OUP user menu

Phylogenetic analysis of fiber-associated rumen bacterial community and PCR detection of uncultured bacteria

Satoshi Koike , Sayo Yoshitani , Yasuo Kobayashi , Keiichi Tanaka
DOI: http://dx.doi.org/10.1016/S0378-1097(03)00760-2 23-30 First published online: 1 December 2003


The fiber-associated rumen bacterial community was phylogenetically examined by analysis of 16S rRNA gene (16S rDNA) sequences. Hay stems of orchardgrass and alfalfa were incubated for 6 and 20 h, respectively in the rumen of two different sheep, and total DNA was extracted from the incubated stems to clone bacterial 16S rDNAs using polymerase chain reaction (PCR). Of 91 such clones, 21 showed more than 97% sequence similarity with known isolates, 32 clones had 90–97% similarity with known sequences, and for the remaining 38 clones, the similarity was less than 90%. The majority of clones fell into the CytophagaFlavobacterBacteroides and low G+C Gram-positive bacterial phyla (43 and 44%, respectively). Prevotella-related and Butyrivibrio fibrisolvens-related sequences formed large clusters in the phylogenetic tree. Unknown sequences were found to form three unique clusters, one of which was suggested by semi-quantitative PCR to be more prevalent in the rumen receiving a high alfalfa diet.

  • Rumen
  • Fiber-associated bacterium
  • 16S rDNA
  • Phylogeny
  • Uncultured bacterium

1 Introduction

Ruminant animals harbor a diverse and dense microbial population in the rumen. This symbiotic microbiota, which consists of bacteria, fungi and protozoa, allows the animals to utilize plant fiber as an energy source. In recent years, the genetic diversity of the rumen bacterial community has been revealed by sequence analysis of 16S ribosomal RNA gene (16S rDNA) [1,2]. Through the process of analyzing the rumen bacterial community, the presence of uncultured bacteria has been recognized. However, the roles of these undescribed bacteria in plant fiber digestion are controversial [1,2].

Rumen bacteria have been classified into four groups according to environment: (1) free-living bacteria associated with the rumen liquid phase; (2) bacteria associated with feed particles; (3) bacteria associated with rumen epithelium; and (4) bacteria attached to the surface of protozoa [3,4]. Bacteria associated with feed particles are considered to be the most important group for fiber degradation because of their predominance in terms of bacterial mass and endoglucanase activity [5,6].

Fibrobacter succinogenes, Ruminococcus flavefaciens and Ruminococcus albus are considered to be representative cellulolytic bacteria of the rumen [7]. In a previous study, we used the newly developed competitive polymerase chain reaction (PCR) assays to determine the kinetics of fiber attachment of these three species and the ecological significance of F. succinogenes in the rumen [8]. However, other bacterial members, including uncultured bacteria, should be monitored for characterizing the whole process of fiber digestion in the rumen.

In the present study, we focused on fiber-associated bacteria, using molecular analysis of 16S rDNA to determine the members of a fibrolytic consortium for degrading grass and legume hays. Then, a preliminary detection of uncultured fiber-associated rumen bacteria was attempted by PCR with newly designed primers based on 16S rDNA sequences unique to these bacteria.

2 Materials and methods

2.1 Animals and sampling

Two ruminally fistulated sheep (average body weight, 78.5 kg) were used in the phylogenetic study. One sheep was fed orchardgrass hay (1.5 kg per day), while the other was given alfalfa hay (800 g per day) and commercial formula feed for beef cattle (200 g per day). Orchardgrass hay and alfalfa hay contained 11.6 and 16.5% crude protein (CP) and 68.2 and 45.6% neutral detergent fiber (NDF) as dry basis, respectively. The feeding was done once daily at 09:00 h. The sheep had free access to mineral block and water. Nylon bags containing chopped stems of orchardgrass hay or alfalfa hay (2 cm in length) were placed into the rumen of sheep immediately prior to feeding. Orchardgrass hay stems (4.2% CP and 77.6% NDF), were incubated in the rumen of the sheep on the orchardgrass diet, whereas alfalfa hay stems (12.1% CP and 40.9% NDF) were in the rumen of the sheep on the alfalfa hay diet. Bags were removed from the rumen at 6 h for the orchardgrass sample and 20 h for the alfalfa sample. The bags were rinsed thoroughly in water (38°C), and then stored at −80°C.

For PCR detection of an unknown bacterial group in the rumen (see Section 3), two ruminally fistulated sheep (average body weight 61.0 kg) were given the diet of alfalfa hay and formula feed as above, but with three different proportions (200 vs. 800, 500 vs. 500 and 800 vs. 200 g per day, respectively) in rotation. Each diet was given for 3 weeks and whole rumen contents were withdrawn prior to the feeding on the last day.

2.2 DNA extraction

Total DNA was extracted from fiber-associated bacteria by a bead-beating method as described previously [8]. Briefly, each hay stem sample (0.35 g) was mixed with 0.35 ml of TE buffer (pH 8.0) and 0.7 ml of Tris-buffered phenol (pH 8.0) in a 2 ml tube containing 0.25 g of glass beads (diameter 425–600 µm; Sigma Chemical). After adding 40 µl of 10% sodium dodecyl sulfate (SDS), the tubes were subjected to shaking three times for 2 min, with 2 min incubation on ice between rounds of shaking. Tubes were centrifuged at 16 000×g for 5 min. DNA in the supernatant was purified by hydroxyapatite chromatography (Hydroxyapatite Bio-Gel HTP Gel, Bio-Rad) followed by gel filtration (MicroSpin S-200 HR Columns, Amersham Pharmacia Biotech).

2.3 16S rDNA amplification and cloning

The 16S rDNA was PCR amplified using primers S-*-Univ-530-a-S-16 (5′-GTGCCAGCMGCCGCGG-3′) and S-*-Univ-1392-a-A-15 (5′-ACGGGCGGTGTG TRC-3′) [9]. PCR tubes containing 0.5 µM each primer, 1.5 mM MgCl2, 0.2 mM each deoxyribonucleotide triphosphate, 1× rTaq DNA buffer, 2.5 U of rTaq DNA polymerase (Toyobo) and 1 µl (∼100 ng) of template DNA in a total volume of 50 µl were prepared. PCR was performed by GeneAmp System 2400 (Perkin-Elmer) using the following program: initial denaturation at 94°C for 30 s, followed by 15 cycles of 30 s denaturation at 94°C, 30 s annealing at 58°C and 1 min extension at 72°C with a final extension at 72°C for 7 min.

PCR products were separated on a 1.5% NuSieve agarose gel (Bio-Whittaker Molecular Applications) containing ethidium bromide. The products were excised from the gel and purified by QIAquick gel extraction kit (Qiagen). Purified PCR products were cloned into pCR2.1 vector (Invitrogen) to transform Escherichia coli INVαF′ (Invitrogen). Transformants from this clone library were randomly selected for extraction of the recombinant plasmids, using Wizard Plus SV Minipreps DNA Purification System (Promega).

2.4 Sequence analysis

The 16S rDNA nucleotide sequences of the inserts in E. coli clones were determined by cycle sequencing using a ThermoSequenase cycle sequencing kit (Amersham) and a DSQ2000L (Shimadzu) automated DNA sequencer.

Sequences retrieved from 16S rDNA libraries of fiber-associated bacteria were compared to the 16S rDNA sequences available in the GenBank database using the BLAST program. The presence of chimeric sequences in the libraries was detected by the CHECK_CHIMERA program from Ribosomal Database Project (RDP) [10]. All reference sequences were obtained from GenBank. Sequences were aligned using the multiple sequence alignment software CLUSTAL W ver.1.81. Phylogenetic analysis was performed using the neighbor-joining method [11]. Gaps were excluded from the phylogenetic analysis. To statistically evaluate the branching, bootstrap analysis was carried out with data resampled 1000 times.

2.5 Primer design and methods for uncultured bacterial clones

PCR primers U3-F (5′-TAAACCCTCGGTGCCGTC-3′) and U3-R (5′-CTTCCCTTTGTCACCGCC-3′) were newly designed for the unknown group 3 (see Section 3) consisting of uncultured bacterial clones AAB7, AAB13 and AAB24. The primer sequences represented a sequence identical in all three bacteria of the unknown group 3, but are different from sequence in related bacteria. To evaluate primer specificity, DNA extracted from the ruminally incubated alfalfa hay used for constructing clone library was PCR amplified using this primer set with the conditions as follows: initial denaturation at 94°C for 30 s, followed by 30 cycles of 30 s denaturation at 94°C, 30 s annealing at 58°C and 30 s extension at 72°C with a final extension at 72°C for 10 min. The PCR products were separated on an agarose gel to confirm specificity of the amplification, and then subjected to sequence analysis. Other PCR products with the designed primer set and serially diluted DNA from the rumen digesta of sheep on three different alfalfa diets were run on an agarose gel to estimate prevalence of the targeted uncultured bacteria. Then, 1 µl inclusion of undiluted template DNA in PCR allowed a threshold of detection of 104 copies of the targeted 16S rDNA per g of rumen digesta. This indicated that the detection at a dilution of 10−3 is, for instance, equivalent to 107 copies of the 16S rDNA per g.

3 Results

3.1 Sequence analysis

Partial 16S rDNA sequences of 92 random clones from fiber-associated bacteria were analyzed. A sequence from one clone (AAB9) seemed to be chimeric; therefore, this clone was excluded from further analysis. Nomenclature and nucleotide sequence accession numbers of the clones were as follows: for the orchardgrass-associated bacterial library, clone names begin with ‘OAB’ followed by the clone number. Clone names in the alfalfa-associated bacterial library begin with ‘AAB’ followed by the clone number. Nucleotide sequences have been deposited in GenBank under the accession numbers AB045738 to AB045746, AB056633 to AB056649 and AB113670 to AB113735.

Based on a 97% sequence similarity criterion [12], 21 of the 91 clones in our libraries were considered to represent known bacterial species. About 35% of the sequences (32 clones) were 90–96% similar to a database sequence (Table 1), and the remaining 42% (38 clones) had identity values less than 90%. Of the clones having more than 97% similarity with known bacterial species, 12 belonged to Butyrivibrio fibrisolvens, whereas two were identified as F. succinogenes. For the clones with 90–96% similarity to known sequences, the nearest known bacterial species were different for bacteria isolated from orchardgrass versus bacteria isolated from alfalfa (Table 1), e.g. Treponema was seen only in the clone library from alfalfa.

View this table:

Known bacteria sharing more than 90% similarity in 16S rDNA sequence with fiber-associated bacterial clones

Nearest known bacteriaNumber of clones from different hay sources
Orchardgrass-associated clonesAlfalfa-associated clones
B. fibrisolvens8-41
P. ruminicola16-4
F. succinogenes1-1-
S. ruminantium21--
S. succinovorans2---
S. dextrinosolvens1---
P. ruminis1---
Prevotella brevis-3--
Selenomonas sputigena-2--
Eubacterium ramulus-1--
Eubacterium oxidoreducens-1--
T. bryantii---5
C. aldrichii---2
R. flavefaciens---1
Anaeroplasma bactoclacticum---1
Streptococcus hansenii---1
Clostridium populeti---1
Clostridium leptum---1
Desulfovibrio fairfieldensis---1
  • Boldface shows representative ruminal species.

  • Percentage shows sequence similarity.

3.2 Phylogenetic analysis

The results of phylogenetic analysis are shown in Fig. 1. The majority of clones sequenced in the present study fell into the CytophagaFlavobacterBacteroides (CFB) and low G+C Gram-positive bacterial (LGCGPB) phyla (43 and 44%, respectively). Within the CFB phylum, a number of sequences were placed into the Prevotella group, while 11 clones formed a novel cluster not affiliated with any known bacterial isolates (unknown group 1, Fig. 1). Branching of this unknown group from Bacteroides distasonis was supported by an 84.7% bootstrap value.


Phylogenetic placement of 16S rDNA sequences of orchardgrass-associated bacteria (begin with OAB followed by a clone number) and alfalfa-associated bacteria (begin with AAB followed by a clone number). Sequences of clones obtained in the present study are shown in boldface. Bootstrap values for 1000 trees are shown at branch points. Only values of 80% or more are shown. The bar represents nucleotide substitutions per sequence position.

Another two unknown groups (unknown groups 2 and 3) branched from fibrolytic species Clostridium aldrichii and Eubacterium ruminantium within the LGCGPB phylum. Other clones were affiliated with known bacterial species such as Selenomonas ruminantium, Schwartzia succinovorans and B. fibrisolvens, showing more than 97% similarity. Eight sequences, all from the alfalfa library, fell into the Spirochaetes phylum.

3.3 Analysis of unknown group 3 bacteria under different diets

To characterize the prevalence of bacteria of unknown group 3 under different feed conditions, we designed primers from the 16S rDNA sequences of fiber-associated bacteria AAB7, AAB13 and AAB24 of this group. First, to verify the specificity of the primers, DNA retrieved from incubated alfalfa was PCR amplified and cloned to construct a mini-library. The sequence similarities among seven randomly chosen clones are shown in Table 2. All seven clones had more than 97% similarity with the sequences of AAB7 and AAB13, while identity values with AAB24 were less than 97% except for one clone. Identity values with RC36 and 5C3d-11, the nearest relatives in GenBank, were 96% for all clones. Therefore, the primers show specificity for AAB7 and AAB13, judging by the criterion of ≥97% similarity.

View this table:

Sequence identity between clones in a mini-library generated by U3-F and U3-R primers targeting uncultured fiber-associated bacteria AAB7, AAB13 and AAB24 that were retrieved from in sacco incubated alfalfa stem

Clone No.Similarity
  • Uncultured rumen bacterial clone from cattle (Tajima et al., 1999 [1]).

  • Uncultured rumen bacterial clone from cattle (Tajima et al., 2000 [21]).

The primers for unknown group 3 were then used to examine ruminal bacteria from sheep fed diets with different proportions of alfalfa. The semi-quantitation of the targeted uncultured bacteria in sheep rumen digesta is shown in Fig. 2. Maximal dilutions of the template DNA to detect the uncultured bacteria in the rumen of two sheep receiving diets with 80, 50 and 20% alfalfa hay were 10−3, 10−2 and 10−1, respectively. These dilutions indicated that 107, 106 and 105 copies of the 16S rDNA of the uncultured bacteria were present in the rumen receiving the respective diet.


Semi-quantitative PCR detection of uncultured bacteria within the unknown group 3 cluster in the rumen of sheep fed three diets containing different proportions of alfalfa hay. DNA extracted from rumen samples was serially diluted and used as PCR template. PCR was carried out with the newly designed primer set which targeted unknown group 3 designated in the present study. Levels of 16S rDNA in each sample were calculated from maximal dilution level (see Section 2). The results from only one sheep are presented in the figure, since the remaining sheep showed similar results. A: Alfalfa hay, 800 g per day; concentrate, 200 g per day. B: Alfalfa hay, 500 g per day; concentrate, 500 g per day. C: Alfalfa hay, 200 g per day; concentrate, 800 g per day.

4 Discussion

Previous studies with molecular biological approaches have partially described microbial flora in the rumen [1,2,13]. Here, we focused on a fiber-associated microflora to study the bacterial populations relating to fiber degradation. Of the 21 clones with more than 97% similarity to known isolates, 12 clones belonged to B. fibrisolvens (Table 1). Although phenotypic and phylogenetic diversity of B. fibrisolvens has been demonstrated [14], some strains of this species are known as cellulolytic and xylanolytic bacteria [15]. Therefore, the abundance of B. fibrisolvens-affiliated clones in the present study suggests the importance of this species for fiber degradation in the rumen. To prove this assumption, however, the functions of these cloned bacteria need to be assessed.

Although only one clone had more than 97% similarity to Prevotella ruminicola, many clones fell into the genus Prevotella (Table 1, Fig. 1). Prevotella species are recognized as the major groups in the rumen [15], and some Prevotella strains from the rumen possess endoglucanase and xylanase activities [16,17] but do not degrade cell wall material extensively in pure culture [15]. The isolation of many Prevotella-like clones from ruminally incubated hay stems suggests indirect concern of these bacteria with ruminal fiber breakdown possibly as oligosaccharide and xylan fermenters.

Again, there is no direct evidence that the above detected clones belonging to Butyrivibrio and Prevotella are functional on hay stems. However, since the hay stems were thoroughly washed before DNA extraction, the remaining bacteria might be specific on the stems. With regard to fibrolytic bacteria, clones affiliated to R. flavefaciens were found only from rumen solid but not from liquid [1]. The roles of these tightly attached bacteria retrieved as 16S rDNA clones should be obviously elucidated with efforts to cultivate them.

Positive interactions for cellulose degradation have been observed between cellulolytic and non-cellulolytic bacteria in vitro [18,19]. In the present study, six clones were affiliated with non-cellulolytic rumen bacteria, such as S. ruminantium, Succinovibrio dextrinosolvens, S. succinovorans and Pseudobutyrivibrio ruminis (Table 1). The results may indicate that the fiber-associated bacterial community consists of not only fibrolytic species but also non-fibrolytic species, and fiber degradation would be accelerated by interactions between these fibrolytic and non-fibrolytic bacteria. One known example of such an interaction is cross feeding of hydrolysis products between Selenomonas and Fibrobacter [20]. Some strains of S. ruminantium have CMCase but do not digest cellulose by themselves. The potential for the synergism of S. ruminantium with F. succinogenes is suggested in the present study (Table 1, Fig. 1). In fact, we have successfully isolated many S. ruminantium strains with high CMCase and fiber-attaching activities from ruminally incubated orchardgrass, one of which is phylogenetically close (98% identity) to the present clone OAB36 (Sawanon et al., unpublished results).

At present, F. succinogenes, R. flavefaciens and R. albus are considered as the representative cellulolytic species in the rumen [7]. However, only one clone belonging to F. succinogenes has been obtained among 318 sequences from rumen digesta in the reported studies [1,2,21]. Among the present 91 sequences, two sequences were assigned to F. succinogenes (Table 1). This frequency in retrieving F. succinogenes indicates that, for analyzing fibrolytic consortia, hay stems might be a good source of bacteria as rumen solids [1,2]. We have already confirmed the abundance of F. succinogenes in whole rumen digesta and hay stems suspended in the rumen of sheep [8].

The set of bacteria isolated from orchardgrass seemed to be different from the set isolated from alfalfa (Table 1). Although experimental conditions such as animals and incubation times were not identical in preparation of the clone libraries from orchardgrass and alfalfa, the hay source may affect the members of fiber-associated rumen bacterial community, e.g. sequences sharing 90–96% similarity with Treponema bryantii were frequently seen in the library from alfalfa, but not at all in the library from orchardgrass (Table 1, Fig. 1). The saccharolytic spirochete T. bryantii has been shown to associate with F. succinogenes, and T. bryantii could enhance cellulose degradation in coculture with F. succinogenes [19]. In addition, the high protein content of alfalfa would affect associated bacterial community, preferentially stimulating growth of proteolytic species such as B. fibrisolvens [15].

The presence of uncultured bacteria has been recognized in the rumen [1] and in the hindgut of monogastric animals [2224]. In the present study, 42% of clones had low similarity (<90%) to known isolates. Therefore, the fiber-associated bacterial community appears to include a considerable proportion of unknown bacteria, as observed in the rumen digesta [1]. Daly et al. [25] carried out quantitative analysis of equine colonic microflora and indicated that the abundance of unknown bacteria was similar to that of known bacteria. Our phylogenetic analysis showed three unique assemblages of unknown sequences within the CFB and LGCGP phyla (Fig. 1). Judging from phylogenetic placement, unknown group 1, located in the CFB phylum, seems to belong to the Bacteroides genus (Fig. 1). Unknown Bacteroides lineages were also found in previous studies [1,26]. Ramsak et al. [26] revealed the high degree of genetic diversity of ruminal Bacteroides. These observations emphasize the need to develop a methodology to isolate and cultivate unknown gut bacteria for characterizing their functions.

There were two unknown groups in the LGCGPB phylum (Fig. 1); these groups branched from the fibrolytic species C. aldrichii and E. ruminantium. Although 16S rDNA sequence does not determine bacterial phenotype, the similarity to fibrolytic species and the origin on ruminally incubated hay stems may argue that these unknown groups are related to digestion of hay stem components.

Uncultured ruminal bacteria of unknown group 3, which branches from the ruminal xylanolytic bacterium E. ruminantium, were preliminarily investigated in the present study. Sequencing analysis of a mini clone library from the PCR products with a new primer set and DNA from ruminally incubated alfalfa demonstrates that the primers satisfy the specificity for the targeted clones AAB7 and AAB13 judging from the criterion of ≥97% similarity (Table 2). This primer set was used for monitoring the targeted uncultured bacteria, unknown group 3, in the rumen during the diet shift from high alfalfa diet to low alfalfa diet. The uncultured bacteria showed a clear response to this diet shift: they increased their ruminal levels with increasing dietary alfalfa level (Fig. 2). Although the semi-quantitative PCR with serially diluted template DNA did not allow accurate quantitation of the targeted uncultured bacteria, the changes in unknown group 3 bacterial levels with diet suggest that these bacteria are directly or indirectly involved in alfalfa digestion in the rumen. Further investigation with isolation of the bacteria as live organisms is necessary to assess the significance of the uncultured bacteria in alfalfa digestion.

In conclusion, we determined bacterial members of the fiber-associated microflora by 16S rDNA sequencing. The use of ruminally incubated hay stems was an effective approach for retrieving the fiber-associated bacteria. The community consisted of known fibrolytic and non-fibrolytic bacteria, and Prevotella-related and B. fibrisolvens-related sequences formed large clusters in the phylogenetic tree. Unknown sequences fell into some assemblages, one of which, unknown group 3, appeared to be more prevalent in the rumen when a high alfalfa diet was given. These results suggest the importance of analyzing fibrolytic consortia, as they consist of undescribed bacteria in a high proportion.


  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].
View Abstract