The e¡ects of stubble retention and nitrogen application on soil microbial community structure and functional gene abundance under irrigated maize

The effects of agronomic management practices on the soil microbial community were investigated in a maize production system in New South Wales, Australia. The site has been intensively studied to measure the impact of stubble management and N-fertilizer application on greenhouse gas emissions (CO 2 and N 2 O), N-cycling, pathology, soil structure and yield. As all of these endpoints can be regulated by microbial processes, the microbiology of the system was examined. Soil samples were taken after a winter fallow period and the diversity of the bacterial and fungal communities was measured using PCR-denaturing gradient gel electrophoresis. Stubble and N shifted the structure of bacterial and fungal communities with the primary driver being stubble addition on the fungal community structure ( P o 0.05 for all effects). Changes in C, N (total and NO 3 ), K and Na, were correlated ( P o 0.05) with variation in the microbial community structure. Quantitative PCR showed that nif H (nitrogen ﬁxation) and nap A (denitriﬁcation) gene abundance increased upon stubble retention, whereas amo A gene numbers were increased by N addition. These results showed that the management of both stubble and N have signiﬁcant and long-term impacts on the size and structure of the soil microbial community at phylogenetic and functional levels. study examined the effect of stubble and N management on the size and structure of dominant microbial communities in the Whitton ﬁeld site. Community structure and diversity of bacteria and fungi were assessed by rRNA gene-based PCR-DGGE, and were related to changes in the properties of the resulting


Introduction
The size and structure of soil-borne microbial communities are intimately linked with physicochemical soil properties (Garbeva et al., 2004;Balser & Firestone, 2005) and soil microbial communities sustain many vital ecosystem processes, such as nutrient cycling, decomposition of organic matter and waste, nutrient availability, degradation of pesticides and contaminants, soil structure, and plant growth and health (van Elsas et al., 1997). Assessing the effect of soil management practices on microbial community structure, diversity and activity in soil is therefore critical to advancing the understanding of the functionality, stability, and resilience of managed and natural ecosystems (Kennedy & Smith, 1995).
The ability to manage inputs to soils in agricultural systems provides an excellent opportunity to investigate many areas of microbial ecology in a directed approach. The information provided may ultimately be used to manage agroecosystems towards production, sustainability and environmental targets. Increasingly, ecological investigations assessing microbial community structure and diversity employ culture-independent techniques such as rRNA gene-based PCR-denaturing gradient gel electrophoresis (PCR-DGGE) profiling (Muyzer & Smalla, 1998). Culturedependent methods are, in general terms, predisposed towards selection and assessment of easily cultured and fast growing microorganisms. Such methods can therefore exert strong experimental bias on r-strategy community assessment, which is not acceptable in broad ecological studies. As such, the use of molecular methods for microbial community analysis has been applied to a wide range of areas in soil microbial ecology.
An irrigated maize monoculture field trial site was established at Whitton, New South Wales, Australia, in 1999 and ran until 2005. The trial was designed to investigate the effect of a range of conservation tillage and N management treatments on maize production, plant health and soil nutrient cycling (C and N), and to investigate environmental end-points, such as greenhouse gas emission (e.g. N 2 O), C-sequestration and soil structural benefits (Harvey et al., 2006;Kirkby et al., 2006;Meyer et al., 2006). Many of these processes are driven by microbial activity, and it is therefore essential to investigate the effect of altering stubble and nitrogen management on the size and structure of the microbial community.
This study examined the effect of stubble and N management on the size and structure of dominant microbial communities in the Whitton field site. Community structure and diversity of bacteria and fungi were assessed by rRNA gene-based PCR-DGGE, and were related to changes in the physicochemical properties of the soil resulting from management treatments. Although rRNA gene-based PCR-DGGE provides information on microbial community structure, it gives little information on functional aspects of these communities. Accordingly, the response to management of key functional groups was measured using real-time PCR. Focus was given towards microbial communities involved in the geochemical cycling of N for the following reasons: (1) N is a key driver for ecosystem fertility but can also act as a pollutant (eutrophication, greenhouse gas); (2) microbial activity controls most N transformations; and (3) methods are available to target conserved genes encoding enzymes involved in N transformations. Amplification of conserved regions of the amoA, nifH and napA genes was therefore used to assess the relative size of the microbial communities involved in N-fixation, nitrification and denitrification.

Materials and methods
Site, soil collection and processing The irrigated continuous maize trial site was located in the Murrumbidgee irrigation area, NSW, Australia. Soil was sampled from under four different management practices: stubble retained, with and without N addition; or stubble burnt, with and without N addition. Where stubble was retained, it was rotary-hoed back into the soil at the end of the growing season. Treatments receiving N had fertilizer applied at district practice rates (329 kg N ha À1 year À1 ). The N application was split: presowing, 112 kg N ha À1 was applied as super phosphate (32% N and 10% P), 5 kg ha À1 of liquid N at sowing, 230 kg ha À1 urea 1 month after sowing, and a second application of urea 1 week after the first . P was applied across treatments without N addition to balance inputs. Soil from each of the management treatments was collected from four different localities (replicates): two true randomized field blocks and two spatially distant (4 50 m) subreplicates within the blocks. The distance between sampling locations within the field blocks was greater than the distance between blocks. At each sampling point, several small samples of topsoil (1-15 cm) from randomly selected different locations were collected using a hand trowel and mixed into a bulk sample. Soils were sampled in September 2004, which followed the winter fallow period, and dried in a glasshouse to c. 3% moisture. Dried soil samples were crushed in a mortar and pestle but were not sieved so as to retain coarse organic matter. The maximum water holding capacity (MWHC) of the soil was determined using Shaw's method (Jenkinson & Powlson, 1976) (MWHC = 58 g water per 100 g soil). The physicochemical properties of soils from the different treatments were determined by the Analytical Services Unit of CSIRO Land and Water, Adelaide. Samples (100 g dry weight) of each of the 16 soils were preincubated at 20 1C for 14 days in the dark at 50% MWHC prior to microbial biomass determination and DNA extraction.

Microbial biomass determinations
Microbial biomass carbon (MBC) was determined using the chloroform fumigation extraction method (Jenkinson & Powlson, 1976). Briefly, for two samples of c. 10 g from each of the soils, one sample was directly extracted to determine dissolved C and the second was extracted following chloroform fumigation-lysis of microbial cells (7 days of fumigation). Extractions were performed by adding 30 mL of K 2 SO 4 (0.5 M; pH 6.3) to each sample, shaking for 1 h and then filtering through two layers of filter paper (Whatman no. 42) into plastic phials. Total C was determined by infrared detection following combustion at 950 1C on a Formacs series combustion TOC/TN analyser (Skalar Analytical Ltd).

DNA extraction and PCR
From each sample of incubated soil, three 1-g subsamples were processed for DNA extraction using the UltraClean Soil DNA isolation kit (MoBio Laboratories) following the manufacturer's instructions but with mechanical disruption by a FastPrep beadbeater (FP120; Qbiogene). Extracts from subsamples were combined into a final volume of water (150 mL), and DNA quantified using a SpectraMAX Spectrofluorometer with PicoGreen double-stranded DNA quantification reagent (Molecular Probes Inc.).
PCR for bacterial and fungal community analyses was conducted in an Eppendorf mastercycler gradient PCR machine using Hot-Star Taq DNA polymerase (1 U per 25 mL) and buffer (Qiagen). Functional gene quantification was conducted on a Stratagene MÂ3000P real-time PCR machine using Qiagen QuantiTect SYBR green 1 mix, with amplification adjusted to background (passive) fluorescence of ROX. C T values were determined from a software-determined baseline; C T is the cycle at which PCR amplification is in exponential amplification above a given threshold. For all genes, PCR amplification was initiated by a hot-start incubation at 95 1C for 15 min, and the amplification cycles were followed by a final elongation step at 72 1C for 5 min. Dissociation curves were conducted at the end of real-time PCR and reaction products were also separated on 1.5% agarose gels to confirm single banding at the expected size. All reactions were conducted in a volume of 25 mL.

Functional gene detection
Functional gene detection for nitrogen fixation was based on PCR of nitrogenase reductase gene (nifH) fragments as described by Widmer et al. (1999). The nested PCR used primers nifH(forA) and nifH(rev) for the first-round amplification, and primers nifH(forB) (Zehr & McReynolds, 1989) and nifH(rev) for the second (Table 1). In addition to the PCR master mix, first-round reactions contained 20 pmol of primers nifH(forA) and nifH(rev) and 2 mL of DNA (1 : 10 dilution). Second-round amplifications used primers nifH(forB) and nifH(rev) and 2 mL of 1 : 10 dilution first-round PCR product as template. First-round PCR cycling conditions were: 30 cycles of 94 1C for 1 min, 48 1C for 1 min and 72 1C 2 min. Second-round PCR cycling conditions were: 30 cycles of 94 1C for 30 s, 57 1C for 30 s and 72 1C for 45 s.
Measurement of genetic potential for denitrification was based on detection of periplasmic nitrate reductase (napA) following Flanagan et al. (1999). The first round of the nested PCR used primers V16 and V17, whereas the second round used primers V66 and V67 (Table 1); primers were used at 15 pmol. In the first round of PCR, 2 mL of 1 : 10 dilution DNA was used as template for amplification. In the subsequent round of PCR, 2 mL of the first-round reaction mixture was used as the DNA template. Both first-and second-round PCR followed the same cycle conditions: 30 cycles of 94 1C for 1 min, 50 1C for 1 min and 72 1C for 2 min.
PCR was carried out to amplify a fragment of the ammonia monooxygenase gene (amoA), following the method and using the primers described in Stephen et al. (1999) as modified from Rotthauwe et al. (1997). PCR amplifications contained 10 pmol each of primers amoA-1F Ã and amoA-2R (Table 1), 10 mg bovine serum albumin (Roche Diagnostics Ltd), and 2 mL of soil-extracted DNA. PCR conditions consisted of 35 cycles of 92 1C for 1 min, 50 1C for 1 min and 68 1C for 45 s.

Community structure analysis
The structure of the soil bacterial and fungal communities was based on rRNA gene-targeted PCR-DGGE. Bacterial community-specific PCR used universal bacterial primers F968-GC and R1401 as described in Duineveld et al. (1998). Each PCR contained primers at 20 pmol and used 2 mL of 1 : 10 diluted DNA extract. The PCR cycle used a touchdown procedure with a reduction in annealing temperature from 67 to 57 1C over the first 20 cycles, and at 57 1C for the 20 cycles thereafter. Each cycle had denaturation at 94 1C for 1 min, annealing for 1 min and extension at 72 1C for 1 min. PCR of fungal internal transcribed spacer (ITS)-RNA genes was conducted with primers ITS1F-GC and ITS2 Ã (Table 1). Each PCR contained 20 pmol of each primer and used 2 mL of 1:10 diluted soil DNA. Thermocycling conditions were 40 cycles of 94 1C for 1 min, 56 1C for 1 min and 72 1C for 1 min. Confirmation of successful bacterial and fungal PCR was achieved by running 2 mL of PCR product on a 1.5% agarose gel, staining with ethidium bromide and visualizing under UV light. The remaining reaction mix was used for DGGE analysis. DGGE analysis of amplified bacterial 16S and fungal ITS rRNA genes was performed on the DCode Universal Mutation Detection System (Bio-Rad). The polyacrylamide gels (7% w/v) contained a linear formamide/urea gradient ranging from 45% to 55% (bacterial) or 20% to 50% (fungal) denaturant. Gels were run for 16 h at 60 V and 60 1C, silverstained (Bio-Rad silver stain kit), dried and scanned (Epson expression 1680 transparency scanner). From the gel images (e.g. Fig. 1), DNA band positions and intensity were detected using TOTALLAB software (Nonlinear Dynamics Ltd). Background signal was removed using the rolling ball method and 43 and 59 different band positions were detected for bacterial 16S rRNA gene and fungal ITS gels, respectively. Each band position was inferred as a distinct operational taxonomic unit (OTU) and relative band intensity was used as an indication of relative abundance (discussed in Fromin et al., 2002).

Statistical analysis
An ecological distance matrix was compiled on fourth root transformed DGGE banding data (to reduce overweighting of dominant species) using the Bray-Curtis dissimilarity algorithm. Nonmetric multidimensional scaling (MDS ordination) was used to investigate the effect of different soil treatments on the structure of the bacterial and fungal communities; on MDS plots and cluster analysis, most similar communities are grouped closer together (Clarke & Warwick, 2001). Two-way (stubble by nitrogen) crossed analysis of similarities (ANOSIM) was performed on the similarity matrix data set (Bray-Curtis). Redundancy analysis (RDA) was used to relate environmental log-transformed variables to microbial community structure, with the Monte Carlo permutation test used to calculate the significance and proportion of the total variance in community structure explained by environmental variables (CANOCO 4.53;Microcomputer Power).
where S = total number of species, N = number of individuals and P i = proportion of the total count arising from the ith species. ANOVA was used to analyse the effect of management treatments, soil physicochemical properties and functional gene abundance (GenStat8; Lawes Agricultural Trust).

Soil physicochemical properties and microbial biomass
Retention of maize stubble significantly increased soil total C, organic C and total N, but decreased NO À 3 N ( Table 2). The only significant effect of application of N to soil (329 kg N ha À1 year À1 ) was an increase in N as NO À 3 . The major treatment factor affecting MBC was stubble retention, which led to a significant increase (P o 0.001) ( Table 2). Nitrogen application reduced MBC where stubble was retained (P = 0.011). The amount of DNA extracted from soil increased with stubble retention (P o 0.001), but not with N addition (Table 2). Overall, there was a strong correlation between extractable DNA and MBC (R 2 = 0.79).

Microbial community structure and diversity
Stubble retention and nitrogen application significantly (P o 0.05) affected the structure of the bacterial and fungal communities ( Fig. 1; ANOSIM; Table 3). R-values show that for both microbial communities, the strongest discriminator of community structure was stubble retention (Table 3). These results were supported by MDS analyses, which again 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Fig. 1. Fungal DGGE gel showing profiles of ITS1 region of the rRNA genes amplified from soil using PCR. Each band is inferred to represent a distinct operational taxonomic unit, and band intensity is inferred as a measure of abundance. Lanes 1-4 = stubble burnt, no N application; lanes 5-8 = stubble burnt, 1N fertilization; lanes 9-12 = stubble retained, no N application; lanes 13-16 = stubble retained, 1N fertilization.
showed that both bacterial and fungal communities differed in composition under stubble burnt and stubble retention regimes ( Fig. 2a and b). The structure of the fungal community was also affected by nitrogen addition (R = 0.573, P = 0.003), but on the MDS plot it is evident that this effect was secondary to stubble treatment (Fig. 2b). Cluster analysis (complete linkage) showed similar effects (data not shown).
Redundancy discriminate analysis (RDA ordination) with Monte Carlo permutation was used to relate environmental variables with microbial community structure. For the bacterial community, RDA described 43.5% of the variation within the first two axes (Fig. 3a), and for the fungal community 39.1% (Fig. 4a). The low explanation of variation over two dimensions was due to the large species numbers (43 OTUs for bacteria; 59 fungal). Nevertheless, the bacterial community structure clearly separated along the x-axis based on stubble treatment (Fig. 3a). Variation in soil carbon (total or organic) explained 12-13% of the changes in bacterial community structure (P o 0.05; Table  4, Fig. 3b). Total N, NO À 3 , K 1 and Na 1 were also significantly related to changes in bacterial community composition (P o 0.05; Table 4, Fig. 3b). Similarly, variation in fungal community composition was most strongly linked to changes in soil total and organic C, NO À 3 (but not total N) and K 1 (Table 4, Fig. 4b).
Both stubble retention (P = 0.001) and nitrogen addition (P = 0.012) increased the richness of fungal, but not bacterial, DGGE banding patterns (Fig. 1, Table 5). Neither the evenness nor the diversity of fungal and bacterial community profiles was significantly affected by stubble or N addition (Table 5).

Functional gene detection
The cycle numbers at which exponential PCR amplification (C T ) of key genes for nitrogen fixation (nifH) and denitrification (napA) were reached were significantly lower in soil in which stubble had been retained (Table 5). As C T is inversely proportional to DNA template, relative abundance of DNA template (copy number) for both the nifH and napA genes was significantly higher under stubble retention. Conversely, C T for amoA detection was not affected by stubble retention, but by nitrogen application (Table 5). With applied N, the C T for amoA detection was significantly lower than for soils with no N application (P o 0.003). Therefore, N application significantly increased the copy number of amoA genes in the maize soils. The copy number of nifH, the gene for nitrogen fixation, increased in abundance by a factor of 2.2 under stubble retention. This takes into account a 4.2-fold increase in nifH copy number per Â 1.9 increase in total microbial population following stubble retention.

Discussion
Changes in farm management practices have a direct impact on soil microbial community structure, thereby affecting processes such as the biogeochemical cycling of major elements (e.g. C, N, P), decomposition processes, disease  suppression and other ecosystem functions (Swift & Anderson, 1994). Such process changes have been observed following retention of maize stubble and application of Nbased fertilizer in a high-value irrigated production system in Australia (Edis et al., 2006;Harvey et al., 2006;Kirkby et al., 2006;Meyer et al., 2006). Using soil from the same site, we have shown a significant change in the size of the total microbial community and in the structure of fungal and bacterial communities between management treatments. Using functional gene-based quantitative PCR, we also showed that agricultural management practices at the site directly affected the abundance of key groups of N-cycling bacteria. As such, this work supports previous evidence for an enduring change in the microbiology of the soil, which is linked to observed and measured changes in microbially driven processes and which has importance for both production and environmental endpoints. Successive years of maize stubble retention and N addition to soil resulted in significant alteration in soil physico-chemical and biological properties, even when determined at the end of the winter fallow period. With stubble retention, levels of total and organic C significantly increased in the soil, as did the size of the soil microbial community (as determined by MBC and extractable DNA). Although there was a strong correlation between the amount of DNA extracted from soil and MBC (R 2 = 0.79), the actual usefulness of using extracted DNA concentration as a method of quantifying microbial biomass between soils is questionable owing to variation in efficiencies of DNA extraction between soil types and between techniques used (e.g. Miller et al., 1999). Furthermore, it is likely that some of the DNA extracted from the soil would have originated from nonmicrobial material, such as plant residues and soil micro-and macrofauna.
Stubble retention also increased soil total N, but mineral N (NH þ 4 ) levels remained unchanged. Nitrogen as NO À Retained, 1 N in the stubble-retained soil system was incorporated into the increased microbial biomass. The microbial biomass was negatively affected by N addition in the stubble retention treatments. Stubble retention and N addition were shown to influence community structure significantly, as determined using PCR-DGGE profiling. Analysis of both bacterial and fungal communities demonstrated that stubble retention was the strongest driver affecting species composition, particularly for the fungal community. Using PCR-DGGE, Clegg et al. (2003) found that N fertilizer exerted a significant effect on the total bacterial and actinomycete community structures. Similarly, we found that nitrogen application affected total bacterial and fungal community structure, but the effect of N was not as strong as that of stubble management. RDA relating bacterial and fungal community structure to changes in soil physicochemical properties explained o 50% of the variation. As such, other environmental drivers not measured in this study were important in driving microbial community structure. Nevertheless, changes in bacterial and fungal communities were correlated with changes in several soil physicochemical properties. Not surprisingly, changes in soil C and N (total or NO À 3 ) accounted for the largest amount of variation in terms of community composition. Changes in K and Na explained low but significant portions of the variation in the bacterial and fungal communities. Concentrations of cations were higher in stubble-burned treatments owing to the deposition of oxidized ions back into the soil (i.e. K, Na, Ca and Mg oxides) (Xu et al., 2002).
Interestingly, the effect of agricultural management practice was greater for the fungal community than the bacterial. Previously, similar studies have focused on the effects of farm management processes on the soil bacterial community. The major drivers affecting structure of total bacterial communities are soil type and plant species (Girvan et al., 2003;Garbeva et al., 2004) although many other factors may have significant influence as shown in this study. The fungal community also appears to be highly responsive to ecosystem change, particularly in relation to high rates of organic inputs such as stubble. Soil fungi play important roles in cycling of C, denitrification and production of N 2 O (e.g. Shoun et al., 1992), they elicit and suppress plant disease, and stabilize soil aggregates, etc. Further studies on management effects on microbial communities should therefore include both bacteria and fungi.
Although stubble and N treatment significantly affected microbial community size and structure, very little effect was observed on diversity estimates. This lack of effect was surprising, as repeated addition of a large amount of maize stubble (10 t ha À1 ) might be expected to increase the community size and dominance of specific fungal decomposers, thus decreasing richness and evenness. However, the richness of the fungal community significantly increased following both application of stubble or N to the soil. It is important to recognize the limitations of ribosomal-based PCR-DGGE when studying microbial diversity. In particular, assessment of band numbers can be a poor surrogate for diversity owing to comigration of bands, multiple bands  13.0 Ã 9.7 Ã K 8 . 7 Ã 9.7 Ã Total N 6.9 Ã Na 6.9 Ã from single species owing to heterogeneity in rRNA gene sequences and amplification bias of PCR. For example, total band positions for the bacterial communities only delimited 43 different OTUs. Given the huge overall diversity of bacterial species in most soils, it is clear that the approach only assesses numerically dominant species. Significant shifts in functional diversity (substrate utilization) following addition of maize stubble or litter to soil have been demonstrated. Using Biolog s GN plates, Sharma et al. (1998) showed that functional diversity of soil bacterial communities increased after addition of maize litter in several different soils. In their incubation experiment, the shift in function was measurable over 32 weeks after maize litter addition. In the present study, we have shown that changes in microbial community size and structure, as well as the abundance of key functional genes, are affected by management of stubble and N into soil. Measurements were made from field soil following the winter fallow period, demonstrating the enduring effect of management treatments and the potential for subsequent impact on the following crop.
At the Whitton field site, the geochemical cycling of N has been studied in detail (Edis et al., 2006;Meyer et al., 2006). In particular, research has focused on the production of N 2 O as a result of either denitrification or nitrification, the latter being a relatively minor process in terms of total N 2 O emissions. It is generally assumed that a microbial community capable of undertaking these processes is both present and potentially active under the given agroecological conditions. PCR amplification of genes involved in nitrate reduction (napA) and nitrification (amoA) was used to demonstrate the microbial capacity for these processes to occur in the soil and showed that the management of stubble and N had a direct effect on the population size of N-cycling bacteria. Under stubble retention, the population size of denitrifying bacteria was much higher than under the stubble-burnt treatment. Although this is inconsistent with the increased N 2 O emissions from the stubble-burnt treat-ment, the results are consistent with denitrification potential of the soil if the physicochemical properties at the time of sampling are considered. At the end of the fallow period, soil NO À 3 concentration was significantly greater following both stubble retention and N fertilization than in stubble retention only, but there was no significant difference in soil NH þ 4 . The genetic capacity to reduce NO À 3 is common to a wide range of soil bacteria, including nonspecialized denitrifying bacteria. As such, an increase in the population size of nitrate-reducing bacteria may be the result of greater overall soil microbial population size in the stubble-retained treatment. To assess the microbial potential for N 2 O emissions in a soil system accurately, other genes involved in the denitrification pathway should be also assessed.
In contrast to the abundance of the denitrification gene napA, the size of the microbial community involved in nitrification was affected only by nitrogen addition. This result is not surprising as ammonia-oxidizing bacteria (AOB) are chemolithotrophic. Although the size of the AOB community is dependent on NH þ 4 , the ecophysiological preferences among AOB means that their activity can be affected by ecosystem change. Thus, agricultural practices, such as stubble retention, can potentially affect AOB communities and impact on N-cycling until a new community structure, adapted for the current environmental conditions, is stabilized.
The abundance of free-living N 2 -fixing bacteria approximately doubled following stubble retention but was not affected by N application. Large inputs of C (e.g. maize stubble) to soil significantly increase the abundance of freeliving N 2 -fixing bacteria (Hegazi et al., 1986). However, it is also generally accepted that application of high rates of N to soils reduces dependence of the ecosystem on free-living N 2fixing organisms. This discrepancy may be explained by considering the timing of stubble addition to the soil and the soil sampling for this study. Maize stubble was incorporated into the soil at the end of the growing season by which time some of the N applied over the growing season had been removed by either biotic or abiotic processes. Thus, the decomposition of stubble during the winter fallow period may still be enhanced through the activity of N 2 -fixing bacteria. Although the contribution of free-living N 2 -fixing bacteria to N-uptake of maize in high-input agricultural sites is likely to be very low, it is clear that these bacteria have ecological importance other than provision of N for plant use. In particular, the activity of these bacteria over the winter fallow period may aid in the decomposition of organic waste with high C : N ratio.
The results of the real-time PCR allowed quantification of functional gene presence under the different soil treatments; absolute quantification of copy number of these genes was not readily achievable in this study. A major problem is that different species of bacteria can hold varying copy number of each functional gene. Nevertheless, the data clearly show that modification of both stubble and N management practices has a significant impact on the overall quantity of bacterial genes involved in key steps of the N cycle.
Establishing links between bacterial or fungal diversity and ecosystem processes in soil is difficult. Studies reporting strong associations between these factors are rare and typically focus on highly stressed or disturbed ecosystems (e.g. Girvan et al., 2005). Underpinning these studies are ecological hypotheses suggesting that due to distribution of functionality traits across broad phylogenetic groups of soil microorganisms, associations between soil microbial diversity and ecosystem processes (function) are tenuous. However, spatial and temporal environmental heterogeneity in soil may require that a multitude of species with physiological variability and habitat preferences occur together for long-term ecosystem stability. This is important for the maintenance of strong ecological resistance to single or multiple stresses (whether natural or anthropogenic).
This work has shown that long-term management of maize stubble and N had significant post-seasonal impacts on the size and structure of bacterial and fungal communities in soils. Such sustained shifts in soil microbial communities can impact on ecosystem function, as shown by overall changes in N-based functional genes, and also on the productivity of a subsequent crop. Future work will focus on the relationship between the structure of ecologically significant functional communities and ecosystem processes at this trial site.