Local adaptation of microbial communities to ... - Wiley Online Library

0 downloads 0 Views 277KB Size Report
heavy metal contaminants on microbial communities, we performed two experi- ... may result in the production of free radicals that disrupt proteins, nucleic acids ...

RESEARCH ARTICLE

Local adaptation of microbial communities to heavy metal stress in polluted sediments of Lake Erie Matthew J. Hoostal, M. Gabriela Bidart-Bouzat & Juan L. Bouzat Department of Biological Sciences, Bowling Green State University, Bowling Green, OH, USA

Correspondence: Juan L. Bouzat, Department of Biological Sciences, Bowling Green State University, Bowling Green, OH 43403-0212, USA. Tel.: 11 419 372 9240; fax: 11 419 372 2024; e-mail: [email protected] Received 25 January 2008; revised 17 April 2008; accepted 18 April 2008. First published online June 2008. DOI:10.1111/j.1574-6941.2008.00522.x ¨ Editor: Max Haggblom Keywords extracellular enzyme activity; community adaptation; heavy metals; microbial communities.

Abstract Microbial communities must balance the assimilation of biologically necessary metals with resistance to toxic metal concentrations. To investigate the impact of heavy metal contaminants on microbial communities, we performed two experiments measuring extracellular enzyme activities (EEA) in polluted and unpolluted sediments of Lake Erie. In the first experiment, inoculations with moderate concentrations of copper and zinc appreciably diminished EEA from uncontaminated sites, whereas EEA from contaminated sediments increased or were only negligibly affected. In the second experiment, we compared the effects of three separate metals (i.e. copper, arsenic, and cadmium) on microbial community metabolism in polluted and unpolluted locations. Although copper and arsenic elicited differential effects by inhibiting EEA only in unpolluted sediments, cadmium inhibited EEA in both polluted and unpolluted sediments. Multivariate analyses of EEA from polluted sediments revealed direct associations among hydrolytic enzymes and inverse or absent associations between hydrolases and oxidases; these associations demonstrated resilience to heavy metal stress. In contrast, addition of heavy metals to unpolluted sediments appeared to have a higher impact on the multivariate pattern of EEA associations as revealed by an increase in the number of associations, more inverse relationships, and potential enzymatic trade-offs. The results of this study suggest community-level adaptations through the development of resistance mechanisms to the types and local levels of heavy metals in the environment.

Introduction Anthropogenic inputs of heavy metals enter the environment by effluent from industry, waste-water treatment plants, landfills, and mining (Eisler, 1998). Although some metals, such as copper and zinc, are necessary cofactors in enzymes and electron transport chains, toxic levels of metals may result in the production of free radicals that disrupt proteins, nucleic acids, and phospholipids (Halliwell & Gutteridge, 1984, 1985). Metals may also displace metal enzyme cofactors, disrupting the structural integrity and function of enzymes (Stadtmann, 1993; Stohs & Bagchi, 1995). Given the critical role of metals in cellular function, bacteria must coordinate the assimilation of low levels of biologically essential heavy metals and their tolerance of toxic concentrations (Silver & Phung, 1996; Silver, 1998). 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

Bacteria have adapted a variety of heavy metal tolerance mechanisms, which are often plasmid-borne and capable of being spread throughout a bacterial community by lateral gene transfer (Coombs & Barkay, 2004; Martinez et al., 2006). Heavy metals may, therefore, act as important selective agents driving the evolution of microbial communities. For example, the incidence of plasmid-borne mercury resistance genes has been positively correlated with mercury concentrations in lake sediments (Barkay & Olson, 1986). The potential for natural selection acting at the community level, while controversial among evolutionary biologists (Sober & Wilson, 1998), has been suggested in a variety of laboratory experiments (Goodnight, 1990; Cullen & Neale, 1994; Swenson et al., 2000a, b; M¨uller et al., 2001; Sun & Friedmann, 2005). Microbial communities, often composed of a vast number of metabolically interdependent species with the potential of horizontal gene transfer, provide an apt FEMS Microbiol Ecol 65 (2008) 156–168

157

Microbial community adaptation to heavy metals

model for studying the possibility for community-level adaptations (Swenson et al., 2000a). Lake Erie, a large freshwater ecosystem with an east–west gradient of increasing contaminants across its three basins (Painter et al., 2001), provides an appropriate system for studying the potential role of contaminants in modulating microbial community adaptations. Multiple river inputs of industrial and agricultural contaminants into the central and western basins led to the identification of 11 areas of concern (AOCs), defined by the United States–Canada Great Lakes Water Quality Agreement as ‘geographic areas that fail to meet the general or specific objectives of the agreement where such failure has caused or is likely to cause impairment of beneficial use of the area’s ability to support aquatic life’ (Painter et al., 2001; Fig. 1). Although the hydrodynamics of Lake Erie leads to the significant mixing of sediments (Burns, 1985), differential inputs of organic matter, metal contaminants, and other xenobiotics may promote microbial community adaptation at the local scale. Given the role of microbial communities in the natural remediation of anthropogenic contaminants (Iwamoto & Nasu, 2001), studies on the local adaptation of microbial consortia to environmental pollutants are warranted. Assays of microbial community metabolism, such as extracellular enzyme activity (EEA) profiles, have been used for monitoring the responses of microbial communities to ecological perturbations. EEA assays allow quantification of the heterotrophic assimilation of organic matter by microbial consortia (Sinsabaugh & Foreman, 2001) independent of culturing methods, which underestimate a large majority of microorganisms in the environment (Hugenholtz et al.,

1998). Thus, EEA can be used as efficient biological indicators of microbial community metabolism. For example, EEA have been used to assess microbial responses to heavy metals (Kandeler et al., 2000), responses to excessive nitrogen deposition (Waldrop et al., 2004), responses to dissolved organic matter (Hoostal & Bouzat, 2008), and microbial infections of plants (Collmer & Keen, 1986; Barras et al., 1994). Because extracellular enzymes are available to a broad range of species within the microbial community, the coordination of these enzymes for heterotrophic metabolism may provide the bases for community-level adaptations. In this study, we assessed the adaptive significance of heavy metals by evaluating patterns of EEA responses of microbial communities to heavy metal stress in both polluted and relatively unpolluted sediments. Although numerous studies have demonstrated heavy metal tolerance mechanisms in microbial taxa from isolated extreme environments (Tetaz & Luke, 1983; Bruins et al., 2000; Cervantes et al., 2001), studies that show local adaptation of indigenous microbial communities associated with a gradient of environmental pollution throughout a major ecosystem are less common. The central basin of Lake Erie is beset with heavy metal contamination in the western and southern regions, while the northern and eastern regions are relatively pristine (Painter et al., 2001; Fig. 1). Therefore, microbial communities from polluted and less polluted regions of the central basin may demonstrate differential resilience to heavy metal stress as a result of local adaptation. Significant mixing of sediments in the central basin, however, may prevent local adaptation and result in physiologically plastic EEA responses of microbial communities to varying metal concentrations.

Fig. 1. Map of Lake Erie showing the distribution of sampling sites in polluted (black squares) and unpolluted (open squares) regions, as well as the location of AOCs associated with Lake Erie’s central basin (light gray circles). Sampling stations are numbered according to US EPA designations. The area of black circles is proportional to the average amount of heavy metals (Cu, Zn, As, Pb, Hg, and Cr) in sediment samples compared with background concentrations (as quantified by Painter et al., 2001). The lowest metal concentrations, prominent in the northern and eastern region of Lake Erie’s central basin, are 0.2–1.7 times the background concentration of metals, as estimated by preindustrial levels of metals detected in sediment cores (Painter et al., 2001). Metal concentrations in the southern and western region of Lake Erie’s central basin are 1.7–2.7 times the background concentration of metals.

FEMS Microbiol Ecol 65 (2008) 156–168

2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

158

To establish whether indigenous microbial communities from polluted sediments are locally adapted to metal stress, we performed two complementary experiments. In the first experiment, we inoculated sediments collected from polluted and unpolluted regions with either moderate or high levels of two biologically necessary metals, i.e., copper and zinc. Our general hypothesis is that EEA profiles of microbial communities from polluted regions will demonstrate greater metal tolerance and, thus, will be more resilient to inoculations with moderate concentrations of heavy metals than microbial communities from nonpolluted regions. However, high concentrations of metals should be toxic and should inhibit EEA in both polluted and unpolluted sediments. In a second experiment, we evaluated microbial community EEA patterns in response to low concentrations of three different metals (i.e. copper, arsenic, and cadmium). Microbial communities must coordinate the assimilation of copper with the tolerance of toxic concentrations by using various multicopper oxidases and ATPases (Cervantes & Gutierrez-Corona, 1994). In contrast, the major means bacteria utilize for arsenic resistance involves the enzymatic reduction of arsenate to arsenite, followed by removal from the cell by an ATPase (Cervantes et al., 1994). Cadmium may be directly removed from the cell by either an ATPase or a cation–proton antiporter (Silver, 1998). Because microbial species have developed distinct tolerance strategies for different metals, we chose to compare differences in microbial resilience to each of these metals in polluted and less polluted sediments. Specific associations among certain EEA may increase the efficacy of heterotrophic organic matter assimilation by microbial species (Judd et al., 2006), potentially facilitating local adaptation at the community level. Heavy metal stress may, therefore, affect EEA by either decreasing overall enzymatic activities or by disrupting patterns of association among specific EEA. If microbial communities are locally adapted to tolerate heavy metal concentrations in the environment, one would expect the association patterns among microbial EEA from polluted sediments to be more resilient to heavy metal stress than those patterns from unpolluted sediments.

Materials and methods Collection of sediment samples Sediment samples were collected from both polluted and unpolluted sediments to assess potential differential responses of microbial EEA to heavy metal stress. Surface sediments were collected in June of 2003 during a cruise of the CCGS Limnos. Samples were collected with a Ponar grab from 10 sampling stations distributed throughout the 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

M.J. Hoostal et al.

central basin of Lake Erie (Fig. 1). At each station, a sediment sample was collected within 5 cm from the sediment surface in sterile 15-mL polystyrene tubes and kept on ice until return to the lab, where they were stored at 20 1C until further analysis. While freezing and thawing may affect EEA, Lee et al. (2007) demonstrated that, among a variety of approaches for storing soil samples, microbial biomass and enzymatic activities were least affected when stored at 4 or 20 1C. Five sampling stations corresponded to polluted regions within the central basin, whereas five stations were from relatively pristine regions. Sampling sites were designated as polluted or unpolluted based on Painter et al.’s (2001) analytical measurements of heavy metal concentrations at corresponding sites throughout Lake Erie sediments (Fig. 1). The central basin contains four Areas of Concern, comprised of the Ashtabula, Cuyahoga and Black Rivers along the southern border of Lake Erie, and Wheatley Harbor along the northern shore, with metal contamination contributing to each site’s designation as an AOC. Polluted sites sampled in this study were located proximate to each of these AOCs (Fig. 1). Polluted locations had metal concentrations exceeding the threshold effect level (TEL) of the Canadian Sediment Quality Guidelines, which is defined as ‘the concentration below which adverse biological effects are expected to occur rarely’ (Painter et al., 2001). TEL concentrations are 35.7, 5.9, and 0.596 mg g 1 of sediment for copper, arsenic, and cadmium, respectively. Unpolluted regions of our sampling corresponded to areas with metal concentrations well below the TEL. Metal concentrations from Painter et al. (2001) corresponding to unpolluted sites in our study were, on average, 22.9, 1.2, and 0.44 mg g 1 of sediment for copper, arsenic, and cadmium, respectively. Metal concentrations corresponding to polluted sites were 53.2, 6.8, and 1.6 mg g 1 of sediment for copper, arsenic, and cadmium, respectively. More recent reports (Ohio Lake Erie Commission, 2004) have continued to classify AOCs proximate to our polluted sampling sites as contaminated sediments based on fish consumption advisories and toxicity to benthic organisms.

Experiment 1: responses of microbial EEA to varying concentrations of heavy metals EEA were used as a metric of the resilience (or sensitivity) of microbial communities. Microbial community EEA from polluted and unpolluted sediments were compared following inoculations with moderate and high levels of two biologically relevant heavy metals: copper and zinc. Sediment slurries for each sampling station were obtained by mixing 1 g of sediment with 100 mL of sterile water. Slurries were inoculated with either moderate (5 mM) or high concentrations (50 mM) of copper sulfate and zinc sulfate FEMS Microbiol Ecol 65 (2008) 156–168

159

Microbial community adaptation to heavy metals

solutions. To account for potential environmental variation among sites within polluted and unpolluted locations, we measured each EEA from an untreated control at each sampling station. Treatment concentrations were 10–100fold higher than those found in Lake Erie; thus, they should be especially noxious to microbial communities and more apt to assess microbial resilience to heavy metal stress. Sediment slurries were incubated for only 16 h before EEA assays. Doubling times of microorganisms in the environment vary widely, ranging from 0.5 to 4 days in estuarine and river water samples (Castillo et al., 2004; Crump et al., 2004) to c. 15 days in soils, lake, and deep-sea sediments (Onstott et al., 1998). Therefore, the incubation period used to assess EEA minimized potential shifts in community composition. Two classes of hydrolytic EEA (esterases and glycosidases), as well as two oxidases, were determined to assess heterotrophic microbial activity. These enzymes were selected on the basis of their direct role in the cycling of carbon, nitrogen, phosphorus, and sulfur, providing appropriate markers of microbial community metabolism ´ ´ & Rai, 1993; Torsvik & Oevreas, (Chrost, 1991; Chrost 2002). Hydrolytic EEA were measured in 96-well microplates based on the methods described by Sinsabaugh & Foreman (2001). A fluorescent methylumbelliferyl (MUF)linked substrate solution (Sigma-Aldrich, St Louis, MO) was added to 200 mL of sediment slurry in each microplate well, with a 40-mM final substrate concentration. The esterases measured included alkaline phosphatase and arylsulfatase, whereas the glycosidases consisted of b-1, 4-glucosidase and b-N-acetylglucosaminidase (NAGase). The maximum reaction velocity (Vmax) for hydrolytic EEA was quantified by measuring EEA at 30-min intervals for 4 h using a Wallac Victor 1420 Multipurpose Counter (Perkin Elmer, Turku, Finland) at an excitation wavelength of 365 nm and an emission wavelength of 450 nm. Enzyme assays, methylumbelliferyl reference standards, and substrate controls were each replicated eight times. Saturation curves were constructed by measuring enzyme activities at substrate concentrations of 0, 10, 20, 40, 50, and 80 mM. All hydrolytic enzymes were saturated below 40 mM, with the exception of sulfatase, which often became saturated at substrate concentrations between 60 and 80 mM. Therefore, all hydrolytic activities were measured at Vmax, with the exception of sulfatase, which was measured as potential activity. Extracellular oxidases were quantified spectrophotometrically in 96-well microplates based on the methods of Saiya-Cork et al. (2002). Phenol oxidase activity was assessed using L-3, 4-dihydroxyphenylalanine (L-DOPA) (final concentration = 1 mM) as the enzyme substrate. Manganese peroxidase activity was measured similarly, with the addition of 10 mL of 3% H2O2 to each well. Plates were incubated at room temperature and measured hourly at FEMS Microbiol Ecol 65 (2008) 156–168

450 nm for 12 h to determine Vmax. As with hydrolytic EEA, treatments and controls for oxidase EEA were assayed in replicates of 8. In both cases, assay concentrations used in the experimental treatments (1 mM) were at saturation levels.

Experiment 2: responses of microbial EEA to various metals A second set of assays was designed to test for differential responses of microbial communities to different heavy metals in polluted vs. unpolluted sediments and assess patterns of association among microbial EEA. To measure EEA, sediment slurries were prepared as described in Experiment 1, but inoculated with copper sulfate, cadmium chloride, or sodium arsenate solutions to a final concentration of 0.50 mM. EEA from untreated controls (independent from Experiment 1) were additionally measured at each site. The lower concentrations used in this experiment are comparable to those found in Painter et al.’s (2001) analysis of heavy metal concentrations in sediments of Lake Erie’s central basin and are, therefore, more ecologically relevant. Following inoculations with either copper, cadmium, or arsenic salts, sediment slurries were again incubated for 16 h before EEA were measured as described above.

Biomass assay Sediments offer unique challenges in measuring biomass due to high levels of plant and microbial detritus. Thus, we used microbial respiration, measured through a dehydrogenase assay adapted from Alef & Nannipieri (1995) and Sigler & Zeyer (2002), as a proxy for microbial biomass (van der Werf & Verstraete, 1987). Measures of dehydrogenase activity have frequently been used to evaluate overall microbial activity (Casida, 1977) and previous studies have demonstrated significant correlations between dehydrogenase activity and microbial biomass carbon (Cochran et al., 1989; Goyal et al., 1993; Batra & Manna, 1997). To compare microbial biomass between contaminated and uncontaminated locations, three to five independent replicates of 2 g of sediment samples with 2 mL of triphenyltetrazolium chloride (TTC) solution (0.1 g TTC per 100 mL of 100 mM Tris buffer, pH = 7.6) were incubated in the dark at 30 1C for 24 h. Acetone (20 mL) was then added to each sediment sample, followed by a 2-h incubation period in the dark at room temperature. The absorbances of solutions were subsequently measured at 546 nm using a Thermospectronic Genesys 20 photospectrometer (ThermoFisher Scientific, Waltham, MA). Autoclaved sediments were analyzed as negative controls. Swabs from autoclaved sediments were plated with Bacto agar (Voigt, Lawrence, KS) media to ensure sterility. 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

160

Data analysis In both experiments, the effects of heavy metal inoculations on each of the microbial extracellular enzymes were evaluated using a nested factorial design with the generalized linear model y = Location1Site (Location)1Treatment1 Location  Treatment1Site (Location)  Treatment. Treatment refers to whether samples were untreated or inoculated with either 5 or 50 mM concentrations of a copper and zinc solution in the first set of experiments, or 0.50 mM copper, arsenic, or cadmium in the second set of experiments. Location refers to whether sediment samples were collected from polluted or unpolluted regions and site refers to the actual sediment sampling stations (Fig. 1). The evaluation of the Treatment  Location interaction is particularly important in this study because it would indicate whether EEA responses of microbial communities to metals are location specific, most likely due to local adaptation of microbial consortia to different inputs of metals into the central basin of Lake Erie. In both experiments, preplanned contrasts were performed to assess the mean differences between control and heavy metal treatments for each location. Differences in the biomass of microbial communities between polluted and unpolluted locations were compared using univariate ANOVAs, with the following generalized linear model: y = Site1Location1(Site  Location). All statistical analyses were performed using SAS, v 8.0 (SAS Institute). Model assumptions for ANOVAs were assessed through residual analyses. Normality of residuals was assessed using Shapiro–Wilk tests (Sokal & Rohlf, 1995), normal probability plots, and residual plots. In most cases, the data violated the assumption of normality; therefore, tests were performed using transformed data when necessary (rankedtransformed data for Experiment 1 and log- or squared root-transformed data for Experiment 2). Homoscedacticity was evaluated by plotting log of residuals vs. log of predicted values and assessing whether these two values were correlated (Spearman’s correlation coefficients). We evaluated multivariate normality by plotting ordered squared Mahalanobis distances against the quantiles of a w2 distribution (Khattree & Naik, 1999), which should fit a straight line with a slope equal to 1. To assess potential shifts in patterns of associations among hydrolytic and oxidase EEA from Experiment 2, as well as the resilience of these associations to heavy metal stress in polluted and unpolluted sediments, we performed principal components analyses (PCA) and Mantel tests of matrix correlations (Marroig & Cheverud, 2001; BidartBouzat et al., 2004). PCA is useful to group variables that are correlated and thus make predictions on potential processes causing these multivariate association patterns (Tabachnick & Fidell, 1989). For this purpose, we calculated 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

M.J. Hoostal et al.

principal component (PC) loadings, which are the correlations between each response variable and the PCs. PC loadings reveal ‘how closely a variable and a PC are related’, and as an extension, patterns of associations among variables that load on the same and/or different PCs (Tabachnick & Fidell, 1989; McGarigal et al., 2000). The first two principal components (PC1 and PC2) were retained for each treatment–location combination based on the Mineigen criterion of Proc Factor (SAS Institute). The significance of PCs was based on a benchmark of 0.55 for the absolute values of PC loadings, as suggested by Tabachnick & Fidell (1989). To formally assess treatment effects on patterns of association among EEA, we performed Mantel tests of matrix correlations (Mantel, 1967). Specifically, we compared the correlation matrix of each untreated microbial community (i.e. control) from a specific location (polluted or unpolluted) with the correlation matrices obtained subsequent to inoculations with copper, cadmium, or arsenic. In addition, matrix correlations of microbial communities from different locations exposed to the same heavy metal treatment were compared to evaluate whether communities differed in their collective functional response to different environments. Mantel tests evaluate the associations among two or more independent dissimilarity matrices using sampled randomization techniques (Sokal & Rohlf, 1995), with 10 000 permutations of the rows and columns of one of the matrices. Dissimilarity matrices were obtained by subtracting the absolute value of the Pearson correlation coefficients from one. Matrix correlation coefficients use values from 1 to 11 to measure structural similarities between matrices. A matrix correlation coefficient of 11 or 1 indicates identical correlation patterns or mirror-imaged matrices, respectively. Conversely, a 0 or a c. 0 value denotes no structural similarity between matrices and, thus, a nonsignificant Mantel correlation coefficient (Marroig & Cheverud, 2001). Mantel tests were performed using the program ARLEQUIN v. 2.0 (Schneider et al., 2000).

Results Experiment 1: responses of microbial EEA to varying concentrations of heavy metals Analyses of variance revealed that the main effects of location, treatment, and site, as well as their interactions were highly significant for each of the six enzyme activities investigated (all P o 0.001). The significant Treatment  Location interactions are particularly important in this study because they indicate that EEA responses of microbial communities to heavy metal treatments are location-specific (i.e. dependent on the location of origin of microbial communities). Preplanned contrasts performed FEMS Microbiol Ecol 65 (2008) 156–168

161

Microbial community adaptation to heavy metals

Control Low High

300

*

200

100

*

40

* 20

*

* *

Polluted

Polluted

200

200

150

150

100

50

*

*

*

0

Unpolluted

NAGase activity

β-Glucosidase activity

60 Sulfatase activity

Phosphatase activity

400

Unpolluted

*

100

50

*

* *

0 Polluted

*

0

Unpolluted

Polluted

Unpolluted

500

Peroxidase activity

Phenol oxidase activity

500

400

300

200

400

* *

*

300

200

100 Polluted

Unpolluted

Polluted

Location

Unpolluted Location

Fig. 2. Response of EEA measured in polluted and unpolluted locations within the central basin of Lake Erie. The interaction between location and treatment is represented by the slope of the line. Control, low, and high copper and zinc treatments are compared. The y-axis represents the mean enzyme activity of five sampling sites in either unpolluted or polluted locations. Error bars represent SEs of enzyme activity means. Activity is measured in Zmol L 1 h 1. An asterisk denotes a significant difference (P o 0.05) between the corresponding treatment and the control in the sampled environment (polluted or unpolluted).

to assess the mean differences between control and inoculation treatments with 5 mM of heavy metals for each location indicated that the activities of hydrolytic enzymes were inhibited in sediments from unpolluted areas (contrasts of treatment vs. control P o 0.001) but not in those from polluted sites (Fig. 2). More specifically, activities of the hydrolytic enzymes phosphatase, sulfatase, b-glucosidase, FEMS Microbiol Ecol 65 (2008) 156–168

and NAGase decreased by 39%, 39.1%, 51%, and 55%, respectively, compared with controls in unpolluted sediments. In contrast, the activities of oxidase enzymes such as phenol oxidase and peroxidase were either not significantly affected or increased (contrast P o 0.001) under moderate levels of heavy metals in both polluted and unpolluted sediments. Alkaline phosphatase was the only enzyme whose 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

162

activity was significantly diminished in polluted sediments by the 5 mM metal treatments (contrast P = 0.001). The effects of high concentrations of copper and zinc (50 mM) were especially deleterious on hydrolytic EEA in both polluted and unpolluted areas but did not significantly affect activities of oxidases (with the exception of manganese peroxidase in unpolluted sediments: P = 0.001; see Fig. 2).

Experiment 2: responses of microbial EEA to various metals In a second set of inoculations, both polluted and unpolluted sediment samples were spiked with 0.5 mM treatments of either copper, arsenic, or cadmium. Enzymatic activities were usually greater in polluted than unpolluted sediments (Fig. 3). The main effects of location and treatment were significant (P o 0.001), as well as most of the interaction terms (P o 0.01). To directly test the adaptive significance of heavy metals in polluted and unpolluted sediments, we performed preplanned contrasts of control vs. heavy metal treatments at each location. Results from these contrasts demonstrated that heavy metal inoculations generally inhibited enzyme activity in unpolluted sediments (Fig. 3). More specifically, copper, arsenic, and cadmium each significantly inhibited phosphatase (all P o 0.001), sulfatase (all P o 0.001), b-glucosidase (all P o 0.001), and NAGase (all but copper P o 0.01) activities in unpolluted sediments. Copper and arsenic generally had either a negligible effect or actually increased enzyme activities in polluted sediments. In particular, copper increased sulfatase and NAGase activities (P o 0.001), while arsenic increased phosphatase (P = 0.035), sulfatase (P o 0.001), and NAGase (P = 0.002) activities in polluted sediments. Notably, cadmium inhibited hydrolytic EEA in both polluted and unpolluted sediments (all P o 0.01; Fig. 3). The response of oxidases to heavy metals was more complex than that of hydrolases (Fig. 3). Phenol oxidase activity in unpolluted sediments increased following copper additions, but decreased in response to cadmium (both contrast P-values o 0.001). Phenol oxidase activity from polluted sediments was not significantly affected by heavy metal inoculations. On the other hand, manganese peroxidase activity was inhibited by each metal in polluted sediments (all contrast P-values o 0.001) and inhibited by arsenic and cadmium in unpolluted sediments (contrast Pvalues o 0.001). PCA graphically revealed pronounced differences in the patterns of association among EEA in polluted compared with unpolluted sediments (see Fig. 4, control treatments). In polluted sediments, hydrolytic enzyme activities were consistently grouped into the first PC and generally showed fairly strong and positive correlations with this PC (i.e. PC 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

M.J. Hoostal et al.

loadings Z0.73). Conversely, the activity of the oxidase enzyme phenol oxidase was negatively correlated with PC1 (and thus, inversely associated with the activity of hydrolases). In addition, the activity of peroxidase consistently loaded on PC2, although only in the arsenic and cadmium treatments, and appeared to be independent of the activity of hydrolases (except in the arsenic treatment; see Fig. 4). These relationships suggest a combined action of hydrolase enzymes as well as potential inverse or absent associations between hydrolases and oxidases. Overall, integrated EEA responses appeared to be resilient to heavy metal inoculations in polluted environments, because minimal changes in these multivariate responses were observed across treatments (compare control vs. heavy metal treatment groups in Fig. 4). On the other hand, in unpolluted areas, hydrolases and oxidases demonstrated a less consistent pattern in their correlations with the PCs and thus in their patterns of associations across treatment groups. Unpolluted environments apparently disrupted patterns of EEA integration (compare controls in polluted and unpolluted sediments in Fig. 4) and led to a further reorganization of EEA associations following heavy metal inoculations (compare control vs. heavy metal treatment groups in unpolluted environments in Fig. 4). Specifically, inoculations with each metal treatment in unpolluted environments induced an increase in the number of variables correlated with PC1 and more inverse relationships (negative PC loadings represented by dotted lines in Fig. 4). These associations, however, were weaker than those from polluted sediments, as denoted by PC loadings usually lower than 0.73 (Fig. 4). The differential effects of heavy metals upon associations among EEA from polluted and unpolluted sediments were corroborated by results from a series of matrix correlation analyses. As explained previously, a close to zero Mantel correlation coefficient indicates no structural similarity among matrices (i.e. matrices are different) and thus, a nonsignificant P-value. Our results revealed marked differences between correlation matrices of control and each of the heavy metal treatments as indicated by nonsignificant Mantel’s correlation coefficients in unpolluted environments (P = 0.487, 0.748, and 0.369 for copper, arsenic, and cadmium, respectively). Conversely, Mantel tests performed to compare matrices of control vs. each of the heavy metal treatments in polluted environments were significant for ‘control vs. copper’ (P = 0.020) and ‘control vs. cadmium’ (P = 0.013), indicating a rejection of the ‘no structural similarity’ null hypothesis (i.e. these matrices were very similar). In other words, inoculations with copper and cadmium did not significantly modify the pattern of association among enzymatic activities of microbial communities from polluted sediments. However, arsenic appeared to have a marginal effect on the patterns of EEA associations in polluted areas (P = 0.1). FEMS Microbiol Ecol 65 (2008) 156–168

163

Microbial community adaptation to heavy metals

Control Cu As Cd

* 300

* 200

* * *

100

Polluted

100

* Sulfatase activity

Phosphatase activity

400

80 60

* *

40

* **

20

Unpolluted

Polluted

Unpolluted

160

120 100

* *

80

NAGase activity

β-Glucosidase activity

140 140

* *

60

Polluted

120

* *

100

*

*

80

*

Unpolluted

Polluted

Unpolluted

400

* 300

200

* 100

Peroxidase activity

Phenol oxidase activity

500

400

300

** * * *

200

100 Polluted

Unpolluted

Polluted

Location

Unpolluted Location

Fig. 3. Response of EEA measured in polluted and unpolluted locations within the central basin of Lake Erie. Untreated (Control), copper (Cu), cadmium (Cd), and arsenic (As) treatments are compared. The interaction between location and treatment is represented by the slope of the line. The yaxis represents the mean enzyme activity of five sampling sites in either unpolluted or polluted locations. Error bars represent SEs of enzyme activity means. Activity is measured in Zmol L 1 h 1. An asterisk denotes a significant difference (P o 0.05) between the corresponding treatment and the control in the sampled environment (polluted or unpolluted).

Estimates of microbial biomass Indirect estimates of microbial biomass, as measured by the dehydrogenase assays, differed between polluted and FEMS Microbiol Ecol 65 (2008) 156–168

unpolluted sediments. Our results suggest that microbial biomass is greater in polluted locations, which showed a mean absorbance of 0.142  0.008 compared with 0.084  0.010 for unpolluted sediments. The main effects of 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

164

M.J. Hoostal et al.

Polluted PC1

Unpolluted PC2

PC1

PC2

Control



Phosphatase

Peroxidase

Phosphatase

Peroxidase

Sulfatase

Phenol oxidase

Sulfatase

Phenol oxidase

β-Glucosidase

Copper

PC1

Arsenic

PC2 Peroxidase

Sulfatase

Phenol oxidase

Sulfatase

Phenol oxidase

NAGase

β-Glucosidase

PC2

PC1

NAGase PC2

Phosphatase

Peroxidase

Phosphatase

Peroxidase

Sulfatase

Phenol oxidase

Sulfatase

Phenol oxidase

β-Glucosidase PC1

Cadmium

PC1 Phosphatase

NAGase

β-Glucosidase

PC2

PC1

NAGase PC2

Phosphatase

Peroxidase

Phosphatase

Peroxidase

Sulfatase

Phenol oxidase

Sulfatase

Phenol oxidase

β-Glucosidase

location and site (nested in location) were significant (P o 0.05).

Discussion The results of this study showed distinct spatial patterns in the functional response of microbial communities to heavy metal stress, which may be associated with the local bioavailability of environmental contaminants. Enzymatic activities of hydrolases from microbial communities from unpolluted sediments of Lake Erie were usually inhibited by both moderate levels of metal solutions, as well as various types of heavy metals (Figs 2 and 3). In contrast, EEA from polluted sediments were either negligibly affected or actually increased in their response to heavy metals. The increase in microbial EEA from contaminated sediments suggests that these communities are more resilient to 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

PC2

NAGase

Peroxidase

PC1

c

β-Glucosidase

Phosphatase

β-Glucosidase

Fig. 4. PCAs on six measured EEA (alkaline phosphatase, arylsulfatase, b-glucosidase, Nb-acetylglucosaminidase, phenol oxidase, manganese peroxidase). Lines indicate variables with significant component loadings, sij 4 0.55 or o 0.55 (Tabachnick & Fidell, 1989; McGarigal et al., 2000). Line thickness corresponds to three ranges of absolute values of loadings: 0.55–0.73 (thinnest), 0.73–0.91 (intermediate), and 4 0.91 (thickest). Solid and dotted lines represent positive and negative loadings, respectively. w denotes a marginal PCA loading (0.50).

NAGase

NAGase

β-Glucosidase

NAGase

heavy metal stress. Microbial taxa express a wide range of utilizations for copper and zinc as enzymatic and electron transport cofactors (Silver, 1998) and a more limited role for arsenic as a terminal electron acceptor (Stolz & Oremland, 1999). Thus, moderate concentrations of these metals could potentially promote overall metabolic activity. In contrast, cadmium is generally considered to be a nonbiologically usable metal, which may help explain the observed deleterious effects of cadmium in both polluted and unpolluted sediments. As expected, very high concentrations of all metals disrupted metabolic function regardless of the pollution history of sediments. The differential EEA responses to heavy metal stress therefore suggest that metal tolerance mechanisms may have been selected in microbial communities from polluted sediments, allowing them to become adapted to the higher levels of heavy metal contaminants in their local environment. FEMS Microbiol Ecol 65 (2008) 156–168

165

Microbial community adaptation to heavy metals

While adaptations are commonly studied at the individual and population levels, several studies have emphasized the hierarchical nature of the evolutionary process and the potential role of natural selection at the community level (Goodnight, 1990; Swenson et al., 2000a, b). Two major characteristics of microbial communities may favor the development of community-level adaptations. First, microbial community metabolism is highly integrated, generating resources of common use among diverse bacterial taxa within communities (Fisher, 2003). Second, genes associated with metabolic responses to environmental selective agents (such as heavy metals, antibiotics, PCBs, etc.) are commonly found in plasmids and, therefore, can be transferred laterally among distantly related taxa within microbial communities (Coombs & Barkay, 2004). Under such a framework, microbial communities in polluted regions would be expected to evolve as integrated systems adapted to local environmental gradients of anthropogenic stress. In fact, the patterns of association among EEA observed in untreated controls were consistent with this idea, showing differential levels of metabolic integration between polluted and unpolluted sediments. For example, microbial communities from polluted sediments revealed a more integrated action of hydrolases, an important group of enzymes for nutrient cycling, when contrasted with unpolluted sediments. This outcome is not unexpected, because substantial riverine inputs into polluted sediments generally result in a more complex pool of organic matter (Wetzel, 1992) and thus necessitate the coordination of multiple EEA for complete heterotrophic assimilation (Fisher, 2003). Multivariate analyses of EEA responses in polluted and unpolluted sediments further demonstrated that the resilience of microbial communities in polluted sediments to heavy metal stress likely results from local adaptive processes. The PCA and the Mantel tests showed that, in contaminated sediments, EEA associations were more resilient to heavy metal stress than those from unpolluted sediments. In other words, while heavy metal inoculations impacted the number and type of associations between EEA and the PCs (and thus among EEA themselves) in unpolluted environments, these patterns remained fairly stable in microbial communities from polluted sediments (Fig. 4). The resilience in the patterns of association reflected by the PCA may suggest enzymatic tradeoffs, which could be based on the taxonomic composition of microbial communities, because individual EEA may be associated with specific microbial taxa (Eiler et al., 2003). The increased biomass of microbial communities in polluted sediments compared with unpolluted sediments could partially explain the observed differences in the levels of EEA in untreated controls. In our experiments, responses of EEA to metal treatments were, however, measured against these untreated controls, demonstrating that metal stress FEMS Microbiol Ecol 65 (2008) 156–168

inhibited EEA mainly in sediments from unpolluted regions. Furthermore, the multivariate analysis showed that the association patterns among EEA were more resilient to heavy metal stress in polluted sediments. While previous studies have demonstrated decreases in microbial enzymatic activity due to heavy metal stress (Kandeler et al., 2000), the resilience of associations patterns among EEA reported in this study represents a relatively novel approach for examining the response of microbial communities to environmental contaminants. The sensitivity of microbial community EEA, especially the hydrolase activities, to heavy metals in unpolluted, but not polluted sediments, suggests that EEA may be used as efficacious biological markers of microbial ecotoxicity to heavy metals. Biomarkers should elicit a biological response to chemicals that gives a measure of exposure and possibly toxic effect (Peakall, 1994). Biological markers offer advantages over analytical chemical techniques that measure environmental concentrations of heavy metals but do not necessarily indicate their bioavailability and biological effects (Handy et al., 2003). For example, increases in the concentration of dissolved organic matter often decrease the bioavailability of heavy metals in aquatic sediments (Landrum et al., 1987), reducing their toxicity. Moreover, the presence of multiple metals can have synergistic or antagonistic effects on overall toxicity (Preston et al., 1999). The results of our study demonstrated that the responses of microbial community EEA to heavy metal stress provide information regarding heavy metal exposure and toxicity and thus supply an ecologically relevant measure of microbial community metabolism. The adaptation of microbial communities to local gradients of anthropogenic stress provides an important conceptual framework for linking small-scale spatial processes, such as local microbial community metabolism, to ecosystem-level processes. For example, we have previously demonstrated ecosystem-level patterns of EEA responses to various carbon substrates, which indicated that the source and availability of carbon may play a modulatory role in the local adaptation of microbial communities (Hoostal & Bouzat, 2008). Similarly, the resilience of microbial communities to heavy metal stress in polluted sediments suggests that microbial communities are able to adapt to toxic levels of these metals. In this study, EEA provided a metric for examining the resilience of microbial communities to heavy metal stress in a large ecosystem with a gradient of anthropogenic impacts. Certainly, the molecular mechanisms underlying the resilience of microbial EEA in polluted sediments, as well as geochemical factors that may affect the bioavailability of heavy metals, should be evaluated to confirm the results of this study. Lake Erie represents one of the largest freshwater lakes in the world, with a population of 11 million people living within its basin (Bolsega & 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

166

Herdendorf, 1993). Moreover, Lake Erie’s central basin offers one of the most historically well-documented instances of eutrophication as well as environmental degradation due to pollution (Burns, 1985; Mortimer, 1987). The present study represents an initial step to assess the potential role of contaminants in generating large-scale patterns of microbial community adaptations. Further studies on the structural and functional responses of microbial consortia to environmental pollutants will be essential for assessing the potential role of microbial communities in the bioremediation of contaminants for ecosystem recovery.

Acknowledgements We would like to thank Drs George Bullerjahn, Zhaohui Xu and Michael McKay, and three anonymous reviewers for their constructive comments. We also thank Dr McKay for the collection of sediment samples. The Department of Biological Sciences at BGSU, the Ohio Lake Erie Protection Fund, and the Professional Association of Diving Instructors (PADI) provided partial funding for this research.

References Alef K & Nannipieri P (1995) Methods in Applied Soil Microbiology and Biochemistry. Academic Press, London, pp. 228–231. Barkay T & Olson BH (1986) Phenotypic and genotypic adaptation of aerobic heterotrophic sediment bacterial communities to mercury stress. Appl Environ Microbiol 52: 403–406. Barras F, van Gijsegem F & Chatterjee AK (1994) Extracellular enzymes and pathogenesis of soft-rot Erwinia. Annu Rev Phytopathol 32: 201–234. Batra L & Manna MC (1997) Dehydrogenase activity and microbial biomass carbon in salt-affected soils of semiarid and arid regions. Arid Soil Res Rehabil 11: 295–303. Bidart-Bouzat MG, Portnoy S, DeLucia EH & Paige KN (2004) Elevated CO2 and herbivory influence trait integration in Arabidopsis thaliana. Ecol Lett 7: 837–847. Bolsega SJ & Herdendorf CE (1993) Lake Erie and Lake St. Clair Handbook. Wayne State University Press, Detroit. Bruins MR, Kapil S & Oehme FW (2000) Microbial resistance to metals in the environment. Ecotoxic Environ Safety 45: 198–207. Burns NM (1985) Erie: The Lake that Survived. Rowman and Allanheld, Totowa, NJ. Casida LE (1977) Microbial metabolic activity in soil as measured by dehydrogenase determinations. Appl Environ Microbiol 34: 630–636. Castillo MM, Allan JD, Sinsabaugh RL & Kling GW (2004) Seasonal and interannual variation of bacterial production in lowland rivers of the Orinoco basin. Fresh Biol 49: 1400–1414.

2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

M.J. Hoostal et al.

Cervantes C & Gutierrez-Corona F (1994) Copper resistance mechanisms in bacteria and fungi. FEMS Microbiol Rev 14: 121–137. Cervantes C, Guanyong J, Ramirez J & Silver S (1994) Resistance to arsenic compounds in microorganisms. FEMS Microbiol Rev 15: 355–367. Cervantes C, Campos-Garc´ıa J, Devars S, Guti´errez-Corona F, Loza-Tavera H, Torres-Guzm´an JC & Moreno-Sanchez R (2001) Interactions of chromium with microorganisms and plants. FEMS Microbiol Rev 25: 335–347. ´ RJ (1991) Environmental control of the synthesis and Chrost activity of aquatic microbial ectoenzymes. Microbial Enzymes ´ RJ, ed), pp. 29–51. Springerin Aquatic Environments (Chrost Verlag, New York. ´ RJ & Rai H (1993) Ectoenzyme activity and bacterial Chrost secondary production in nutrient-impoverished and nutrientenriched freshwater mesocosms. Microb Ecol 25: 131–150. Cochran VL, Elliott LF & Lewis CE (1989) Soil microbial biomass and enzyme activity in subarctic agricultural and forest soils. Biol Fertil Soils 7: 283–288. Collmer A & Keen NT (1986) The role of pectic enzymes in plant pathogenesis. Ann Rev Phytopathol 24: 383–409. Coombs JM & Barkay T (2004) Molecular evidence for the evolution of metal homeostasis genes by lateral gene transfer in bacteria from the deep terrestrial subsurface. Appl Environ Microbiol 70: 1698–1707. Crump BC, Hopkinson CS, Sogin ML & Hobbie JE (2004) Microbial biogeography along an estuarine salinity gradient: combined influences of bacterial growth and residence time. Appl Environ Microbiol 70: 1494–1505. Cullen JJ & Neale PJ (1994) Ultraviolet radiation, ozone depletion, and marine photosynthesis. Photosynthesis Res 39: 303–320. Eiler A, Langenheder S, Bertilsson B & Tranvik LJ (2003) Heterotrophic bacteria growth efficiency and community structure at different natural organic carbon concentrations. Appl Environ Microbiol 69: 3701–3709. Eisler R (1998) Copper hazards to fish, wildlife, and invertebrates: a synoptic review. Contaminant hazard reviews. USGS Report No. 33. Fisher H (2003) The role of biofilms in the uptake and transformation of dissolved organic matter. Aquatic Ecosystems: Interactivity of Dissolved Organic Matter (Findlay S & Sinsabaugh RL, eds), pp. 285–313. Academic Press, San Diego. Goodnight CJ (1990) Experimental studies of community evolution II: the ecological basis of the response to community selection. Evolution 44: 1625–1636. Goyal S, Mishra MM, Dhankar SS, Kapoor KK & Batra R (1993) Microbial biomass turnover and enzyme activities following the application of farmyard manure to field soils with and without previous long-term applications. Biol Fertil Soils 15: 60–64. Halliwell B & Gutteridge JM (1984) Oxygen toxicity, oxygen radicals, transition metals and disease. Biochem J 219: 1–14.

FEMS Microbiol Ecol 65 (2008) 156–168

167

Microbial community adaptation to heavy metals

Halliwell B & Gutteridge JM (1985) The importance of free radicals and catalytic metal ions in human diseases. Mol Aspects Med 8: 89–193. Handy RD, Galloway TS & Depledge MH (2003) A proposal for the use of biomarkers for the assessment of chronic pollution and in regulatory toxicology. Ecotoxicology 12: 331–343. Hoostal MJ & Bouzat JL (2008) The modulating role of dissolved organic matter on spatial patterns of microbial metabolism in Lake Erie sediments. Microb Ecol 55: 358–368. Hugenholtz P, Goebel BM & Pace NR (1998) Impact of cultureindependent studies on the emerging phylogenetic view of bacterial diversity. J Bacteriol 180: 4765–4774. Iwamoto T & Nasu M (2001) Current bioremediation practice and perspective. J Biosci Bioeng 92: 1–8. Judd KE, Crump BC & Kling GW (2006) Variation in dissolved organic matter controls bacterial production and community composition. Ecology 87: 2068–2079. Kandeler E, Tscherko D, Bruce KD, Stemmer M, Hobbs PJ, Bardgett RD & Amelung W (2000) Structure and function of the soil microbial community in microhabitats of a heavy metal polluted soil. Biol Fertil Soils 32: 390–400. Khattree R & Naik DN (1999) Applied Multivariate Statistics. SAS Institute, NY. Landrum PF, Nihart SR, Eadie BJ & Herche LR (1987) Reduction in bioavailability of organic contaminants to the amphipod Pontoporeia hoyi by dissolved organic matter of sediment interstitial waters. Environ Toxicol Chem 6: 11–20. Lee YB, Lorenz N, Kincaid Dick L & Dick RP (2007) Cold storage and pretreatment incubation effects on soil microbial properties. Soil Sci Soc Am J 71: 1299–1305. Mantel NA (1967) The detection of disease clustering and a generalized regression approach. Cancer Res 27: 209–220. Marroig G & Cheverud JM (2001) A comparison of phenotypic variation and covariation patterns and the role of phylogeny, ecology, and ontogeny during cranial evolution of new world monkeys. Evolution 55: 2576–2600. Martinez RJ, Wang Y, Raimondo MA, Coombs JM, Barkay T & Sobecky PA (2006) Horizontal gene transfer of PIB-Type ATPases isolated from radionuclide- and metal-contaminated subsurface soils. Appl Environ Microbiol 72: 3111–3318. McGarigal K, Cushman S & Stafford S (2000) Multivariate Statistics for Wildlife and Ecology Research. Springer-Verlag, New York. Mortimer CH (1987) Fifty years of physical investigations and related limnological studies on Lake Erie, 1928–1977. J Gt Lakes Res 13: 407–435. M¨uller AK, Westergaard K, Christensen S & Sørensen SJ (2001) The effect of long-term mercury pollution on the soil microbial community. FEMS Microbiol Ecol 36: 11–19. Ohio Lake Erie Commission (2004) 2004 State of the Lake Report: Lake Erie quality index. http://www.epa.state.oh.us/ oleo/reports/leqi/leqi2004/leqiz.htm Onstott TC, Phelps TJ, Kieft T, Colwell FS, Balkwill DL, Frederickson JK & Brockman FJ (1998) A global perspective on the microbial abundance and activity in the deep

FEMS Microbiol Ecol 65 (2008) 156–168

subsurface. Enigmatic Microorganism and Life in Extreme Environments (Seckbach J, ed), pp. 489–499. Kluwer Academic Publishers, the Netherlands. Painter S, Marvin C, Rosa F, Reynoldson TB, Charlton MN, Fox M & Estenik JF (2001) Sediment contamination in Lake Erie: a 25-year retrospective analysis. J Gt Lakes Res 27: 434–438. Peakall DB (1994) The role of biomarkers in environmental assessment (1). Introduction. Ecotoxicol 3: 157–160. Preston S, Coad N, Townsend J, Kilham K & Paton GI (1999) Biosensing the acute toxicity of metal interactions: are they additive, synergistic, or antagonistic. Environ Toxicol Chem 19: 775–780. Saiya-Cork K, Sinsabaugh R & Zak D (2002) The effects of long-term nitrogen deposition on extracellular enzyme activity in an Acer saccharum forest soil. Soil Biol Biochem 34: 1309–1315. SAS Institute Inc. (1999) SAS/STAT User’s Guide, Version 8. SAS Institute Inc., Cary, NC. Schneider S, Roessli D & Excoffier L (2000) Arlequin: A software for population genetics data analysis. v. 2.0. Genetics and Biometry Lab, Department of Anthropology, University of Geneva. Sigler WV & Zeyer J (2002) Microbial diversity and activity along the forefields of two receding glaciers. Microb Ecol 43: 397–407. Silver S (1998) Genes for all metals – a bacterial view of the Periodic Table. J Industrial Microbiol Biotech 20: 1–12. Silver S & Phung LT (1996) Bacterial heavy metal resistance: new surprises. Ann Rev Microbiol 50: 753–789. Sinsabaugh RL & Foreman CM (2001) Activity profiles of bacterioplankton in a eutrophic river. Freshwater Biology (Findlay S & Sinsabaugh RL, eds), pp. 471–489. Academic Press, San Diego. Sober E & Wilson DS (1998) Unto Others: The Evolution and Psychology of Unselfish Behavior. Harvard University Press, Cambridge. Sokal RR & Rohlf FJ (1995) Biometry. WH Freeman and Co., New York. Stadtmann ER (1993) Oxidation of free amino acids and amino acid residues in proteins by radiolysis and by metal-catalyzed reactions. Annu Rev Biochem 62: 797–821. Stohs SJ & Bagchi D (1995) Oxidative mechanisms in the toxicity of metal ions. Free Radical Biol Med 18: 321–336. Stolz JF & Oremland RS (1999) Bacterial respiration of arsenic and selenium. FEMS Microbiol Rev 23: 615–627. Sun HJ & Friedmann EI (2005) Communities adjust their temperature optima by shifting producer-to-consumer ratio, shown in lichens as models: II. Experimental verification. Microb Ecol 49: 528–535. Swenson W, Arendt J & Wilson DS (2000a) Artificial selection of microbial ecosystems for 3-chloroaniline biodegradation. Environ Microbiol 2: 564–571.

2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

168

Swenson W, Wilson DS & Elias R (2000b) Artificial ecosystem selection. Proc Natl Acad Sci 97: 9110–9114. Tabachnick BG & Fidell LS (1989) Using Multivariate Statistics. Harper and Row, New York. Tetaz TJ & Luke RK (1983) Plasmid-controlled resistance to copper in Escherichia coli. J Bact 154: 1263–1268. Torsvik V & Oevreas L (2002) Microbial diversity and function in soil: from genes to ecosystems. Curr Opin Microbiol 5: 240–245.

2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

c

M.J. Hoostal et al.

Waldrop MP, Zak DR & Sinsabaugh RL (2004) Microbial community response to nitrogen deposition in northern forest ecosystems. Soil Biol Biochem 36: 1443–1451. Wetzel RG (1992) Gradient-dominated ecosystems: sources and regulatory functions of dissolved organic matter in freshwater ecosystems. Hydrobiologia 229: 181–191. van der Werf H & Verstraete W (1987) Estimation of active soil microbial biomass by mathematical analysis of respiration curves: calibration of the test procedure. Soil Biol Biochem 19: 261–265.

FEMS Microbiol Ecol 65 (2008) 156–168

Suggest Documents