Phylogenetic Analysis of 16S rDNA Sequences Manifest Rumen Bacterial Diversity in Gayals (Bos frontalis) Fed Fresh Bamboo Leaves and Twigs (Sinarumdinaria)

Six male Gayal (Bos frontalis), approximately two years of age and with a mean live weight of 203±17 kg (mean±standard deviation), were housed indoors in metabolism cages and fed bamboo (Sinarundinaria) leaves and twigs. After an adjustment period of 24 days of feeding the diet, samples of rumen liquor were obtained for analyses of bacteria in the liquor. The diversity of rumen bacteria was investigated by constructing a 16S rDNA clone library. A total of 147 clones, comprising nearly full length sequences (with a mean length of 1.5 kb) were sequenced and submitted to an on-line similarity search and phylogenetic analysis. Using the criterion of 97% or greater similarity with the sequences of known bacteria, 17 clones were identified as Ruminococcus albus, Butyrivibrio fibrosolvens, Quinella ovalis, Clostridium symbiosium, Succiniclasticum ruminis, Selenomonas ruminantium and Allisonella histaminiformans, respectively. A further 22 clones shared similarity ranging from 90-97% with known bacteria but the similarity in sequences for the remaining 109 clones was less than 90% of those of known bacteria. Using a phylogenetic analysis it was found that the majority of the clones identified (57.1%) were located in the low G+C subdivision, with most of the remainder (42.2% of clones) located in the Cytophage-Flexibacter-Bacteroides (CFB) phylum and one clone (0.7%) was identified as a Spirochaete. It was apparent that Gayal have a large and diverse range of bacteria in the rumen liquor which differ from those of cattle and other ruminants. This may explain the greater live weights of Gayal, compared to cattle, grazing in the harsh natural environments in which Gayal are located naturally. (


INTRODUCTION
The Gayal or Mithun (Bos frontalis) is a rare semi-wild bovine species distributed throughout Bangladesh, Bhutan, China, India, Malaysia and Myanmar (Mondal et al., 2004;Rajkhowa et al., 2006).They belong to the family Bovidae (Giasuddin and Islam, 2003) and have adapted to harsh environmental conditions attaining higher live weights than cattle kept in similar environments (Cheng, 1984;Giasuddin et al., 2003;Mao et al., 2005).In these environments Gayal consume bamboo, reeds and woody plants which are not consumed by other local ruminants.
Ruminants are able to digest a range of plant materials by virtue of the extensive population of microbes, including bacteria, fungi and protozoa, which are found in the rumen, and provide the host with nutrients predominantly in the form of volatile fatty acids and microbial protein (Van Soest, 1994).The rumen bacteria have been shown by traditional procedures to belong to some 22 predominant species (Krause and Russell, 1996) but based on recent observations it is clear that this prior estimation represents only a small portion of the total diverse population that may colonise the rumen (Amann et al., 1995;Von Wintzingerode et al., 1997;Zoetendal et al., 1998;Krause et al., 2003;Shin et al., 2004a).
Molecular techniques based 16S/18S ribosomal RNA (rRNA) and their encoding genes (rDNA) have been used extensively to study rumen microbes both quantitatively (Stahl et al., 1988;Koike et al., 2003a;Sylvester et al., 2004) and quantitatively (Kocherginskaya et al., 2001;Shin et al., 2004b;Regensbogenova et al., 2004).These techniques have been used to construct a library of 16S rDNA clones of rumen microbes and to demonstrate the considerable diversity of rumen bacteria (White et al., 1999).
The present study was conducted to examine the diversity of rumen bacteria in lured Gayal offered a diet typical of that consumed under natural conditions.The molecular techniques now available were used to construct a library of 16S rDNA clones of rumen bacteria, and a phylogenetic tree for the clones isolated.

Animals, diet and collection of rumen liquor
A herd of Gayal was lured into the Gayal Research Station (located at N 25°47′02.6″,E 099°05′56.5″at an altitude of 2,260m above sea level) using salt.Six male Gayal approximately two years of age and with a live weight of 203±17.0kg (mean±standard deviation) were selected from the herd and confined to metabolism cages in an open shed lit naturally.The animals were allowed to adjust to the environment for 24 days.They were fed fresh bamboo leaves and twigs (Sinarundinaria sp.) typical of the forage grazed naturally, which were harvested daily and offered ad libitum in equal portions at c 08:00 and 18:00 h.
At the end of the adjustment period representative portions of the diet offered were collected each day over four days for determination of chemical composition using the procedures published by Goering and Van Soest (1970) and the Association of Official Analytical Chemists (AOAC, 1990).The components measured were dry matter (DM), crude protein (CP), crude fibre (CF), ether extract (EE), ash, neutral detergent fibre (NDF), acid detergent fibre (ADF) and acid detergent lignin (ADL).
Rumen liquor was collected on day 25 immediately before morning feeding.Approximately 500 ml of liquor was collected via a stomach tube located in the mid part of the rumen and connected to a vacuum pump (Yuangklang et al., 2005;Khampa et al., 2006a).The first 100 ml collected were discarded and the remaining liquor was passed through four layers of cheese cloth to remove particulate matter.The filtered liquor was transferred to previously sterilised tubes which were snap frozen in liquid nitrogen then stored at -80°C pending extraction of DNA.

Extraction of DNA
The frozen samples of rumen liquor were thawed and equal portions of the liquor from the six animals were pooled to obviate differences which may have arisen because of animal-to-animal variation.Total DNA in the pooled sample then was extracted as follows.An aliquot of 0.5 ml of homogenised rumen liquor was added to a tube (2 ml capacity) containing 0.5 g of sterilised glass beads (425-600 µm diameter: Sigma Chemical Company, USA) which then was centrifuged (10 minutes, 12,000 rpm, 4°C).The pellet was rinsed twice, with 0.5 ml saline-EDTA (0.15 M NaCl and 0.5 M EDTA, pH 8.0) followed by centrifugation (10 minutes, 12,000 rpm, 4°C), before re-suspension in 0.5 ml saline-EDTA.Mechanical disruption was effected using a Mini-Bead Beater TM Mill (Cole-Parmer International, USA) at maximum speed for 3 minutes on 3 occasions separated by periods of 2 minutes of chilling on ice.The bacterial DNA then was extracted using a Bacteria DNA Mini-Kit (Watson Biotechnologies Inc., Shanghai, P. R. China) in accordance with the manufacturer's guidelines.Finally, the pellet of DNA was dried then dissolved in TE buffer (pH 8.0) containing DNase-free RNase (100 µg/ml), prior to storage at -20°C.

PCR primers and amplification
Amplification of the 16S rDNA was achieved using: • a forward primer F27 (5'-AGAGTTTGATCMTGGCTCAG-3'); and • a reverse primer R1492 (5'-TAGGYTACCTTGTTACGACT-3').A total of 20 µl of reaction mixture consisted of 20 pmol of each primer, c 100 ng of template DNA, 2 µl of 10× reaction buffer (Takara), 0.1 mM of each dNTP (Roche) and 2.5 U of Taq DNA polymerase (Promega).The PCR amplification was performed by Gene Thermal Cycle (Hangzhou, P. R. China) using the following program: denaturing at 95°C for 5 minutes, followed by 15 cycles of 30 seconds of denaturing at 95°C, 30 seconds of annealing at 55°C and 2 minutes of elongation at 72°C with a final extension at 72°C for 7 minutes.The PCR products were separated by electrophoresis in 1.0% (w:v) agarose gel and stained with ethidium bromide.Products were excised from the gel and purified with UNIQ-10 column DNA Gel Extraction Kits (Sangoo, Shanghai, P. R. China) in accordance with the directions of the manufacturer.

Cloning and sequencing
The purified PCR products were transferred directly into the TA Cloning ® Kit with One Shot ® TOP10 competent cells (Invitrogen, San Diego, USA) according to the protocol of the manufacturer.The recombinant plasmids then were extracted by the alkaline lysis mini-prep procedure (Birnboim and Doly, 1979) using a Mini-Plasmid Rapid Isolation Kit (BioDev-Tech, Beijing, P. R. China).The 16S rDNA was sequenced using an ABI PRISM Big Dye Terminator Cycle Ready Reaction Kit and an ABI PRISM 3730 DNA Sequencer (Applied Biosystems) by Invitrogen Co. (Shanghai, P. R. China).

Sequence analyses and phylogenetic tree constructing
All reference sequences were obtained from the GenBank/EMBL/DDBJ/RDP.Sequences from the current study were analysed by the CHECK_CHIMERA program (Maidak et al., 2001) to remove any chimeric rDNA clone.Similarity with data base entries was assessed using an online BLAST search (Madden et al., 1996).Sequence alignment was achieved using multiple sequence alignment software CLUSTAL W Version 1.81 (Thompson et al., 1994).The criterion used to define a clone sequence as being that for a particular species of rumen bacteria was that the similarity of the sequence should be 97% or greater with that of the known species (Stackebrandt and Goebel, 1994).The phylogenetic tree was constructed by the neighbourjoining method (Saito and Nei, 1987) and the tree was evaluated using the bootstrap test based on 1,000 resamplings (Felsenstein, 1985) based on the recommendation of Paster et al. (1991).

Nomenclature and nucleotide sequence accession numbers
The prefix GRC was used to denote clones identified.All nucleotide sequences have been deposited in GenBank under the accession numbers DQ673466-DQ673612.

Sequence similarity
A total of 147 16S rDNA clones of nearly full length and averaging 1,500 bp were isolated from the rumen liquor of the Gayal.These clones were classified into 86 Operational Taxonomy Units (OTU) (Table 1).

Phylogenetic analysis
Because the similarity for most of the sequences with those of known rumen bacteria was too low to identify the sequence as representing a particular taxon, a phylogenetic tree was constructed to investigate the taxonomic placement.The results of this phylogenetic analysis are shown in Figure 1.The majority of the clones (57.1%) were positioned within the low G+C subdivision.Most of the remaining clones (42.2%) were classified as belonging within the Cytophage-Flexibacter-Bacteroides (CFB) phylum and only one clone (GRC40), comprising 0.7% of the total, was classified as belonging within the Spirochaets.
Within the low G+C phylum seven OTU formed a novel cluster in which Q. ovalis was located.The bootstrap value for the cluster was 92%.A further eight OTU clustered with Ruminococcus bromii, R. flavifaciens and R. albus, three OTU were related to Eubacterium celluloselvens, nine were affiliated with B. fibrisolvens, and three were clustered with R. gnavus and Clostridium fusiformis.There were, however, groups of three, two, four, six and three OTU which formed five independent clusters which were not linked to any known rumen bacteria.A further two OTU (GRC25 and GRC66) constituted a new cluster deeply branched in the low G+C phylum not related to any known bacterium even though the bootstrap confidence level was 100% (Figure 1).
In the CFB phylum only two OTU clustered with Prevotella spp., sharing 94 and 95% similarity and having a bootstrap value of 100% indicating the stability of the group.Four and a separate single OTU constituted new clusters located between Cytophaga hutchinsoni and C. fermentans.The remaining 17 OTU constituted three separate clusters which were remotely associated with Bacteroides spp.and Prevotella spp.(Figure 1).
A single clone (GRC40) located within the Spirochaetes shared 93% similarity with Treponema bryanti and had a bootstrap value of 100%.

DISCUSSION
It is clear that the rumen of the Gayal contained an extensive and diverse array of bacterial species.The majority of the species (57.1%) were classified as belonging to the low G+C subdivision with most of the remaining species (42.2%) being located in the CFB subdivision.Only one organism classified as a Spirochaete was identified and a Clones having ≥97% similarity of 16S rDNA with a known species were referred to the bacterial names.
b Clones having >97% similarity of 16S rDNA among each other were defined as 1 operational taxonomy unit (OTU).c Clones having <90% similarity of 16S rDNA with a known species and with >90% similarities among each other were defined as an uncultured group.
no Proteobacteria were identified.The proportions of bacteria located on the low G+C and CFB phyla for the Gayal, whilst exceeding those previously reported, approximated the proportions which have been reported.Thus Tajima et al. (1999) reported that 52.4% of 42 bacteria identified in the rumen liquor of Holstein cow (Bos taurus) fed a diet of hay belonged to the low G+C phylum and 38.1% to the CFB phylum.Similarly, An et al. (2005) reported that of the 194 clones isolated from the rumen liquor of yak (Bos grunniens) grazing on the Qinghai-Tibetan Plateau, 54.1% were low G+C organisms and 30.9% were CFB organisms.Moreover, Edwards et al. (2004) in summarising the published data for rumen bacteria reported that on average 54% of rumen bacteria were members of the low G+C phylum and 40% were from the CFB phylum.
Amongst the bacteria comprising the CFB subdivision, only two OTU (GRC48 and GRC51) were closely related to Prevotella runinicola which has been identified previously as the most numerous of the bacteria in the rumen (Stewart et al., 1997).The remaining 22 OTU located in the CFB subdivision was clustered loosely and their identification requires further investigation.Overall, the present results were consistent with previous reports that CFB bacteria were less numerous in the rumen in wild animals (Nelson et al., 2003;An et al., 2005) compared to domestic cattle.In connection with the latter, Whitford et al. (1998) reported a predominance (101 of 133 clones) of 16S rDNA sequences similar to that for P. ruminacola in isolated material from domestic cattle.
The majority (59 of 86 or 68.6% of the OTU, comprising 109 of 147 or 74.1% of clones) of the bacteria identified in the present study had a similarity in sequence of 90% or less compared to known rumen bacterial species.Similar observations have been reported by Nelson et al. (2003) who found that only 24 of 279 clones isolated from the digestive tracts of Zebu cattle, zebra (Equus quagga) and three wild antelope (Eland, Tauotragus oryx; Thompson's gazelle; and Grant's gazelle, Gazella granti) were similar to known species of bacteria.An et al. (2005) reported that 49 of 91 (53.9%)OTU, comprising 106 of 194 (54.6%) clones, isolated from the rumen of the yak and 27 of 88 (30.7%)OTU consisting 68 of 197 (34.5%) clones isolated from Chinese Yellow Cattle were similar to identified species.On contrast, for domestic sheep fed orchardgrass or alfalfa hay diet (CP 11.6 or 16.5%, NDF 68.2 or 45.5%), Koike et al. (2003b) cloned 91 16S rDNA sequences from the fibre-associated bacteria in the rumen and found that 38 of these (41.8%) were less than 90% similar to known species.As a result, our finds further supported the result as shown by Kobayashi (2006), who pointed out that rumen samples of wild and semi-wild ruminants could contain novel bacteria to a greater extent than those of domestic counterpart.
It might be argued that the variations in the species of bacteria in the rumen are affected by diet.Indeed, Latham et al. (1972) and Tajima et al. (2001) have shown that diet affects the composition of the microbial population of the rumen.However, differences between the present observations and those published previously are considered to be related, at least in part, to animal species.For example, Q. ovalis and A. histaminiformans, which have been associated with sheep fed diets high in readily fermented carbohydrates (Krumholz et al., 1993) and cattle and horses (Garner et al., 2002), were retrieved from the rumen liquor of Gayal in the present study.However, these two kinds of bacteria were not emerged in bacterial 16S rDNA clone libraries of rumen fluid from cow fed hay diet, transition diet and high-grain diet (Tajima et al., 2000).Furthermore, another species commonly found in the rumen of ruminants fed grain based diets, S. ruminatium (Stewart et al., 1997), was identified in the rumen liquor of the Gayal, but the starch digesting bacterium Ruminobacter amylophilus (Stewart et al., 1997) was not identified in the present study.Moreover, in previous comparative studies between Gayal and Yunnan Yellow Cattle (Bos taurus) fed pelleted lucerne (Medicago sativum), Deng et al. (2007) reported that there was a significantly higher population of total as well as cellulolytic and amylolytic bacteria compared to Yunnan Yellow Cattle, which was associated with higher concentrations of volatile fatty acids in the rumen of the Gayal compared to the cattle.Further evidence that Gayal and Yunnan Yellow Cattle have different populations of rumen bacteria has been obtained in vitro where enhanced gas production and digestibility of DM have been measured for the Gayal under same condition (Xi et al., 2007).
In interpreting the results of the present study some caution needs to be expressed because of possible flaws in the analytical techniques utilised.As noted by Von Wintzergerode et al. (1997), technical biases during amplification of bacterial 16S rDNA from total DNA should be considered.First, bacteria for which genomic DNA has high G+C contents may be less readily amplified in the PCR process with universal primers (Reysenbach et al., 1992).Indeed Tajima et al. (2001) confirmed this by amplifying F. succinogenes (not identified in the present study) and showing a prolonged lag phase as compared to other bacteria under identical amplification conditions.Secondly, genome size and the copies of rrn genes also affect the amount of PCR amplification products which will be used as the target for cloning, even though the cloning step is random (Farrelly et al., 1995).Thirdly, the number of PCR cycles influences the bacterial diversity (Suzuki and Giovannoni, 1996) and to overcome this effect the number of PCR cycles employed was restricted to 15 with the view to improving the efficiency of cloning.

CONCLUSIONS AND RECOMMENDATIONS
In spite of these possible deficiencies in analytical techniques, it is concluded that Gayal have a large and diverse range of bacteria in the rumen.The present results are in conformity with the observations of Deng et al. (2007) that rumen liquor of Gayal contains greater numbers of total bacteria than similarly fed cattle.This would offer an explanation for the greater live weights of Gayal, compared to cattle, under field conditions.Clearly, further studies should be undertaken to confirm and extend the present observations.

Figure 1 .
Figure 1.(i) Phylogenetic tree of 16S rDNA sequences from the Gayal rumen clones (begin with GRC followed by OTU number).The Methanobacteriun formicicum (AY196659) and Methanobrevibacter ruminantium (AY196666) are used as the out-group for rooting the tree.Numbers in parentheses represent the sequence accession numbers in GenBank.Numbers in square brackets indicate the clone number out of the total clones.The number at each branch points is the percentage supported by 1000 bootstrap trees.Only values of 60% or above are shown.The bar represents 5% sequence divergence.

Figure 1 .
Figure 1.(ii) Phylogenetic tree of 16S rDNA sequences from the Gayal rumen clones (begin with GRC followed by OTU number).The Methanobacteriun formicicum (AY196659) and Methanobrevibacter ruminantium (AY196666) are used as the out-group for rooting the tree.Numbers in parentheses represent the sequence accession numbers in GenBank.Numbers in square brackets indicate the clone number out of the total clones.The number at each branch points is the percentage supported by 1000 bootstrap trees.Only values of 60% or above are shown.The bar represents 5% sequence divergence.

Table 1 .
(i) Similarity values of 16S rDNA sequences retrieved from the rumen fluid of Gayal OTU a Phylum Clones (entries) Nearest relative (GenBank accession No) Similarity

Table 2 .
Distributions of 16S rDNA clones and operational taxonomy units (OTU) retrieved from the rumen fluid of Gayal