Postinoculation Protozoan Establishment and Association Patterns of Methanogenic Archaea in the Ovine Rumen

Abstract

Association patterns between archaea and rumen protozoa were evaluated by analyzing archaeal 16S rRNA gene clone libraries from ovine rumen inoculated with different protozoa. Five protozoan inoculation treatments, fauna free (negative control), holotrich and cellulolytic protozoa, Isotricha and Dasytricha spp., Entodinium spp., and total fauna (type A) were tested. We used denaturing gradient gel electrophoresis, quantitative PCR, and phylogenetic analysis to evaluate the impact of the protozoan inoculants on the respective archaeal communities. Protozoan 18S ribosomal DNA clone libraries were also evaluated to monitor the protozoal population that was established by the inoculation. Phylogenetic analysis suggested that archaeal clones associated with the fauna-free, the Entodinium, and the type A inoculations clustered primarily with uncultured phylotypes. Polyplastron multivesiculatum was the predominant protozoan strain established by the holotrich and cellulolytic protozoan treatment, and this resulted predominantly in archaeal clones affiliated with uncultured and cultured methanogenic phylotypes (Methanosphaera stadtmanae, Methanobrevibacter ruminantium, and Methanobacterium bryantii). Furthermore, the Isotricha and Dasytricha inoculation treatment resulted primarily in archaeal clones affiliated with Methanobrevibacter smithii. This report provides the first assessment of the influence of protozoa on archaea within the rumen microbial community and provides evidence to suggest that different archaeal phylotypes associate with specific groups of protozoa. The observed patterns may be linked to the evolution of commensal and symbiotic relationships between archaea and protozoa in the ovine rumen environment. This report further underscores the prevalence and potential importance of a rather large group of uncultivated archaea in the ovine rumen, probably unrelated to known methanogens and undocumented in the bovine rumen.

Methanogens are classified under the kingdom Archaea and are divided into five major orders, Methanobacteriales, Methanosarcinales, Methanococcales, Methanomicrobiales, and Methanopyrales. Among these, Methanobacteriales dominates the rumen environment (36, 41). Methanogens live under strictly anaerobic conditions and are the only organisms that derive all their metabolic energy from the reduction of CO₂ by hydrogen to produce methane. Based on phylogenomic analysis, 31 proteins are uniquely present in all methanogens, strongly indicating that all methanogenic archaea form a monophyletic group exclusive of other archaea and that this lineage likely evolved from Archaeoglobus (15). Methanogenesis serves as the terminal electron sink process during organic matter decomposition in the rumen (16, 17) and has long been considered a metabolic waste process accounting for 5 to 15% of metabolizable energy loss in ruminants (17). Anthropogenic methane production is of environmental concern, and efforts to reduce it have focused on the reduction of methane eructation from ruminants, with little success. Understanding the ecology of rumen methanogens may provide clues to their control.

The rumen microbial ecosystem comprises diverse interdependent populations of obligatory anaerobic prokaryotes, protozoa, and fungi and accounts for most of the fermentative activity in the rumen (18). An episymbiotic association with methanogenic bacteria was described for rumen ciliate protozoa of the family Ophryoscolecidae (40) and other rumen ciliate protozoa (13, 39). Protozoa in anaerobic habitats rich in hydrogen, such as the rumen, bear hydrogenosomes and are often associated with methanogenic bacteria (11, 12, 35). The polymorphic nature of protozoa and the difficulty of cultivating specific species and strains have slowed the effective evaluation of protozoal ecology and taxonomy (8) and have further accentuated the lack of knowledge of the ecological relationships with other members of the rumen microbial community. Studies with isolated protozoa (13, 19, 27) have documented associations between methanogenic archaea and specified protozoa of the rumen. However, the interrelationship between protozoa and total archaeal communities is less well understood. Using the ovine rumen model, our objective was to characterize the association patterns of methanogenic archaeal communities with specific inoculated protozoan populations. We evaluated archaeal diversity in response to selected combinations of Isotricha spp., Dasytricha spp., Entodinium spp., cellulolytic, and typical “type A” protozoan inoculations. To maintain amplification efficiency for the quantitation of changes in the rumen methanogenic archaeal populations in response to the treatments, we adapted previously validated primers and hybridization probes to generate a ca. 250-bp amplicon for real-time PCR quantitation. By identifying the association of methanogens with specified protozoa, this study provides data toward understanding the role of selected protozoa in rumen methanogenesis.

MATERIALS AND METHODS

Inoculation and rumen sampling.

Ten rumen-cannulated Canadian Arcott wethers (castrates) between 42 and 51 kg of body weight and approximately 1 yr old from a naturally fauna-free (FF) sheep flock (21) were used to establish different rumen protozoan populations. One group (FF) of two wethers were maintained without fauna (negative control), while the other four groups of two wethers each were inoculated intraruminally as follows: the “ILP” group was inoculated with holotrich and cellulolytic protozoan species (Isotricha, Dasytricha, Polyplastron, Diplodinium, and Ostracodinium), the “ID” group with Isotricha and Dasytricha spp., the “ENT” group with Entodinium sp., and the “TA” group with type A (10) fauna. The rumen fluids used for protozoal inoculations were obtained from different donor sheep containing specific rumen protozoan populations as described previously (20). The donor sheep had been maintained with particular protozoan populations for several years. The five groups of wethers were fed an experimental diet composed of corn silage, 93.6% (dry weight); soybean meal, 5.5% (dry weight); and a vitamin-mineral mixture, 0.9% (dry weight). Rumen sampling for DNA extraction and analysis was done before the morning feeding at 0800 h, 90 days postinoculation and 30 days after initiation of feeding the experimental diet. Each sample was a composite of equal volumes of rumen fluid from two wethers from the same treatment. All animals were housed and treated in accordance with the guidelines of the Canadian Council on Animal Care (2).

DNA extraction and purification.

Total DNA was extracted from 400 μl of rumen fluid, using a Power Soil DNA kit (MoBio Laboratories, Inc., Carlsbad, CA) according to the manufacturer's protocol. Genomic DNA concentration was measured using a Beckman spectrophotometer, DU640 series.

Evaluation of protozoan consortia.

Due to the polymorphic nature of protozoa (9), the results of the inoculation treatments were evaluated using molecular methods to ascertain the establishment of the intended protozoan communities. The 18S ribosomal DNA (rDNA) of protozoan communities was amplified using 50 ng of purified template DNA. The PCR mixture contained 1 μl of template, 2.5 μl of 10× dilution buffer, 10 pmol of each primer, and 1 unit of Ex-Taq polymerase (Takara Shuzo, Japan) in a final volume of 25 μl. The protozoan specific universal primer P-SSU-342F (Table 1) and the universal eukaryote primer Medlin B (Table 1) were used for the amplification. The numbering of protozoan primer sites was based on Saccharomyces cerevisiae numbering. The thermal profiles for the amplification were as follows: an initial denaturation at 95°C for 3 min, followed by 25 cycles of 95°C for 30 s, 50°C for 30 s, and 72°C for 1.5 min. Purification of the PCR fragment (∼1,300 bp), cloning, and transformation were as described below. A clone library of protozoan 18S rDNA from each pooled sample, based on treatments, was evaluated to confirm colonization by the intended protozoa.

TABLE 1.

Oligonucleotides used in respective experiments in this study

Target	Experiment	Primer name	Oligonucleotide sequence	Reference
Protozoa	PCR	P-SSU-342F	5′-CTTTCGATGGTAGTGTATTGGACTAC-3′	22
		Medlin B	5′-TGATCCTTCTGCAGGTTCACCTAC-3′	24
Archaea	Primary PCR	21F	5′-TTCCGGTTGATCCYGCCGGA-3′	33
		1406-1389R	5′-ACGGGCGGTGTGTGCAAG-3′	23
	Nested PCR	344F	5′-ACGGGGYGCAGCAGGCGCGA-3′	27
		915R	5′-GTGCTCCCCCGCCAATTCCT-3′	34
		G+C clamp	5′-CGCCCGCCGCGCCCCGCGCCCGTCCCGCCGCCCCCGCCCC-3′	25
	qPCR	896-915F	5′-AGGAATTGGCGGGGGAGCAC	34
		MB 1174F	5′-GAGGAAGGAGTGGACGACGGTA-3′	27

PCR amplification of archaeal 16S rRNA genes.

The 16S rRNA gene sequences of the archaeal communities were amplified by nested PCR using 50 ng of purified template DNA. The final volume of the PCR contained 1 μl of template, 2.5 μl of 10× dilution buffer, 10 pmol of each primer, and 1 unit of Taq polymerase (Takara Shuzo, Japan) in a final volume of 25 μl. Archaeal primers used for the primary PCR were 21F and 1406-1389 R (Table 1). Primers used for the nested PCR were 344F and 915R (Table 1). The forward primer used in the nested PCR had an added 40-bp G+C clamp (25). The PCR amplification was performed in a 96-well ICycler (Bio-Rad Laboratories, Hercules, CA). A step-down protocol (26) was used in which the initial annealing temperature, 60°C, was reduced by 0.5°C per cycle for the first 10 cycles. The amplification conditions involved denaturation at 95°C for 1 min, followed by 25 cycles of 95°C for 30 s, 60°C (decreasing to 55°C) for 30 s, and 72°C for 1 min. The nomenclature of all primers used for archaeal amplifications was based on that of the Escherichia coli numbering system.

Denaturing gradient gel electrophoresis.

Parallel denaturing gradient gel electrophoresis (DGGE) (25, 26) was performed using a DGGE-2401 system (C.B.S. Scientific, Del Mar, CA). PCR-amplified 16S rRNA gene fragments were separated by using a polyacrylamide gel with 0.5× TAE buffer (20 mM Tris-acetate, 10 mM sodium acetate, 0.5 mM EDTA) with a 35 to 60% linear gradient of denaturant (100% corresponds to 7 M urea and 40% formamide in deionized water). The gel was run at 60°C and 150 V for 7 h and then placed in 200 ml of fixing solution (10% ethanol and 0.5% acetic acid) overnight. The gel was stained in 200 ml of 0.1% (wt/vol) silver nitrate for 15 min and then developed in 200 ml of 1.5% sodium hydroxide (wt/vol), 0.1% sodium borohydride (wt/vol), and 0.4% (vol/vol) formaldehyde for 7 min to reveal the DGGE fingerprints of the archaeal communities represented in the respective samples. DGGE banding patterns and clustering were analyzed using BioNumerics software (Applied Maths, Inc., Austin, TX).

Cloning DGGE fragments.

The resolved bands from the DGGE gel were aseptically excised into 1.5-ml Eppendorf tubes. Each excised band was washed with 200 μl of distilled water. The bands were incubated in 50 μl of TE buffer, pH 7.8, at 37°C overnight and further PCR amplified using the 344F (without the G+C clamp) and 915R (Table 1) primers as mentioned previously. The reaction setup and amplification conditions were as previously stated, using 1 μl of the extracted template. The PCR products were evaluated using agarose (1%) gel. Correctly sized bands were further purified using a PCR purification kit (QIAGEN Sciences, MD). The purified bands were cloned into pGEM-T Easy vector (Promega, San Louis Obispo, CA), and the ligation product was used to transform E. coli JM109 cells by heat shock (42°C for 45 s). The transformed cells were then plated on LB-ampicillin (100 mg/liter) plates and incubated overnight at 37°C. Individual colonies were randomly picked and grown overnight at 37°C in 3 ml of LB medium supplemented with ampicillin (100 mg/liter). All clones were checked for the correct inserts by PCR using universal M13F and M13R primers. The cycling conditions consisted of an initial denaturation at 95°C for 3 min, followed by 25 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s. Randomly picked plasmids from clones which harbored inserts of the correctly estimated length were isolated and sequenced.

Real-time quantitative PCR.

Quantitation of total and methanogenic archaea was done using real-time quantitative PCR (qPCR). To obtain standards for the qPCR, purified genomic DNA from Methanosphaera stadtmanae was amplified using primers Arch 344F and Arch 1406-1389R (Table 1). The reaction mixture (25 μl) contained 1 μl of genomic DNA, 10 pmol of each primer, 200 nmol of each deoxynucleoside triphosphate, and 1 unit of Ex-Taq polymerase (Takara-Shuzo, Japan). The cycling conditions were as follows: initial denaturation at 94°C for 3 min, followed by 25 cycles of 94°C for 30 s, 50°C for 30s, and 72°C for 1.5 min. The amplification product was electrophoresed on 0.8% agarose gel in 1× TAE buffer, followed by ethidium bromide staining to confirm the production of a single product of the expected molecular weight. An approximately 1,100-bp fragment was gel purified using a QIAquick PCR purification kit (QIAGEN Sciences, MD). The purified PCR product was then cloned into pGEM-T Easy (Invitrogen, Carlsbad, CA). A single colony, verified for the expected insert using PCR, was grown in 3 ml of LB medium supplemented with ampicillin (100 μg/ml) overnight. The culture was centrifuged at 5,000 × g to pellet the cells. Plasmid was extracted using a QIAprep Spin miniprep kit according to the manufacturer's instructions (QIAGEN Sciences, MD). The purified plasmid was quantified using a Beckman spectrophotometer (DU640 series). The number of 16S rRNA gene copies present in the plasmid preparation was calculated using the DNA concentration and the molecular mass of the vector with the insert. The concentrated plasmid was serially diluted (10-fold) to provide a range of 10⁸ to 10 copies·μl⁻¹. Serially diluted samples were used to generate a standard curve. Quantitative PCR was performed on a 96 well iCycler thermal cycler fitted with an optical module (Bio-Rad Laboratories, Hercules, CA) using an IQ SYBR green supermix (Bio-Rad Laboratories Hercules, CA) fluorophore. Each reaction mixture (20 μl) contained 20 ng of genomic DNA and 10 pmol of each primer. The Arch 896-915F and the Arch 1406-1389R (Table 1) primer set was used for the quantitation of total archaea. The Methanobrevibacter, Methanosphaera, and Methanobacterium sp.-specific primer, MB 1174F (Table 1), and Arch 1406-1389R (Table 1) were used to estimate the copy numbers for archaea related to the mentioned groups and their close relatives. The cycling conditions consisted of 40 cycles of 94°C for 30 s, 63°C for 30 s, and 72°C for 30 s.

Phylogenetic analysis.

Sequences were checked for chimeras using Ribosomal Database Project CHECK-CHIMERA (version 2.4) software (5) and aligned using CLUSTAL X (37). The phylogenetic tree was based on 572 bp and was inferred from the calculated evolutionary distances. Phylo_Win software (14) was used to generate the phylogenetic tree, using a neighbor-joining algorithm, a global gap removal option, and a bootstrap analysis of 1,000 replicates. Reference sequences for both protozoan and archaeal trees were retrieved from the GenBank database (1). The 18S rDNA of Paramecium tetraurelia and 16S rRNA gene sequence of Aquifex aeolicus were used as out groups for the rooted protozoan and archaeal trees, respectively. Phylogenetically distinct clusters were judged based on a bootstrap value of >50%.

Statistical analysis of libraries.

The total numbers of operational taxonomic units (OTUs), richness, and diversity were calculated for the pooled sequences from each inoculation treatment. Relative distances between sequences were calculated using DNAdist software of the PHYLIP package. DOTUR software (29) was used to assign sequences to OTUs based on the farthest distance algorithm. An OTU at the species level was defined as a ≤3% difference in base positions. Rarefaction curves which related the number of OTUs with the number of sequences were plotted from the output from DOTUR (data not shown). The maximum number of OTUs for each archaeal consortium was determined from DOTUR analysis at a 3% distance difference. Species richness was estimated using a ChaoI richness index (3), and community diversity was estimated using the Shannon-Weaver index of diversity (32) as implemented in DOTUR. The statistical significance of the differences between mean values was determined using confidence interval estimates associated with predicted values from DOTUR. The 16S rDNA archaeal libraries from inoculation treatments were compared with those of control treatments (FF) using ∫-Libshuff (30) to determine whether a pair of the libraries was drawn from the same population. This test confirmed whether differences in the libraries were due to treatments or chance. Coverage curves were plotted with the FF treatment set as the homologous library and those from inoculation treatments heterologous. A homologous coverage refers to the number of sequences in a given library without homologs or replicates, and it was defined as <3% difference in sequence positions in this study. Heterologous coverage refers to the number of sequences not found in another library at the same level of difference. P values assigned to the multiple comparisons were adjusted using Bonferroni adjustment. Furthermore, SONS (31) was used to estimate the fraction of OTUs shared between respective libraries. The recommended output from DOTUR was used together with a tab-delineated file with sequence names and respective library designations.

Sequencing.

All clones were sequenced at Lethbridge Research Centre Sequencing Laboratory.

Nucleotide sequence accession numbers.

All sequence data have been submitted to the GenBank database under accession numbers DQ832550 to DQ832582 and DQ836487 to DQ836625 for protozoa and archaea, respectively.

RESULTS AND DISCUSSION

Protozoan inoculation.

An analysis of PCR-amplified protozoal 18S rDNA, using agarose gel electrophoresis (results not shown), indicated that the FF treatment was in fact protozoan free and further confirmed specific amplification of ∼1,300 bp of protozoan DNA from each inoculation treatment. Phylogenetic analysis of protozoan 18S rDNA sequences (Fig. 1) suggested that sequences obtained from ILP inoculation (cluster I; n, 5) clustered within the family Ophryoscolecidae and were closely related to Polyplastron multivesiculatum (sequence identity, >99%). All TA protozoan sequences (n, 12) were of the family Ophryoscolecidae. Cluster II consisted of clones (3/12) from TA inoculation (sequence identity, >98.5%). Two clones clustered with CRG11 (AF502941) isolated from cow rumen. Supported by a bootstrap value of 100, over 80% (5/6) of ENT protozoan sequences were related to CRG11 (AF502941) (sequence identity, >99%) in cluster III. The closest Entodinium caudatum (U57765) strain-related clone was TA_P515 (see Fig. 3, cluster IV), exhibiting 98.7% identity. Over 80% of ENT protozoa showed a high percentage of identity (97.0% to 99.7%) with nearly 60% (7/12) of TA protozoan clones. Last, the ID inoculation resulted in two clusters (V [n, 3] and VI [n, 6]) closely related to Dasytricha ruminantium and Isotricha intestinalis, respectively, both of the family Isotrichidae.

FIG. 1.

Phylogenetic tree of protozoan communities based on 18S rDNA showing postinoculation protozoan establishment in rumen. Treatments are as follows: ILP, holotrich and cellulolytic protozoa; ID, Isotricha/Dasytricha; ENT, small entodiniomorph; TA, faunated (positive control). FF, fauna-free (negative control), was excluded due to no amplification with protozoan-specific primers. Paramecium tetraurelia was used as an outgroup for the rooted tree. Bootstrap values (>50%) are indicated. Scale bar represents 10% estimated sequence difference. Roman numerals represent clusters within families.

FIG. 3.

Phylogenetic tree of archaeal communities in rumen samples inoculated with different protozoan cultures. The analysis was based on partial 16S rRNA genes. Inoculations (in bold) are indicated after clone labels and are as follows: FF, fauna-free (negative control); ILP, holotrich and cellulolytic protozoa; ID, Isotricha/Dasytricha; ENT, small entodiniomorph; TA, faunated (positive control). Aquifex aeolicus was used as an outgroup for the rooted tree. Bootstrap values (>50%) are indicated on the tree. Scale bar represents a 10% estimated sequence difference. Roman numerals represent clusters within groups.

The fact that ILP treatment resulted in only large entodiniomorphid protozoa affiliated with P. multivesiculatum implied a limitation in the establishment of the simultaneously inoculated Isotricha, Dasytricha, Diplodinium, and Ostracodinium spp. to below the detection limits (<10 copies/ng DNA) of the primers used. Starting quantities of each species were not known, but their presence in the donor rumen samples was confirmed by microscopy. Cannibalistic predation of Polyplastron spp. has been reported (42). Although no evidence for Polyplastron engulfment of Isotricha has been documented, all strains of P. multivesiculatum engulf Epidinium spp., Diplodinium spp., Ostracodinium spp., and Eudiplodinium magii (6). The difficulty in detecting other protozoa in the ILP treatment suggests that the protozoan sequences affiliated with P. multivesiculatum may have predated on other protozoan species and, to a large extent, account for the diversity in the associated archaeal sequences. Furthermore, with >99% identity observed among the P. multivesiculatum species-related sequences, it is likely that a single phylotype was involved. In the ILP treatment, although the inocula contained Isotricha, Dasytricha, Polyplastron, Diplodinium and Ostracodinium spp., cloning and sequencing indicated the prevalence of Polyplastron sp. Therefore, we anticipate that in a scenario where a number of protozoan species are inoculated, the resultant archaeal species may depend on the outcome of competition between inoculated protozoa. However, the resulting archaeal consortium may differ significantly from, for example, that resulting from a stabilized natural protozoan succession, as observed for TA treatment in which the typical type A fauna was characterized by Entodinium sp.-related sequences.

Furthermore, based on the phylogenetic analysis (Fig. 1), significant differences exist in phylogenetic placement of the Entodinium group, but limited 18S rDNA sequences of related cultured protozoa exist in the GenBank for comparison. The tree also revealed a greater diversity among various protozoan subfamilies than had been previously demonstrated. For example, several clones from the TA treatment group do not exhibit strong affiliations to identifiable groups and appear to be distinct. The phylogenetic analysis suggested that with the exception of the predominance of P. multivesiculatum in the ILP treatment group, the inoculation treatments resulted in the establishment of the expected protozoa.

Archaea. (i) DGGE banding patterns.

Two major clusters (A and B) were observed for the analysis of DGGE banding patterns derived from amplified archaeal fragments (Fig. 2). The analysis suggested >50% similarity between banding patterns from TA- and ENT-associated archaea and >60% similarity between ID- and ILP-associated communities. Furthermore, banding patterns for the FF group treatment were unique, although it shared some similarities with ID and ILP clusters. These may suggest significant phylogenetic relatedness of archaeal species from ENT and TA, based on the G+C content. Suggested relatedness between ID and ILP was also apparent. Finally, bands were observed for the FF group treatment, which indicated the presence of an archaeal community that is independent of protozoa. The primer pair (Arch 344F and Arch 915R) used for the DGGE analysis (Table 1) was chosen because of the primers' universality and selectivity for archaea, the provision of an adequate fragment length for resolution on the gel, and the resulting fragments being informative for phylogenetic analysis. However, we do not interpret the results of the DGGE analysis to imply phylogenetic relationships but as predictive of community similarity based on G+C content. Theoretically, if any two phylogenetically distinct fragments with the same G+C content are examined using DGGE, the fragments will migrate similarly.

FIG. 2.

DGGE profiles of archaeal communities in sheep rumen samples inoculated with different groups of protozoa. Treatments are as follows: FF, fauna-free (negative control); ILP, holotrich and cellulolytic protozoa; ID, Isotricha/Dasytricha; ENT, small entodiniomorph; TA, faunated (positive control). The similarity index (%) of banding patterns is indicated on the scale above. Two major clusters (A and B) are also indicated.

(ii) Phylogenetic analysis.

An analysis of the phylogenetic tree (Fig. 3) for archaeal sequences (n, 139) resulted in the identification of two major groups. These were the uncultured archaea, which contained 73% of the total sequences analyzed, and sequences affiliated with the Methanobacteriaceae (27%). Based on a bootstrap cutoff of >50%, three clusters (A, B, and C) were identified within the uncultured group, and two clusters (D and E) were identified within Methanobacteriaceae. Predominantly, sequences from the FF (negative control; n, 23), ENT (n, 36), and TA (n, 41) treatments had no known cultured relatives. Over 69% of the archaeal sequences from FF, 88% from TA, and 72% from ENT treatments clustered with sheep rumen clones Qld_09 (AY995282) and Qld_03 (AY995277) in cluster A.

From the phylogenetic analysis, TA treatment-derived protozoa were closely related to ENT protozoa (Fig. 1). This correlated with the observations for the archaeal DGGE analysis (Fig. 2). Ninety percent of TA treatment-associated archaea clustered with 92% of those from ENT in clusters A and C (Fig. 3). In comparison with the bovine rumen, the preeminence of Entodinium-like sequences in TA treatment was similar to that in a previous study (33) which reported that the predominant protozoa (81%) identified in bovine rumen belonged to the Entodinium group. Although Entodinium represents the major type A fauna in the bovine and ovine rumens, the presence of this large group of uncultured archaea (Fig. 3) has not been documented in the bovine rumen. Methanobacteriales was the only cultured methanogenic group identified in our study, to the exclusion of other methanogenic families such as Methanosarcinales and Methanomicrobiales, that have been found in bovine rumen (28, 38, 41). This suggests differences in the diversity of archaea that colonize the bovine and ovine rumen.

In cluster B (Fig. 3), two clones from FF treatment clustered with close relatives of Qld16 (AY995288) from the ovine rumen and Ar28 (AF157523) from a swine waste pit. Relative to the total number of sequences analyzed for each treatment, cluster C comprised sequences from FF (26%), ILP (70%), ID (30%), ENT (5%), and TA (2%). These clustered with M2 (AB034186) and M7 (AB034184), associated with rumen samples. Six clones from ILP treatment clustered with M. thermautotrophicus (AY196661), M. stadtmanae (AY196684), M. ruminantium (AY196666), and M. bryantii (AY196658) in cluster D. This cluster included one clone from ID treatment representing 2% of sequences (n, 37) examined from this treatment. Nearly 90% (23/26) of archaeal clones within cluster E were derived from ID treatment and were uniquely clustered with Methanobrevibacter smithii. Additionally, only two clones from ENT and one from TA, representing, respectively, 5% and 2% of the number of clones analyzed from the respective treatments, were found in this cluster. The majority (95%) of TA-associated archaea were uncultured. Low external bootstrap values precluded the identification of distinct archaeal clusters from ENT and TA treatments. ENT-associated archaeal clones were associated predominantly with uncultured archaea and closely related to those associated with FF (negative control) treatment. None of the sequences from the FF (negative control) clustered within the Methanobacteriaceae.

These results suggest strong association patterns between different archaeal groups and specified protozoal inocula and imply that dominant protozoan species influence archaeal communities in the ovine rumen. Therefore, selected protozoa may alter the rumen environment (e.g., via predation) to alter the establishment of archaeal species in the rumen. The presence of archaea in FF treatment implies an archaeal association with the rumen microbial community independent of protozoa. This suggests that FF-associated archaea do not share an obligate/symbiotic relationship with rumen protozoa and are most likely free living. These observations may relate to the evolution of commensal and symbiotic associations between archaea and protozoa in the rumen.

A previous study (45) observed that archaeal clones from a cannulated Corriedale sheep rumen, with naturally occurring protozoan (type A) fauna, were affiliated with Methanomicrobium mobile, Methanobrevibacter ruminantium, and M. smithii. In the present study, only one clone, representing less than 5% of the total clones from TA inoculation, was related to M. smithii. Furthermore, our results suggested that clones related to M. ruminantium were associated with ILP inoculation and M. smithii with ID inoculation. These differences in methanogenic archaeal associations between the two studies may be due to differences in the dominating protozoan fauna, which, in turn, may be influenced by differences in host diet (7, 8) and geographical location (9, 42).

Although, we could not partition the individual roles of the holotrichs Isotricha sp. and Dasytricha sp. in the archaeal assemblage, we provide evidence to show that the archaeal group(s) affiliated with M. smithii primarily associates with members of the family Isotrichidae. The specificity of such an association raises the possibility of a stronger relationship, most likely symbiotic. Furthermore, archaeal groups related to M. bryantii, M. stadtmanae, and M. ruminantium predominantly associated with ILP inoculation, in which large entodiniomorphs closely related to P. multivesiculatum predominated. Similarly, a symbiotic association is likely.

From the phylogenetic analysis, the only previously published rumen-associated archaeal sequences that clustered with the uncultured group were M2 (AB034186) and M7 (AB034184) and phylotypes observed in a recent study from Australia (43). This group of uncultured archaea was dominantly associated with FF, ENT, ILP, and TA treatments. Their association with most treatments may suggest ubiquity in the ovine rumen. The relative importance of this archaeal lineage may have been underestimated in the past due to their paucity or absence in various studies (36, 41) of clone libraries. This is also the first report that links this prevalent yet less understood uncultured archaeal lineage to specified protozoa in the ovine rumen archaeal community.

(iii) Quantitative PCR.

A primer pair specific for Methanobacterium, Methanosphaera, and Methanobrevibacter spp. was used in the qPCR assay, due to the primers' relatedness to the retrieved dominant groups based on the phylogenetic analysis. In general, the ID inoculation resulted in the lowest number of total and methanogenic archaea compared to that of the rest of the treatments (Fig. 4). Between-treatment comparisons suggested that the total archaea observed for TA was significantly (P ≤ 0.05) higher than that of ID inoculation, although copy numbers for methanogens in both treatments were not significantly (P ≤ 0.05) different (Fig. 5). A comparison of the copy numbers for archaea and methanogens indicated no significant (P ≤ 0.05) differences within each inoculation treatment, and estimated copy numbers for total archaea and methanogens in the FF treatment were not different from those of protozoan inoculation treatments.

FIG. 4.

Comparisons of archaeal 16S rRNA gene libraries from respective protozoan inoculations using ∫-Libshuff analysis. Comparisons were made using the integral form of the Cramér-von Mises statistic as implemented in ∫-Libshuff with 10,000 randomizations. The P value indicated in each plot was corrected using Bonferroni multiple comparisons adjustment. Libraries from respective inoculation treatments were compared with the FF library. Filled circles represent the homologous library. Open circles represent the heterologous library. Plots A, B, C, and D compare results for FF versus those of ILP, ID, TA, and ENT treatments, respectively. ILP, holotrichs and cellulolytic protozoa; ID, Isotricha and Dasytricha; TA, type A; ENT, entodinia.

FIG. 5.

Copy numbers of total and methanogenic (Methanobacterium-, Methanosphaera-, and Methanobrevibacter-related) archaea using real-time quantitative PCR. The abscissa shows the treatments. FF, fauna-free (negative control); ILP, holotrich and cellulolytic protozoa; ID, Isotricha/Dasytricha; ENT, small entodiniomorph; TA, faunated (positive control).

The significant decrease in the number of archaea associated with ID inoculation compared to that of the type A treatment may be due to different rates of establishment of inoculated protozoa. Although initial numbers were not known, diet may influence growth of protozoa (4). With no differences observed between the estimates of total archaea and methanogens, we infer that archaea in the ovine rumen are essentially methanogenic and are dominated by an uncultured group. Although we were able to amplify the uncultured group using Methanobacterium, Methanosphaera, and Methanobrevibacter cluster primers, the extent of their relationship with cultured groups should be examined beyond the 16S rRNA gene level.

(iv) Statistical analysis of libraries.

To understand the composition of archaea associated with different protozoa, community indices for the respective inoculation treatments were estimated (Table 2). Archaea associated with TA inoculation had the lowest number of OTUs, followed by those associated with ENT inoculation. The ID inoculation had the highest number of OTUs. No differences in OTUs were observed between FF- and ILP-associated archaea. Archaea associated with ID inoculation had the highest OTUs. TA-associated archaea had the lowest OTUs, followed by ENT inoculation. However, the observed differences did not result in differences in diversity and richness indices (Table 2). It is evident that the archaeal community in the rumen offers very limited OTUs. This may be due to the fact that methanogens in the rumen are monophyletic.

TABLE 2.

Estimated maximum OTUs and diversity and richness indices^a

Treatment	OTUs	Shannon-Weaver diversity index (CL)	ChaoI richness estimate (CL)
FF	9	1.86 (1.50-2.23)	14.0 (11.6-33.0)
ILP	9	1.54 (1.11-1.98)	10.5 (9.2-21.5)
ID	12	1.66 (1.23-2.09)	30.0 (15.9-93.3)
ENT	8	1.83 (1.49-2.16)	9.0 (8.1-18.6)
TA	7	1.25 (0.81-1.69)	10.0 (7.4-29.9)

^aFF, fauna-free (negative control); ILP, Isotricha/large protozoa; ID, Isotricha/Dasytricha; ENT, small entodiniomorph; TA, faunated (positive control). Confidence limits (CL) are indicated in parentheses.

The ∫-Libshuff analysis (Fig. 5) suggested that compared to that of the negative control, the inoculation treatments resulted in significantly different archaeal libraries. Therefore, changes observed for the respective libraries were not due to chance but to the effects of the treatments. Also, the SONS analysis (Table 3) indicated that the FF-associated archaeal library shared only 9% of its OTUs with the ID and ILP libraries and none with the ENT library. The ID library shared 27% of its OTUs with ILP and 7% with both ENT and TA. Also, the ILP library shared 13% and 7% of its OTUs with ENT and TA libraries, respectively. The highest proportion of OTUs (45%) was shared between TA and ENT treatments. These observations are confirmed by clustering patterns of clones in the phylogenetic analysis (Fig. 3) and to a lesser extent in the DGGE analysis.

TABLE 3.

Comparison of 16S rRNA gene libraries using SONS to estimate the fraction of OTUs shared among libraries^a

Library		Fraction of shared OTUs
A	B	Ab	Ba
FF	ID	0.09	0.07
FF	ILP	0.09	0.07
FF	ENT	0.00	0.00
FF	TA	0.18	0.18
ID	ILP	0.27	0.27
ID	ENT	0.07	0.06
ID	TA	0.07	0.09
ILP	ENT	0.13	0.13
ILP	TA	0.07	0.09
ENT	TA	0.31	0.45

^aAb, fraction of library A shared with B. Ba, fraction of library B shared with A. FF, fauna-free (negative control); ILP, Isotricha/large protozoa; ID, Isotricha/Dasytricha; ENT, small entodiniomorph; TA, faunated (positive control).

A plausible explanation for TA sharing a higher fraction of OTUs with ENT would be that most of the TA and ENT protozoan sequences fall in the same cluster (Fig. 1). Therefore, similarity in postinoculation protozoan sequences is indicative of similarity in corresponding archaeal libraries.

It is noted, however, that although most of the ENT-related clones were in the same cluster as those of FF, SONS analysis indicated that they were not close enough to be placed in the same OTUs. This is because the SONS analysis sets limits on sequence differences (0.03). However, using phylogenetic analysis, strong external bootstrap values could place sequences with significant variance in base positions in a cluster. Furthermore, there was a significant lack of similarity between FF- and ID-derived archaeal clones in the phylogenetic analysis. This observation was further confirmed with SONS analysis. The body of evidence suggests quantitative and qualitative changes in the archaeal community and implies that protozoal populations could possibly be exploited to alter the rumen environment to select for or against specific archaeal species.

Conclusions.

This study shows that specific methanogen populations may associate with specific protozoal populations in the ovine rumen. The study also provides evidence to suggest a phylogenetic difference between a very large cluster of uncultured archaeal sequences with no known cultured relatives and cultured methanogenic archaea. Isolation, characterization, and further physiological tests will be necessary to validate the group's physiological differences compared to cultured members of the Methanobacteriaceae. This study demonstrates that this large group of diverse uncultured rumen archaea is an important part of the rumen ecosystem of sheep. These organisms are now shown to be present in sheep from different geographical areas (Australia and Canada). With the exception of members of the family Isotrichidae, these diverse uncultured rumen archaea are mainly associated with members of the family Ophryoscolecidae. Since these sequences have yet to be discovered or quantified in cattle, their role or presence in the broader rumen ecosystem is not clear. The predominance of clones not related to any cultured archaea may make steps toward mitigating methane production from ruminants, such as the development of vaccines targeting rumen methanogens (44), difficult. A critical determination has to be undertaken to unravel the underlying genetic and/or physiological differences between this large unknown group and cultured methanogens before targeted methane reduction strategies may be successful.