and also thanks the Trans Antarctic Association for support in funding this research. K. Ryan acknowledges the support of the Foundation of Research, Science & Technology contract (VICX0706).
Published ahead of print on 2 October 2009.