Susceptibility to additional stressors in metal-tolerant soil microbial communities from two pollution gradients

Susceptibility to additional stressors in metal-tolerant soil microbial communities from two pollution gradients

Applied Soil Ecology 98 (2016) 233–242 Contents lists available at ScienceDirect Applied Soil Ecology journal homepage: www.elsevier.com/locate/apso...

2MB Sizes 5 Downloads 65 Views

Applied Soil Ecology 98 (2016) 233–242

Contents lists available at ScienceDirect

Applied Soil Ecology journal homepage: www.elsevier.com/locate/apsoil

Susceptibility to additional stressors in metal-tolerant soil microbial communities from two pollution gradients Hamed Azarbada,b,* , Nico M. van Straalenb , Ryszard Laskowskia , Karolina Nikiela ,  skaa Wilfred F.M. Rölingc, Maria Niklin a

Institute of Environmental Sciences, Jagiellonian University, Gronostajowa 7, 30-387 Krakow, Poland Department of Ecological Science, Faculty of Earth and Life Sciences, VU University Amsterdam, de Boelelaan 1085, 1081 HV Amsterdam, The Netherlands Department of Molecular Cell Physiology, Faculty of Earth and Life Sciences, VU University Amsterdam, de Boelelaan 1085, 1081 HV Amsterdam, The Netherlands b c

A R T I C L E I N F O

A B S T R A C T

Article history: Received 16 February 2015 Received in revised form 15 July 2015 Accepted 21 October 2015 Available online 4 November 2015

Soil microbial communities exposed to long-term metal pollution have been shown to maintain their functionality by developing tolerance under field conditions. This study investigated how the cost of acquiring metal tolerance affected the responses of such communities to secondary stressors (stress-onstress). We studied soil microbial communities living in soils polluted by metals to different degrees, sampled along two metal pollution gradients in Poland. We hypothesized that communities adapted to high concentrations of metals would be more susceptible to additional stressors that they had not previously encountered (benzo[a]pyrene and salt), compared to communities exposed to low concentrations of metals. In contrast, metal-adapted communities were expected to be more stable when exposed to stress factors within the same category of toxicity or to stress factors that periodically occur naturally, such as arsenic and flooding. In 60-day microcosm experiments, phospholipid fatty acid (PLFA) profiles were used to compare changes in microbial communities sampled from the field and exposed to four different additional stressors. Our results showed that, regardless of existing pollution, additional stress restructured the microbial communities similarly in all soils. However, salt and flooding stress had stronger effects on community structure than arsenic and benzo[a]pyrene. PLFAs indicative of fungi, Gram-negative and Gram-positive bacteria generally declined under stress relative to unspecific PLFAs (such as 16:0 and 18:0). This trend may be caused by the inhibition or suppression of sensitive microbial groups and selection of tolerant groups by the stressors applied and also because of the proliferation of fast-growing species under stress conditions. Overall, our study showed that metaltolerant communities that have been selected in the field over many years are not more susceptible to additional stress than communities exposed to low concentrations of metals; thus, adaptation in these microbes has evolved without apparent costs. ã 2015 Elsevier B.V. All rights reserved.

Keywords: Secondary stressors Metal pollution PLFA Forest soil Microbial communities

1. Introduction Soil microbial activity has a substantial impact on ecosystems because microorganisms play a key role in biogeochemical cycles. Metabolized carbon leaves the soil primarily as carbon dioxide (soil respiration) and represents one of the greatest fluxes in the global carbon cycle (Schindlbacher et al., 2011). Pollution-induced shifts in the structure of microbial communities in soil organic layers may affect decomposition rates and CO2 production (Azarbad et al.,

* Corresponding author at: Institute of Environmental Sciences, Jagiellonian University, Gronostajowa 7, 30-387 Krakow, Poland. E-mail address: [email protected] (H. Azarbad). http://dx.doi.org/10.1016/j.apsoil.2015.10.020 0929-1393/ ã 2015 Elsevier B.V. All rights reserved.

2013). Long-term exposure to metals may result in selection for distinct soil microbial communities, and these community changes may affect responses of ecosystem functions to additional stressors. How different microbial groups (i.e., fungi, Grampositive and Gram-negative bacteria, etc.) selected for metal tolerance, respond to additional stressors, however, is not well understood. This study focuses on two organic forest soils with a history of metal contamination (since the 1970s), predominantly by zinc (up to 4300 mg kg 1 total Zn) and lead (up to 2900 mg kg 1 total Pb) (Azarbad et al., 2013). Two point sources, Olkusz and Miasteczko  ˛skie in Poland, have caused separate yet similar gradients of soil Sla metal concentrations. We previously showed that long-term exposure to metal contamination has negatively influenced

234

H. Azarbad et al. / Applied Soil Ecology 98 (2016) 233–242

microbial biomass (as revealed by phospholipid fatty acids, PLFAs) and basal respiration rate, whereas such exposure has not resulted in large shifts in community structure (determined by Illumina sequencing of 16S rRNA genes) or the composition of functional capabilities (established by Functional Microarray analysis: GeoChip 4.2) along these two gradients (Azarbad et al., 2013, 2015a). Our recent study on these gradients revealed that long-term exposure of soil microbial communities to metal pollution improved their functionality (as determined by basal respiration rates) and structural stability (studied with denaturing gradient gel electrophoresis, DGGE) in response to additional metal (arsenic) and salt (NaCl) stress, whereas no difference between heavily polluted and less polluted soils was found in the responses to benzo[a]pyrene and flooding stress (Azarbad et al., 2015b; see Supplementary Fig. S1). DNA has a relatively long persistence in the environment, even when the associated organisms are no longer viable (CórdovaKreylos et al., 2006). This persistence can be a disadvantage when analyzing community shifts using DNA-based molecular approaches, such as DGGE. In contrast, PLFAs represent the viable component of a community because phospholipids outside the cell are rapidly dephosphorylated and, thus, no longer detectable when the associated organisms die. This feature is particularly important when using PLFAs as an indicator of the entire potentially active microbial community (including Bacteria, Archaea and microbial Eukarya, such as fungi) as opposed to non-living biomass. Community-broad changes in the relative abundance of microbial groups and in physiological status can be easily be detected using PLFA profiles (Bååth et al., 2005). This is due to both physiological and selective responses. When exposed to stressors, bacterial cells are able to adapt to changes in their membrane fluidity mainly by modifying their membrane fatty acid composition (degree of saturation) to maintain membrane fluidity at a constant level (Heipieper et al., 1992; Petersen and Klug, 1994). In addition, susceptible groups may be inhibited and resistant groups increase, giving rise to switches in the abundance of PLFAs specific to microbial groups. Given the increasing multiplicity of environmental stressors associated with global change, there is an urgent need to develop a better understanding of the interactive effects of multiple stressors on ecosystems to better predict their responses to a changing environment (Vinebrooke et al., 2004). The combined impact of pollutants and additional environmental stressors on natural soil communities is generally understudied. To increase our understanding of such interactive effects, including ‘stress-on-stress,’ it is essential to know which microbial groups are selected in microbial communities under intense selective pressure by metals and how this selection affects their responses to additional stressors. Differences between communities adapted to high concentrations and low concentrations of metals are expected because of trade-offs among the various adaptive responses, especially when these responses involve stress factors of different natures. Such trade-offs are often referred to as ‘costs of tolerance’ (Rundle and Nosil, 2005). The question of increased tolerance to secondary stress in stress-adapted communities is of great relevance to risk assessment. If adaptation to one stress factor leads to the loss of resistance against other stress factors, adaptation should not be considered a positive phenomenon and cannot, for instance, be taken as an argument to raise maximum acceptable concentrations of toxicants. If, however, adaptation provides a general improvement of vigor and resistance to many different stress factors, concerns about chronic pollution in the field might be moderated. The aim of this study was to test the hypotheses that (1) microorganisms in highly polluted organic forest soils differ in their tolerance to additional stressors compared to those in less-

polluted soils and (2) the response to a stressor is consistent across different forests. Previously, we observed that simultaneously studying these two gradients allowed for much stronger inference about effects of metal pollution (Azarbad et al., 2013, 2015a). The first hypothesis was tested by exposing microbial communities inhabiting soils differing vastly in metal pollution levels to four different stressors; the second was studied by using two heavy metal pollution gradients in Scots pine forests near Olkusz and  ˛skie in southern Poland. We also predicted that the Miasteczko Sla different types of stressors used in this study would differentially affect microbial communities and the composition of their cell membranes. 2. Materials and methods 2.1. Site description and soil sampling Characteristics of the collected soils, including total and soluble metal levels, toxicity indices (TI), pH and organic matter contents have been reported previously (Azarbad et al., 2013). Briefly, TI was calculated as S(Ci/EC50i), where Ci is the concentration of metal i in soil and EC50i is the concentration of that metal causing 50% reduction in dehydrogenase activity as reported by Welp (1999). We used two different measures of metal pollution: TItot is based on total metal concentrations, and TIwe is based on waterextractable concentrations. By summing the relative concentrations, a measure is obtained for the cumulative metal loading, each metal weighted by its toxicity. For the sake of completeness, basic soil properties of the sampling sites (e.g., TI, pH, organic matter content and water holding capacity) are reproduced in Supplementary Table S1. In the Olkusz area (O), a transect was established extending  ˛skie area 1.9 to 31.8 km from a smelter, and in the Miasteczko Sla (M), a transect was established from 2.1 to 52.6 km from the source of pollution. Along each transect, six sampling locations were chosen (running from high to low metal pollution), all in Pinus sylvestrisforest stands. At each location, an approximately 100 m2 plot (based on our earlier study which provided baseline knowledge on microbial community structure and general soil functioning in relation to soil characteristics) was designated (Azarbad et al., 2013, 2015a). During three sampling times (May, June and August 2013), ten samples (approximately 400–500 g wet mass per sample) of the organic topsoil O layer (approximately 10 cm thick) were randomly taken with a 5 cm-diameter soil auger, and the top 10 cm of each sample was used for analysis, sieved (mesh size 2 mm), and mixed to obtain a single representative sample per location and per month. Soil samples were stored at 4  C for 10 days before any microbiological analyses were performed. 2.2. Soil microcosms experiments for evaluating the effects of different stressors In the experiment, a total of 360 microcosms comprising soils originating from 12 sampling locations along the two field gradients were used. Microcosms were subjected to one of five treatments, and each combination of soil, treatment and time interval was run in triplicate. Observations were conducted at the beginning (day 1) and the end (day 60) of the experiment. The measurements required sacrificing the microcosms; thus, the observations were all independent of one another. The five treatments were as follows: (1) untreated soil (control, treated only with deionized water); (2) As2O5 to test the tolerance of soil microbial communities in polluted soils to a metal different from those to which the communities have been chronically exposed in nature (Cd, Zn and Pb); (3) benzo[a]pyrene as an

H. Azarbad et al. / Applied Soil Ecology 98 (2016) 233–242

organic pollutant, which is different from the primary stress (metal pollution); (4) NaCl as a strong osmotic stress, new to all studied communities; (5) flooding, a stressor commonly experienced by all communities because prolonged rain events have frequently occurred in the past. Arsenic (As2O5) was dissolved in deionized water and added to the soils at a concentration equivalent to 2 g kg 1 dry weight. Because of the release of arsenic from industrial sources and in mining areas, its concentration in soil can reach 20 g kg 1 dry weight. Therefore, the concentration chosen in our study was within the range of adverse effects in polluted soils (Smith et al., 1998). Benzo[a]pyrene was dissolved in dichloromethane (1 g L 1) and added to the soil to obtain a final concentration of 0.5 g kg 1 dry weight following the recommendations of Brinch et al. (2002). 20% of a soil sample was spiked with the benzo[a]pyrene solution and thoroughly mixed. After evaporation of the dichloromethane (overnight at 22  C), the spiked sub-sample (20%) was mixed with the rest (80%) of the soil and shaken thoroughly. The final benzo[a] pyrene concentration (0.5 g kg 1 dry weight) corresponds to a level of PAH triggering action, according to UK regulation (Jones et al., 1996; Klimkowicz-Pawlas and Maliszewska-Kordybach, 2003). The salt concentration was chosen based on demonstrated stress effects (changes in respiration rate) in previous research (ToborKaplon et al., 2005). Flooding was simulated by soaking soil samples (moisture level of 100% of WHC) in deionized water to the saturation point with distilled water to a level of 2–3 cm above the soil surface.

235

Microcosms consisting of soil samples equivalent to 20 g dry weight were placed in sterile 150-mL plastic jars. Soil moisture content was adjusted to 50% of the water-holding capacity (except for the flooding stress samples) to obtain maximum respiration  ska et al., 2005). After thorough mixing, all jars responses (Niklin were closed (not tightly) with caps to limit moisture loss while allowing for aeration. The microcosms were incubated at 22  C (1  C) for 60 days. Moisture levels were maintained at 50% WHC (except for the flooding stress samples) by monitoring sample weights every third day and compensating for evaporative water loss with deionized water. Soil subsamples were taken for PLFA analysis at days 1 and 60 after applying the stressors. All microbial analyses described in the following sections were performed in triplicate, and the results are presented as the mean values with standard deviations. 2.3. Phospholipid fatty acid analysis (PLFA) The response of soil microbial community structure to stressors was determined by phospholipid fatty acid (PLFA) analysis as described by Frostegård and Bååth (1996). Briefly, total soil lipids were extracted from 1 g of fresh soil using a phosphate buffer: methanol:chloroform (1:2:0.8, v/v/v) mixture. After splitting the extracts into two phases by adding chloroform and buffer, the lipid-containing phase was dried under a stream of nitrogen and stored at 20  C. The lipid material was fractionated on a silica column into neutral and glycol/phospholipid-containing polar

Fig. 1. Arsenic (As(V): 2 g kg 1 dry weight) as a secondary stressor in metal-polluted forest soils. The samples were derived from twelve locations along two metal pollution  ˛skie (M1–M6; running from high to low metal pollution) and Olkusz (O1–O6; also running from high to low metal pollution), southern Poland. (A) gradients in Miasteczko Sla Changes in the biomass of Gram-positive (Gram+) bacteria, Gram-negative (Gram–) bacteria, unspecific PLFAs, fungal PLFAs and total PLFAs. Microbial biomass PLFAs in stressed samples are expressed relative to the control (%). (B) Principal component analysis (PCA) ordination biplots of the microbial PLFA relative abundance (mol%) in the stressed soil samples (blue) and their controls (red). The percentage of the total variance explained by each axis is indicated next to the axis name. The PLFAs were measured after 1 day (left) and 60 (right) days of exposure. Error bars indicate standard deviation (n = 3). See the Materials and Methods section for the stress doses, identification of the PLFAs used to calculate the biomass for specific microbial groups. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

236

H. Azarbad et al. / Applied Soil Ecology 98 (2016) 233–242

lipids. The phospholipid fraction was dried under a flow of nitrogen and saved for derivatization. The phospholipids were subjected to a mild alkaline methanolysis, and the methylated lipids were identified by their relative retention times using gas chromatography (PerkinElmer1 Clarus1 600 Gas Chromatograph, Shelton, U. S.A.) with flame ionization detection. Methyl nonadecanoate (19:0) was used as an internal standard. The abundance of individual PLFAs was expressed as the mole percentage of the total PLFA. Total microbial biomass was estimated from the total PLFAs. Certain PLFAs are unique biomarkers of particular functional groups of microorganisms, allowing for quantification of the biomass of specific groups. Based on literature, PLFAs were grouped into bacterial (Gram-positive and Gram-negative bacteria), fungal and unspecific origins. The PLFAs 18:2v6c and 18:2v9c were used as indicators of fungal species; i15:0, a15:0, i16:0 and i17:0 and branched PLFAs were used as indicators of Gram-positive bacteria (Gram+); and cy17:0, cy19:0, and 16:1v9c were used as indicators of Gram-negative bacteria (Gram–). Straight-chain PLFAs, including 14:0, 16:0, 18:0 and 20:0, were used as unspecific origins PLFAs, mainly non G+ and G– bacteria (non-specific bacteria) (Vestal and White, 1989; Frostegård et al., 1991; Zelles et al., 1992; Frostegård and Bååth, 1996; Zogg et al., 1997; Grayston et al., 2001; Phillips et al., 2002; Potthoff et al., 2006; Liang et al., 2008). To assess the effects of stressors, not only individual PLFAs but also the ratios of fungal PLFAs to bacterial PLFAs (fungal/sum of Gram+ and Gram–) and the ratio of Gram+/Gram– were calculated (Zhang et al., 2013). Adaptation in membrane composition, in terms of changes in the degree of saturation, including the relative abundances of total

saturated/total monounsaturated acids (SFAs/MUFAs): (16:0 + 17:0 + 18:0 + 20:0)/(16:1v9c + 18:1v9c), were tested as possible bioindicators of environmental stress (Fierer et al., 2003; Mazzella et al., 2007). 2.4. Statistical analyses Any PLFAs detected in less than 10% of the samples were excluded from statistical analysis (Potthoff et al., 2006). Mole percentage values were arcsine transformed to obtain a normal data distribution (Fry, 1996). Arcsine-transformed mole percentages of individual PLFAs were used as input values in principal components analysis (PCA) based on Pearson correlations to explore the variability among stressors and control treatments. Biplots of the first two principal components were used to visualize dissimilarity among microbial communities. The loading scores for the individual PLFAs were used to assess the relative contribution of each individual PLFA to the principal component axes (Fierer et al., 2003). The percentage change in (1) PLFA-based biomass, (2) the relative abundance of microbial groups, (3) fungal/bacterial ratio, (4) Gram+/Gram– and (5) SFAs/MUFAs ratios in each treatment were expressed relative to the control on days 1 and 60. A paired t-test was used to identify significant differences (p < 0.05) in PLFA ratios between control and stress-treated soils across all locations separately in each transect. Simple regression analysis was performed to determine the influence of primary pollution (expressed as toxicity index, TI) on the percent change of biomass and relative abundances of the above-mentioned microbial groups across all locations in each transect.

Fig. 2. Benzo[a]pyrene (0.5 g kg 1 dry weight) as a secondary stressor in metal-polluted forest soils along two gradients. (A) Percentage change in biomass of different microbial groups relative to the corresponding control (%). (B) Principal component analysis (PCA) ordination biplots of the microbial PLFA relative abundance (mol%) in the stressed soil samples (blue) and their corresponding controls (red). See the description of Fig. 1 for more details. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

H. Azarbad et al. / Applied Soil Ecology 98 (2016) 233–242

Statistical analyses were performed using PAST (PAleontological STatistics) (Hammer et al., 2001) and Statgraphics Centurion XV software (StatPoint, Herndon, VA, USA). 3. Results PCA of the PLFA profiles suggested no significant differences in control soils between day 1 and day 60 of the incubation (Fig. S2); hence, any observed significant variation in PLFA profiles between treated soils and controls can be attributed to the applied stressor. 3.1. Response to arsenic (additional metal stress) Contamination levels (TI) at each pollution gradient had no significant impact on either total microbial biomass (PLFAtot) or relative abundances of the different microbial groups in arsenicstressed samples. PLFAtot increased in almost all arsenic-stressed samples, in comparison to their controls, after one day of exposure  ˛skie and 53% for Olkusz; (on average, 52% for Miasteczko Sla Fig. 1A). After 60 days, the PLFAtot declined in arsenic-treated samples by 38% on the average (relative to PLFAtot in the stressed soils at day 1) and approached the value of the corresponding  ˛skie controls in almost all locations along the Miasteczko Sla transect. However, the PLFAtot remained higher along the Olkusz transect (Fig. 1A). PCA of the PLFA profiles suggested substantial differences in microbial community structure between stressed samples and controls one day after arsenic addition. The first principal component (PC1) explained 32% of the variance in the data, and the second (PC2) explained the 24%. The arsenic-stressed samples

237

were separated along PC2 from the control samples (Fig. 1B). The greatest variation along PC1 was not directly related to the stressor, indicating that the effect of natural variation (e.g., concentrations of heavy metals and pH values) was also somewhat stronger than the impact of stressors on the PLFA profiles. The PLFAs, which showed large negative loadings (loading score < 0.3) on PC2 and increased substantially in relative abundance in arsenic-treated samples, were PLFAs 18:0, 16:1v9, cy17:0 and cy19:0. In contrast, i16:0, i15:0 had large positive loadings on PC2 (loadings >0.3) and were found in relatively large proportions in the control samples. A similar pattern was observed after 60 days of exposure to arsenic. To more thoroughly investigate shifts in the arsenic-stressed microbial communities in comparison with the control, PLFAs indices, such as Gram+/Gram–, fungal/bacterial and saturated and mono-unsaturated fatty acid ratios (SFAs/MUFAs), were calculated (Fig. S4A, S4B). No significant association between TI and PLFAs indices were observed in both transects. 3.2. Response to benzo[a]pyrene (organic pollution stress) In general, no significant effect of benzo[a]pyrene on the total microbial biomass was detected at day 1 or day 60. However, benzo [a]pyrene had an impact on fungal biomass (Fig. 2A). Fungal markers exhibited significantly higher relative abundance (t = 4.1, p = 0.001) in almost all stressed samples in comparison with their controls, immediately after the stress (day 1) (on average, 39% at  ˛skie and 50% at Olkusz; Fig. S3C). In general, relative Miasteczko Sla abundance of fungal markers for day 60 remained high, but there was no significant difference between stressed samples and their controls, showing a certain degree of recovery (Fig. S3D).

Fig. 3. Salt (6.5 g kg 1 dry weight) as a secondary stressor in metal-polluted forest soils along two gradients. (A) Percentage change in biomass of different microbial groups relative to the corresponding control (%). (B) Principal component analysis (PCA) ordination biplots of the microbial PLFA relative abundance (mol%) in the stressed soil samples (blue) and their corresponding controls (red). See the description of Fig. 1 for more details. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

238

H. Azarbad et al. / Applied Soil Ecology 98 (2016) 233–242

PCA revealed differences between stressed and control samples for both day 1 and day 60 (Fig. 2B). In general, stressed samples had higher negative scores on the PC2 axis compared to the control samples. PLFAs 18:1v9c and 18:2v6c were more abundant in the stressed communities, both immediately after the stress (day 1) and after 60 days (Fig. 2B). The increase in fungal abundance was also reflected in the ratio of fungal-to-bacterial PLFAs, which increased in almost all stressed soils in comparison with the controls on day 1 (t = 4.8, p = 0.0004) and, to some extent, on day 60, regardless of existing pollution. Other ratios remained stable and near the control ratio (Figs. S4C, S4D). 3.3. Responses to salt (osmotic stress) No significant relationship was found between either the absolute or relative abundances of the different microbial groups and pollution levels. A small increase in the level of total microbial biomass in the salt-stressed samples was observed in comparison to their controls after one day of exposure (on average, 13% for  ˛skie and 11% for Olkusz; Fig. 3A). After 60 days of Miasteczko Sla exposure, PLFAtot remained elevated (relative to PLFAtot in the stressed soils at day 1), increasing by 13% on average for  ˛skie and by 37% for Olkusz (Fig. 3A). One day of Miasteczko Sla salt stress led to increased unspecific PLFAs in all soils compared to  ˛skie and 49% at the controls (on average, 66% at Miasteczko Sla Olkusz; Fig. 3A). In contrast, concentrations of fungal, Gram– and Gram+ bacteria PLFAs decreased considerably in salt-stressed samples compared to control soil. A consistent trend was also observed on day 60.

PCA revealed that the salt-stressed samples at day 1 were separated from control samples along PC1 (Fig. 3B). PC1 explained 60% of the variance. This was not the case for the previous two stressors, for which the greatest variation along PC1 was not directly related to the stressor and stressed samples were separated along PC2 from the controls (Figs. 1 and 2B). The PLFAs, which showed large negative loadings (loading score < 0.3) on PC1 and increased substantially in proportional abundance in salttreated samples, were PLFAs 16:0 and 18:0. Most of the PLFAs, e.g., 16:1v9c and 18:1v9c, had high positive loadings on PC1 (loadings >0.3) and were found in relatively large proportions in the control samples. A similar pattern was observed after 60 days of salt exposure (Fig. 3B). The ratios of Gram+/Gram– (day 1: t = 6.2, p < 0.0001; day 60: t = 6.1, p < 0.0001) and total saturated/total monounsaturated fatty acids (day 1: t = 8.8, p < 0.0001; day 60: t = 6.5, p < 0.0001) increased significantly in stressed samples compared to the controls on days 1 and 60, regardless of the pollution level (Figs. S4E, S4F). 3.4. Responses to flooding (common stress) The pollution level had no significant impact on the change in PLFAtot and relative abundances of the different microbial groups in flooding-stressed samples. The PLFAtot decreased in almost all flooding-stressed samples compared to the control after one day of  ˛skie and 20% for exposure (on average, 13% for Miasteczko Sla Olkusz; Fig. 4A). A similar pattern was observed after 60 days of exposure to flooding. Flooding stress dramatically impacted fungal,

Fig. 4. Flooding (100% WHC) as a secondary stressor in metal-polluted forest soils along two gradients. (A) Percentage change in biomass of different microbial groups relative to the corresponding control (%). (B) Principal component analysis (PCA) ordination biplots of the microbial PLFA relative abundance (mol%) in the stressed soil samples (blue) and their controls (red). See the description of Fig. 1 for more details. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

H. Azarbad et al. / Applied Soil Ecology 98 (2016) 233–242

Gram+ and Gram– bacterial biomass in comparison to controls in both transects (Fig. 4A). The first principal component (PC1) explained 68% of the variance in the data, and the second (PC2) explained the next 13%. Similar to the salt stress, the flooding-stressed samples were separated from the control samples along PC1 (Fig. 4B). The PLFAs, which showed large negative loadings (loading score < 0.3) on PC1 and increased substantially in relative abundance in floodingtreated samples, were PLFAs 16:0 and 18:0. PLFAs 16:1v9c, cy17:0, 18:1v9c, and cy19:0, which had large positive loadings in PC1 (loadings >0.3), were relatively more abundant in the control samples. Generally, a pattern similar to that of day 1 was observed after 60 days of flooding exposure (Fig. 4B). As observed for salt stress, Gram+/Gram– (day 1: t = 6.2, p < 0.0001; day 60: t = 6.2, p < 0.0001) and total saturated/total monounsaturated fatty acids (day 1: t = 10.6, p < 0.0001; day 60: t = 6.5, p < 0.0001) increased substantially after flooding compared to the control on days 1 and 60, regardless of existing pollution (Figs. S4G, S4H). 3.5. Combination of all stressors To evaluate whether the applied stressors resulted in similar community responses, we further analyzed community structures by comparing PLFA data for all four stressors and their controls. The resulting PCA for day 1, shown in Fig. 5A, indicates that 42% of the variation was accounted for by PC1 and 21% by PC2. Along PC1, the salt- and flooding-stressed samples overlapped but were separated from those for benzo[a]pyrene, arsenic and the control samples, which were grouped (Fig. 5A). PC1 had a clear negative relationship (correlation < 0.3) with two PLFA markers (16:0 and 18:0) that were associated with flooding and salt stress. The PLFAs cy17:0, 16:1v9 and 18:1v9c had large positive loadings in PC1 (loadings >0.3). These PLFAs described control, benzo[a] pyrene and, to some extent, arsenic-stressed samples. PC2 had a negative relationship (correlation < 0.3) with PLFAs 18:2v6c and 18:1v9t, and PLFAs 17:0, i15:0 and i16:0 had large positive loadings in PC2 (loadings >0.3). In general, the PCA for day 60 showed a similar pattern to that of day 1 (Fig. 5B).

239

4. Discussion The aim of this study was to test whether soil microbial communities from sites differently polluted with metals differ in their responses to additional stressors. Our results show that the background of metal levels did not affect the response degree of the communities to additional stressors, measured either as changes in microbial biomass or community structure. Thus, the hypothesis that communities adapted to high metal concentrations would be more susceptible to additional stressors to which they had not been previously exposed (benzo[a]pyrene and salt) and more stable in response to additional metal stress (arsenic) and flooding stress than communities adapted to low metal concentrations was not supported. The present study allowed for clear conclusions regarding the effects of additional stressors on bacteria and fungi in metal-polluted soils because we considered changes in microbial communities that were common to two distinct field gradients and polluted by different smelters. PLFA profiles demonstrated that, although the four stress factors caused direct qualitative and quantitative effects, the effects were independent of the origin of the communities. 4.1. Increases in microbial biomass in response to some stressors A remarkable observation was that microbial biomass increased in arsenic- and salt-treated soils compared to the controls, regardless of the background pollution level. This is supported by other studies, which have shown that metal stress does not always reduce biomass (Knight et al., 1997; Feris et al., 2003). This means that density compensation occurred rapidly, i.e., tolerant species quickly replaced sensitive species (Van Der Wurff et al., 2007). This phenomenon was contrary to our expectation that the microbial biomass would decrease in treated soils as a result of additional stress and that this effect would be more pronounced in highly polluted soils. One possible explanation for this result could be related to the availability of easily decomposable carbon derived from sensitive organisms killed by arsenic and salt stress. This would benefit microbial populations able to realize rapid growth and

Fig. 5. Principal component analysis (PCA) ordination biplots of the microbial PLFA relative abundance (mol%) of a combination of all stressed-samples (three sampling times: May, June and August 2013) after 1 (A) and 60 days (B) of exposure. Arsenic (blue), benzo[a]pyrene (brown), salt (green), flooding (pink) and control (red). The percentage of the total variance explained by each axis is indicated next to the axis name. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

240

H. Azarbad et al. / Applied Soil Ecology 98 (2016) 233–242

reproduction in face of stress, while species that have successfully managed to evolve the complex adaptations needed to resist stress cannot profit so quickly from these new sources. 4.2. Changes in microbial community structure and PLFAs indices after applying different classes of stressors All stressors applied in our experiments caused shifts in the structure of the microbial communities studied. We did not observe an increase in specific fatty acids classes. However, there was a substantial increase in the abundance of unspecific PLFAs (such as 16:0 and 18:0) under flooding, salt and, to some extent, arsenic stress compared to the corresponding control (see Supplementary Fig. S3), suggesting that certain soil microorganisms are able to quickly occupy new niches or rapidly adapt to new conditions. These fatty acids are not good indicators for (a) specific group(s) of microorganisms, since they occur in a wide range of microorganisms (Lechevalier, 1977). The most likely explanation for the increase in the abundance of unspecific PLFAs is the possible presence of a large pool of microorganisms from which species might be selected (by growth and/or adjustment) under new conditions. The phylogenetic and functional richness of microbial communities were previously observed to be unaffected by historical pollution levels (Azarbad et al., 2015a), which may explain why TI did not affect the relative changes in community structure. 4.2.1. Arsenic Arsenic stress has an effect on the membrane adaptive response, as shown by the high degree of saturation found in arsenic-exposed microorganisms. An increase in the saturation ratio at high arsenic concentrations in the current study might help to reduce arsenic uptake and minimize its toxic effects. Similar effects have been reported previously. For example, Pepi et al. (2008) observed an increase in the degree of fatty acid saturation as a result of the simultaneous effect of toluene and arsenic on the PLFA profiles in Pseudomonassp. ORAs5 and Bacillussp. ORAs2. 4.2.2. Benzo[a]pyrene As shown by the high fungal-to-bacterial ratio in most stressed soils, it appears that fungi make up a greater proportion of the microbial community in benzo[a]pyrene-stressed soils compared to the controls. Other studies have shown similar responses (Joynt et al., 2006; Langworthy et al., 1998). It has been suggested that PAH degradation in nature is a consequence of sequential breakdown by fungi and bacteria, with the fungi performing the initial oxidation step (Meulenberg et al., 1997; Kotterman et al., 1998). This dynamic would explain why fungal biomass rapidly increased. 4.2.3. Salt The increase in the ratios of Gram+/Gram– and SFAs/MUFAs in salt-treated soils indicates differences in tolerance to salt between Gram+ and Gram– bacteria. Gram+ bacteria contain strong cell walls in contrast to a single-layer cell wall in Gram– bacteria (Schimel et al., 2007). Therefore, Gram– bacteria appear to be more sensitive to salt stress, as shown in this study. The greater sensitivity of fungi to salt stress compared to bacteria on day 1 is consistent with previous studies (Chowdhury et al., 2011). Salt stress is known to affect the fluidity of the cell wall, and lipid regulation plays critical roles in cell wall fluidity (Los and Murata, 2004). As shown by PCA, the fatty acids predominating under saltstress conditions were saturated fatty acids (unspecific PLFAs), such as 16:0 and 18:0, indicating that salt stress gave rise to an increase in both long-chain fatty acids and straight-chain fatty acids (Fig. 3B). With increasing environmental salinity, the internal

osmotic pressure of cells also increases. The increase in saturated fatty acids with increasing salinity suggests a reduction in membrane fluidity (with a more rigid membrane) and permeability. Such a change would improve the ability of microorganisms to prevent solute leakage from the cell at high salinity to control osmotic pressure (Nicolaus et al., 2001). 4.2.4. Flooding The study showed that inundation had strong effects on the abundance and biomass of the microbial community, as shown by various markers (e.g., SFAs/MUFAs ratio). Flooding decreased the abundance of fungi and Gram– and Gram+ bacteria. However, the abundance of unspecific PLFAs markers (such as 16:0 and 18:0) increased substantially in the flooded soils (Fig. S3). The decreased presence of fungi under flooded conditions, also observed in previous studies (Bossio and Scow, 1998; Drenovsky et al., 2004; Mentzer et al., 2006), confirms the sensitivity of fungi to inundation (Unger et al., 2009). Flooding stress is a rapid event, and in microorganisms without effective stress response mechanisms, water will flow into the cell unless it has strong cell walls such as in Gram+ bacteria (Schimel et al., 2007). The observed increased fatty acid saturation and remarkable decrease in biomass and abundance of Gram– bacteria under flooding conditions suggest that, aerobic Gram– bacteria were particularly affected by inundation. Previous studies have shown that mono-unsaturated fatty acids are usually associated with aerobic growth (Bossio and Scow, 1998; Bossio et al., 2006). The increase in SFAs (straight-chain PLFAs) in flooded soils indicates that cell membranes under flood stress adapted by increasing the rigidity of the membrane, similar to the results from salt stress, causing lower water-diffusive permeability compared to the controls. 4.3. Risk assessment of contaminated forest soils in response to additional stressors Our results showed that the historical level of metal pollution did not determine stability of microbial biomass nor abundances of microbial groups. Regardless of the existing pollution, additional stress restructured the microbial communities in all soils in a similar manner. In addition, microbial communities selected for metal resistance did not show a loss of their capacity to deal with additional stress. This observation may be taken to argue that adaptation should not be considered a negative phenomenon in risk assessment. Our previous study revealed that the respiration rate (measured as functional responses) remained, to some extent, stable in the case of flooding and benzo[a]pyrene (except for day 1) (Azarbad et al., 2015b; Fig. S1); however, the PLFA results revealed substantial changes in community structure. Functional redundancy might explain the phenomenon of limited changes in the activities of microorganisms despite evident changes in community structure. It is expected that if the community structure is altered it will not necessarily be reflected in total microbial activity, as more resistant species replace the more sensitive ones (Nannipieri et al., 2003). These results suggest that functional redundancy within microbial communities in stress-impacted soils, which, by buffering the selective effects of stressors, enables microbial communities to perform similar functions (e.g., catabolize a similar suite of carbon substrates) despite the observed shifts in community structure. Therefore in risk assessment studies it would be best to consider the responses of ecological function after applying stressors and to classify a case as high risk if it shows a reduction of function (activity), rather than a change in community structure.

H. Azarbad et al. / Applied Soil Ecology 98 (2016) 233–242

5. Conclusion Overall, unspecific fatty acids (such as 16:0 and 18:0), rather than the markers of specific, well-defined groups, were indicative of stress-induced changes in the microbial communities. Although the changes in microbial communities in response to the four additional stress factors were clear, the effects did not differ between soils of different pollution levels. This finding indicates the absence of significant trade-offs between adaptation to high metal levels in the soil and sensitivity to additional stressors. With the expected increase in extreme weather events due to climate change and industrial pressures on nature, we can certainly anticipate an increase in multi-stressed ecosystems. It is therefore worth carrying out research using other types of stressors than we used in this study e.g., heat shock and also longer periods incubation in order to promote a standardized microbial stability tool to study the effects of secondary stressors on metal-tolerant microbial communities. Acknowledgements Dr Wilfred Röling, associated professor from Molecular Cell Physiology department, died on Friday the 25th of September 2015, at the age of 48 years. Our thoughts wander to the many moments of inspiration, joy and discovery we shared, and to a person who has been a massive pillar under systems biology and ecology. Wilfred was passionate about science. We will continue to travel the avenues that he has outlined though. We will miss him dearly.  ski for his help in laboratory work. We thank Maciej Jan Choczyn This study was performed within a PRELUDIUM Grant from the Polish National Science Center (No. UMO-2012/05/N/NZ8/00925) and the ‘Environmental stress, population viability and adaptation’ project (No. MPD/2009-3/5) and was supported by grant DS759 of the Institute of Environmental Sciences, Jagiellonian University. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. apsoil.2015.10.020. References  ska, M., Van Gestel, C.A.M., Van Straalen, N.M., Röling, W.F.M., Azarbad, H., Niklin Laskowski, R., 2013. Microbial community structure and functioning along metal pollution gradients. Environ. Toxicol. Chem. 32, 1992–2002.  ska, M., Laskowski, R., van Straalen, N.M., van Gestel, C.A.M., Azarbad, H., Niklin Zhou, J., He, Z., Wen, C., Röling, W.F.M., 2015a. Microbial community composition and functions are resilient to metal pollution along two forest soil gradients. FEMS Microbiol. Ecol. 91, 1–11.  ska, M., Nikiel, K., Van Straalen, N.M., Röling, W.F.M., 2015b. Azarbad, H., Niklin Functional and compositional responses in soil microbial communities along two metal pollution gradients: does the level of historical pollution affect resistance against secondary stress? Biol. Fertil. Soils doi:http://dx.doi.org/ 10.1007/s00374-015-1033-0. Bååth, E., Díaz-Raviña, M., Bakken, L.R., 2005. Microbial biomass, community structure and metal tolerance of a naturally Pb-enriched forest soil. Microb. Ecol. 50, 496–505. Bossio, D.A., Scow, K.M., 1998. Impacts of carbon andflooding on soil microbial communities: phospholipid fatty acid profiles and substrate utilization patterns. Microb. Ecol. 35, 265–278. Bossio, D.A., Fleck, J.A., Scow, K.M., Fujii, R., 2006. Alteration of soil microbial communities and water quality in restored wetlands. Soil Biol. Biochem. 38, 1223–1233. Brinch, U.C., Ekelund, F., Jacobsen, C.S., 2002. Method for spiking soil samples with organic compounds. Appl. Environ. Microbiol. 68, 1808–1816. Córdova-Kreylos, A.L., Cao, Y., Green, P.G., Hwang, H.M., Kuivila, K.M., LaMontagne, M.G., Van De Werfhorst, L.C., Holden, P.A., Scow, K.M., 2006. Diversity, composition, and geographical distribution of microbial communities in California salt marsh sediments. Appl. Environ. Microbiol. 72, 3357–3366. Chowdhury, N., Marschner, P., Burns, R.G., 2011. Response of microbial activity and community structure to decreasing soil osmotic and matric potential. Plant Soil 344, 241–254.

241

Drenovsky, R.E., Vo, D., Graham, K.J., Scow, K.M., 2004. Soil water content and organic carbon availability are major determinants of soil microbial community composition. Microb. Ecol. 48, 424–430. Feris, K., Ramsey, P., Frazar, C., Moore, J.N., Gannon, J.E., Holben, W.E., 2003. Differences in hyporheic-zone microbial community structure along a heavymetal contamination gradient. Appl. Environ. Microbiol. 69, 5563–5573. Fierer, N., Schimel, J.P., Holden, P.A., 2003. Variations in microbial community composition through two soil depth profiles. Soil Biol. Biochem. 35, 167–176. Frostegård, A., Bååth, E., 1996. The use of phospholipid fatty acid analysis to estimate bacterial and fungal biomass in soil. Soil Biol. Fertil. Soils 22, 59–65. Frostegård, Å., Tunlid, Å., Bååth, E., 1991. Microbial biomass measured as total lipid phosphate in soils of different organic content. J. Microbiol. Methods 14, 151– 163. Fry, J.D., 1996. Evolution of host specialization: are trade-offs overrated? Am. Nat. 148, 84–107. Grayston, S.J., Griffith, G.S., Mawdsley, J.L., Campbell, C.D., Bardgett, R.D., 2001. Accounting for variability in soil microbial communities of temperate upland grassland ecosystems. Soil Biol. Biochem. 33, 533–551. Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: paleontological statistics software package for education and data analysis. Palaentol. Electon. 4 (1), 9. Heipieper, H.J., Diefenbach, R., Keweloh, H., 1992. Conversion of cis unsaturated fatty-acids to trans, a possible mechanism for the protection of phenoldegrading Pseudomonas putida p8 from substrate toxicity. Appl. Environ. Microbiol. 58, 1847–1852. Jones, K.C., Alcock, R.E., Johnson, D.L., Northcott, G.L., Semple, K.T., Woolgar, P.J., 1996. Organic chemicals in contaminated land: analysis, significance and research priorities. Land Contam. Reclam. 4, 189–197. Joynt, J., Bischoff, M., Turco, R., Konopka, A., Nakatsu, C.H., 2006. Microbial community analysis of soils contaminaed with lead, chromium and petroleum hydrocarbons. Microb. Ecol. 51, 209–219. Klimkowicz-Pawlas, A., Maliszewska-Kordybach, B., 2003. Effect of anthracene and pyrene on dehydrogenases activity in soils exposed and unexposed to PAHs. Water. Air Soil Pollut. 145, 169–186. Knight, B.P., McGrath, S.P., Chaudri, A.M., 1997. Biomass carbon measurements and substrate utilization patterns of microbial populations from soils amended with cadmium, copper, or zinc. Appl. Environ. Microbiol. 63, 39–43. Kotterman, M.J.J., Vis, E.H., Field, J.A., 1998. Successive mineralization and detoxi5cation of benzo(a) pyrene by the white rot fungus Bjerkanderasp. strain BOS55 and indigenous microMora. Appl. Environ. Microbiol. 64, 2853–2858. Langworthy, D.E., Stapleton, R.D., Sayler, G.S., Findlay, R.H., 1998. Genotypic and phenotypic responses of a riverine microbial community to polycyclic aromatic hydrocarbon contamination. Appl. Environ. Microbiol. 64, 3422–3428. Lechevalier, M.P., 1977. Lipids in bacteria taxonomy –a taxonomist's view. Crit. Rev. Microbiol. 7, 109–210. Liang, C., Fujinuma, R., Balser, T.C., 2008. Comparing PLFA and amino sugars for microbial analysis in an Upper Michigan old growth forest. Soil Biol. Biochem. 40, 2063–2065. Los, D.A., Murata, N., 2004. Membrane fluidity and its roles in the perception of environmental signals. Biochim. Biophys. Acta 1666, 142–157. Mazzella, N., Molinet, J., Syakti, A.D., Bertrand, J.C., Doumenq, P., 2007. Assessment of the effects of hydrocarbon contamination on the sedimentary bacterial communities and determination of the polar lipid fraction purity: relevance of intact phospholipid analysis. Mar. Chem. 103, 304–317. Mentzer, J.L., Goodman, R.M., Balser, T.C., 2006. Microbial response over time to hydrologic and fertilization treatments in a simulated wet prairie. Plant Soil 284, 85–100. Meulenberg, R., Rijnaarts, H.H.M., Doddema, H.J., Field, J.A., 1997. Partially oxidized polycyclic aromatic hydrocarbons show an increased bioavailability and biodegradability. FEMS Microbiol. Lett. 152, 45–49. Nannipieri, P., Ascher, J., Ceccherini, M.T., Landi, L., Pietramellara, G., Renella, G., 2003. Microbial diversity and soil functions. Eur. J. Soil Sci. 54, 655–670. Nicolaus, B., Manca, M.C., Lama, L., Esposito, E., Gambacorta, A., 2001. Lipid modulation by environmental stresses in two models of extremophiles isolated from Antarctica. Polar Biol. 24, 1–8.  ska, M., Chodak, M., Laskowski, R., 2005. Characterization of the forest humus Niklin microbial community in a heavy metal polluted area. Soil Biol. Biochem. 37, 2185–2194. Pepi, M., Heipieper, H.J., Fischer, J., Ruta, M., Volterrani, M., Focardi, S.E., 2008. Membrane fatty acids adaptive profile in the simultaneous presence of arsenic and toluene in Bacillus sp. ORAs2 and Pseudomonas sp. ORAs5 strains. Extremophiles 12, 343–349. Petersen, S.O., Klug, M.J., 1994. Effects of sieving, storage, and incubationtemperature on the phospholipid fatty-acid profile of a soil microbial community. Appl. Environ. Microbiol. 60, 2421–2430. Phillips, R.L., Zak, D.R., Holmes, W.E., White, D.C., 2002. Microbial community composition and function beneath temperate trees exposed to elevated atmospheric carbon dioxide and ozone. Oecologia 131, 236–244. Potthoff, M., Steenwerth, K., Jackson, L.E., Drenovsky, R.E., Scow, K.M., Joergensen, R. G., 2006. Soil microbial community composition as affected by restoration practices in California grassland. Soil Biol. Biochem. 38, 1851–1860. Rundle, H., Nosil, P., 2005. Ecological speciation. Ecol. Lett. 8, 336–352. Schimel, J., Balser, T.C., Wallenstein, M., 2007. Microbial stress-response physiology and its implications for ecosystem function. Ecology 88, 1386–1394. Schindlbacher, A., Rodler, A., Kuffner, M., Ktzler, B., Sessitch, A., ZechmeisterBoltenstern, S., 2011. Experimental warming effects on the microbial

242

H. Azarbad et al. / Applied Soil Ecology 98 (2016) 233–242

community of a temperate mountain forest soil. Soil Biol. Biochem. 43, 1417– 1425. Smith, E., Naidu, R., Alston, A.M., 1998. Arsenic in the soil environment: a review. Adv. Agron. 64, 149–195. Tobor-Kaplon, M.A., Bloem, J., Römkens, P.F.A.M., de Ruiter, P.C., 2005. Functional stability of microbial communities in contaminated soils. Oikos 111, 119–129. Unger, I.M., Kennedy, A.C., Muzika, R.M., 2009. Flooding effects on soil microbial communities. Appl. Soil Ecol. 42, 1–8. Van Der Wurff, A.W.G., Boivin, M.-e.Y., Van Den Brink, P.J., Kools, S.A.E., Van Megen, H., Riksen, J., Kammenga, J., 2007. Type of disturbance and ecological history determine structural stability. Ecol. Appl. 17, 190–202. Vestal, J.R., White, D.C., 1989. Lipid analysis in microbial ecology. Quantitative approaches to the study of microbial communities. Bioscience 39, 535–541. Vinebrooke, R.D., Cottingham, K.L., Norberg, J., Scheffer, M., Dodson, S.I., Maberly, S. C., Sommer, U., 2004. Impacts of multiple stressors on biodiversity and

ecosystem functioning: the role of species co-tolerance. Oikos 104, 451–457. Welp, G., 1999. Inhibitory effects of the total and water-soluble concentrations of nine different metals on the dehydrogenase activity of a loess soil. Biol. Fertil. Soils 30, 132–139. Zelles, L., Bai, Q.Y., Beck, T., Beese, F., 1992. Signature fatty acids in phospholipids and lipo polysaccharides as indicators of microbial biomass and community structure in agricultural soils. Soil Biol. Biochem. 24, 317–323. Zhang, B., Wang, H., Yao, S., Bi, L., 2013. Litter quantity confers soil functional resilience through mediating soil biophysical habitat and microbial community structure on an eroded bare land restored with mono Pinus massoniana. Soil Biol. Biochem. 57, 556–567. Zogg, G.P., Zak, D.R., Ringelberg, D.B., MacDonald, N.W., Pregitzer, K.S., White, D.C., 1997. Compositional and functional shifts in microbial communities due to soil warming. Soil Sci. Soc. Am. J. 61, 475–481.