Specific PCR Assay for a Tannin-Tolerant Selenomonas ruminantium Isolate, Derived from Helicase Coding Sequences

Abstract

Sequences from a tannin-tolerant Selenomonas ruminantium isolate (EAT2) that hydrolyzes gallic acid were identified. Two exhibited identity to helicases with a wide phylogenetic distribution. PCR amplification by using primers from one helicase gene detected 2,000 to 5,000 EAT2 genome equivalents but did not amplify total gastrointestinal microbial DNA of nine other ungulate species.

Tannins diminish protein and dry matter digestibility in ruminants (4). The tolerance of certain mammalian species for hydrolyzable tannins may be attributable to tannin-resistant bacteria in the rumen. Tannin-tolerant bacteria have been isolated from Enterobacteriaceae and the genera Selenomonas and Streptococcus (7, 9, 12). We have previously reported the characterization of Selenomonas ruminantium isolates, capable of growing in 70 g of tannic acid per liter, from African ruminants, bush duiker (Sylvicapra grimmia), goat (Capra hircus), and sheep (Aries ova). Five of these isolates were identified as S. ruminantium on the basis of fermentation of carbohydrates and of comparison of the 16S ribosomal DNA sequences to those of reference isolates (9, 10). Due to our interest in transferring tannin-tolerant bacteria from wildlife into domestic ruminants to improve their ability to digest tannin-rich feeds, we isolated DNA sequences unique to individual S. ruminantium isolates for development of probes to monitor persistence in the rumen following inoculation into novel hosts. We used genome-wide suppression subtractive hybridization (SSH) PCR (2, 3) as implemented in the PCR-Select Bacterial Genome Subtraction Kit (Clontech). S. ruminantium EAT2 from bush duiker was selected as the “tester” because it exhibited the highest levels of tannin tolerance and unlike the other isolates could hydrolyze gallic acid (9). The “driver” DNA was from an Ethiopian goat isolate (EG19) genetically closely related to EAT2 (10). Nine sequences were obtained by applying SSH to genomic DNA according to the manufacturer's instructions, except that the first and second subtractive hybridizations were performed at 62°C instead of 63°C, and were cloned into a pGem T vector (Promega). The sequences were evaluated for specificity by hybridization to slot blots of DNA (100 ng) from five tannin-tolerant and two S. ruminantium reference isolates. At a washing stringency of 0.5× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-0.1% sodium dodecyl sulfate, all nine cloned sequences hybridized most strongly to EAT2 and variably to the other six isolates. Two of the nine sequences hybridized specifically to EAT2 DNA. Results for clone number 1 are shown in Fig. 1A.

FIG. 1.

Specificity of a helicase-derived sequence for a bush duiker isolate (EAT2) of S. ruminantium. (A) Slot blot of total genomic DNA (100 ng) from seven different S. ruminantium isolates hybridized with a cloned sequence from EAT2 isolated by SSH. The slots are 1, EG 4.2; 2, EAT2; 3, reference isolate D; 4, ES 3; 5, EG19; 6, reference isolate 108; and 7, ES 4. (B) Agarose gel of PCR products generated from S. ruminantium genomic DNA by using primers derived from a cloned helicase-like sequence from EAT2. Lanes: M, size standard; 1, EG 19; 2, EG 4; 3, ES 4; 4, ES 3; 5, reference isolate 108; 6, reference isolate D; 7, EAT2.

The nucleotide sequences of the two EAT2-specific sequences and of a third clone that hybridized more strongly to EAT2 than to other isolates were determined. Blast X searches of the NCBI databases revealed that clones 1 (524 bp) and 3 (941 bp) contained open reading frames (ORFs) with significant amino acid identity to different mobile element or prophage-encoded ORFs. For clone 1 there was very high identity to a superfamily II RNA/DNA helicase gene of Lactobacillus gasseri (accession no. ZP_00046684) and a phage-associate helicase of Streptococcus pyogenes (1) (NP_665244). These corresponded to 72% identity over 123 amino acids (aa) and 56% identity over 122 aa, respectively (E values: 2e-47 and 1e-33). The most significant match of clone 3 was with a transposase of Fusobacterium nucleatum (5) (NP_602338), corresponding to 38% identity over 270 aa (4e-42). Prophages have been described previously from rumen bacteria, including S. ruminantium (6). Our data indicate that coding sequences within prophages can have a wide phylogenetic distribution among different genera but can also be quite restricted among isolates within a single species. The sequence of clone 2 (722 bp) exhibited identity with ORFs of human endosymbionts, the highest identity being to a conserved hypothetical protein of Bacteroides thetaiotaomicron (13) (NP_809516).

To investigate the potential of these sequences for monitoring bacteria introduced into the rumen, PCR primers for amplification of approximately 200 bp from clones 1 and 2 were designed by using Primer3 software (http://frodo.wi.mit.edu/). The sequence 1 primers were (5′ TCACTTGGAGCCTGGAACTT 3′) and (5′ TCATTTCTGCTCCCCTCCTA 3′). Sequence 2 primers were (5′ GGTACAAAAGCTGGGCATGT 3′) and (5′ GTCAGCCTGCCTGGTATCAT 3′). These primers were applied to DNA (50 ng) from seven S. ruminantium isolates (10) in a standard PCR by using Taq polymerase (Promega) in the manufacturer's buffer under the following parameters: 1 min at 94°C, 1 min at 55°C, and 1 min at 72°C for 30 cycles, followed by a 5-min extension at 72°C. Both sets of primers generated products only from EAT2 DNA. Representative results are shown for the PCR primers derived from clone 1 in Fig. 1B. The primers were also applied to total DNA that had been prepared from the gastrointestinal contents of wild and domestic ungulates by a physical disruption method with zirconia silicon beads and a bead beater (11), collected as described previously (8, 9, 10). Neither set of primers generated any products from 300 ng of total DNA prepared from 10 μl of rumen fluid from five different Bos indicus cattle. However, a product was generated from EAT2 genomic DNA mixed with 300 ng of rumen fluid DNA. Representative results from using the clone 1 primers are shown in Fig. 2A. A similar PCR was performed by using total gastrointestinal fluid DNA from one sheep, one goat, and six wildife species, Thompson's (Gazella thompsoni) and Grant's (Gazella granti) gazelles, eland (Taurotragus oryx), impala (Aepyceros melampus), Coke's hartebeest (Alcephalus busephalus), and one nonruminant, i.e., zebra (Equus burchelli). No products were generated from these samples. To evaluate the sensitivity of detection by PCR, decreasing quantities of serially diluted EAT2 DNA were combined with 300 ng of total rumen fluid DNA from cattle and were amplified with the primers derived from clone 1. It was possible to detect 10 pg of EAT2 DNA under these conditions (Fig. 2B, lane 5). Approximately 2,000 to 5000 genome equivalents of S. ruminantium (assuming a genome size of 2 to 5 megabases) could be detected against a background of total DNA generated from approximately 10⁸ bacteria, indicating the potential utility of the probe for tracking EAT2 introduced into domestic ruminants.

FIG. 2.

Specificity and sensitivity of PCR amplification of EAT2 genomic DNA. Shown are agarose gels of PCR products generated by using EAT2-derived primers. (A) Lanes 1 to 5 show PCR products generated from a DNA preparation derived from 10 μl of rumen fluid (300 ng/reaction) from five different B. indicus cattle; lane 6 shows a PCR product generated from 10 ng of EAT genomic DNA mixed with 300 ng of B. indicus total rumen fluid genomic DNA. (B) EAT2-specific PCR products generated from EAT2 DNA serially diluted into 300 ng of B. indicus total rumen microbial DNA; quantities of EAT2 DNA used are as follows: lane 1, 100 ng; lane 2, 10 ng; lane 3, 1 ng; lane 4, 0.1 ng; lane 5, 0.01 ng; and lane 6, 0.001 ng. Lane M is a size standard.

Our results confirm the effectiveness and ease of application of the SSH technique (2) for isolating sequences differing between bacterial isolates within a species, relative to earlier subtractive methods. The data also suggest that the technique may selectively result in isolation of coding sequences. The 16S ribosomal DNA sequences of the S. ruminantium isolates used in this study were 99% identical, and they clustered closely in dendrograms derived from genome-wide amplified fragment length polymorphism analysis (10).

Nucleotide sequence accession number.

The nucleotide sequence of the helicase-like sequence (524 bp) was submitted to GenBank with accession number AY465877.