Ecology, 83(11), 2002, pp. 3152–3166 2002 by the Ecological Society of America
EXOTIC PLANT SPECIES ALTER THE MICROBIAL COMMUNITY STRUCTURE AND FUNCTION IN THE SOIL
PETER S. K
OURTEV,1,3 JOAN G. EHRENFELD,1
¨ MAX HAGGBLOM2
of Ecology, Evolution and Natural Resources, Cook College, Rutgers University, 14 College Farm Road, New Brunswick, New Jersey 08901 USA 2Department of Biochemistry and Microbiology, Cook College, Rutgers University, 76 Lipman Drive, New Brunswick, New Jersey 08901 USA
Abstract. Exotic plant species are increasingly becoming the focus of research and have been identi?ed as a component of human-induced global change. Successful invaders may alter soil conditions, but the effect of exotic species on soil microbial communities has not been studied. We studied two exotic understory plant species (Japanese barberry [Berberis thunbergii] and Japanese stilt grass [ Microstegium vimineum]) in hardwood forests in northern New Jersey, USA. We sampled bulk and rhizosphere soils under the two exotic species, as well as under a co-occurring native species (blueberry [ Vaccinium spp.]). We indexed the structure (by measuring phospholipid fatty acid [PLFA] pro?les) and function (by measuring enzyme activities and substrate-induced respiration [SIR] pro?les) of microbial communities in the sampled soils. Soils under the three species differed in microbial community structure and function. These differences were observed in both the rhizosphere and bulk soil samples. Differences in the structural variables were correlated to differences in the functional variables as demonstrated by canonical correlation analysis. These results indicate that successful exotic invasive species can have profound effects on the microbial community of the soil.
Key words: Berberis thunbergii; enzymes; exotic plant species; hardwood forests; Microstegium vimineum; phospholipid fatty acid (PLFA); soil microbial communities; substrate-induced respiration (SIR).
INTRODUCTION The composition of microbial communities differs in soils beneath different species of plants (Grayston and Campbell 1996, Westover et al. 1997, Priha et al. 1999, Grayston et al. 2001). These differences have been demonstrated using methods re?ecting membrane chemistry (e.g., PLFA [Borga et al. 1994, Sundh et al. 1997, Bossio et al. 1998, Kelly et al. 1999]), DNA pro?les (Pankhurst et al. 1996, Duarte et al. 1998, Marilley and Aragno 1999), and methods using metabolic pro?les (e.g., BIOLOG and substrate-speci?c respiration [Degens and Harris 1997]). Presumably, the different microbial communities are the result of differences among species in both root inputs, such as rhizosphere exudates and root turnover (Coleman et al. 2000), and in the quantity and chemical quality of aboveground litter inputs, among other factors. The functional capacity of the soil microbial community, as re?ected in the activities of the enzymes involved in nutrient mineralization processes, also varies among soils dominated by different plant species (Waldrop et al. 2000). In most of these studies, however, differManuscript received 5 November 2001; revised 25 March 2002; accepted 2 April 2002. 3 Present address: Department of Biochemistry and Microbiology, Cook College, Rutgers University, 76 Lipman Drive, New Brunswick, New Jersey 08901 USA. E-mail: email@example.com
entiation of microbial communities with respect to the vegetation has been documented for well-established plant species and associations. The displacement of native plant species by exotics is an increasingly common event (Vitousek 1990, D’Antonio and Vitousek 1992), yet the effect of such displacements on soil microbial communities is unknown. We report here a study of changes in both microbial community structure, as indexed by phospholipid fatty acid (PLFA) pro?les, and function, as indexed by soil enzyme activities and substrate-induced respiration (SIR) responses, in response to the invasion of deciduous forest understories by two different exotic species. Whereas evidence of microbial community differentiation with respect to plant species is well established, two questions remain to be answered. First, we explore the extent to which effects of plants on rhizosphere community structure and function are transferred to the bulk soil (Steer and Harris 2000). As rhizosphere-based effects on microbes are mediated by localized production of exudates and mucilages, it is possible that roots exert only a local effect on the microbiota in the densely rooted portion of the soil pro?le. It is unknown how far through the pro?le a plant-mediated effect may extend. Second, we address the relationship between the structure of the microbial communities and their function, which has been rarely explored (Zogg et al. 1997, Waldrop et al. 2000). Exotic plant species invasions provide a natural ex-
EFFECT OF EXOTICS ON SOIL MICROORGANISMS
PLATE. 1. Barberry bushes are scattered in a lawn under a mixed hardwood canopy in Worthington State Forest, New Jersey (USA). This scene is typical of invaded areas throughout the region. Photograph by John G. Ehrenfeld.
periment in which to examine these questions. In recently invaded areas, the dominant native species has been replaced by a new plant species. By examining the structure and function of the soil microbiota near the invasion front, it is possible to compare communities developed on the same soil under two different plant species. We studied hardwood forests in New Jersey where understory plant communities have recently been invaded by two exotic plants: Japanese barberry (Berberis thunbergii, D.C.), a shrub, and Japanese stilt grass (Microstegium vimineum (Trin.) A. Camus.), a C4 grass. These two plants have replaced the native blueberries (Vaccinium spp.) and huckleberry (Gaylussaccia baccata), which are normally found in these forests (Ehrenfeld 1997, Kourtev et al. 1998). In previous work, we have shown that soils beneath these species differ in pH, organic horizon thickness, and net nitri?cation rates, with markedly higher nitri?cation rates beneath both of the exotics (Kourtev et al. 1998, 1999). This difference in nitri?cation rates can be reproduced in greenhouse studies by growing the exotics in previously uninvaded soils (Ehrenfeld et al. 2001). Thus we compared differences in the microbial community under the two exotic species to the microbial community under native shrubs. In our study, we used PLFA analyses to describe the structure of the microbial communities and two measures of the microbial functional diversity in soil, the respiratory response to the addition of different simple substrates in the soil (based on Degens and Harris ) and soil enzymatic activities (Sinsabaugh et al. 1993, Sinsabaugh 1994, Sinsabaugh and Moorhead 1994). Soil enzyme activities have been used as effective indicators of the capacity of the microbiota to mineralize carbon and mineral nutrients, and thus they can be used as measures of the functionality of the microbiota (Kourtev et al. 2002). Respiratory pro?les, an approach modeled on the use of multiple substrates in BIOLOG analyses, is predicated on idea that different species have different capacities to metabolize a range
of simple substrates (sugars, amino acids, fats, etc.), and so the response of whole soil communities to such a range of substrates is an indicator of the diversity of the organisms present and their potential functional capabilities. We hypothesized that soil microbiota will have responded to the change in vegetation by differing in both structure and function beneath the three plant species. We further predicted that these differences would be found in rhizosphere soil but not in the bulk soils beneath the main rooted soil volume. Finally, we predicted that the correlations between structural indices and functional indices would be stronger in the rhizosphere than in the bulk samples. MATERIALS
We studied soil microbial communities in Worthington State Forest, located on glaciated Paleozoic sandstones of the ridges and valleys province in northwestern New Jersey (Wolfe 1977). This site is characterized by Oquaga and Steinsburg soils (both Typic Dystrochrepts). They have a high content of rocks (10– 30% fragments 6 cm) and are acidic (pH 4.5–5.5; Fletcher 1979). The sampled locations had a closed canopy of mature hardwood trees, including oaks (Quercus spp.), hickories (Carya spp.), black birch (Betula lenta), and others. In noninvaded stands, the understory was composed mostly of blueberries ( Vaccinium spp.) and huckleberries (Gaylussaccia spp). In adjacent invaded stands, the Berberis bushes formed dense thickets, while Microstegium vimineum formed thick lawns (see Plate 1). In some invaded stands, the two exotic plant species co-occurred as scattered bushes within the lawns, although the grass was not found directly under the canopy of the barberry shrubs.
We identi?ed three adjacent locations in our study site, on ?at to moderate slopes: an uninvaded stand
PETER S. KOURTEV ET AL.
Ecology, Vol. 83, No. 11
(hereafter called ‘‘blueberry’’), a dense Berberis thicket (hereafter called ‘‘barberry’’), and a dense M. vimineum lawn with near 100% cover by the exotic species (hereafter called ‘‘stilt grass’’). We identify the soil samples by the name of the plants because the purpose of our study was to identify differences in soil communities under different plant species. Fourteen soil samples were taken from each location: seven from the soil that was attached to and therefore immediately affected by the roots of the plants and seven from soil not in immediate contact with plant roots, deeper into the soil pro?le. All three species root densely in the surface 5–10 cm of the soil. We de?ne the ‘‘rhizosphere’’ soil as that soil within the densely rooted portion (top 5 cm) of the soil pro?le (based on the observation that most roots are within 2 mm of other roots) and ‘‘bulk’’ soil as the soil immediately below the root mat. All bulk soils samples belonged to the same soil horizon. The soil sampling protocol differed for the three plant species, due to differences in their growth forms. B. thunbergii grows as large, multi-stemmed bushes with a large shallow root system and many ?ne roots radiating from a root crown. No other understory plants are found growing within the crown area of a bush. Seven bushes of similar size were randomly selected and two soil samples were collected from each one. First, each bush was carefully excavated and soil adhering to the roots was collected as a rhizosphere sample. A bulk soil sample was collected immediately below the layer of roots. No tree roots were observed in any soil sample (the same was true for Microstegium and Vaccinium soil samples). Vaccinium plants grow as a spatially dispersed network of small, clonal stems, from which grow a diffuse network of very ?ne roots. We chose a starting point by throwing a small object in a random direction. Then, we outlined a transect from the starting point and established seven plots at 10-m intervals along the tran50 cm and contained 10 sect. Each plot was 50 blueberry stems. In each plot, all the plants were excavated and the soil that adhered to the ?ne Vaccinium roots was collected. The soil collected in this way was combined into one rhizosphere sample. After carefully removing any remaining roots, mineral soil was collected from ?ve points deeper in the square (the four corners and the center of the plot) and combined into a bulk soil sample. M. vimineum culms are long and decumbent, rooting at the nodes, but the root systems produced by both the main stem and the nodes are short and sparse. However, in the dense lawns, the surface 5 cm of soil is exclusively and densely ?lled with stilt grass roots. No other plant species co-occurs in these dense lawns. In the stilt grass locations, we again used seven randomly 50 cm plots (randomly determined as in placed 50 the case of Vaccinium plants); each plot contained 100 culms. The stilt grass plants were removed in
small batches ( 10 at a time), and the soil that adhered to them was collected by shaking it in a large plastic tray. All the soil that was collected in this way was combined to give the rhizosphere sample of a given plot. The bulk soil was collected from the 5 cm of soil immediately below the rooted layer. All soils were collected in a single day in September of 2000 and transported within 2 h to the laboratory under cooled conditions, where they were sifted through a 5-mm sieve. A subsample of each soil was stored at 20 C for later analysis of fatty acids. The remainder of each soil sample was stored in a large plastic container for 2 wk at 20 C and 40% moisture content. We were able to process only a few samples in a day; therefore without preliminary storage, some of the soils would have been immediately studied, while others would have been incubated for several days. Sifting through sieves 0.5 mm and conditioning does not signi?cantly affect SIR pro?les (Degens 1998). After incubation, soils were taken in a random order and SIR responses were measured at a rate of 4 samples/d. After SIR responses were recorded for all 42 samples, we measured the enzyme activities in them, with 14 samples processed per day.
Microbial community structure
We studied the structure of microbial communities in soil samples using PLFA analysis. The method we used was adapted from White (1979) and slightly modi?ed (after a pilot run) to maximize extraction of fatty acid from the soil. Fatty acids were extracted with a one-phase solvent consisting of a 1:2:0.8 mixture of chloroform, methanol, and 50 mmol/L phosphate buffer (pH 7.4). Ten grams of soil from each soil sample were extracted with 20 mL of the solvent in a shaker for 24 h. The samples were centrifuged at 1000 g (9806 m/s2) for 1 h, and the supernatant was removed. The remaining soil was re-extracted with 10 mL of the same extraction solvent for 12 h. The supernatant was removed after centrifuging and the combined extract was then evaporated under nitrogen (N2) to 1 mL. Lipids in the concentrated extract were separated on silicic acid columns (King et al. 1977). The lipids were fractionated into neutral lipids, glycolipids, and polar lipids. The polar lipid fraction was eluted with methanol, and the sample was evaporated to dryness under N2. The polar lipids were then subjected to saponi?cation and methylation according to the MIDI protocol (MIDI 1995). Individual fatty acid methyl esters were identi?ed and quanti?ed using the MIDI Sherlock Microbial Identi?cation System (MIDI, Newark, Delaware, USA). The results for each individual fatty acid were expressed as a percentage of the total amount of fatty acids found in a given sample. A total of 99 PLFAs were identi?ed in the different soil samples. In our analyses, we only used fatty acids present in proportions 0.5%. Certain fatty acids listed
EFFECT OF EXOTICS ON SOIL MICROORGANISMS
TABLE 1. Fatty acids used in the analysis of microbial communities structure in Worthington State Forest, northwestern New Jersey, USA. Organisms Gram negative bacteria Diagnostic fatty acids? 10:0 3OH, 12:0, 12:0 2OH, 12:0 3OH, 14:0 15:0 16:1 7c, 15:0i 3OH, 15:0 3OH 17:0cy 17:0 16:1 2OH, 18:1 7c, 18:1 5c, 11Me 18:1 7c, 19:0cy 12:0i, 12:0a, 13:0i, 13:0a, 14:0i, 14:0a, 15:01, 15:0a, 16:0i, 16:0a, 17:0i, 17:0a 16:1 5c 16:0 10Me 16:0 10Me, 18:0 10Me 18:3 6,9,12c, 18:2 6,9c, 18:1 9c 20:4 6,9,12,15c
Gram positive bacteria VAM fungi? SO4 -reducers Actinomycetes Fungi Protists
? Sources: Drijber et al. 2000, Hill et al. 2000, Olsson and Alstrom 2000, Waldrop et al. ¨ 2000. ? Vesicular–arbuscular mycorrhizal fungi.
in Table 1 are found in large amounts in speci?c groups of microorganisms and therefore have diagnostic meaning. We follow standard nomenclature rules when referring to different fatty acids.
Microbial community function
We used two measures of the function of soil microbial communities: enzyme activities in the soil and respiratory response of soil to the addition of different substrates. We measured the activities of eight enzymes, which can be related to the cycling of carbon (C), nitrogen (N), phosphorus (P), and soil organic matter (SOM): C, -glucosidase (E.C. 126.96.36.199) and endocellulase (E.C. 188.8.131.52); N, -N-acetylglucosaminidase (chitobiase; E.C. 184.108.40.206) and aminopeptidase (E.C. 3.4.11.x); P, acid phosphatase (E.C. 220.127.116.11) and alkaline phosphatase (E.C. 18.104.22.168); SOM, phenol oxidase (E.C. 22.214.171.124 and E.C. 126.96.36.199) and peroxidase (E.C. 188.8.131.52). A subsample of the soil was used to make a slurry with 50 mmol/L acetate buffer (pH 5). -glucosidase, N-acetylglucosaminidase, aminopeptidase, and phosphatase activities were measured using the p-nitrophenol (pNP) method. (Sinsabaugh et al. 1993). The substrates were pNP- -D-glucopyranoside ( -glucosidase), glycinep-anilide (aminopeptidase), pNP- -N-acetylglucosaminide (chitobiase), pNP-phosphate for both phosphatases. Substrates were dissolved in acetate buffer (pH 5), except the substrates for the alkaline phosphatase and aminopeptidase, which were dissolved in 50 mmol/L TRIS buffer (pH 8). Two milliliters of the soil slurry were incubated with 2 mL of the corresponding substrate at 20 C for the recommended time for each enzyme. After incubation, the test tubes containing the samples were centrifuged and 1 mL of the supernatant was transferred to a tube containing 0.2 mL of 1mol/L sodium hydroxide (NaOH). The solutions were then brought up to 10 mL with distilled water and their absorbance was measured at 410 nm. Phenol oxidase and peroxidase activities were mea-
sured using L-3,4-dihydroxyphenylalanine (L-DOPA) and hydrogen peroxide for peroxidase (Sinsabaugh et al. 1993). Activities were processed in the same way as in the pNP method, with the exception that absorbance was directly measured on the supernatant after centrifugation at 460 nm. All of the above enzyme activities were expressed as micromoles per liter of substrate per gram per hour. Endocellulase activity was measured viscometrically using carboxymethylcellulose (CMC; Almin and Eriksson 1967). This method determines the rate of viscosity decrease, which results from polymer degradation by the enzyme action. One milliliter of the soil slurry was incubated with 2 mL of 1.25% CMC for 2 h at 20 C. Afterwards, the samples were centrifuged and the viscosity of the supernatant was measured as the fall velocity in a small-bore pipette. The activity was expressed as units proportional to absolute activity per gram per hour. We used a modi?ed version of the method developed by Degens and Harris (1997) for the measurement of substrate-induced respiration. Approximately 10-g (dry mass) subsamples of each soil sample were added to 26 previously prepared vials that contained a 2-mL solution of a given substrate in 25 of the vials and water in one of the vials (a control). The substrates were prepared in concentrations as indicated by Degens and Harris (1997). The 25 substrates were as follows: ?ve amino acids: L-asparagine, glutamine, L-lysine, arginine, soy protein hydrolizate (contains a number of aminoacids); eight carboxylic acids: -ketobutyric acid, -ketoglutaric acid, fumaric acid, L-ascorbic acid, malic acid, malonic acid, oxalic acid, citric acid; three simple sugars: glucose, manose, m-inositol; one lipid: linoleic acid; four complex plant polymers: cellulose, starch, tannic acid, lignin; and four plant litters (ground): barberry, stilt grass, oak (Quercus velutina), tulip poplar (Liriodendron tulipifera). After soil was added to the vials, it was thoroughly mixed and the vials were closed with airtight caps with septa for gas
PETER S. KOURTEV ET AL.
Ecology, Vol. 83, No. 11
sampling. The samples were then incubated for 2–3 h. The addition of soil to vials was staggered in time to allow for a similar incubation time for all the samples. After incubation, the carbon dioxide (CO2) in the headspace of each vial was measured using a Shimadzu 14C Gas Chromatograph (Shimadzu Scienti?c Instruments, Columbia, Maryland, USA). The results were expressed as micrograms C-CO2 per hour per gram. We subtracted the respiration of the control from the respiration in response to substrates to adjust for the effects of the added water and the stirring of the samples. The CO2 evolved in samples incubated with substrates was higher than that of controls in 99% of the incubations. When analyzing the data, we used both the individual responses to substrates as well as average responses for carboxylic acids, sugars, amino acids, polymers, and litters.
All data were analyzed using SAS for Windows, version 8 (SAS 2000). We analyzed the data using both multivariate (to detect patterns) and univariate (to detect signi?cant differences in individual variables) methods. We used two multivariate techniques to analyze our data. We ?rst applied principal components analysis (PCA) separately on PLFA proportions, enzyme activities, and SIR responses, including all of the variables in each case. We also applied discriminant analysis (DA) to our data. DA strictly requires that at least twice as many samples as variables are studied. We applied DA to all of the enzyme activities (eight variables) and average SIR responses to classes of compounds (six variables). In the case of PLFAs, we ?rst reduced the number of variables by running a stepwise DA. This is a stepwise selection method that uses partial F statistics and Wilk’s lambda criterion to select those variables from a dataset that contribute the most to the canonical discriminant functions. DA was then applied using only the variables selected by the preliminary stepwise DA. The two multivariate methods analyze different aspects of the variation in our datasets. PCA ordinates the samples in a way that maximizes the variation in the samples, regardless of their origin. We then analyzed the scores of the different soil samples using MANOVA and two effects: plants (barberry, stilt grass, or blueberry) and soils (bulk or rhizosphere samples). In the case of a signi?cant MANOVA, we proceeded by analyzing the data with DA. DA results in several canonical discriminant functions that separate a priori determined groups in a given dataset, and it ordinates the samples by minimizing within-group variation and maximizing between-groups variation. Therefore, DA was used as a supplement to PCA and provided information on which variables are best suited for discriminating between the prede?ned groups of soils.
Individual microbial variables were ?rst studied using MANOVA. Since all MANOVAs were signi?cant 3.67, F 2.87, P 0.001; (PLFA, Pillai’s trace enzymes, Pillai’s trace 3.54, F 9.68, P 0.0001; 3.79, F 1.87, P 0.01), we SIR, Pillai’s trace proceeded by using two-way ANOVAs with location (plant species growing on the soil) and soil (bulk vs. rhizosphere) as factors. Means for the main effects and interactions were compared using Tukey’s honestly signi?cant differences (HSD) test and least square differences (LSD). PLFA proportions were arcsine transformed, while SIR and enzyme activities were log transformed so that normality requirements of ANOVA were met. To study the relationship between microbial community structure and function, we used canonical correlation analysis (CANCOR). This method is applied to two sets of variables, and it extracts pairs of canonical functions from each set that are correlated with each other. As with DA, CANCOR requires more samples than variables be used. This analysis was performed separately on bulk and rhizosphere soils. In order to comply with the sample size requirement, we conducted analyses correlating the ?rst four principal components extracted from the analysis of PLFAs (separately for bulk and rhizosphere soils) with (1) the eight enzyme activities and (2) the mean SIR responses for the six types of substrates. A scree plot of the principle components’ eigenvalues was used prior to the CANCOR analysis to ensure that the ?rst four axes extracted in the PLFA analysis were signi?cant. RESULTS
The two-way analysis of variance of individual PLFA resulted in signi?cant differences for most of the fatty acids (Fig. 1). In general, there were more signi?cant effects of plants than of soils. In most cases, when signi?cant interactions were observed, differences in the rhizosphere were more pronounced than those in bulk samples. Diagnostic fatty acids (Table 1) were used to calculate two indexes: bacterial to fungal fatty acids index and gram-negative to gram-positive bacterial PLFAs index. The ratio of gram-negative to gram-positive bacteria was close to 1 in all soils and not signi?cantly different in any of them (data not shown). In contrast, the bacterial to fungal fatty acids index differed under the three plant species in both bulk (11.00 0.33 for barberry, 12.58 0.58 for stilt grass, 1.21 for blueberry soils; means 1 SE) and 11.21 0.37 for barberry, 10.89 and rhizosphere (10.05 0.46 for stilt grass, and 6.98 0.20 for blueberry soils) soils. Two-way ANOVA of these results yielded a high5.73, P 0.01); ly signi?cant interaction term (F therefore, we contrasted means using Fisher’s LSD. There were no signi?cant contrasts in the bulk soils,
EFFECT OF EXOTICS ON SOIL MICROORGANISMS
FIG. 1. Presence and proportion of major phospholipid fatty acids (PLFAs) in the studied soils in Worthington State Forest, northwestern New Jersey, USA (means 1 SE). Sum 3 16:1 7c 15:0i 2OH; Sum 5 18:2 6,9c 18:0a; Sum 7 19:1 6c 19:0cy. Letters (top graph) indicate signi?cant ANOVAs: S, soils; P, plants; S P, their interaction.
while the two exotic species had higher bacterial to fungal indexes than the Vaccinium soils in the rhizosphere samples. The PCA analysis of the fatty acid data clearly separated the six soils (Fig. 2 and Table 2). The ?rst two principle components (PC) explained 54% of the variation in the data. PC1 separated the soils based on plant species, while PC2 separated bulk from rhizosphere samples (Fig. 2). For PC1, the important fatty
acids were 15:0, 15:0 3OH, and 16:1 2OH (all typical of gram-negative bacteria) in the positive direction and 14:0i, 15:0a (typical in gram-positive bacteria), and 16: 1 5c (typical in vesicular-arbuscular mycorrhizal [VAM] fungi) in the negative direction. For PC2, respectively, important fatty acids were 12:0, 14:0 (gramnegative bacteria), 13:0a (gram-positive bacteria), and 18:0 and 19:1 6c in the positive direction and 18:1 9c (fungal) and 17:0i 3OH and 19:0cy (both typical in
PETER S. KOURTEV ET AL.
Ecology, Vol. 83, No. 11
FIG. 2. Results from the multivariate analysis on the three types of microbial community variables. The ?rst column presents results from principal components (PC) analysis, while the second column presents results from a discriminant analysis (DF, discriminant function). Abbreviations are: PLFA, phospholipid fatty acid; SIR, substrate-induced respiration.
gram-negative bacteria) in the negative direction. The stepwise DA analysis of PLFA pro?les isolated 13 important fatty acids and very clearly separated all six kinds of soil samples (Fig. 2). The ?rst two canonical axes explained 80% of the variation in the data. The correlations of the individual fatty acids with the ?rst
two canonical variables are shown in Table 3. For the ?rst discriminant function, the important PLFA variables were typical of gram-negative bacteria and fungi. In contrast, the second discriminant function was strongly related to PLFAs typical of gram-positive bacteria.
EFFECT OF EXOTICS ON SOIL MICROORGANISMS
TABLE 2. MANOVA results comparing the principal components analysis scores along the ?rst two axes for the three types of microbial community variables. Pillai’s trace 1.02 0.53 1.54 0.87 0.52 0.59
Variable PLFA Enzymes SIR
Effect plants soils plants soils plants soils
15.23 15.5 58.31 114.6 6.16 24.4
0.0001 0.0001 0.0001 0.0001 0.001 0.0001
Notes: Three species of plants (Berberis, Microstegium, Vaccinium) and two kinds of soils (bulk and rhizosphere) were included. Abbreviations are:PLFA, phospholipid fatty acid; SIR, substrate-induced respiration.
Two-way ANOVAs of individual enzyme activities showed signi?cant effects of plants and soils on all enzymes, except peroxidase (Table 4 and Fig. 3). Rhizosphere activities were always higher than bulk activities. Cellulolytic activities and phosphatase activities were higher in blueberry samples, while N-related activities were higher in barberry and stilt grass samples. In addition, in all of the signi?cant ANOVAs, there were also signi?cant interactions between the two main effects (Table 4). In two of the cases effects were much more pronounced in bulk than in rhizosphere samples. In the rest of the cases, both rhizosphere and bulk soils showed similar and signi?cant differences (when individual least squares [LS] means were compared); however, the effects were better expressed in one of the soil types. Both PCA and DA clearly separated the six soil samples, with a stronger separation from the discriminant analysis (Fig. 2 and Table 2). The ?rst two axes of PCA explained 64% of the variation in the data. PC1 was a gradient from low to high endocellulase, -glucosidase, and acid phosphatase activities, while PC2 was a gradient of high to low chitobiase and phenol oxidase activities. The ?rst two axes of DA explained 95% of the variation in the data. The correlations of individual enzyme activities with the ?rst two canonical functions are shown in Table 5. Acid phosphatase and phenol oxidase activities were most important in the ?rst canonical function, while aminopeptidase and endocellulase activities were most correlated with the second canonical axis.
mineum soils to sugars and amino acids than the other plant types and a higher response to carboxylic acids in Berberis soils. In most cases, when interaction terms were signi?cant, effects were more pronounced in rhizosphere soil (data not shown). PCA based on SIR responses did not clearly separate soils under the different plant species, and there was a slight separation of rhizosphere soils from bulk soils along axis 1 (Fig. 2); however, MANOVA analysis showed that both plants and soil type had signi?cant effect on the SIR pro?les of the microbial community (Table 2). Since the MANOVA on the PCA scores was signi?cant, we proceeded with the DA of SIR responses, which clearly separated the different soil samples. The ?rst two canonical functions of the analysis explained 74.92% of the variation in the data. The ?rst canonical function was a gradient of low to high utilization of carboxylic acids, polymers, and litters and high to low utilization of amino acids and sugars (Table 6). All SIR responses were positively correlated with the second canonical function and the most important variables were all carboxylic acids: -ketobutyric, ketoglutaric, citric, L-ascorbic, and malonic acids. Correlations between structure and function variables
The ?rst pair of canonical functions explained 50% of the variation in each analysis (data not shown), and therefore we report the cross-loadings for only this function in Table 7. The high degree of explanatory power in the analysis suggests that both measures of microbial functionality (enzyme activities and respiratory responses to substrates) are well correlated with community structure, as indexed by the PLFAs. Redundancy analysis is used in canonical correlation analyses to determine if one of the data sets explains the variation in the second dataset: the cumulative redunTABLE 3. Correlation coef?cients of phospholipid fatty acids (PLFAs) initially selected through stepwise selection with the ?rst two discriminant functions (DF 1 and DF 2) resulting from the subsequent discriminant analysis. Correlated with PLFA 17:0cy 18:1 9c 18:1 7c 15:1i G 17:0i 16:1 5c Sum 3 18:1 2OH 10Me 18:0 18:0 3OH 16:0 13:0i Sum 5 DF 1 0.70 0.54 0.81 0.70 0.12 0.82 0.76 0.14 0.55 0.27 0.47 0.15 0.57 DF 2 0.09 0.10 0.05 0.32 0.45 0.18 0.33 0.44 0.16 0.14 0.19 0.03 0.10 Marker value gram ( ) fungi/gram ( ) gram ( ) ··· gram ( ) VAM fungi/Cytophaga gram ( ) gram ( ) actinomycetes ··· ubiquitous gram ( ) fungi
Substrate-induced respiration responses
In general, soils responded to different substrates in the following order: carboxylic acids, sugars, litters, amino acids, polymers, linoleic acid (Fig. 4). Rhizosphere soils had higher activities than the corresponding bulk soil for all substrates (two-way ANOVAs, analyses not shown), but there were few differences among plants (Fig. 4). The only patterns among classes of substrate were a higher average response by M. vi-
Note: Sum 3 indicates a co-eluting mixture of 16:1 7c and 15:0i 2OH. Sum 5 indicates a co-eluting mixture of 18:2 6,9c and 18:0a.
3160 TABLE 4.
PETER S. KOURTEV ET AL.
Results from the two-way ANOVAs of individual enzyme activities. Contrast
Ecology, Vol. 83, No. 11
Interaction§ Bulk soil Rhizosphere
Enzyme activity Endocellulase -glucosidase Chitobiase Aminopeptidase Acid phosphatase Alkaline phosphatase Phenol oxidase
Effect? S P S S P S S P S S P S S P S S P S S P S
212.89*** 32.36*** 6.98** 111.41*** 52.43*** 23.32*** 149.26*** 28.55*** 18.59*** 24.73*** 100.37*** 35.62*** 301.49*** 136.80** 45.17*** 9.35** 8.76***
Soil R R R R R R R B
M B B M, V B B
B V B, M M B M B M, V B M V V, M V
P P P P P P P
B B M
M B V M B M B, V M B V V
V B V M B, V
49.20*** 72.19*** 12.27***
* ? ? §
P 0.05; ** P 0.01; *** P 0.001. S, soil; P, plants. B, bulk soil; R, rhizosphere soil. Species abbreviations are: B, B. thunbergii, M, M. vimineum; V, Vaccinium.
dancy for all of the computed pairs of canonical functions is analogous to R2 values computed in multiple regression analyses. Based on the redundancy analysis, the relationship between structure and function was stronger for the index of function by enzyme activity than by respiratory response to selected substrates, and this pattern was seen in both rhizosphere and bulk soils. The speci?c relationship of the PLFA axes to the individual enzymes and substrate classes, as measured by the cross-loading values, differed in rhizosphere and bulk soils. However, -glucosidase and acid phosphatase were strongly correlated with the structural index of PLFAs in both soils, as were the respiratory responses to carboxylic acids and plant polymers. DISCUSSION Our results clearly show that soils under three different plant species have microbial communities that differ from each other in both structure and function and that the plants’ effects extend from the densely rooted soil horizon to sparsely rooted bulk soils. The differences in both structure (PLFA composition) and function (enzymes, SIR), however, are greater in the rhizosphere soils than in the bulk soils, suggesting that they are associated with the growth of the plants, either through the activity and growth of the roots and/or through the inputs to the soil surface of litters of different chemical qualities. We point out to the reader that our study was only conducted at one site. Our design is thus pseudo-replicated and the generality of our conclusions is limited. We justify our decision to work in one site in the fol-
lowing way. First, we have gathered extensively replicated data describing the soil under the two exotic species at several different study areas within three separate sites located over a 80-km transect in northern New Jersey (Kourtev et al. 1998, 1999, Ehrenfeld et al. 2001). These studies show very similar patterns in nitrogen cycling and litter decomposition associated with the two exotic plants. They suggest, therefore, that the soils at this one site are likely to have microbial properties similar to other invaded and noninvaded soils in the region and that therefore the results from the one site used in this study are not site-speci?c. Second, the SIR and enzyme samples must be run on fresh soils, so that analysis cannot be delayed in time. This number of samples (1092 for SIR, 335 for enzyme activity, as well as 42 samples subjected to the extraction procedure and analysis for PLFAs) and types of analyses are suf?ciently time-consuming that it is not
TABLE 5. Correlation coef?cients of the individual enzyme activities with the ?rst two discriminant functions (DF 1 and DF 2) produced by the discriminant analysis. Correlated with Enzyme Cellulase -glucosidase Chitobiase Aminopeptidase Acid phosphatase Alkaline phosphatase Phenoloxidase Peroxidase DF 1 0.17 0.63 0.29 0.42 0.97 0.09 0.61 0.12 DF 2 0.73 0.57 0.32 0.70 0.15 0.38 0.43 0.37
EFFECT OF EXOTICS ON SOIL MICROORGANISMS
FIG. 3. Enzyme activities in the six different studied soils (means 1 SE). Signi?cant differences in the activities are presented in Table 4. Note that endocellulase was measured in relative units per gram of soil per hour 1000. SOM soil organic matter.
possible to sample multiple sites within a short enough period of time to avoid seasonal changes in environmental conditions. Based on the above, we believe that we are presenting a strong initial case study, which suggests that exotic species may be affecting ecosystem processes in the soil in ways that have not been examined before. PLFA analyses have proven to be a very effective technique for studying the structure of microbial communities, as they summarize the presence of major microbial groups, as well as providing an index of diversity through the total number of fatty acids isolated
(Bossio et al. 1998). We isolated a total of 99 PLFAs in our samples, which is higher than the number reported from other soils (e.g., Cavigelli et al. 1995). However, more than half of these each represented 0.5% of the total PLFA isolated and may not be included among the total number of fatty acids reported in other studies. The discriminant analyses showed that although rhizosphere and bulk soils under each species had distinctive microbial communities, the differences associated with the plant species occupying the soil were much greater than the rhizosphere-bulk soil differenc-
PETER S. KOURTEV ET AL.
Ecology, Vol. 83, No. 11
FIG. 4. Substrate-induced respiration (SIR) responses in the six studied soils (means 1 SE). Letters in parentheses indicate signi?cant ANOVAs as in Fig. 1. Note the difference in scale for the ?rst two carboxylic acids.
es. Steer and Harris (2000) found that bulk and rhizosphere soils were differentiated by PLFA pro?les after several months’ growth of Agrostis stolonifera. Although the bulk and rhizosphere communities in this study were distinct, they were both different from the community of the initial soil, which supports our in-
ference that the effect of the plant on the soil microbial community extends through the soil beneath the root mass. This effect could be mediated by the movement of soluble organic compounds leached or exuded from the roots and/or decaying litter or might re?ect changes in physical conditions (water movement through ma-
EFFECT OF EXOTICS ON SOIL MICROORGANISMS
TABLE 6. Correlation coef?cients of the individual substrate-induced respiration (SIR) responses with the ?rst two discriminant functions (DF 1 and DF 2) produced by the discriminant analysis. Correlated with Substrate -ketobutyric acid -ketoglutaric acid Citric acid Fumaric acid L-ascorbic acid Malic acid Malonic acid Oxalic acid L-asparagine Glutamine L-lysine Soy hydrolizate Arginine Glucose Manose m-inositol Cellulose Starch Tannic acid Lignin Barberry litter Stilt grass litter Oak litter Tulip poplar litter Linoleic acid DF 1 0.03 0.06 0.31 0.33 0.24 0.34 0.20 0.34 0.33 0.24 0.23 0.20 0.01 0.15 0.16 0.36 0.20 0.25 0.08 0.38 0.21 0.37 0.15 0.04 0.17 DF 2 0.84 0.88 0.89 0.34 0.63 0.58 0.85 0.66 0.48 0.53 0.36 0.61 0.52 0.48 0.55 0.39 0.42 0.02 0.62 0.48 0.62 0.42 0.41 0.55 0.49
enzymes may be stabilized on clays and remain active for long periods of time (Sinsabaugh 1994). However, the patterns of enzyme activity closely mirrored those of the PLFA analyses: Vaccinium rhizosphere soils were strongly separated from the others, barberry and stilt grass soils were clearly distinguished from one another, and bulk and rhizosphere soils under each plant were also distinguished from one another. These results suggest that despite the potential long-term persistence of active enzymes in clays and organic matter, the suite of activities that characterize a given soil re?ect the current vegetation. The higher activities of N-related enzymes (chitobiase and aminopeptidase) activities beneath both of the exotics are particularly striking. Previous studies (Kourtev et al. 1999, Ehrenfeld et al. 2001) have shown higher availability of nitrogen in exotic-invaded soils; therefore, we expected to ?nd very low N-related activities under the two exotics as these enzyme activities are generally thought to be inversely proportional to N availability (Sinsabaugh and Moorhead 1994, Sims and Wander 2002). Several explanations for the observed patterns are possible. First, it is possible that microorganisms in the exotic soils may be competing with plant roots for nitrogen. Recent ?ndings suggest that plants and microbes may compete for available inorganic nitrogen in the soil and that competition may be more severe between nitri?ers and plants for the availTABLE 7. Summary of canonical correlation analyses of the enzyme and substrate-induced respiration (SIR) patterns with the phospholipid fatty acid (PLFA) pro?les in the two kinds of soil. Variables Cross-loading with PLFA-CF1 Enzymes Cellulase -glucosidase Chitobiase Aminopeptidase Acid phosphatase Alkaline phosphatase Phenoloxidase Peroxidase Redundancy analysis Redundancy CF1 Redundancy total Cross-loading with PLFA-CF1 SIR responses Amino acids Carboxylic acids Simple sugars Plant polymers Ground litters Linoleic acid Redundancy analysis Redundancy CF1 Redundancy total Bulk Rhizosphere
cropores, soil moisture content). It might also re?ect the downward movement of microbes through macropores facilitated by water ?uxes through the sur?cial roots. The most notable change in the composition of the microbial community was the large increase in the ratio of bacterial:fungal fatty acids under both exotics, compared to the rhizosphere soils of the native Vaccinium, and the higher bacterial:fungal ratio in all bulk soils compared to their respective rhizosphere soils. While the rhizosphere–bulk soil comparisons may in part re?ect the higher organic matter content of the surface (rhizosphere soils), which typically supports a high saprophytic fungal biomass (Ruzicka et al. 2000), the difference may also re?ect the presence of plant-supported mycorrhizal fungal hyphae. Indeed, rhizosphere soils from the M. vimineum stands were distinguished by the large amounts of VAM fatty acids. In previous studies, we have found that stilt grass produces relatively little root biomass (Ehrenfeld et al. 2001), and the large amount of VAM biomass indicated by the PLFA analysis suggests that mycorrhizae may be particularly important to this species as a substitute for roots. Vaccinium roots are known to support large amounts of ericoid mycorrhizae (Goulart et al. 1993), and this might contribute to the higher absolute amount of fungal PLFA, and resulting lower bacterial:fungal ratio, than in the other soils. Soil enzymes provide an integrated index of microbial activity over long periods of time, because the
0.18 0.65 0.85 0.33 0.87 0.06 0.85 0.49 0.38 0.65
0.49 0.90 0.37 0.28 0.86 0.02 0.24 0.22 0.27 0.62
0.21 0.64 0.13 0.43 0.21 0.01 0.12 0.32
0.22 0.76 0.35 0.65 0.31 0.32 0.23 0.42
Note: PLFA-CF1 indicates the PLFA function from the ?rst pair of canonical functions (CF1) extracted by the correlation analysis.
PETER S. KOURTEV ET AL.
Ecology, Vol. 83, No. 11
able ammonium (Hart and Stark 1997). Alternatively, even though soils under the exotics have higher absolute nitrogen availability, the relative proportions of available nitrogen to available carbon may be low, which could lead to a nitrogen limitation for the soil micro?ora. Differences in oxidative activities closely mirrored aminopeptidase activity, which provides support for the latter hypothesis. Our results show, nevertheless, that high levels of extractable N and high mineralization rates do not necessarily indicate that N is freely available to the microbial community. Vaccinium soils supported higher activities of both cellulolytic enzymes and also higher amounts of acid phosphatase activity. Blueberries typically root in organic-rich horizons with large amounts of recalcitrant compounds; the higher cellulolytic activities may thus re?ect a higher degree of carbon limitation in this microbial community. The higher rates of acid phosphatase may also re?ect the organic-rich surface soils that accumulate below blueberries that typically have a signi?cantly lower pH than do the soils beneath the exotics (Kourtev et al. 1998). Conversely, the striking higher alkaline phosphatase beneath the stilt grass may re?ect the more alkaline conditions that we have found are associated with populations of this plant (Kourtev et al. 1998) and which suppress acid phosphatase and promote alkaline phosphatase activities (Acosta-Mart?nez and Tabatabai 2000, Kramer and Green 2000). ? ¨ The Degens and Harris (1997) method of indexing functional diversity by measuring the respiratory response of whole soil preparations to simple and complex substrates proved to be a useful technique in this study. We found that discriminant analysis based on the addition of 25 different substrates effectively separated both soils beneath different plant species and also bulk and rhizosphere soils within each plant cover type. Thus, the method proved as effective as BIOLOGbased studies of microbial diversity under different plants (Westover et al. 1997, Grayston et al. 1998, 2001). Our results also suggest that differences in utilization of simple substrates are less pronounced among soils than are differences in the activities of extracellular enzymes associated with nutrient and carbon mineralization. Although many of the substrates did not produce different respiratory responses in the different soils, the carboxylic acids, which were also the substrates yielding the highest respiratory responses in all soils, did elicit signi?cantly different responses. Studies using BIOLOG have usually found that amino acids and sugars elicit the highest average response from microbial communities. We believe that the SIR method captures the activity of a wider range of microorganisms (including fungi) than BIOLOG and other culture-related methods, which could re?ect a broader capacity to use these compounds than is found among the culturable organisms. Degens and Harris (1997) also found that carboxylic acids showed the most variable
response in soils under different agricultural management, as well as the highest responses on the average. There is evidence that microbial communities can condition their SIR response after the addition of simple substrates in the soil, for example amino acids (Hopkins and Ferguson 1994). Thus, our results suggest that, in these soils, carboxylic acids were a major easily decomposable C source. For all of the signi?cantly different responses to added substrates, the exotic species soils had higher responses, which indicates a higher availability of these substrates in exotic-affected soils. The relationship between community diversity, as indexed by PLFA pro?les, and functional capacity, as indexed by enzyme activities and SIR responses, is hard to establish, as is evident from the relatively small number of studies that have attempted the comparison. Widmer et al. (2001), for example, found that DNA, PLFA, and BIOLOG methods were all successful at distinguishing different soils, but the three methods suggested different patterns of relationship among the soils. Our results support these ?ndings. While all three methods effectively separated both the plant cover types from one another and the rhizosphere and bulk soils from each cover type, each method displayed somewhat different patterns of similarities and differences, as illustrated in the discriminant analyses (Fig. 2). Waldrop et al. (2000) compared PLFA pro?les, BIOLOG pro?les, and enzyme activities in ?ve tropical soils. Using comparisons with both individual fatty acids and principal component axes extracted from the PLFA data, they found that BIOLOG activity was not well correlated with PLFA patterns, while enzyme activity was well correlated. Similarly, we found that enzyme activities were better correlated with PLFA pro?les than were SIR responses to multiple substrates as shown in the redundancy analysis of the canonical correlation results. The PLFA pro?les only provide an index of community structure rather than speci?c information about species. A correlation of structure (PLFA) and function, e.g., in enzymes, shows that a particular set of species is associated with a particular pattern of enzyme activities but does not tell us which species are responsible. Nevertheless, our results clearly indicate that pro?les of enzymatic and catabolic capacity in the soil are associated with different communities of microorganisms. In summary, our study was aimed at measuring the effects of three plant species on microbial community structure and function in the soil and the relation between these two measures of microbial diversity. The generality of our conclusions is limited by our experimental design; however, our results clearly suggest that exotic species may cause changes in the soil ecosystem that have not been studied before. From the data, several distinct trends emerged. Microbial communities under the different plant species studied were
EFFECT OF EXOTICS ON SOIL MICROORGANISMS
different in their structure and function in both rhizosphere and bulk soil samples. It is striking that, with the same land-use history and in adjacent soils, under a uniform tree canopy, the change in understory ?ora could cause such a pronounced differentiation of the soil microbial community. Our results are also of special interest, since we compared two exotic species to a native species and thus we reveal the potential for exotic plants to signi?cantly alter microbial communities in the soil. Rhizosphere and bulk soil were similar in their structure and function under the two exotic species, underlining the potential for long-term effects of the introduced species. Functional microbial diversity pro?les were well related to soil microbial community structure. The relation held for both bulk and rhizosphere soils and was much stronger for enzyme activities than for the response of soil to added substrates.
ACKNOWLEDGMENTS We wish to thank M. Kourteva, L. Masker and K. Sedia for their help in the ?eld and in the laboratory. This work was partly supported by a USDA grant. LITERATURE CITED Acosta-Mart?nez, V., and M. A. Tabatabai. 2000. Enzyme ? activities in a limed agricultural soil. Biology and Fertility of Soils 31:85–91. Almin, K. E., and K. E. Eriksson. 1967. Enzymic degradation of polymers 1. Viscometric method for the determination of enzymic activity. Biochimica et Biophysica Acta 139: 238–247. Borga, P., M. Nilsson, and A. Tunlid. 1994. Bacterial communities in peat in relation to botanical composition as revealed by phospholipid fatty acid analysis. Soil Biology and Biochemistry 7:841–848. Bossio, D. A., K. M. Scow, N. Gunapala, and K. J. Graham. 1998. Determinants of soil microbial communities: effects of agricultural management, season, and soil type on phospholipid fatty acid pro?les. Microbial Ecology 36:1–12. Cavigelli, M. A., G. P. Robertson, and M. J. Klug. 1995. Fatty acid methyl ester (FAME) pro?les as measures of soil microbial community structure. Plant and Soil 170:99–113. Coleman, M. D., R. E. Dickson, and J. G. Isebrands. 2000. Contrasting ?ne-root production, survival and soil CO2 ef?ux in pine and poplar plantations. Plant and Soil 225: 129–139. D’Antonio, C., and P. Vitousek. 1992. Biological invasions by exotic grasses, the grass/?re cycle, and global change. Annual Review of Ecology and Systematics 23:63–87. Degens, B. P. 1998. Decreases in microbial functional diversity do not result in corresponding changes in decomposition under different moisture conditions. Soil Biology and Biochemistry 30:1989–2000. Degens, B. P., and J. A. Harris. 1997. Development of a physiological approach to measuring the catabolic diversity of soil microbial communities. Soil Biology and Biochemistry 29:1309–1320. Drijber, R. A., J. W. Doran, A. M. Parkhurst, and D. J. Lyon. 2000. Changes in soil microbial community structure with tillage under long-term wheat-fallow management. Soil Biology and Biochemistry 32:1419–1430. Duarte, G. F., A. S. Rosado, L. Seldin, A. C. Keijzer-Wolters, and J. D. Van Elsas. 1998. Extraction of ribosomal RNA and genomic DNA from soil for studying the diversity of the indigenous bacterial community. Journal of Microbiological Methods 32:21–29.
Ehrenfeld, J. E. 1997. Invasion of deciduous forest preserves in the New York Metropolitan region by Japanese barberry (Berberis thunbergii DC.). Journal of the Torrey Botanical Society 124:210–215. Ehrenfeld, J. G., P. S. Kourtev, and W. Z. Huang. 2001. Changes in soil functions following invasions of exotic understory plants in deciduous forests. Ecological Applications 11:1287–1300. Fletcher, S. J. 1979. Soil survey of Warren County, New Jersey. U.S. Department of Agriculture, Soil Conservation Service, Somerset, New Jersey, USA. Goulart, B. L., M. L. Schroeder, R. L. Darnell, J. R. Clark, and W. F. Wilcox. 1993. Blueberry mycorrhizae: current knowledge and future directions. Acta Horticulturae 346: 230–239. Grayston, S. J., and C. D. Campbell. 1996. Functional biodiversity of microbial communities in the rhizosphere of hybrid larch (Larix eurolepis) and Sitka spruce (Picea sitchensis). Tree Physiology 16:1031–1038. Grayston, S. J., G. S. Grif?th, J. L. Mawdsley, C. D. Campbell, and R. D. Bardgett. 2001. Accounting for variability in soil microbial communities of temperate upland grassland ecosystems. Soil Biology and Biochemistry 33:533– 551. Grayston, S. J., S. Wang, C. D. Campbell, and A. C. Edwards. 1998. Selective in?uence of plant species on microbial diversity in the rhizosphere. Soil Biology and Biochemistry 30:369–378. Hart, S. C., and J. M. Stark. 1997. Competition for nitrogen between plants and soil microorganisms. Trends in Ecology and Evolution 385:61–64. Hill, G. T., N. A. Mitkowski, L. Aldrich-Wolfe, L. R. Emele, D. D. Jurkonie, A. Ficke, S. Maldonado-Ramirez, S. T. Lynch, and E. B. Nelson. 2000. Methods for assessing the composition and diversity of soil microbial communities. Applied Soil Ecology 15:25–36. Hopkins, D. W., and K. E. Ferguson. 1994. Substrate induced respiration in soil amended with different amino acid isomers. Applied Soil Ecology 1:75–81. Kelly, J. J., M. Haggblom, and R. L. Tate III. 1999. Changes ¨ in soil microbial communities over time resulting from one time application of zinc: a laboratory microcosm study. Soil Biology and Biochemistry 31:1455–1465. King, J. D., D. C. White, and C. W. Taylor. 1977. Use of lipid composition and metabolism to examine structure and activity of estuarine detrital micro?ora. Applied and Environmental Microbiology 33:1177–1183. Kourtev, P. S., J. G. Ehrenfeld, and W. Z. Huang. 1998. Effects of exotic plant species on soil properties in hardwood forests of New Jersey. Water, Air and Soil Pollution 105: 493–501. Kourtev, P. S., J. G. Ehrenfeld, and W. Z. Huang. 2002. Enzyme activities during litter decomposition of two exotic and two native plant species in hardwood forests of New Jersey. Soil Biology and Biochemistry, in press. Kourtev, P. S., W. Z. Huang, and J. G. Ehrenfeld. 1999. Differences in earthworm densities and nitrogen dynamics in soils under exotic and native plant species. Biological Invasions 1:237–245. Kramer, S., and D. M. Green. 2000. Acid and alkaline phos¨ phatase dynamics and their relationship to soil microclimate in a semiarid woodland. Soil Biology and Biochemistry 32:179–188. Marilley, L., and M. Aragno. 1999. Phylogenetic diversity of bacterial communities differing in degree of proximity of Lolium perenne and Trifolium repens roots. Applied Soil Ecology 13:127–136. MIDI. 1995. Sherlock microbial identi?cation system operating manual. Version 5. MIDI, Newark, Delaware, USA. Olsson, S., and Alstrom. 2000. Characterisation of bacteria ¨
PETER S. KOURTEV ET AL.
Ecology, Vol. 83, No. 11
in soils under barley monoculture and crop rotation. Soil Biology and Biochemistry 32:1443–1451. Pankhurst, C. E., K. Ophel-Keller, B. M. Doube, and V. V. S. R. Gupta. 1996. Biodiversity of soil microbial communities in agricultural systems. Biodiversity and Conservation 5:197–209. Priha, O., S. J. Grayston, T. Pennanen, and A. Smolander. 1999. Microbial activities related to C and N cycling and microbial community structure in the rhizospheres of Pinus sylvestris, Picea abies and Betula pendula seedlings in an organic and mineral soil. FEMS Microbiology Ecology 30: 187–199. Ruzicka, S., D. Edgerton, M. Norman, and T. Hill. 2000. The utility of ergosterol as a bioindicator of fungi in temperate soils. Soil Biology and Biochemistry 32:989–1005. SAS. 2000. SAS/STAT user’s guide, version 8. SAS Institute, Cary, North Carolina, USA. Sims, G. K., and M. M. Wander. 2002. Proteolytic activity under nitrogen or sulfur limitation. Applied Soil Ecology 19:217–221. Sinsabaugh, R. L. 1994. Enzymic analysis of microbial pattern and process. Biology and Fertility of Soils 17:69–74. Sinsabaugh, R. L., R. K. Antibus, A. E. Linkins, C. A. McClaugherty, L. Rayburn, D. Repert, and T. Weiland. 1993. Wood decomposition: nitrogen and phosphorus dynamics in relation to extracellular enzyme activity. Ecology 74:1586–1593. Sinsabaugh, R. L., and D. L. Moorhead. 1994. Resource allocation to extracellular enzyme production: a model for nitrogen and phosphorus control of litter decomposition. Soil Biology and Biochemistry 26:1305–1311. Steer, J., and J. A. Harris. 2000. Shifts in the microbial community in rhizosphere and non-rhizosphere soils during the
growth of Agrostis stolonifera. Soil Biology and Biochemistry 32:869–878. Sundh, I., M. Nilsson, and P. Borga. 1997. Variation in microbial community structure in two boreal peatlands as determined by analysis of phospholipid fatty acid pro?les. Applied and Environmental Microbiology 63:1476–1482. Vitousek, P. 1990. Biological invasions and ecosystem processes: towards an integration of population biology and ecosystem studies. Oikos 57:7–13. Waldrop, M. P., T. C. Balser, and M. K. Firestone. 2000. Linking microbial community composition to function in a tropical soil. Soil Biology and Biochemistry 32:1837– 1846. Westover, K. M., A. C. Kennedy, and S. E. Kelley. 1997. Patterns of rhizosphere microbial community structure associated with co-occurring plant species. Journal of Ecology 85:863–873. White, D. C., W. M. Davis, J. S. Nickels, J. D. King, and R. J. Bobbie. 1979. Determination of the sedimentary microbial biomass by extractable lipid phosphate. Oecologia 40: 51–62. Widmer, F., A. Fiel?bach, E. Laczko, J. Schulze-Aurich, and ? J. Zeyer. 2001. Assessing soil biological characteristics: a comparison of bulk soil community DNA-, PLFA-, and BIOLOG -analyses. Soil Biology and Biochemistry 33: 1029–1036. Wolfe, P. E. 1977. The geology and landscapes of New Jersey. Crane Russak, New York, New York, USA. Zogg, G. P., D. R. Zak, D. B. Ringelberg, N. W. MacDonald, K. S. Pregitzer, and D. C. White. 1997. Compositional and functional shifts in microbial communities due to soil warming. Soil Science Society of America Journal 61:475– 481.