OUP user menu

Molecular analysis of bacterial community succession during prolonged compost curing

Michael Danon, Ingrid H. Franke-Whittle, Heribert Insam, Yona Chen, Yitzhak Hadar
DOI: http://dx.doi.org/10.1111/j.1574-6941.2008.00506.x 133-144 First published online: 1 July 2008

Abstract

The compost environment consists of complex organic materials that form a habitat for a rich and diverse microbial community. The aim of this research was to study the dynamics of microbial communities during the compost-curing phase. Three different methods based on 16S rRNA gene sequence were applied to monitor changes in the microbial communities: (1) denaturing gradient gel electrophoresis of PCR-generated rRNA gene fragments; (2) partial rRNA gene clone libraries; and (3) a microarray of oligonucleotide probes targeting rRNA gene sequences. All three methods indicated distinctive community shifts during curing and the dominant species prevailing during the different curing stages were identified. We found a successional transition of different bacterial phylogenetic groups during compost curing. The Proteobacteria were the most abundant phylum in all cases. The Bacteroidetes and the Gammaproteobacteria were ubiquitous. During the midcuring stage, Actinobacteria were dominant. Different members of nitrifying bacteria and cellulose and macromolecule-degrading bacteria were found throughout the curing process. In contrast, pathogens were not detected. In the cured compost, bacterial population shifts were still observed after the compost organic matter and other biochemical properties had seemingly stabilized.

Keywords
  • biosolids compost
  • clone library
  • community composition
  • oligonucleotide microarray
  • organic matter degradation
  • PCR-DGGE

Introduction

Composting is an aerobic process by which organic materials are degraded through the activities of successive groups of microorganisms. Soil amendment with composted organic material is an ancient practice that is applied throughout the world, and the long-term benefits of compost application to fields are well documented (Ros., 2006). Previous studies have emphasized the importance of achieving compost maturity to ensure balanced plant nutrition and for the biological control of soil-borne plant disease (Fuchs, 2002; Noble & Coventry, 2005). Compost maturity is achieved during the curing process. The duration of curing in the industry varies according to a number of factors, including source materials, composting process and facility, climate, and planned utilization of the final product. Noncured composts may be phytotoxic, whereas extensively cured composts may lose their plant-disease-suppressive properties. For instance, loss of compost suppressiveness toward Sclerotium rolfsii has been demonstrated during prolonged curing time (Danon., 2007). Compost of an appropriate age should therefore be used to control disease in infested soils before sowing. The successful application of compost is considerably dependent on the selection of an appropriate curing period.

The biochemistry of the compost-curing process has been studied extensively. Chen. (1989) found that the original cellulose and hemicellulose contents of cattle manure were reduced by one-third during a 5-month composting and maturing process. Hemicellulose and cellulose may be the main substrates for microorganisms during the maturation process because these components are present in large quantities in cattle manure and the less complex carbon sources are consumed early on in the process (Chen., 1989). Recently, Tang. (2006) reported that organic matter decomposition during the maturation of cattle-manure compost resulted in decreased C/N ratios, microbial biomass, and microbial diversity. The C/N ratio in the compost may level off much before the compost stabilizes (Chefetz., 1996) and although the concentration of the dissolved organic carbon (DOC) of municipal solid-waste compost did not change at maturation, the relative concentrations of different hydrophilic and hydrophobic fractions in the dissolved organic matter continued to change concomitant with the decrease in phytotoxicity (Chefetz., 1998). Tang. (2006) showed that during the maturation period, the proportion of Actinobacteria increases slightly, and this was associated with the disappearance of phytotoxicity.

Considerable effort and a variety of techniques have been applied to the study of compost microbial populations (Cahyani., 2003; Schloss., 2003; Ryckeboer., 2003b). The initial phase of composting is thought to be the most dynamic part of the process and is characterized by rapid increases in temperature, large changes in pH, and the degradation of simple organic compounds. Schloss. (2003) reported two significant shifts in the composition of the microbial community: one occurring between 12 and 24 h and the other between 60 and 72 h into the process. Ryckeboer. (2003b) attempted to determine the microbial succession of the dominating taxa and functional groups of microorganisms, as well as the total microbial activity during composting of biowaste, using incubation, isolation, and enumeration techniques. They reported that bacteria dominated during the thermophilic phase while fungi, Streptomycetes, and yeasts were below detection limits. Different bacterial populations were found in the thermophilic and mesophilic composting phases. During the peak-heating phase of fresh wastes, the only bacteria isolated were bacilli; however, during the cooling and maturation phases, the bacterial diversity of both Gram-positive and Gram-negative bacteria increased (Ryckeboer., 2003b). Using cultivation-independent methods, Cahyani. (2003) studied the bacterial communities in composting of rice straw. They reported a successional transition of bacterial community members: Alphaproteobacteria in the raw materials, Bacillus and Actinomycetes at the thermophilic stage, and Cytophaga and clostridial members at the middle and curing stages. These authors reported that the microbial community remained stable during the curing phase. In contrast, Steger. (2007) revealed compositional changes within the Actinobacteria community in a full-scale composting process of organic household waste over a period of 57 weeks.

Despite the important role of microbial populations in compost quality, only a few studies have thoroughly examined the dynamics of the total microbial populations during compost curing. The aim of this study was to characterize changes in the microbial community structure during prolonged curing of biosolids compost, using molecular techniques. For this purpose, we used three different cultivation-independent techniques based on 16S rRNA gene sequences: PCR-denaturing gradient gel electrophoresis (DGGE), clone libraries, and an oligonucleotide microarray (COMPOCHIP). Molecular techniques are becoming increasingly useful for the detection of different microorganisms without first having to isolate and cultivate them (Dees & Ghiorse, 2001; Green., 2004; Franke-Whittle., 2005; Kelly., 2005). Although they share the same basis, i.e. exploiting conserved and variable regions in the ribosomal small subunit sequence, each of the techniques used in our study has its inherent advantages and disadvantages. A combined analysis was therefore expected to produce more reliable information on the qualitative and quantitative succession of bacterial populations during compost curing.

Materials and methods

Prolonged compost-curing process and compost sampling

Compost samples were obtained from a commercial composting facility (Shacham, Givaat Ada, Dlila Facility, Israel) that prepared compost from a mixture of sewage sludge and yard waste (1 : 1, v/v) in windrows (2.5 m wide and 2 m high). Aeration was achieved by bi-weekly turning for the entire 12-week composting process. The biosolids compost, considered to be mature by the producers, i.e. appropriate for field application, was sampled and further treated as follows: c. 700 L of bulk compost was collected and placed in a cubic, 700-L bin, rewetted and turned every 2 weeks for the first month, then every month for the next 4 months. The resultant compost pile was then left unturned for seven additional months (in total, 1 year of prolonged curing). Compost temperature in the container increased after rewetting to 42 °C in the first week, and then decreased over the next 2 weeks to 32 °C (c. 5 °C above ambient). Approximately 20 L of compost was sampled from the bin at time 0 (beginning of prolonged curing), then after each turning, and stored at 4 °C until analysis.

DNA extraction, general molecular procedures, and replications

DNA from 0.25 g of compost subsamples was extracted using a ‘PowerSoil’ kit (Mobio, Solana Beach, CA). The DNA extracts were used as templates for PCR amplification of bacterial partial 16S rRNA genes for DGGE, clone libraries, and oligonucleotide microarray analysis. For DGGE, we used two replicate DNA samples. The DNA was amplified using the 341-907 primer set with a GC clamp (Muyzer., 1998) and the products were run in separate lanes. For clone libraries we used three replicate DNA samples. The DNA was amplified using the 341-907 primer set without a GC clamp. The PCR products were pooled and mixed before ligation to a vector.

PCR-DGGE analysis

PCR amplification and DGGE were conducted as described previously (Danon., 2007). Briefly, DGGE was prepared according to Muyzer. (1993) using an IngenyPhor-U2 system (Ingeny, Goes, The Netherlands) with 6% (w/v) polyacrylamide gel [acrylamide/bisacrylamide (37 : 1)] in 1 × Tris acetate–EDTA (TAE) buffer and a 20–60% denaturing gradient (80% denaturant corresponding to 7 M urea and 32%, v/v formamide). Dominant DGGE bands were excised and reamplified. Reamplified bands that migrated identically to the excised bands in the DGGE were cloned and sequenced.

Cloning, sequencing, and phylogenetic analysis

Two clone libraries, for noncured (0 days) and cured (336 days) compost samples, were constructed using the pGEM-T easy vector system (Promega, Madison, WI). Ligation and transformation were performed according to the manufacturer's directions. Colonies were screened for the presence of the correctly sized insert, and plasmid DNA was reamplified before sequencing. Plasmids from cloned PCR-DGGE bands were extracted and purified using a miniprep DNA purification kit (Genomed, Löhne, Germany).

Sequencing was performed at the Macrogen Inc. Sequencing Center (Seoul, Korea). Sequences were examined using the check_chimera program located at the Ribosomal Database Project (Cole., 2005), and chimeric sequences were removed from phylogenetic analyses.

Phylogenetic analysis of sequences derived from clones and from DNA of excised bands was performed by alignment to known bacterial sequences using the ‘greengenes’ 16S rRNA gene database and alignment tool (DeSantis., 2003) (http://greengenes.lbl.gov/). Aligned sequences and close relatives were imported to the mega software package version 3.1 (Kumar., 2004). Similarity was tested to sequences available at the National Center for Biotechnology Information (NCBI) using blast analysis (Altschul., 1997). The phylogenetic tree was constructed using the mega software with the Kimura two-parameter method for distance matrix calculations and the neighbor-joining method for tree design. Tree topologies were evaluated by performing bootstrap analysis of 1000 data sets. The rRNA gene sequences were submitted to the GenBank database under accession numbers EU215227EU215310.

Oligonucleotide microarray analysis

The COMPOCHIP microarray, spotted with 369 probes targeting microorganisms that have been reported previously in the composting process, as well as plant, animal, and human pathogens, and plant-disease-suppressive bacteria, was applied to DNA extracted from different composts. The specificity of all probes was assessed in silico, using the ARB program (Ludwig, 2004), and the array was tested with pure cultures of microorganisms, and was shown to work well with only a low percentage of nonspecific hybridizations (Franke-Whittle et al., 2005). For most target organisms, at least three probes were spotted on the slide, and for a few organisms there were only two probes. All the probes included on the COMPOCHIP microarray were designed so as to have similar melting temperatures, and probe sequences ranged in length from 17 to 25 nucleotides. Five or six replicate DNA samples were analyzed. Fluorescence labeling of target DNA, hybridization, scanning of arrays, and image analysis were conducted as described by Franke-Whittle et al. (2005).

Statistical analysis

Statistical analysis of the DGGE data was conducted using Dice correlation coefficients and the unweighted-pair-group method with arithmetic averages (UPGMA) to form a complete linkage dendrogram (Fingerprinting II Informatix, BioRad Laboratories, Hercules, CA).

Statistical analysis of clone libraries and coverage determination were conducted using the procedure developed by Kemp & Aller (2004) to confirm that an asymptotic accumulation curve in both libraries had been reached. Unifrac (Lozupone., 2006) was used to define phylotypes at 3% similarity cutoff.

Statistical analyses of microarray data were performed using the program canoco 4.5 (ter Braak & Šmilauer, 2002). Data from the microarray analysis, the physicochemical analysis, and the cloning of DGGE bands were subjected to principal component analysis (PCA). As differing numbers of replicates were used for each set of experiments, an average was used for each set of replicates in the statistical analysis. Because a strong correlation between SD and means of replicate samples was found, a log-transformation of microarray data was conducted to equalize variances. For the covariance-based redundancy analysis (RDA), the following settings were used: inter-sample distance scaling, no post-transformation of scores, log data transformation (no offset), and center by species.

Results

PCR-DGGE

The community composition of bacteria in biosolids compost samples subjected to different curing times is shown in Fig. 1. UPGMA analysis of the DGGE profiles generated from individual PCR products showed that the duplicate samples of specific ages clustered together. The noncured (0 days) samples were distinctly different from those representing longer curing times. Among the cured compost samples, the 19- and 41-day samples clustered separately from the longer curing times. The oldest sample (336 days) was distinctly different from the 67- to 205-day samples. The bands were numbered (1–44) and identified according to similarity to sequences in the GenBank and the greengenes database. The identifications of DGGE bands are listed in Table 1.

1

PCR-DGGE analysis of the bacterial community in composts from different stages of curing. A UPGMA algorithm was applied to a similarity matrix of Dice and Pearson correlation coefficients generated from the DGGE banding patterns. The numbers correspond to bands identified by 16S rRNA gene sequence analysis (Table 1).

View this table:
1

Identification of PCR-DGGE bands by 16S rRNA gene sequence analysis

Band numberAccession numberPhylum/classFamily
8EU215311BacteroidetesCryomorphaceae
10EU215312BacteroidetesSphingobacteriaceae
13EU215313AlphaproteobacteriaPhyllobacteriaceae
15EU215314AlphaproteobacteriaCaulobacteraceae
19EU215315GammaproteobacteriaXanthomonadaceae
21EU215316GammaproteobacteriaXanthomonadaceae
23EU215317GammaproteobacteriaXanthomonadaceae
26EU215318GammaproteobacteriaXanthomonadaceae
29EU215319BacteroidetesFlavobacteriaceae
30EU215320ActinobacteriaCorynebacterineae
37EU215321AlphaproteobacteriaCaulobacteraceae
38EU215322ActinobacteriaPromicromonosporaceae
  • * Dominant DGGE bands were excised, reamplified and cloned. Amplicons and clones were compared with the original bands in DGGE. Amplicons and clones that did not migrate to the identical positions of the excised bands in the DGGE were not identified.

Clone libraries

Clone libraries were constructed from cured (0 days) and noncured (336 days) composts, each library including a total of 41 and 43 clones, respectively. The complete phylogenetic analysis is given in the supplementary information. Figure 2 illustrates the phylogeny of compost bacteria members affiliated with (a) Actinobacteria and (b) Betaproteobacteria. The clone libraries were found to differ significantly with respect to phylotype composition (P value ≤0.01). Clone rRNA gene sequences were found to belong to seven different phyla: Actinobacteria, Bacteroidetes, Chloroflexi, Deinococci, Gemmatimonadetes, Firmicutes, and Proteobacteria. Most of the clones, derived from both the cured and noncured composts, were classified as belonging to the Alpha-, Beta-, Gamma-, and Deltaproteobacteria classes. The Gammaproteobacteria class contained the highest number of clones: five from the noncured compost and 12 from the cured samples. Eight clones from the noncured compost and five clones from the cured compost were found to belong to Alphaproteobacteria. Only clones from the noncured compost grouped within the Caulobacteraceae family. Other groups found only in the noncured compost were Promicromonosporaceae (Actinobacteria) (Fig. 2a), Comamonadaceae (Betaproteobacteria) (Fig. 2b) and members of the Chloroflexi phylum. Members of the Deinococcus phylum were present only in the cured compost clone library (Supplementary Fig. S1).

2

16S rRNA gene phylogeny of bacteria found in composts at different curing stages (a) Actinobacteria and (b) Betaproteobacteria phyla. Neighbor-joining phylogenetic tree drawn to scale, with branch lengths computed using the maximum composite likelihood method. The scale bar represents 0.01 base substitutions per site. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) is shown next to the branches. Sequences of PCR-DGGE bands are marked by triangles. Sequences from noncured and cured compost are marked by black and white circles, respectively.

Microarray

The COMPOCHIP 16S rRNA gene microarray was applied in order to directly determine which bacteria were present in the different compost samples analyzed. Because a linear correlation between target concentrations and signal intensities of the various probes has been reported by others (Taroncher-Oldenburg., 2003; Tiquia., 2004), the results obtained in this study were analyzed semi-quantitatively.

Figure 3 shows an ordination graph: the two relevant axes explained 34.1% of the variance. Multivariate analysis showed that the bacterial-community compositions of the composts changed with curing time, namely, the noncured and cured composts of different ages clustered separately. Bacteria belonging to the low G+C and Alphaproteobacteria groups were found in high numbers in composts from all curing stages, as well as noncured compost. The probe targeting the genus Actinomyces also revealed high levels of these organisms in both cured and noncured compost samples. The microarray included probes specific for members of the genus Chryseobacterium. Higher levels of this genus were found in the younger composts than in the more mature ones. Higher signals for the probes targeting Microbacterium and Sphingobacterium were detected in the noncured composts. The EUB 338II probe specific for the Verrucomicrobia group was found to correlate well with the more cured composts. The Pseudomonas genus-specific probes gave signals above the threshold values for samples from days 0, 41, and 336, but not for those from days 67 and 130.

3

Two dimensional ordination plot of compost communities from different curing stages analyzed by the COMPOCHIP microarray and canonical analysis (redundancy analysis). Individual replicates are represented by letters (a–f).

High signal-to-noise ratios were obtained for composts from all curing stages with probe 270 specific for Alcaligenes faecalis/defragrans. Although the signal intensities were found to decrease with time, Alcaligenes was still present in the 336-day samples. These results were supported by strong hybridization signals for probe 350, specific for A. faecalis. Despite its name, A. faecalis is not typical to feces, but is a common, nonpathogenic, environmental bacterium that has been reported to be present in composts by other authors (Droffner., 1995; Ryckeboer., 2003a).

The 295 and 296 probes, specific for members of the genera Nitrosovibrio and Nitrosospira, indicated the presence of these bacteria in composts throughout the curing process. However, in the noncured samples, the probes exhibited higher signal levels.

Discussion

The microbial community in compost during prolonged curing was studied using three different culture-independent techniques. We followed the dynamics of the microbial population from the point at which the commercial producer considered the compost appropriately mature for field application and at which point it was found to be suppressive to plant pathogens (Danon., 2007). The results obtained from analysis using DGGE, clone libraries, and a microarray revealed that under the conditions tested here, the compost community continued to change during the curing phase. Recognizing the inherent advantages and disadvantages of each technique, we attempted to integrate the data analysis to produce a comprehensive picture of the microbial community dynamics. Microarrays allow the parallel detection of up to several thousand microbial strains, species, genera, or higher taxonomic groups in a single experiment, depending on the availability of the probe set used on the microarray. However, because different probes have different affinities for their targets (Loy & Bodrossy, 2006), the information on the relative abundance of the different microorganisms derived from the microarray analysis needs to be interpreted with caution. We found higher signals (indicative of better hybridization) for most probes (both higher taxonomic group level probes as well as genus- and species-specific probes) upon hybridization with the noncured compost samples when compared with the cured compost samples. This would indicate that the longer the curing time, the less likely it is that the resulting compost product will have a bacterial community that matches the expected typical compost bacteria targeted by the microarray. As a result of this, clone libraries were more reliable for the comparison of higher taxonomic groups comprising the cured compost microbial community (Fig. 4).

4

Percentage contribution of various bacterial phyla to the total bacterial community in cured and noncured compost, as detected by clones in libraries vs. microarray probes. Groups below the diagonal line were more abundant in the noncured compost, and those above it were more abundant in the cured compost.

In a previous report (Danon., 2007), we have described how biosolids compost undergoes a transition from a highly active to a stabilized biosolid during prolonged curing. Samples of noncured compost were characterized by intense activity, with rapid decomposition of organic matter, as reflected by a higher respiration rate, a faster utilization of glucose, and a higher concentration of DOC. The solubilized material became depleted in carbohydrates and enriched in aromatic compounds. As the available carbon was depleted, the microbial activity, indicated by respiration, decreased, the response to added glucose was slower, and competition for NH4+ decreased, allowing an increase in nitrification. It was proposed that as curing progresses, the remaining substrate in the compost is more lignocellulosic in nature (Danon., 2007). The biochemical data of the previous report were used for comparison of the same samples with microbial population analysis conducted in the current study (Fig. 5). Some of the bacteria associated with the changes in the chemical properties of compost during prolonged curing were then identified.

5

Principal component analysis of chemical and molecular data. Circles indicate the clustering time of curing in days. The principal DGGE bands and microarray probes are highlighted (in red and blue, respectively).

Noncured compost populations

Differences between the noncured and cured composts in relative abundances of bacterial groups are illustrated in Fig. 4. Using both the cloning and the microarray approaches, we found that the members of the Bacteroidetes phylum were more abundant in the noncured composts than in the cured composts (Fig. 4). With a more detailed principal component analysis including DGGE and microarray data as well as process parameters (NH4+, NO3, DOC, and sugar concentration), we found that the dominant population in the noncured compost consisted of a variety of Bacteroidetes species, including Chryseobacterium, Sphingobacterium, and Flavobacterium (Fig. 5). We found members of Sphingobacterium, both by DGGE (band 10) and with the microarray (probe 550) in the noncured compost.

The phylum Bacteroidetes includes a wide variety of bacteria known for their utilization of macromolecules such as proteins, starch, cellulose, and chitin (Manz., 1996), and its members have been detected previously by molecular methods in various composts (Alfreider., 2002; Michel., 2002; Verkhovtseva., 2002). Green. (2004) reported that the most frequently detected sequences in compost-amended potting mixes belong to Chryseobacteria.

Nitrosovibrio and Nitrosospira were predominantly found in the noncured composts in close proximity to the NH4 vector (Fig. 5). These bacteria were detected by the microarray only, indicating that they represent a minor population in the compost total microbial community (Fig. 4). Nevertheless, minor populations may play important roles in their environment, and ammonium-oxidizing bacteria, including members of the above genera, are often reported in composts (Kowalchuk., 1999; Innerebner., 2006).

We found bacteria belonging to the Betaproteobacteria to be more abundant in noncured than cured composts (Fig. 4). Members of the Comamonadaceae family were detected only in the noncured compost clone library (Fig. 2b). These bacteria may originate from the activated sludge process during wastewater treatment, as they were related to Acidovorax sequences isolated from wastewater-treatment plants. Interestingly, Valle. (2004) reported that in a phenol-degrading activated sludge system, the population size of Acidovorax sp. AHL-5 (AY379977.1) changes in association with the phenol-degradation rate. Acidovorax species have also been shown to be the dominant 3-hydroxybutyrate (PHB)-degrading bacteria in soils, composts, and freshwater (Mergaert & Swings, 1996). Other Betaproteobacteria, namely members of the Oxalobacteraceae, were found in both libraries (Fig. 2b).

We found bacteria of the phylum Chloroflexi only in the noncured compost clone library (Fig. S1). Filamentous members of the phylum Chloroflexi have also been found in activated sludge, and they have occasionally been associated with incidences of bulking (Bjornsson., 2002). As the compost used in this study was derived from sewage sludge, the detection of these bacteria is not surprising. Chloroflexi filaments appear to be specialized in polysaccharide degradation.

Members of the genus Brevundimonas (Alphaproteobacteria; Caulobacteraceae) were present and dominant only in the noncured compost clone library (Fig. S1). However, PCA of the DGGE bands (Fig. 5) showed that band 15, identified as Caulobacteraceae, appears to be a dominant microorganism in the compost sample with a curing time of 41 days. Pedro. (2001) found Brevundimonas present in the mesophilic phase of industrial and agricultural waste composting. These bacteria may play an important role in the composting process, and it is possible that they originated from the biosolids, as Caulobacteria members have been isolated previously from sewage and activated sludge (Baker., 1983).

Populations at intermediate compost-curing times

PCA of microarray data showed that the Actinobacterium Microbacterium dominated composts with intermediate curing times of 41–130 days (Fig. 5). Manickam. (2006) reported the ability of Microbacterium isolated from contaminated soil to degrade the persistent and toxic hexachlorocyclohexane. Members of the Actinomycetes genus have been reported to develop more slowly than most other microorganisms and are comparatively ineffective competitors under high-nutrient conditions. The proportion of Actinomycetes relative to other bacteria has been suggested as an indicator of compost maturity (Palmisano & Barlaz, 1996). Thus, it is not surprising that microarray analysis revealed high levels of Actinomycetes in the composts from longer curing times.

DGGE analysis showed the presence of Promicromonosporaceae members between 130 and 205 days of curing. We also found Promicromonosporaceae in the noncured composts clone library (Fig. 2a). Promicromonospora sukumoe and related organisms of the Cellulomonas group have been characterized as cellulose-degrading Actinobacteria (Bakalidou., 2002). Cellulomonas has been reported previously in mature-compost bacterial populations (Ryckeboer., 2003b). Cellulolytic bacteria are common in compost (Herrmann & Shann, 1997) and are known to occur mainly at the end of the thermophilic stage (Gray., 1971). At the end of the composting process, however, the cellulose is often inaccessible to enzymatic attack because of low water content or association with protective substances such as lignin (Stutzenberger., 1970). This results in a decrease in the number of cellulolytic organisms. It has also been reported that in materials with high cellulose content, thermotolerant microorganisms with the ability to degrade cellulose can dominate the end stages of the composting process (Ryckeboer., 2003b). Actinobacteria may survive the thermophilic phase of composting and become active during compost maturation. The decomposition of cellulose by cellulolytic bacteria such as Cellulomonas occurs during the maturing process (Tang., 2006).

Cured compost populations

Gammaproteobacteria were also abundant in the cured compost, and were detected in the clone library as well as by microarray (Fig. 4). It is possible that different species of Pseudomonas were responsible for hybridization with the genus-level probe, and that certain species were present in higher numbers at the start of the curing process, while other species were present later in the curing process. Pseudomonas species are widely distributed in the environment and have been reported previously in mature-compost bacterial communities (Ryckeboer., 2003a). Pseudomonas strains are known to be able to participate in N2 fixation, denitrification, and degradation of pollutants or interaction with toxic metals (Lalucat., 2006), and several strains are known to confer plant-disease suppressiveness (Haas & Défago, 2005). We found sequences affiliated to the Gammaproteobacteria, Xanthomonadaceae families both in the cured compost clone library and among the DGGE bands of cured compost samples (Table 1).

The dominant Actinobacteria found in the cured compost were Corynebacterineae (i.e. Mycobacterium species), as detected in the clone library and by DGGE. Many of the Mycobacteria are potential pathogens. Mycobacterium avium, an environmental opportunistic pathogen, has been isolated from environmental water samples. Furthermore, growth of M. avium in natural water was stimulated by the addition of humic and fulvic acids in acidic, microaerobic environments (Kirschner., 1999).

We found only a few Firmicutes sequences in the compost clone libraries; we did not detect Firmicutes in DGGE and their detection rate in the microarray was low. This finding is surprising considering the dominance previously attributed to this phylum in compost. Ryckeboer. (2003b) reported Bacillus as the most dominant bacterial taxon recovered from compost feedstock, and it was also the most abundant group of bacteria recovered from compost during the thermophilic phase and throughout the entire composting process. Dees & Ghiorse (2001) reported that the two most abundant cultivated isolates from hot synthetic compost belong to the genera Aneurinibacillus and Brevibacillus. We suggest that Firmicutes members do not always persist in compost after the thermophilic phase has ended.

It is apparent that members of different bacterial species capable of similar functions were found throughout most of the process. For example, at the beginning of the curing process, the NH4 concentration was high (Fig. 5) and supported a population specializing in nitrification (i.e. Nitrosovibrio, Nitrosospira); thus, NH4 concentrations decreased rapidly and NO3 increased until 41 days. From 41 to 130 days, NO3 levels continued to increase, suggesting that nitrification activity was maintained by more versatile bacteria (Danon., 2007). Biodegradation of macromolecules, especially of complex materials such as lignocellulose, is a much slower process. A population specializing in cellulolytic activity (i.e. Promicromonosporaceae) became established during the intermediate curing time (130–205 days). A retarded response to added glucose in the cured compost (Fig. 5) indicates a shift from microbial r-strategists to K-strategists. The r-strategists are adapted to intervals of rapid growth, depending on the availability of their substrate, as opposed to the K-strategists, which typically have slow growth rates, presumably benefiting from polymerized substrates that have a long residence time (Fontaine, 2003).

In conclusion, we found a successional transition of bacterial populations during compost curing. Comparison of the results obtained by the three methods was used for identification of different bacterial phylogenetic groups that changed during prolonged compost curing. The Proteobacteria were the most abundant phylum in all cases. The Bacteroidetes and the Gammaproteobacteria were ubiquitous but opposite in their relative dominance. At the beginning of the compost-curing process, Bacteroidetes were dominant. Later, during the midcuring stage, Actinobacteria were dominant. Finally, in the cured compost, the Gammaproteobacteria were more abundant. Despite using a microarray that targets many pathogens (Franke-Whittle., 2005; Franke-Whittle et al., 2005), we did not detect any human, animal, or plant pathogens by this method. This suggests efficient eradication of pathogens during the composting. Furthermore, this observation indicates that no reinfestation with pathogens occurred during the prolonged curing.

The initial phase of composting is thought to be the most dynamic part of the process (Schloss., 2003). Nevertheless, we found that the microbial community continued to change long after the compost organic matter and other biochemical properties had stabilized. It is assumed that bacterial populations, playing their specific role in changing the chemical environment, were replaced in turn by selection pressures driven by these changes in the microhabitat. Changes in complex lignocellulose-rich organic matter may have been too minute to detect using standard physicochemical analysis. However, bacterial community response to these changes was more easily observed.

Supplementary material

The following supplementary material is available for this article:

Fig. S1. 16S rRNA gene phylogeny of bacteria found in composts at different curing stages.

This material is available as part of the online article from: (This link will take you to the article abstract).

Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Supplementary material

The following supplementary material is available for this article:

Fig. S1. 16S rRNA gene phylogeny of bacteria found in composts at different curing stages.

This material is available as part of the online article from: http://www.blackwell-synergy.com/doi/abs/10.1111/j.1574-6941.2008.00506.x (This link will take you to the article abstract).

Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Acknowledgements

We gratefully acknowledge the assistance of Dana Levinson, and thank Sharon Nahum-Zmora for her significant role in carrying out the compost-curing process and providing the samples, the chemical data, and critical remarks. This research was supported by The Negev Foundation and the Israeli Ministry of Agriculture and Rural Development, the Wastewater Treatment Program.

Footnotes

  • Editor: Alfons Stams

References

View Abstract