OUP user menu

Microbial diversity in hot synthetic compost as revealed by PCR-amplified rRNA sequences from cultivated isolates and extracted DNA

Peter M. Dees, William C. Ghiorse
DOI: http://dx.doi.org/10.1111/j.1574-6941.2001.tb00805.x 207-216 First published online: 1 April 2001


High-temperature (≥60°C) synthetic food waste compost was examined by cultivation-dependent and -independent methods to determine predominant microbial populations. Fluorescent direct counts totaled 6.4 (±2.5)×1010 cells gdw−1 in a freeze-dried 74°C compost sample, while plate counts for thermophilic heterotrophic aerobes averaged 2.6 (±1.0)×108 CFU gdw−1. A pre-lysis cell fractionation method was developed to obtain community DNA and a suite of 16S and 18S rDNA-targeted PCR primers was used to examine the presence of Bacteria, Archaea and fungi. Bacterial 16S rDNA, including a domain-specific 1500-bp fragment and a 300-bp fragment specific for Actinobacteria, was amplified by PCR from all compost samples tested. Archaeal rDNA was not amplified in any sample. Fungal 18S rDNA was only amplified from a separate dairy manure compost that reached a peak temperature of 50°C. Amplified rDNA restriction analysis (ARDRA) was used to screen isolated thermophilic bacteria and a clone library of full-length rDNA fragments. ARDRA screening revealed 14 unique patterns among 63 isolates, with one pattern accounting for 31 of the isolates. In the clone library, 52 unique patterns were detected among 70 clones, indicating high diversity of uncultivated bacteria in hot compost. Phylogenetic analysis revealed that the two most abundant isolates belonged in the genera Aneurinibacillus and Brevibacillus, which are not commonly associated with hot compost. With the exception of one Lactobacillus-type sequence, the clone library contained only sequences that clustered within the genus Bacillus. None of the isolates or cloned sequences could be assigned to the group of obligate thermophilic Bacillus spp. represented by B. stearothermophilus, commonly believed to dominate high-temperature compost. Amplified partial fragments from Actinobacteria, spanning the V3 variable region (Neefs et al. (1990) Nucleic Acids Res. 18, 2237–2242), included sequences related to the genera Saccharomonospora, Gordonia, Rhodococcus and Corynebacterium, although none of these organisms were detected among the isolates or full-length cloned rDNA sequences. All of the thermophilic isolates and sequenced rDNA fragments examined in this study were from Gram-positive organisms.

  • Compost
  • Bacillus
  • Phylogeny
  • Thermophile
  • Ribosomal RNA
  • Actinomycete

1 Introduction

Composting is an aerobic, microbial decomposition process in which the production and subsequent insulation of excess metabolic heat in the composting mass can rapidly lead to sustained temperatures above 50°C, with core temperatures in insulated zones between 60°C and 80°C. A typical composting process goes through a series of stages, including rapid temperature increase, sustained high temperatures and gradual cooling of the composting mass. Each successive stage in composting is expected to be accompanied by specific populations of bacteria [1]. Our current knowledge of microbial community structure and bacterial diversity in hot composts (≥50°C) is based largely on isolation and plate count studies [25]. A traditional view of compost microbial community structure suggests that microbial diversity is greatly limited by the high temperatures and that the high-temperature compost community is dominated by low-G+C Gram-positive spore-forming bacteria like Bacillus stearothermophilus[4,5]. The discrepancy between culturable bacterial counts and total direct counts, well known in environmental samples of all types, has also been documented for composted materials [6,7]. Accordingly, previous studies of hot compost community structure and diversity that relied solely on cultivation techniques have probably failed to detect and properly identify a significant portion of the hot compost microflora.

The analysis of compost community structure and dynamics with culture-independent approaches has reinforced this probability. Analyses of cell membrane phospholipids [811] and community metabolic capabilities [813] have revealed changes in community profiles indicative of population succession during the different stages of composting, but these methods only indirectly reflected changes in phylogenetic diversity. Molecular-genetic techniques have been used to directly examine compost community DNA and RNA for the presence of novel groups of microorganisms [1416]. Blanc et al. showed that several cloned 16S rDNA sequences from high-temperature (64–84°C) kitchen and garden waste composts were related to Thermus species that have only recently been reported in hot compost [1418]. Other rDNA sequences from that study closely resembled sequences from bacteria that have not been previously associated with hot compost, including Saccharococcus thermophilus and Rhodothermus marinus, isolated from high-temperature (70°C) sugar beet refining operations and marine hydrothermal vents, respectively [14]. Furthermore, Kowalchuk et al. showed that several different composted materials contained ammonia-oxidizer-like 16S rRNA and rDNA sequences from β-subgroup Proteobacteria, using denaturing gradient gel electrophoresis and a competitive PCR method [15]. Most recently, Peters et al. found that genetic profiles based on compost community rDNA showed population succession and increasing phylogenetic diversity during the composting of mushroom growth substrate [16]. Prominent rDNA fragments detected in that study confirmed the presence of low- and high-G+C Gram-positive bacteria and further indicated the presence of γ-subdivision Proteobacteria and yeast species related to the genus Candida.

The need for a combination of culture-based and cultivation-independent methods in basic studies of compost microbial ecology has been emphasized [1416]. Here, we report results comparing prominent thermophilic compost isolates with a library of amplified and cloned 16S rRNA genes from extracted compost community DNA. We also present a detailed phylogenetic analysis of 16S rRNA sequences from the isolates and clone libraries.

2 Materials and methods

2.1 Compost samples

The hot compost samples were obtained from 30-l synthetic food waste (SFW) composting reactors charged with non-sterile tap water, dog food (Big Red Puppy Food, Pro-Pet Inc., Syracuse, NY, USA) and maple wood chips (∼0.5×1.5 cm; Coastal Lumber, Cayuta, NY, USA). For each reactor, ∼4.3 kg dog food, 5.2 kg wood chips and 9.5 kg H2O were mixed to obtain a C/N ratio of 18 and an initial moisture content of 55% wet basis (mass of H2O/mass of wet solids) [19]. For microbiological analyses, 20–25-g grab samples were obtained from individual SFW reactors at ∼60°C and 74°C and immediately stored at −20°C, until samples were used for cell fractionation, cell lysis and DNA extraction (see below). One sample was obtained from a similar 30-l reactor containing dairy manure following a peak temperature of ∼50°C. This sample was chosen because it clearly contained fungi, based on microscopic analysis.

2.2 Enumeration of bacteria

Bacterial cells were enumerated by an acridine orange direct counting method [20] in which 100 mg of freeze-dried compost cell extract (described below) was diluted 100-fold in 0.1% sodium pyrophosphate, pH 7.0 (NaPP) [21], and shaken for 30 min at room temperature on an orbital shaker at ∼150 rpm. The mixture was then amended with enough 37% formaldehyde to bring the final concentration to 0.5% formaldehyde. The sample was further diluted 100-fold in sterile NaPP before molten Noble agar (1%) was added to a final concentration of 0.1% agar. 5-μl aliquots of the final mixture were spread on precleaned glass microscope slides over a circular area of 1.0 cm2 and dried in air. The dried smears were stained by adding several drops of 0.01% (w/v) acridine orange solution for 2 min, rinsed briefly with 1 M NaCl solution, followed by 20 ml of sterile distilled water. The rinsed slides were drained, blotted dry, mounted in 10 μl of sterile distilled water, sealed with vaspar (50% vaseline:50% paraffin) and examined under a 100× phase contrast oil immersion lens (numerical aperture 1.4) on a Zeiss Standard 18 epifluorescence microscope equipped with a 50-W HBO mercury arc lamp (Osram, Munich, Germany) and a Zeiss 09 filter.

Freeze-dried compost cell extracts were serially diluted in NaPP, plated on five different diluted growth media and incubated at 60°C. The growth media were R2A (Difco, Detroit, MI, USA), plate count agar (Difco), nutrient broth (Difco), trypticase soy agar (BBL, Cockeysville, MD, USA) and Biolog Universal Growth Medium (Biolog, Hayward, CA, USA). Each was diluted with distilled water to 1/10 the manufacturer's recommended concentration and amended with 20 g l−1 bacteriological agar (Difco).

2.3 Cell extraction, lysis and DNA purification

A modified cell extraction protocol was developed, based on the procedures described by Holben [22]. A 10-g sample of the frozen compost was added to 90 ml of 0.1% NaPP and shaken for 30 min at room temperature on an orbital shaker at ∼150 rpm. The mixture was transferred to a sterile Waring blender jar and homogenized using three 30-s bursts with a 30-s pause between blending steps. Wood chips and debris were left behind in the blender jar, while the remainder of the homogenized slurry was transferred to 35-ml centrifuge vials and centrifuged for 20 min in a Sorvall Model RC-5B using an SS-34 rotor at ∼12 000×g. The blender cup containing wood chips and debris was rinsed with an additional 35 ml of NaPP. The resulting slurry was transferred to a fresh 35-ml centrifuge vial and centrifuged for 20 min at 12 000×g. The pelleted material from the rinsate was combined with the pellet from the first step. Small (<5 mm) wood chip particles and dog food debris were intentionally left in the slurry in order to include organisms that might remain tightly attached to compost particle surfaces. The combined pellets from the centrifugation steps were frozen at −80°C and then freeze-dried overnight in a lyophilizer (Lyph-Lock 4.5, Labconco, Kansas City, MO, USA). The freeze-dried material was stored in sterile plastic scintillation vials at −20°C.

For cell lysis and DNA purification, 50 or 100 mg of the freeze-dried pellet containing the extracted cells was added to a 2.2-ml screw-top vial (Lab Product Sales, Rochester, NY, USA) containing ∼1.2 g of 0.1 mm diameter silica–zirconia beads (Biospec Products, Bartlesville, OK, USA). The vials received 300 μl each of: sodium phosphate buffer (100 mM NaH2PO4, pH 8.0); lysis buffer (100 mM NaCl, 500 mM Tris, pH 8.0, 10% sodium dodecyl sulfate); and chloroform–isoamyl alcohol (24:1) mixture and then were agitated for 5 min at maximum speed (∼2500 rpm) in a Mini-Beadbeater 8 (Biospec Products). The vials were centrifuged briefly at 13 000×g in a microcentrifuge to settle the contents and the aqueous supernatant containing the community DNA was transferred to a sterile 1.5-ml microcentrifuge vial. Some freeze-dried cell samples were lysed in the bead beater using the FastDNA SPIN Kit for Soil (Bio101, Vista, CA, USA), following the manufacturer's instructions. In both cases, the aqueous solutions containing crude DNA were further purified using the FastDNA SPIN Kit. Purified DNA was eluted in a final volume of 50 μl sterile deionized H2O, diluted 100-fold for PCR amplification and stored at −20°C.

2.4 PCR amplification, cloning and sequencing

A 10-μl aliquot of 100-fold diluted, purified DNA sample was incorporated in 25 μl or 50 μl PCR reaction mixture containing the following: 1×PCR buffer (10×buffer is 200 mM Tris–HCl, pH 8.4, 500 mM KCl), MgCl2, and primers. Primer and MgCl2 concentrations and annealing temperatures used with the individual primer pairs are listed in Table 1. The PCR protocol began with a 5-min denaturation step at 94°C followed by 35 cycles of denaturation at 94°C for 30 s, annealing at primer-specific temperature (Table 1) for 1 min and extension at 72°C for 1 min. The protocol was concluded with an additional 5 min of extension at 72°C. PCR products were electrophoretically separated and visualized in 1.5% agarose gels stained with ethidium bromide. PCR products obtained with the primer pairs 27f-1492r [23] and F243-R513 [24] were cloned into pCR2.1 (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. The amplified rDNA fragments from five individual reaction vials were pooled and an aliquot of that mixture was used in the ligation step. White transformant colonies were selected and the vector insert size was verified using the 27f-1492r primer pair. Clones with unique amplified ribosomal DNA restriction analysis (ARDRA) patterns (procedure described below) were grown overnight in 5 ml Luria broth plus kanamycin (50 μg ml−1). The resulting cells were pelleted and sent for plasmid isolation with a Qiagen Biorobot 9600 using Qiaprep Turbo columns. Purified plasmid DNA was used as template for automated DNA sequencing with a 96-lane ABI 377 sequencing instrument (Perkin-Elmer Applied Biosystems, Foster City, CA, USA) using Big Dye Terminator fluorescent nucleotides. Primers used for automated sequencing [23,25] are listed in Table 1. Individual sequence fragments were assembled and edited using the GeneMan program (Lasergene, DNASTAR, Madison, WI, USA).

View this table:

Oligonucleotide primers used for amplification and sequencing of ribosomal RNA genes

Primer namePrimer sequence (5′→3′)Primer concentration (μM)MgCl2 concentration (μM)Targeted groupAnnealing temperature (°C)Reference
927RCCSTTGTGGTGCTCCC0.24.0Archaea52S. Barns, personal communication
  • a M=C or A, Y=C or T, S=G or C, R=A or G, K=G or T, W=A or T, H=A or C or T. Lower case letters in the 27f and 1492r primers indicate linker arms that were attached to the primers for subcloning experiments. Alternative versions of the 27f and 1492r primers were synthesized without the linker arm and also used in this study.

  • b Primer R513 was used with and without a GC clamp: CGCCCGGGGCGCGCCCCGGGCGGGGCGGGGGCACGGGGGG.

  • c na=not applicable.

  • d Primer 787R is the reverse complement of primer P3 Mod. [25]

2.5 ARDRA screening of 16S rDNA fragments

ARDRA [26] patterns for cultivated bacteria and cloned rDNA fragments were determined by picking cells from individual colonies with a platinum needle and transferring the cells into 10 μl of sterile deionized water in 0.2-ml thin-walled PCR vials. Cells were lysed during 5 min of incubation at 94°C. The cell lysates were then directly incorporated in 25 μl PCR mixture and amplified as described above. Cloned fragments of 16S rDNA were re-amplified with primer pair 27f-1492r (∼1500 bp) specific for Bacteria, and with primer pair F243-R513 (∼300 bp) specific for Actinobacteria. Amplification product size was verified in 1.5% agarose stained with ethidium bromide before the amplicons were digested overnight in a double restriction enzyme digestion with the tetrameric restriction enzymes HaeIII and HhaI [27]. Double restriction digests consisted of 20-μl reactions containing 12 μl PCR product, 1×buffer (10 mM Tris–HCl, 10 mM MgCl2, 50 mM NaCl, 1 mM dithiothreitol, pH 7.9) and 2 U of each restriction enzyme, incubated overnight at 37°C. Restriction digestion fragments were separated by gel electrophoresis for 1.5 h at ∼4 V cm−1 using 3% MetaPhor agarose (FMC BioProducts, Rockland, ME, USA). Digital images of the ethidium bromide-stained gels were recorded and compared using the Kodak Digital Science 1D image analysis system (Kodak, Rochester, NY, USA). Band sizes in the ARDRA patterns were estimated by using a 100-bp ladder (Gibco BRL, Gaithersburg, MD, USA) as a size standard. Computer-assisted ARDRA analysis [27] was performed with the Webcutter software (http://www.firstmarket.com/cutter/cut2.html).

2.6 Phylogenetic analysis

The computational tools of the Ribosomal Database Project (http://www.cme.msu.edu/RDP/html/index.html) were used to calculate the similarity (Sab) values for individual rDNA sequences using the SEQUENCE_MATCH program [28]. A BLAST search (http://www.ncbi.nlm.nih.gov.blast) was used to identify additional related sequences and the closest relatives determined in both searches were included in further phylogenetic analyses. 16S rDNA sequences were manually aligned using the ARB_EDIT tool in the ARB software package (http://www.mikro.biologie.tu-muenchen.de) [29] and phylogenetic trees were estimated using software incorporated in the ARB package. Distance matrices were estimated using the Kimura correction and trees were constructed using the neighbor-joining algorithm. Maximum-likelihood trees were constructed using the fastDNAml program as incorporated in the ARB software package. Homologous nucleotides were included in the calculations by using the lgcw_rr5_nov98 filter for Bacillus spp. and the hgc_rr5_nov98 filter for Actinobacteria species. The CHECK_CHIMERA program [28] was used to identify potential chimeric sequences among the environmental clones.

2.7 Nucleotide sequence accession numbers

The 22 16S rDNA sequences described in this study have been deposited in the GenBank sequence database under accession numbers AF252308AF252329.

3 Results

3.1 Enumeration and isolation of compost bacteria

Acridine orange direct counts of freeze-dried high-temperature (74°C) SFW compost samples indicated the presence of 6.4 (±2.5)×1010 cells gdw−1. There was no significant variation in the numbers of CFU that grew on the five different growth media (n=3, results not shown). Plate counts incubated at 60°C from the freeze-dried 74°C compost samples averaged 2.6 (±1.0)×108 CFU gdw−1 for the five different growth media. This indicated an ∼250-fold difference between the total direct count and the culturable heterotroph counts. The freeze-dried compost cell extract material used in the direct count and plate count studies was the same as that used in subsequent DNA extraction methods. Several attempts were made to cultivate bacteria on the five diluted heterotrophic growth media incubated at 72°C. All were unsuccessful.

3.2 Domain-level screening of hot compost rDNA

Using a domain-level phylogenetic approach [30], we first targeted members of the domain Bacteria, which were detected in all of the compost community DNA samples, based on positive PCR amplifications using the primer pair 27f-1492r (Table 1). Actinomycete-specific PCR primers amplified rDNA sequences in all of the compost samples tested, including the hottest samples collected at 74°C. In contrast to the domain Bacteria, the S3EUK-UF1 primer pair specific for fungal rDNA did not amplify the expected product from the 60°C and 74°C samples, although fungal rDNA was amplified in 50°C samples known to contain fungi (Table 2). Archaeal rDNA fragments were not amplified in any of the compost samples tested using primer pair 23FPL and 927R.

View this table:

Amplification of ribosomal DNA in hot compost from Bacteria, actinomycetes, fungi and Archaea using the PCR primers listed in Table 1

Primer setSpecificityCompost temperature
50°C dairy manure60°C SFW74°C SFW

3.3 Compost isolates

Prominent thermophilic compost isolates from the highest dilutions in the plate count experiments were examined by ARDRA screening of 16S rRNA genes in order to determine whether the assemblage of culturable thermophilic isolates was dominated by a particular organism or group of organisms. Thermophilic bacteria isolated from SFW compost samples yielded 14 unique ARDRA patterns among 63 isolates tested (Fig. 1A). One isolate ARDRA pattern clearly dominated the samples, accounting for 31 of the 63 colonies tested. The remainder of the ARDRA patterns from isolated bacteria were present at low frequency, i.e., 10% or less of the total number of isolates examined. An effort was made to survey all colonies present at the highest dilutions in the plate count experiments. Enumeration plates at the lower dilutions were frequently overgrown by fast-spreading colonies of Bacillus strains, making it difficult to survey all colonies on those plates. Virtually all of the ARDRA patterns derived from thermophilic compost isolates could be related to one another based on a pattern of common ARDRA bands indicative of Bacillus species (discussed in Section 3.4), with minor shifts of the bands allowing for discrimination between unique patterns (results not shown). The two most common ARDRA types observed on high-dilution enumeration plates (designated as HC isolates) were selected for 16S sequencing and phylogenetic analysis.


Frequency distribution of ARDRA patterns from isolated bacteria (A) and extracted DNA (B). Each bar on the x-axis represents a unique ARDRA pattern. A: Sixty-three isolated strains from thermophilic compost. B: Seventy 16S rDNA sequences amplified and cloned from extracted compost.

3.4 Cloned compost rDNA sequences

The ∼1500-bp PCR product amplified from extracted compost community DNA was used to generate a clone library of 16S rDNA fragments. Of 70 clones analyzed, 52 unique ARDRA patterns were detected, with 43 of the patterns appearing only once in the library (Fig. 1B). With the exception of isolate HC15 and clone pPD6, there was no overlap, based on comparison of ARDRA patterns, between the rDNA fragments from isolated thermophilic strains and those amplified and cloned directly from community DNA. Many hot compost ARDRA patterns could be related to one another by virtue of common bands shared by all the clones. The cloned rDNA fragments showed a common ‘skeletal’ digestion pattern featuring bands of 290, 245, 165, and 135 bp in length (all approximate values). This skeletal RFLP pattern can also be associated with Bacillus species based on a computer-assisted restriction digestion of rRNA gene sequences deposited in the Ribosomal Database Project [28].

3.5 Phylogenetic analysis of isolates and cloned sequences

Fig. 2 shows a neighbor-joining tree with the estimated phylogenetic placement of thermophilic compost bacteria and cloned rDNA sequences derived from hot compost. Isolate HC6 represents the ARDRA pattern that was observed most frequently (31 times) among the 63 isolates tested, and its 16S sequence was most closely related to Brevibacillus borstelensis[31]. Isolate HC5 was also a common colony type on enumeration plates, and its DNA sequence most closely resembled rDNA from the thermophilic organism Aneurinibacillus thermoaerophilus (formerly B. thermoaerophilus[31]), isolated from the high-temperature stages of sugar beet refining [32]. Comparison of ARDRA patterns suggested that isolate HC15 was highly similar to one of the cloned rDNA fragments (discussed below) and phylogenetic analysis indicated that both sequences were very similar to rDNA sequences from the organism Bacillus coagulans. Based on comparison of ARDRA patterns, isolates HC5 and HC6 were not considered to be identical to any of the cloned sequences in this study.


Phylogenetic tree showing the relationship of thermophilic bacteria (HC isolates, see Fig. 1A) and 16S rDNA sequences cloned from hot compost (pPD clones, see Fig. 1B) compared with previously described Bacillus species rDNA sequences deposited in public databases. The scale bar represents 0.10 change per nucleotide position.

Ten clones representing several of the most commonly observed ARDRA patterns (Fig. 1B) were selected for sequencing and phylogenetic analysis (Fig. 2). With the exception of clone pPD5, the phylogenetic analysis indicated that the cloned sequences all grouped within the broad phylogenetic radiation of Bacillus species [33]. Clone pPD5 was most closely related to Weissella confusa (formerly Lactobacillus confusus[34]). In addition to the 10 cloned 16S rDNA fragments for which full-length sequence information was obtained, one clone, pPD11, was discarded after sequencing showed it was a chimeric molecule with the first 360 nucleotides of the rDNA identical to the sequence from clone pPD5. The remainder of the chimeric molecule was highly similar to Lactococcus lactis.

Clone pPD14 corresponds to the ARDRA pattern detected most frequently in cloned rDNA fragments from compost community DNA. Phylogenetic analysis of the full-length 16S rDNA clone sequences (Fig. 2) indicates that the majority of the Bacillus-like sequences detected in SFW compost clustered together in two groups. Clones pPD10 and pPD15 were most closely related to B. thermoamylovorans, which was isolated from a fermented palm beverage [36]. Five other sequences (pPD4, pPD7, pPD12, pPD13, and pPD14) made up a discrete cluster that was not closely related to any described Bacillus species. The sequences in the cluster containing clone pPD14 were also clearly distinct from the obligately thermophilic group 5 Bacillus species first described by Ash et al., which are represented by B. stearothermophilus[33]. The remaining sequences, pPD6 and pPD8, aligned with B. coagulans and B. licheniformis respectively, species that can exhibit either mesophilic or thermophilic growth optima, depending on the strain [37,38].

3.6 Detection of Actinobacterial rDNA sequences

Despite the conspicuous presence of Actinobacteria[39] in many types of composted materials [2,3], actinomycete-type full-length sequences were not detected in the clone library of full-length rDNA sequences. Partial 16S rDNA fragments amplified from compost community DNA with the actinomycete-specific primer pair (Table 2) were used to generate a group-specific clone library. ARDRA screening of cloned partial 16S rDNA fragments was used to identify unique actinomycete-type sequences. Eight distinct ARDRA patterns were observed in a preliminary library containing 26 clones, and these patterns were chosen for sequencing and phylogenetic analysis. Fig. 3 shows a maximum-likelihood tree based on an alignment of partial 16S rDNA fragments from compost actinomycetes and closely related bacterial species. An effort was made to sequence representatives of all the ARDRA patterns detected among the actinomycete-specific partial rDNA fragments. The partial sequences related to Saccharomonospora viridis (clones TA10 and TA11) were the only clones that were closely affiliated with thermophilic actinomycete species previously isolated from hot compost. The remaining sequences were most closely related to species which have not been previously reported in compost, including Gordonia terrae (clone TA6) and two clusters related to Corynebacterium (clones TA3, TA8 and TA9) and Rhodococcus strains (clones TA2, TA4 and TA7).


Phylogenetic tree showing the relationship of cloned actinomycete-type 16S rDNA partial sequences (TA clones) compared with related 16S rDNA sequences from Actinobacteria[40] deposited in public databases. The cluster of organisms including Thermobifida fusca represents Actinobacteria commonly reported in compost [2,3]. The scale bar represents 0.10 change per nucleotide position.

4 Discussion

The combination of cultivation-dependent and -independent approaches used in this study yielded complementary, non-overlapping information about the composition of the microbial community in hot SFW compost. Composts are known to contain an assemblage of low-G+C Gram-positive bacteria, but this study suggests that the majority of the Bacillus-like sequences detected by our molecular approaches were not closely related to species commonly isolated from hot compost. Our results are consistent with previous culture-based studies showing that hot compost microbial communities are dominated by Gram-positive bacteria including Bacillus spp. and Actinobacteria. However, our 16S rDNA analyses revealed a collection of Gram-positive sequences that was far greater than expected from culture-based studies.

The results from our PCR-based, domain-level community analysis confirmed that Bacteria, including Actinobacteria, were present in all compost samples between 50°C and 74°C. Fungal 18S rDNA sequences were only detected in the 50°C compost sample. The latter finding is consistent with previous studies that have shown an absence of fungal growth in compost samples with temperatures of 60°C and hotter [10,11]. Methanogenic Archaea have been previously isolated from hot compost [40]; however, we did not amplify Archaeal rDNA sequences in this study despite the use of DNA extraction and PCR amplification methods which have revealed the presence of diverse methanogen-type sequences in peat samples (J. Yavitt, unpublished results). Because the extracted community DNA was diluted 100-fold to obtain reliable PCR amplification, it is possible that any Archaea that may have been present at low abundance remained undetected.

The substantial number of unique 16S rDNA sequences observed in our study suggests that a diverse assemblage of Gram-positive bacteria was present in hot SFW compost, which may be a reflection of the complex nature of the substrate and the heterogeneous conditions within the composting reactor. The rDNA sequences fall into two categories: (1) sequences related to, but not identical with, organisms that previously have been isolated from compost, and (2) sequences related to organisms which have not been previously isolated from hot compost. The first category contained a large number of unique Bacillus-type rDNAs detected by ARDRA screening and sequencing of selected isolates and clones (Fig. 2). The second category contains Gordonia-, Corynebacterium-, and Rhodococcus-type partial sequences (Fig. 3), suggesting that SFW compost samples harbored Actinobacteria that have not been reported in previous cultivation- or molecular-based studies of compost communities. These findings contribute to a growing body of evidence indicating that the diversity of the Gram-positive bacteria in hot compost is not severely limited during the high-temperature phase of composting, as was originally suggested by cultivation studies [4,5]. This point is supported by recent culture-independent studies of hot compost showing increasing complexity in compost community genetic profiles during the successive stages of composting [14,16].

While the clone library ARDRA patterns showed that many of the full-length rDNA sequences we detected were related to one another, phylogenetic analysis of selected sequences from the clones and isolates indicated clear distinctions between them and many of the aerobic, endospore-forming Bacillus spp. that have traditionally been reported in hot compost [38]. Previous 16S rRNA sequence analyses of cultivated thermophilic Bacillus spp. showed that many of the obligate thermophiles (i.e., those unable to grow below ∼40°C, represented by B. stearothermophilus) belong in the group 5 cluster [33,41]. Recent studies using cultivation-independent methods have identified other putative group 5 Bacillus-type rDNA sequences in a variety of composted materials [14,16]. In our study, we failed to detect any rDNA sequences that were conclusively related to the obligately thermophilic group 5 Bacillus species. Instead, our cloned sequences were distributed within the broader radiation of Bacillus-type organisms, including a cluster of sequences distantly related to B. thermoamylovorans, isolated from a fermented African palm beverage and another strain, Bacillus sp. strain 115898, recently isolated from silage [42]. Two prominent thermophilic strains isolated in this study were most closely related to organisms that to our knowledge have not been reported previously in compost, including members of the genera Aneurinibacillus and Brevibacillus[31].

In this study we identified one strain (isolate HC15), which was detected by both isolation and cultivation-independent rDNA cloning methods (clone pPD6). It is interesting to note that in a similar study of hot compost, Peters et al. identified an isolate, strain Sko08 (GenBank accession number AF213284), which has a partial rDNA sequence similarity to our isolate HC15 and clone pPD6 [16]. All three rDNA sequences clustered closely with sequences from B. coagulans, a ubiquitous, acid-tolerant species with thermophilic growth capabilities [38]. Strain HC15 differed from clone pPD6 by only four nucleotides across the length of nearly 1500 sequenced positions in the 16S rRNA gene. Thus, it seems possible that the cloned pPD6 sequence is identical to that of strain HC15, and that the discrepancy merely represents nucleotide misincorporation during PCR or variability within multiple operons of the same strain. Such interoperon sequence heterogeneity in ribosomal RNA genes has been observed on several occasions, with a Bacillus relative being one of the most striking examples [43].

The cloning and sequencing results obtained with actinomycete-specific PCR primers clearly indicate the presence of unique, uncultivated Actinobacteria species in our hot compost samples. However, lacking isolates or the analysis of full-length 16S rDNA sequences, the phylogenetic placement of these sequence fragments can only be presumptive [44]. The absence of Actinobacterial sequences from the library of full-length 16S rRNA genes is puzzling, but not unusual. The most likely explanation is that the abundance of Bacillus-type DNA in the extracted community nucleic acids obscured the presence of DNA from less prevalent bacteria. A recent study examined the sensitivity of the PCR to detect minority members in complex communities, finding that a population comprising 0.1–1% of the total DNA could be detected [45].

The combined results of this work suggest that the current view of community diversity in hot compost needs to be expanded to acknowledge the dominance of assemblages of related, but phylogenetically distinct low-G+C Gram-positive bacteria. None of the isolates or cloned sequences we detected could be conclusively assigned to a single obligately thermophilic Bacillus species, such as B. stearothermophilus. With the exception of isolate HC15 and clone pPD6, there was virtually no overlap between the rDNA sequences from the isolated bacteria and amplified and cloned sequences from community DNA. It is likely that more extensive screening of isolates and cloned rDNA sequences would reveal a greater overlap between the two approaches to population detection. Apparently, the bacterial communities in hot composts are composed of a broad spectrum of closely related species which are best identified by the combined use of isolation, ARDRA, and sequencing techniques as employed in this work.


This project was supported in part by the United States Department of Agriculture Food and Agricultural Sciences National Needs Graduate Fellowships Program, USDA/CSRS 93-38420-8743. We thank L. Walker, M. Howeler, and P. Schloss for providing the compost samples, as well as expert help and information about the composting reactors. Eugene Madsen and Richard Devereaux kindly provided several suggestions that helped to improve the manuscript. We are grateful to Patti Durfey for secretarial assistance. Preliminary compost studies were carried out by B. Eaglesham, R. Garen, C. Thomas, and P. Trutmann.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
View Abstract