| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Departamento de Biología Animal, Vegetal y Ecología, Universidad de Jaén, 23071 Jaén, Spain
b School of Natural Resources, Ohio State University, 2021 Coffey Road, Columbus, OH 43210-1085
* Corresponding author (Richard.Dick{at}snr.osu.edu)
Received for publication December 10, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
6c), the arbuscule mycorrhizae marker (16:15c), and iso- and anteiso-branched PLFAs (Gram positive bacteria) were suppressed with increasing pollution whereas 17:0cy (Gram negative bacteria) increased with metal pollution. For both extraction methods, richness and diversity were greater in nonpolluted soils and lowest in polluted soils. The ELFA method was sensitive for reflecting metal pollution on microbial communities and could be suitable for routine use in ecological monitoring and risk assessment programs because of its simplicity and reproducibility.
Abbreviations: ELFA, ester-linked fatty acid • FAME, fatty acid methyl ester • NMS, nonmetric multidimensional scaling • PLFA, phospholipid-linked fatty acid
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Soil microorganisms are very sensitive to environmental change (Turco et al., 1994), and significant changes of the microbial community can occur following disturbance, both in terms of total biomass and species composition (Harris et al., 1991, 1993; Insam and Domsch, 1988; Stahl et al., 1988). Measures of the microbial community following the initiation of reclamation efforts could be used as an indicator of restoration progress (Harris et al., 1991) and may give insights into potential ways to accelerate a restoration.
The recent development of molecular and biochemical techniques has enabled a better understanding of microbial community structures in soil ecosystems (Kennedy and Gewin, 1997), with most of the species being unknown and unculturable (Torsvik et al., 1998). One widely used approach is the analysis of microbial phospholipid-linked fatty acid (PLFA) composition. In this method, microbial lipids are extracted from environmental samples in a phase mixture of chloroform, methanol, and water (Bligh and Dyer, 1959). Lipids associated with the organic phase are then fractionated into neutral, glyco-, and phospholipids on silicic acid columns (Vestal and White, 1989). Finally, the phospholipids are subjected to alkaline methanolysis to produce fatty acid methyl esters (FAMEs) for gas chromatography analysis.
Phospholipid-linked fatty acids are major cell membrane constituents, and their polar head groups and ester-linked side chains vary in composition between eukaryotes and prokaryotes, as well as among many prokaryotic groups. These compounds rapidly degrade on cell death (Pinkart et al., 2002); as a result, the variation in types of PLFAs provides a "fingerprint" of the living microbial community and has been used to study microbial community changes in soil. Therefore, changes in the community structure in response to an environmental stressor can be monitored by the comparison of PLFAs that differ among specific groups of microorganisms. Phospholipid-linked fatty acid methods have been used to characterize microbial communities from heavy metal–contaminated soils (Pennanen et al., 1996; Bååth et al., 1998; Shi et al., 2002; Rajapaksha et al., 2004). However, the only studies of microbial community structures on remediated metal-polluted soils are Kelly and Tate (1998) and Kelly et al. (2003), in which a mixture of municipal sewage sludge and power plant fly ash was applied to remediate polluted soil. Thus, our field study provides a good opportunity to determine the effect of different remediation treatments on the microbial community structure.
The PLFA method is time consuming and does not lend itself for practical environmental monitoring. In contrast to PLFA, a simpler method is direct extraction of fatty acids (MIDI, 1995), originally designed to extract fatty acids and identify pure cultures of bacteria (Cavigelli et al., 1995; Schutter and Dick, 2000; Petersen et al., 2002). With the MIDI method, microbial cells in soil are saponified by heating with a strong aqueous alkali. Once fatty acids are cleaved from lipids, they are methylated to form FAMEs, which are extracted in an organic solvent. Unlike PLFAs that can only come from viable cell membranes, extracted FAMEs may come from cellular storage compounds or dead microbial, animal, or plant cells. If these are extracted in significant amounts from soils, it is more difficult to draw conclusions about changes in the extant microbial community based on these profiles.
To date, there are several studies comparing MIDI and PLFA methods (Petersen et al., 2002; Steger et al., 2003; Drenovsky et al., 2004) showing that, although both extraction methods were able to differentiate among communities of different soil types, the MIDI method included a significant background of nonmicrobial material and in some cases was less sensitive to soil environmental conditions than PLFA.
Because of this concern, less harsh methods for the direct extraction of fatty acids from soil organisms have been developed. In this sense, some authors have demonstrated the effectiveness of the ester-linked fatty acid (ELFA) procedure for assessing community structure (Drijber et al., 2000; Schutter and Dick, 2000). This method uses a mild alkaline reagent to lyse cells and release fatty acids as methyl esters from lipids. In theory, only ester-linked and not free fatty acids are extracted with this method. For instance, the ELFA method has been successfully used to characterize microbial communities in grass seed field soils and placed communities into groupings similar to those generated by a DNA-based method (Ritchie et al., 2000).
The ability of the ELFA method to profile the microbial community structure in soil has also been evaluated by comparison with PLFA analysis on the same samples in several studies. Drijber et al. (2000) compared both methods for their ability to discriminate between plots with different crop managements. They found that both PLFAs and ELFAs were in large part comparable. Steger et al. (2003) reached the same conclusion when both methods were applied to compost samples of different ages. Both papers also mention the need of more studies to confirm whether the ELFA method is as efficient as PLFA in other environments.
There is little information comparing ELFA and PLFA profiles microbial communities in heavy metal–polluted soils. The Guadiamar basin (southwestern Spain) provided a unique opportunity to study microbial communities in an actual metal contamination accident. In 1998 this watershed experienced a major environmental catastrophe when a containment dam of mine spoil failed. More than 4000 ha of alluvial soils of the Guadiamar basin were flooded with heavy metal– and acid-polluted sediments and waters. A general overview of the incident can be found in Grimalt et al. (1999), Gallart et al. (1999), Alastuey et al. (1999), and López-Pamo et al. (1999). Since that time, remediated sites have been established that could be compared to polluted soils and adjacent uncontaminated soils. This suite of treatments provides an exceptional situation to study microbial communities and processes along a heavy metal and pH gradient.
The objectives of this study were to (i) investigate the effect of contaminated and remediated acid heavy metal soils on microbial community structure and composition by lipid profiling, (ii) conduct a comparison of ELFA and PLFA methods, and (iii) identify keystone biomarkers for their usefulness to restoration ecologists during reclamation practices.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
We selected two sections of the basin for the study, "upper watershed" (sandy loam soil), which was nearest the original holding pond, and "lower watershed" (loam soil). Each stretch of the watershed had three treatments: (i) nonpolluted control plots in an area adjacent to the spill (two plots in the upper and one in the lower watershed); (ii) polluted and reclaimed (two plots in each watershed); and (iii) polluted plots (one plot in the upper and two in the lower watershed). The polluted treatment was the original contaminated soil that had pyrite mine spoil deposited on the soil surface that was left by the Regional Environmental Authorities for research purposes. Reclamation of polluted and reclaimed soils was done by removal of the pyrite slurry deposited and addition of calcium carbonate (sugarbeet lime) and iron oxy-hydroxides to prevent metal solubilization.
The statistical design of the study was a randomized block design where the upper and lower watersheds were the blocks and the level of pollution was the main factor.
Soil Sampling
Soils were sampled in September 2002. Each soil sample was composited from four subsamples randomly collected in each plot, using a 7-cm-diameter and 5-cm-deep core sampler. Soil samples of polluted plots were collected just underneath the pyrite mud layer, which was removed before sampling. Soil samples were sieved (2 mm) and stored at 4°C until further analysis.
Physicochemical Analysis
Soil water content (105°C for 24 h) and percent of organic matter (estimated as ignition loss, 550°C for 2 h; Nelson and Sommers, 1982), were analyzed on a 20-g subsample. Soil pH was measured in a 1:1 soil to CaCl2 0.01 M solution (McLean, 1982). The DTPA-extractable Cd, Cu, Pb, and Zn (1:2 soil to extractant ratio; Lindsay and Norvell, 1978) were analyzed by atomic absorption instrument.
Phospholipid-Linked Fatty Acids Analysis
Lipids were extracted by a modified Bligh and Dyer (1959) procedure as described by Butler et al. (2003). Briefly, soil was extracted with 50 mM phosphate, chloroform, and methanol (0.8:1:2) and the next day centrifuged, filtered, and then resuspended (chloroform and methanol, 1:2 ratio). After addition of 2 M NaCl there was a phase separation (chloroform) removal and immediate drying under a stream of N2. Dried lipid extracts were redissolved in chloroform and separated into neutral lipids, glycolipids, and phospholipids on solid-phase extraction columns containing 500 mg of silica (Supelco, Bellefonte, PA) with chloroform, acetone, and methanol, respectively. The methanol, containing phospholipids, was immediately dried under N2.
Dried phospholipids were converted to fatty acid methyl ester (FAMEs) through mild alkaline methanolysis (1:1 methanol-toluene and 0.2 M KOH in methanol) and heating for 15 min (38–42°C). The FAMEs were extracted (deionized water, 1 M acetic acid, and hexane) with hexane phase being removed, dried under N2, and redissolved in hexane and analyzed by gas chromatography (Agilent, Palo Alto, CA).
Ester-Linked Fatty Acids Analysis
Fatty acids were extracted from soil samples by using the ester-linked methods described in detail by Schutter and Dick (2000). Briefly, 3 g of soil was mixed with 15 mL of 0.2 M KOH in methanol, and the preparation was incubated for 1 h at 37°C, during which ester-linked fatty acids were released from soil microorganisms and methylated. After the incubation 3 mL of 1.0 M acetic acid were added to neutralize the pH. The FAMEs were extracted with 10 mL of hexane, and the sample was centrifuged at 2000 rpm for 20 min to separate the aqueous and hexane phases. The hexane layer was transferred to a clean tube, and evaporated under N2, after which FAMEs were resuspended in 2 mL hexane for analysis.
Quantification and Identification of Fatty Acid Methyl Esters
Gas chromatography analyses were performed using a 5890 Series II instrument (Hewlett-Packard, Palo Alto, CA) equipped with an Agilent 19091B-102 Ultra-2 column (length, 25 m; internal diameter, 0.2 mm; film thickness, 0.33 µm). The carrier gas was He, and the oven temperature was ramped from 120 to 270°C at a rate of 5°C per min with a 5-min hold at 300°C between samples.
Individual PLFA peaks were determined by comparison with retention times of commercial standards: a mixture of 37 FAMEs (FAME 37 47080-U; Supelco), a mixture of 24 bacterial FAMEs (P-BAME 24 47080-U; Supelco), 18:26,9c, 10Me16:0, and 10Me18:0 (Matreya, Pleasant Gap, PA), and MIDI standards (Microbial ID, Newark, DE). The PLFA peaks were confirmed by comparing peaks run on 25-m HP Ultra-2 columns (internal diameter, 0.2 mm; film thickness, 0.33 µm) with the MIDI system (Microbial ID). Standard curves for quantification were produced with tridecanoic FAME (Supelco). Unknown peaks were noted and listed by retention time only.
Data are expressed as relative amount, calculated as the area of each PLFA peak relative to the summed area of all PLFA peaks, after first adjusting for the number of C atoms per mole of PLFA.
Three analytical replicates of each fresh soil were extracted by both methods.
Fatty Acid Methyl Esters Nomenclature
Standard nomenclature is used to describe FAMEs detected by both extraction methods. Fatty acids were designated as the total number of carbon atoms followed by a colon, the number of double bonds followed by the position of the double bond from the aliphatic () end of the molecule. The prefixes a and i indicate anteiso and iso branching, respectively. Other notations are "Me" for a methyl group, "OH" for hydroxyl, and "cy" for cyclopropane groups.
Fatty acids were also grouped by structural classes, including saturated straight chain, monounsaturated, polyunsaturated, branched, and hydroxy fatty acid classes. These classes and selected individual fatty acids were used as indicators (biomarkers) for particular microbial groups (Bossio and Scow, 1998; Zelles et al., 1992).
Statistical Analysis
The experimental design was RCB 2 where the main factor was pollution and the blocking factor was the location of each set of treatments (upper and lower watersheds). Means separation was done with a Tukey HSD test for post hoc analysis of the main factor.
The FAMEs present in only one replicate of a soil sample within the entire data set were deleted before any analysis of data. The deleted FAMEs were 14:15c, 15:18c, i16:0 3OH, and 17:0 3OH. This approach prevented fatty acids that were only sporadically detected or unreliably quantified from influencing the analyses (Bossio and Scow, 1998).
The FAMEs with retention time less than i14:0, which were few in number, were also deleted from the data set. Fatty acids with more than 20 carbons were not included in the analysis because these are generally not of microbial origin (Zelles et al., 1995).
As the relative peak areas were not normally distributed, a nonmetric multidimensional scaling (NMS) analysis was used as an alternative method for sample ordination purposes. Nonmetric multidimensional scaling is an iterative ordination method that is well suited to data that are non-normal or are on arbitrary, discontinuous, or otherwise questionable scales. To assess overall effects of extraction method and degree of pollution, NMS analysis was performed on a data set containing relative FAME amounts extracted with the PLFA and ELFA methods for all soil types. Subsequently, NMS analysis was applied to data from each extraction method separately, to compare their ability to discriminate overall effects of pollution on microbial communities. In both cases, we performed the NMS analysis based on Sørensen's distance, and the "slow and thorough" autopilot mode of NMS in PC-ORD (McCune and Mefford, 1999) using randomized data for a Monte Carlo test of significance. Final stability of each run was evaluated by examining plots of stress (a measure of the dissimilarity between ordinations in the original n-dimensional space and in the reduced dimensional space) versus number of iterations.
To establish a functional relationships between NMS axis scores (dependent variable) and pH and DTPA heavy metal fraction (independent variables), after controlling for their covariation, we used a forward stepwise multiple regression procedure. A new variable was not included when its addition did not produce a significant (P = 0.05) increase in the variance accounted for.
The comparison of the different measured richness, evenness, and diversity was done by using a one-way analysis of variance, and Tukey HSD test for post hoc analysis. Richness (S) refers to the number of FAMEs detected in a given soil sample. The FAME evenness (E), a measure of how evenly FAMEs were distributed in a given soil sample, was calculated as E = H/ln(S). Diversity was calculated as the Shannon index where H' = –
pi ln(pi), and pi is the proportional amount of each FAME. Dominance was calculated as the Simpson index where D = 1 – p2i.
A Mantel test was performed to examine similarities between PLFA and ELFA patterns using PC-ORD (McCune and Mefford, 1999). The Mantel test evaluates the null hypothesis of no correlation between two distance matrices of FAME profiles, one based on PLFA data and the other based on ELFA data. The Mantel test was performed by using Sørensen's distance measure and the randomization (Monte Carlo) procedure with 1000 randomized runs.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
A fuller description of soil condition at time of sampling can be found in Hinojosa et al. (2004a).
Effect of Extraction Method on Microbial Community Structure in Polluted Soils
The number of different peaks detected was 55 using PLFA method, and 61 for the ELFA method. Fifty-four fatty acids were common to both methods. Some fatty acids were detected only with ELFA extraction method (15:16c, i16:1, i16:0, i17:0 3OH, 17:0 2OH, i19:1, and one unknown peak), and the 15:0 2OH was detected only with the PLFA method.
Total fatty acid content was over 19-fold higher with ELFA than with PLFA method (Table 2). However, we found a similar reduction of the amount of fatty acids in reclaimed and polluted soils with respect to nonpolluted soils, using both methods. Total amounts of fatty acids in reclaimed and polluted soils were, respectively, about 45 and 10% of the amount of fatty acids in nonpolluted soils.
|
The relative areas of peaks of fatty acids from both PLFA and ELFA extraction methods were used for the ordination of the whole dataset by using a NMS analysis (Fig. 1) . Soil samples were separated mainly by their degree of pollution (Axis 1) and second by the extraction method (Axis 2). Thus, by using either PLFAs or ELFAs we were able to distinguish the microbial communities as a function of degree of pollution and both showed very similar treatment separation patterns.
|
6, and 20:46 FAMEs. Those FAMEs were found in relatively higher proportion with the ELFA method than with the PLFA method. When NMS analysis was performed with PLFA patterns (Fig. 2 , top), the variability explained by Axes 1 and 2 was 97.1%. This multivariate analysis indicated a clear discrimination of microbial communities between polluted and nonpolluted soils. However, ordination of reclaimed plots depends on the level of restoration achieved for particular watershed sites. In general, reclaimed soils from the lower watershed showed PLFA patterns similar to those of nonpolluted soils, according to similar values in pH. On the other hand, the PLFA patterns of reclaimed soils from the upper watershed were more similar to the soils from polluted plots.
|
7c, 16:15c, 16:0 10Me, 17:18c, i18.0, 18.26,9c, 18:19c, 18:16c, 11Me18:17c, 18:0 10Me, and 18:3OH, and negatively with 17:0cy. Axis 2, which mainly discriminates between the group of polluted soils and the group of reclaimed soil from the upper watershed, was highly positively correlated with 16:19c, i17:1, and 17:0, and negatively with 19:111c, 19:19c, and one unknown peak. None of the pH and heavy metal variables were included in the regression equation as independent variables for the NMS Axis 2 scores.
|
In this case, stepwise regression analysis between NMS scores and soil properties results in the inclusion of pH, Zn, and Pb with Axis 1 and only pH with Axis 2, as variables that may have a functional relationship with microbial communities (Table 3). The ELFAs strongly and negatively correlated with Axis 1 were i14:0, i15:1, a15:0, i16:1, 16:17c, 16:15c, 16:0 10Me, 17:18c, i18:0, 18:26,9c, 11Me18:17c, 19:16c, 19:0 10Me, 18:0 3OH, and one unknown peak. On the other hand, 17:0cy, 19:111c, and 19:19c were strongly and positively correlated with Axis 1. Axis 2 was highly and positively correlated with 17:18c, and negatively with 16:0. This axis mainly discriminated between the group of polluted soils and the group of reclaimed soil from the upper watershed.
Simple correlations were done between fatty acid concentrations and pH (data not shown) and extractable metal content (Table 4). Most of the fatty acids extracted by either method had a significant (P < 0.01) positive r value with pH (exceptions for PLFA being: 14:020H = NS; 15:020H = 0.77*; 17:0cy = 0.93*; 19:1 11c + 9c = 0.89*; 19:010Me = NS; 20:0 = NS) (exceptions for ELFA being: 15:0 = NS; 16:17c = NS; 15:020H = NS; 17:0cy = 0.91*; 18:010Me = NS; 19:111c + 9c = 0.83*; 20:26, 9c = NS, where * indicates significance at the 0.05 probability level and NS is not significant). Correlations with Pb were not significantly correlated with any individual fatty acid or structural group of fatty acids (data not shown).
|
5c, 17:18c, 18:19c, 18:16c, i14:0, i18:0, 18:0 3OH, 16:0 10Me, 18:0 10Me, a15:1, i16:1, and 18Me18:17c. Conversely, the common fatty acids most positively correlated with heavy metals and negatively with pH were 19:16c and 17:0cy.
It is interesting to note that PLFA extractions had three fatty acids showing significant positive correlations with extractable heavy metals (Cd, Cu, and Zn) whereas only one ELFA had any significant positive correlations with metal contents. Conversely, for the important marker 18:26,9c, when extracted by PLFA method there were significant negative correlations with all metals but when extracted as ELFA it was not significant (Table 4).
Taking into account the structural groups of fatty acids from PLFA analysis, we found a significant negative correlation of methyl branched and monounsaturated fatty acids with each extractable heavy metal (Cd, Cu, and Zn), whereas the correlation with pH was significantly positive. However, the methyl branched fatty acids group did not show such high correlation with heavy metal ELFA extractions.
On the other hand, some groups of fatty acids, such as polyunsaturated and branched, showed significant positive correlation with pH using ELFA data, although they did not show this correlation with the PLFA extraction method.
Specific differences among microbial communities from soil with different degrees of contamination were clarified by comparison of the proportions of fatty acid structural classes in Table 5. The PLFA method resulted in higher proportions of monounsaturated and branched classes of fatty acids than did the ELFA method. However, the PLFA method showed lower proportions of saturated and polyunsaturated fatty acids. The PLFA method, in comparison with the ELFA method, resulted in a higher ratio of Gram positive to Gram negative bacteria, a lower ratio of fungi to bacteria, and lower ratio of saturated to unsaturated fatty acids (Table 5). The unsaturated, branched, and hydroxy fatty acids were found in the highest proportion in nonpolluted soils and the lowest in polluted soil, whereas reclaimed soils showed intermediate values and high variability. Consequently, nonpolluted soils had the highest Gram positive to Gram negative and fungi to bacteria ratios with polluted soils having the lowest, and reclaimed being intermediate.
|
|
The effect of heavy metal pollution in the decrease of richness and diversity could be due to the high concentrations of Cu and Zn, as confirmed by the highly significant negative correlation between these indices and these heavy metals (data not shown).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
However, even though multivariate analysis gives similar microbial community results for ELFA and PLFAs, it is important in the case of ELFAs to investigate specific markers and their taxonomic associations (Haack et al., 1994) because all fatty acids are extracted; not only those closely associated with microbial membranes, but also fatty acids from nonmicrobial sources (such as organic matter and plants). Structural classes and biomarkers have been important and useful indicators of trends in soil microbial communities in numerous studies using profiles based on both PLFA and whole soil fatty acids profiles (Cavigelli et al., 1995; Bossio and Scow, 1998).
Monounsaturated fatty acids have been used as indicators of aerobic and high substrate conditions (Larkin, 2003). In our study, nonpolluted soils had the higher proportions of monounsaturated and both reclaimed and polluted soil showed lower values when we used both PLFA and ELFA extraction methods. Moreover, this statement is supported by the significant positive correlations between monosaturated and heavy metals (mainly Cu and Zn).
Zelles et al. (1992)(1995) considered the monounsaturated to saturated fatty acid ratio to be very important for evaluating communities and environmental conditions, and generally observed ratios of >1 for grassland and cultivated soils with high C content and organic inputs, approximately 1 for other cultivated soils (such as potatoes), and less than 1 for soils characterized by low substrate, low organic inputs. Bossio and Scow (1998) also suggested that high values of this ratio indicate high substrate availability. Our values for the ratio of monounsaturated to saturated were consistent with those of Zelles et al. (1992)(1995) when we analyzed data from PLFA method, averaging approximately 1.3 for nonpolluted soils, and approximately 0.8 and 0.9 for reclaimed and polluted soil, respectively. Other authors (Ellis et al., 2001) have suggested that high ratios of monounsaturated to saturated fatty acids may be better apportioned to the intrinsic rate of nutrient turnover. That is, nutrients may be equally available in each sample but the toxic effects of the metals inhibit their utilization. This hypothesis was supported by the high positive correlation between this ratio and pH and negative correlation with Cu, Zn, and Cd found in our study. On the other hand, when this ratio (monounsaturated to saturated fatty acids) was calculated with data resulting from ELFA method, our values were inconsistent and not useful as an overall indicator of the community disturbance.
In both PLFA and ELFA methods the fatty acid 16:15c was most responsible for differentiation among FAME profiles from soil with different degree of pollution. This fatty acid is known to be a major component of arbuscular mycorrhizal (AM) fungi (Haack et al., 1994; Graham et al., 1995), and has been proposed as a valuable biomarker for estimating AM biomass and distribution (Olsson, 1999). However, 16:15c is also produced by a select group of bacteria, including Cytophaga, Flavobacterium, Flexibacter, and others (Nichols et al., 1986). Drijber et al. (2000) observed high levels of 16:15c in a native mixed prairie sod and decreasing levels in wheat-fallow plots with increasing tillage intensity (no-till, sub-till, plow). The prevalence of 16:15c in less disturbed systems is consistent with AM fungi as the probable origin of the fatty acid, as tillage or other soil disturbance is known to adversely affect AM fungi. Frostegård et al. (1993) also found quantities of this fatty acid to increase dramatically in limed (high pH) soils. This fatty acid has decreased in response to additions of zinc and other heavy metals to soils (Frostegård et al., 1993, 1996; Kelly et al., 1999). Our results were consistent with these findings for both extraction methods, where this fatty acid was significantly positively correlated with pH (PLFA, r = 0.87; ELFA, r = 0.89) and negatively with heavy metals (mainly with Cu and Zn). Thus, 16:15c appears to be particularly responsive to environmental changes, and may be a good indicator of changes in the microbial community structure.
The polyunsaturated fatty acids (represented mainly by 18:26c) are associated primarily with fungi (Guckert et al., 1985), but are also a constituent of plant membranes. We found a decrease in the proportion of fungi associated with an increase of heavy metal pollution (significant negative correlation between 18:26c and heavy metals, mainly with Cu and Zn). This is consistent with metal contamination in a forest soil in the vicinity of a zinc smelter in the Mid-Atlantic United States (Kelly and Tate, 1998) and for soils near metal smelter in Finland (Pennanen et al., 1996).
These results run contrary to the general assumption that fungi, as a group, are more tolerant than bacteria to heavy metals (Doelman, 1985; Hiroki, 1992). This assumption was initially inferred by comparison of metal tolerance of pure culture and isolated microorganisms, which may not represent the bioavailability of the metals to microorganisms in soils. Bacteria dominate in smaller pores, aggregates, and particles, which may better protect them. Conversely, branching organisms are associated with larger pores and aggregates, which would make them more vulnerable to metals leaching into soils. Another factor is that contact with metal contamination of one hyphae on a branching organism likely affects the whole organism and a large biomass. This may be unlike unicellular organisms where metal toxicity to a single cell may not affect the whole colony or affect a much smaller portion of the biomass.
Second, the recent study by Rajapaksha et al. (2004) in a 60-d laboratory experiment supports the idea that fungal activity was less affected by heavy metals than bacteria in a short-term period. However, they found that bacterial activity then slowly recovered to values similar to those of the control soil. Thus, heavy metal–sensitive bacteria were probably responsible for the decrease in bacteria in a short term. Nevertheless, the competitive advantage of more tolerant bacteria may result in change in community structure and even an increase of bacterial biomass in a long-term study, as it is in our case. Conversely, fungal biomass increased with metal load within a few days after metal addition. This may be due to the fact that fungal growth is carbon limited. The extra carbon released from dead bacteria would thus trigger increased fungal growth, overriding the negative impact of the heavy metals. Bacteria, which were negatively affected by the metals, apparently could not take advantage of this extra carbon.
On the other hand, a significant effect on fungal biomass may relate to the multiple effects on mycorrhizal fungi (as noted above). Pennanen et al. (1996) suggested that damage to fine roots by metals may reduce rhizosphere habitats for mycorrhizal fungi. This theory is supported by findings that elevated heavy metal concentrations decrease (Koomen et al., 1990) or in some cases prevent (Gildon and Tinker, 1983) the mycorrhizal infection of plant roots. Therefore, the decrease in the proportion of fungi (mainly represented by 18:26c) that we found in polluted sites may be due to inhibition of mycorrhizal infection of plant roots by heavy metals. This is supported by the significant positive correlation between the general fungal marker (18:26c) and the mycorrhizal marker (16:15c). Using a rough estimation as described by Olsson (1999), we found that the mycorrhizal marker (16:15c) comprised >60% of the total fungal biomass represented by the marker 18:26c.
Since fungi are central for decomposition, heavy metal contamination may have a significant impact on ecosystem function by reducing fungi and indirectly reduce the ability of soils to perform decomposition. This would fit with our previous research at this same site where the activity of key enzymes involved in decomposition and nutrient mineralization (phosphatases, arylsulfatase, ß-glucosidase, urease, and dehydrogenase) were suppressed (Hinojosa et al., 2004a, 2004b). The reclaimed soils had higher activities than polluted soils but, typically, 1.5 to 3 times lower levels of activity than the nonpolluted soil with ß-glucosidase, an important enzyme during decomposition, showing the greatest discrimination. However, conclusions derived from both mycorrhizal and fungal markers should be considered carefully, since both markers have nonfungal sources (Olsson, 1999; Frostegård and Bååth, 1996; Nichols et al., 1986).
In our study we found an increase of Gram negative over Gram positive bacteria in polluted soils. This change was indicated by a decrease in several iso- and anteiso-branched PLFAs, commonly found in Gram positive bacteria, and mainly by an increase in 17:0cy (significant positive correlation with Cd, Cu, and Zn concentrations and negative correlation with pH) considered to be typically from Gram negative bacteria. Gram negative bacteria are considered to be fast growing microorganisms that utilize a variety of C sources and can adapt quickly to a variety of environmental conditions (Ponder and Tadros, 2002), which is consistent with our polluted conditions. We found a strong positive correlation between the Gram positive to Gram negative ratio and pH, and negative correlation between this ratio and Cd, Cu, or Zn. This agrees with the results of Frostegård et al. (1993), who found that PLFA patterns were reflected in the different levels of metal concentrations in two contrasting soil types (i.e., arable versus forest) that were experimentally contaminated with heavy metals. Evidence for a similar shift in response to heavy metals pollution has been found in studies reported by Doelman and Haanstra (1979) and Hiroki (1992).
On the other hand, the relative proportion of fungal and bacteria biomass is an important indicator of reestablishment of soil microbial community stability and, hence, ecosystem self-regulation (Bardgett and McAlister, 1999; Zeller et al., 2001). Soil microbial communities of undisturbed terrestrial ecosystems tend to be dominated by fungal microbial biomass (Bardgett and McAlister, 1999). Disturbance is known to be especially detrimental to fungal populations and a number of studies have shown that physical disturbance results in increased bacterial dominance (Beare et al., 1992; Stahl et al., 1988; Schutter et al., 2001). Therefore, the fungi to bacteria ratio provides a more accurate indication of these relative populations among soils with different degrees of pollution. In our study, we consistently found both extraction methods had higher fungal to bacterial ratios for nonpolluted and reclaimed soils than polluted soils, although it should be noted there was high variability in the reclaimed soils. This result suggests that the fungal to bacterial ratio may be a good indicator for assessing the success of reclamation which is consistent with Mummey et al. (2002) and may reflect degree of ecosystem stability.
The decrease of all the methyl-branched fatty acids (almost exclusively from actinomycetes) with increasing metal pollution and significant positive correlation with pH and negative with Cd, Cu, and Zn agrees with the findings of Lechevalier (1977). Conversely, Frostegård et al. (1993) found 10Me16:0, 10Me17:0, and 10Me18:0 increased in metal-polluted forest soil while for arable soil these decreased. Other results in the literature also indicate that different actinomycetes can respond differently to elevated heavy metal concentrations (Williams et al., 1977; Hiroki, 1992). As mentioned for fungi above, the branching nature of actinomycetes may make them more positionally susceptible to heavy metals but this may vary with soil type and the particular type of pollution, as reflected in the mixed responses of this organism in studies so far.
We found that the abundance of several structural classes was dependent on the extraction method. For example, the relative amount of saturated and polyunsaturated fatty acids was higher when the ELFA method was used. Conversely, the monounsaturated and branched fatty acids were more abundant when soil was extracted with the PLFA method. Hydroxy and methyl branched fatty acids showed similar proportions with both extraction methods. Thus, these results show that inferences regarding community structure may vary according to the method employed (Schutter and Dick, 2000).
However, there are practical considerations which suggest that the ELFA method has advantages compared with the PLFA method. Approximately twice as many samples can be extracted by the ELFA method over the PLFA method in a given period and it is simpler and less expensive. It is, therefore, a useful tool for routine use in ecological soil monitoring.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
5c marker that is associated with AM fungi. The only disadvantage of the ELFA method was that it did not discriminate between metal treatments based on the monosaturated to saturated fatty acid ratio.
The data demonstrate that the ELFA method has potential as a practical method to monitor changes in soil quality. Such a method could help public agencies to establish guidelines for assessing soil pollution to protect soil resources. However, care should be taken when interpreting the results especially where site-specific parameters may affect the response to the assays used. For example, microbial responses to soil contamination may vary, at least in part, due to variations in metal bioavailability among different soils.
More detailed studies on diverse soils and calibration of the method to establish thresholds are needed to fully determine the potential of ELFAs to monitor restoration of polluted soils to develop recommendations and thresholds for evaluating metal-polluted or remediated soils based on microbial community composition.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
| |||||||||||
