Correspondence: Jan S. Suchodolski, Gastrointestinal Laboratory, Department of Small Animal Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Texas A&M University, 4474 TAMU, College Station, TX 77843-4474, USA. Tel.: +1 979 458 0933; fax: +1 979 458 4015; e-mail: firstname.lastname@example.org
The aim of this study was to describe the microbial communities along the gastrointestinal tract in healthy cats based on analysis of the 16S rRNA gene. Gastrointestinal content (i.e. content from the stomach, duodenum, jejunum, ileum, and colon) was collected from four healthy conventionally raised colony cats and one healthy specific pathogen-free (SPF) cat. Bacterial 16S rRNA genes were amplified using universal bacterial primers and analyzed by comparative sequence analysis. A total of 1008 clones were analyzed and 109 nonredundant 16S rRNA gene sequences were identified. In the four conventionally raised cats, five different bacterial phyla were observed, with sequences predominantly classified in the phylum Firmicutes (68%), followed by Proteobacteria (14%), Bacteroidetes (10%), Fusobacteria (5%), and Actinobacteria (4%). The majority of clones fell within the order Clostridiales (54%), followed by Lactobacillales, Bacteroidales, Campylobacterales, and Fusobacteriales (14%, 11%, 10%, and 6%, respectively). Clostridiales were predominantly affiliated with Clostridium clusters I (58%) and XIVa (27%). The intestinal microbiota of the SPF cat displayed a reduced bacterial diversity, with 98% of all clones classified in the phylum Firmicutes. Further classification showed that the Firmicutes clones belonged exclusively to the class Clostridiales and were predominantly affiliated with Clostridium cluster I.
It has been recognized that the intestinal microbiota plays an important role in gastrointestinal health and disease. The resident bacterial flora has a physiologic effect on gastrointestinal motility, the development of the intestinal epithelium, and the immune system (Falk et al., 1998; Hooper et al., 2001). The microbiota further supplies nutrients such as vitamins, lactate, and short-chain fatty acids to host tissues (Macfarlane & Macfarlane, 2003). The residential bacteria also provide a natural defense mechanism against invading pathogens (Gibson & Roberfroid, 1995).
In contrast, alterations in the commensal intestinal microbiota have been implicated in gastrointestinal disease in humans and many animal species, including cats (Johnston et al., 2001; Linskens et al., 2001; Janeczko et al., 2008). It has been suggested that an ineffective clearance of enteric pathogens or a loss of tolerance to the residential intestinal microbiota may be a contributing factor in the pathogenesis of inflammatory bowel disease (IBD) in humans (Linskens et al., 2001). In cats, FISH analysis revealed differences in the composition of the duodenal mucosa-associated microbiota between healthy cats and cats with IBD, with the latter having a significantly higher number of mucosa-associated organisms belonging to the family Enterobacteriaceae (Janeczko et al., 2008). Given the impact of the microbiota on gastrointestinal health, knowledge about the composition of the gastrointestinal microbiota of healthy cats is important for future clinical studies exploring differences in microbial communities between healthy cats and cats with gastrointestinal disease.
It has now been recognized that traditional culture-based techniques, as used in previous studies, fail to accurately characterize bacterial communities and underestimate biodiversity in complex biological ecosystems, such as the intestine (Itoh et al., 1984; Amann et al., 1995; Langendijk et al., 1995; Greetham et al., 2002). Recent studies aiming to characterize the intestinal microbiota in many mammalian species have used molecular methods based on the amplification of the 16S rRNA gene (Greetham et al., 2002; Hold et al., 2002; Wang et al., 2003; Delgado et al., 2006). It is likely that a molecular-based approach may identify greater bacterial diversity in the intestinal tract of cats than reported previously. Therefore, in the present study, we aimed to characterize the residential bacteria found in all segments of the intestine of healthy cats using 16S rRNA gene sequences analysis.
Materials and methods
The protocol for sample collection was approved by the University Laboratory Animal Care Committee at Texas A&M University.
Five healthy colony cats (ages 13–18 months), euthanized for an unrelated project, were used for molecular characterization of the intestinal microbial communities. One of the five healthy cats was a specific pathogen-free (SPF) cat. The SPF cat was born and raised in a barrier-maintained facility, theoretically free from all pathogens, until the age of 7 weeks. After 7 weeks, the cat was transferred to a nonbarrier laboratory facility. The other four cats were raised conventionally. All five cats were fed Hill's Science Diet Original (Hill's, Topeka, KA) and received no treatment that would be expected to alter the intestinal microbiota.
Immediately after death, luminal intestinal content was collected by needle aspiration. All samples were collected separately, and no samples were pooled. Samples were transferred into sterile tubes, snap-frozen in liquid nitrogen, and stored at −80 °C until analysis. It was attempted to collect samples from several segments along the gastrointestinal tract. However, in some cats not all segments could be sampled due to an insufficient amount of luminal content. For cats 1 and 2, samples from the jejunum, ileum, and colon were collected. For cat 3, only a sample from the colon was available. For cat 4, samples from the duodenum, ileum, and colon were obtained. From the SPF cat, samples were collected from the stomach, duodenum, jejunum, ileum, colon, and rectum.
Extraction of DNA
Each sample was extracted and analyzed separately. DNA extraction was carried out as described previously, using a bead-beating method, followed by phenol : chloroform : isoamylalcohol extraction (Suchodolski et al., 2008). Purified DNA was stored at −80 °C until further use. A negative control containing H2O instead of a sample was purified in parallel to each extraction batch to screen for contamination of extraction reagents.
16S rRNA gene amplification by PCR
Extracted DNA was used as a template for PCR amplification of a c. 450-bp amplicon of the 16S rRNA gene with universal bacterial primers F-341 (5′-CCTACGGGAGGCAGCAG-3′) and R-786 (5′-GACTACCAGGGTATCTAATC-3′). Each reaction mixture (25 μL) consisted of a reaction buffer (GeneAmp 10 × PCR Gold buffer, Applied Biosystems, Foster City, CA) [final concentrations: 15 mM Tris-HCl, 50 mM KCl, 3 mM MgCl2 (pH 8.0)], 1.25 U of Taq DNA polymerase [Amplitaq Gold Low DNA (LD), Applied Biosystems], 50 μM each of the deoxynucleoside triphosphates, 0.24 μM each primer, and c. 100 ng of DNA template. To screen for potential contamination of PCR reagents, a negative PCR control using H2O instead of a DNA template was used. The samples were amplified in a thermocycler (Mastercycler Gradient, Eppendorf AG, Hamburg, Germany), using the following PCR protocol: initial denaturing at 95 °C for 3 min, 30 cycles of denaturation at 95 °C for 30 s, annealing at 54 °C for 30 s, extension at 72 °C for 1 min, and a final extension at 72 °C for 10 min. Two samples (one duodenal and one ileal sample each) collected from cat 4 did not have enough amplified DNA after 30 PCR cycles for a successful ligation. Therefore, these samples were reamplified using the same PCR protocol with 35 cycles. For all samples, between four and eight independent PCR reactions were performed. The independent PCR reactions for the corresponding compartments of cat 1 and 2 were analyzed together. PCR products belonging to the same sample were pooled and concentrated, using the DNA Clean & Concentrator-5™ (Zymo Research, Orange, CA), following the manufacturer's instructions. The purity of the PCR amplicons was assessed on 1% agarose electrophoresis gels stained with Gel Red™ (Biotium Inc., Hayward, CA).
Cloning of bacterial 16S rRNA gene amplicons
Amplified PCR products were ligated into pCR®4-TOPO® linearized cloning vectors and the products were transformed into chemically competent DH5α™-T1REscherichia coli by heat shock following the manufacturer's instructions (TOPO TA, Invitrogen, Carlsbad, CA). Transformed products were grown overnight on Luria–Bertani (LB) medium with ampicillin (75 μg mL−1) at 37 °C. The pCR®4-TOPO® vector allows direct selection of recombinant cells via disruption of the lethal E. coli gene ccdB. Up to 96 colonies for each sample were randomly selected and clones were grown for 24 h in 1.4 mL LB broth treated with ampicillin (75 μg mL−1) in 96-well blocks (Perfectprep® BAC 96, Eppendorf, North America Inc., Westbury, NY).
Plasmid extraction and sequencing of 16S rRNA genes
Plasmid DNA was purified using the Perfectprep® BAC 96 plasmid purification kit (Eppendorf). Plasmid DNA was then eluted using 50 μL of DNA grade water. The purified inserts were sequenced using the ABI PRISM BigDye Terminator Cycle Sequencing Kit (Applied Biosystems), and the products were analyzed with an automated sequence analyzer (ABI 3100 Capillary Sequencer, Applied Biosystems).
Each sequence was edited to exclude the PCR primer-binding sites, and then tested for possible chimeric structures using the check_chimera and pintail software available through the Ribosomal Database Project (RDP). Identified chimeras were excluded from further analysis.
Cloned sequences were compared with existing 16S rRNA gene sequences using GenBank and RDP (release 9.59, c. 489 840 16S rRNA gene sequences), and the closest neighbor for each sequence was obtained. Sequences were aligned using the multiple sequence alignment program clustal_w and the alignment was inspected and manually adjusted using the alignment editor in the bioedit software. A phylip distance matrix was generated and used as the input file for the dotur software to determine operational taxonomical units (OTUs) (Schloss & Handelsman, 2005). Groups of sequences with <2% sequence divergence (98% similarity) to each other were defined as an OTU. Phylogenetic trees were generated based on the neighbor-joining algorithm using the treecon software package (version 1.3b) and the Jukes–Cantor model for inferring evolutionary distances (Van de et al., 1996). Branch stability was assessed by bootstrap analysis (100 replicates) using the algorithms available in the treecon package. Aquifex pyrophilus was used as an outgroup. A dendrogram was constructed based on the unifrac distance metric, to illustrate the phylogenetic affiliation of the microbiota from the different compartments obtained from each cat.
The coverage of the individual clone libraries (i.e. the probability that any additional analyzed clone is different from any previously obtained single clone) was calculated according to Good (1953) using the formula [1−(n/N)] × 100, where n is the number of molecular species represented by one clone and N is the total number of sequences. Bacterial diversity was calculated using the Shannon–Weaver diversity index, defined as −Σpi ln(pi), where pi is the proportion of individual bacteria found in a certain species (Atlas & Bartha, 1998).
Nucleotide sequence accession numbers
The 16S rRNA gene sequences obtained were deposited into the GenBank database with accession numbers EU877804–EU877912.
A total of 1332 clones were selected randomly. Of these, 1071 clones contained an insert with a sequence of adequate quality. Sixty-three of these clones were identified as possible chimeras and excluded from further analysis. A total of 109 OTUs, representing a total of 1008 clones, were used for subsequent phylogenetic analysis. Table 1 summarizes the number of analyzed samples, analyzed clones, identified OTUs, coverage, and bacterial diversity for each intestinal segment.
Table 1. Number of analyzed samples, selected clones, identified OTUs, coverage, and bacterial diversity index (H) constructed from samples obtained from various segments of the feline intestinal tract
Twenty-one OTUs (19%) showed <98% similarity to existing 16S rRNA gene sequences in the NCBI database. Five different bacterial phyla were identified, with the majority of OTUs being classified as Firmicutes, followed by Proteobacteria, Bacteroidetes, Fusobacteria, and Actinobacteria, respectively.
A total of 21 clones representing seven OTUs were identified within the phylum Actinobacteria. Two clones were isolated from the jejunum (10%), 11 clones were isolated from the ileum (52%), and eight clones were isolated from the colon (38%). One OTU showed 98% similarity to Actinomyces hyovaginalisAF489584 and representative clones were obtained from the jejunum (two), ileum (six), and colon (two).
A total of 95 clones were classified within the phylum Bacteroidetes representing 17 OTUs (Supporting Information, Fig. S1). The majority of clones were isolated from the ileum and colon (43% and 50%, respectively), followed by the rectum (5%) and the jejunum (<2%). Four different bacterial families were identified: Bacteroidaceae, Porphyromonadaceae, Prevotellaceae, and Rikenellaceae. A total of 54 clones representing 10 OTUs were identified within the family Bacteroidaceae. A total of 35 clones representing four OTUs were classified within the family Prevotellaceae. The Rikenellaceae and Porphyromonadaceae families were represented by one and two OTUs, respectively.
The majority of all clones (82%) were classified within the phylum Firmicutes representing 67 OTUs. Figures S2 and S3 show the OTUs classified within the bacterial class Clostridiales, and Fig. S4 illustrates the OTUs identified within the Bacilli and Mollicutes class.
A total of 754 clones were classified in the bacterial class Clostridiales representing 56 different OTUs. Clones belonging to this bacterial class were affiliated with six different Clostridium clusters: clusters I, III, IV, XI, XIVa, and XIVb. The percentages of the 16S rRNA gene sequences affiliated with these clusters are illustrated in Fig. 1. A total of 541 clones representing 13 OTUs were affiliated with Clostridium cluster I. Four OTUs were affiliated with the Clostridium perfringens subgroup, and one OTU was affiliated with Sarcina ventriculiAF110272 (Clostridium subcluster Ia). One OTU was affiliated with Clostridium cluster III. Six OTUs were affiliated with Clostridium cluster IV, and two OTUs were affiliated with Clostridium cluster XI. A total of 54 clones were affiliated with Clostridium cluster XIVa representing 20 OTUs. Three OTUs were affiliated with Clostridium cluster XIVb.
A total of 70 clones representing 10 different OTUs were classified within the class Bacilli. Thirty-four clones were isolated from the jejunum (48%), nine from the ileum (13%), and 27 from the colon (39%). Further classification showed that 65 clones (92%) representing eight OTUs were classified in the bacterial order Lactobacillales. One OTU consisting of one clone from the colon was classified in the class Mollicutes.
A total of 26 clones representing four OTUs were classified within the phylum Fusobacteria. Four clones were isolated from the jejunum (15%), four from the ileum (15%), and 18 from the colon (69%). One OTU in this phylum had 99% similarity to Fusobacterium equinumAJ295750, and its representative clones were predominantly obtained from the colon (10), followed by the jejunum (three) and the ileum (two), respectively. One OTU had 99% similarity to Fusobacterium russiiX55409 and had one representative clone from the jejunum, ileum, and colon each.
A total of 78 clones representing 14 OTUs were classified within the phylum Proteobacteria (Fig. S5). The majority of clones were isolated from the duodenum (41%), followed by the ileum, colon, jejunum, and rectum (32%, 19%, 5%, and 3%, respectively). Four bacterial classes were identified: Betaproteobacteria, Deltaproteobacteria, Epsilonproteobacteria, and Gammaproteobacteria. The two predominant classes were Gammaproteobacteria and Epsilonproteobacteria. Within the class Gammaproteobacteria, six OTUs were identified consisting of 13 clones from the colon, five from the ileum, and two from the jejunum. The Epsilonproteobacteria class was comprised of four OTUs representing 32 clones isolated from the duodenum and 14 clones from the ileum. OTUs from the duodenum were only classified in the class Epsilonproteobacteria. Two OTUs were classified as Betaproteobacteria. One OTU from the ileum was classified as a Deltaproteobacteria, with 99% similarity to Desulfovibrio pigerAF192152.
Comparison of phylogenetic groups between the SPF cat and the conventionally raised cats
Three bacterial phyla were identified in the SPF cat: Firmicutes, Proteobacteria, and Bacteroidetes, with 98% of clones classified within the phylum Firmicutes. Two OTUs belonged to the phylum Bacteroidetes, class Bacteroidales. Three OTUs fell in the phylum Proteobacteria. Two of these were classified as Gammaproteobacteria and one was classified as Betaproteobacteria. Further classification showed that the Firmicutes clones belonged exclusively to the class Clostridiales. Clostridium cluster I was the most predominant cluster (96%), followed by cluster XIVa (2%) and cluster XIVb (1%) (Fig. 1). Less than 1% of sequences were affiliated with Clostridium clusters III and IV. Sequences affiliated with Clostridium cluster I were isolated from all compartments while sequences affiliated with clusters III, XIVa, and XIVb were only isolated from the rectum. Sequences affiliated with Clostridium cluster IV were isolated from the colon.
In contrast, five different bacterial phyla were identified in the four conventionally raised cats: Firmicutes (67%), Proteobacteria (14%), Bacteroidetes (10%), Fusobacteria (5%), and Actinobacteria (4%). Further classification within the phylum Firmicutes showed clones in three bacterial classes: Clostridiales, Bacilli, and Mollicutes. Clones isolated from the conventionally raised cats were predominantly affiliated with Clostridium clusters I (58%) and XIVa (27%) (Fig. 1). Sequences belonging to Clostridium cluster I were isolated from three compartments (i.e. jejunum, ileum, and colon). The number of clones affiliated with cluster I increased along the intestinal tract, with the highest number of clones isolated from the colon. Sequences affiliated with Clostridium cluster XIVa were predominantly isolated from the colon (Fig. 2).
Spatial differences within the feline intestinal tract
Because of the above-described differences in the microbiota between the SPF cat and the four conventionally raised cats, the SPF cat was excluded for this part of the analysis. Also, as mentioned above, not all intestinal segments could be sampled from all four conventionally raised cats. Therefore, a separate analysis was performed describing only the intestinal segments for which samples were obtained from at least two individual cats (i.e. jejunum, ileum, and colon). Fourteen different bacterial orders were identified in these three intestinal segments. The proportions of the predominant bacterial orders within the selected compartments are shown in Fig. 3. The majority of clones were classified within the order Clostridiales. The second most predominant order was Lactobacillales in the jejunum and Bacteroidales in the ileum and colon. Clones classified within the order Bacteroidales were predominantly isolated from the distal intestine (i.e. ileum and colon), with only one clone isolated from the proximal portion of the intestine (i.e. jejunum).
Comparison of the clone libraries using unifrac analysis
The dendrogram constructed based on the unifrac distance metric revealed that the samples tended to cluster by the individual cats rather than the collection site (Fig. 4). A separate analysis was performed to analyze the phylogenetic similarities of the OTUs that were characterized within the phylum Firmicutes. The results of this analysis were similar to the overall unifrac analysis, where the clone libraries clustered by individual cat, rather than by the intestinal site from where the sample was obtained.
The intestinal bacterial microbiota of the feline gastrointestinal tract was characterized using comparative 16S rRNA gene analysis. This study illustrated that phylogenetic similarities were found between the different compartments obtained from each cat. The microbiota in different intestinal compartments tended to be more similar within the individual than between corresponding compartments of different cats. These findings are similar to the results of one study that analyzed the phylogenetic composition of human colon and fecal samples and reported that samples clustered by the individual they came from, and not by the biopsy site (Eckburg et al., 2005).
Firmicutes were the most abundant phylum in the feline intestinal tract (825 clones) and were also the most diverse (67 OTUs). This finding is consistent with previous studies that used microbiological culture techniques to analyze the microbiota in the proximal portion of the small intestine (i.e. duodenum and jejunum) and the colon in healthy cats. Clostridium spp. were the most common bacterial species identified in duodenal aspirates, occurring in over 90% of cats (Papasouliotis et al., 1998; Johnston et al., 2001). The predominant bacterial species in the jejunum were Enterococcus spp., Streptococcus spp., and Lactobacillus spp., which all belong to the phylum Firmicutes (Osbaldiston & Stowe, 1971). The most common bacterial species isolated in colonic and fecal samples were Enterococcus spp. and Lactobacillus spp. (Osbaldiston & Stowe, 1971; Itoh et al., 1984).
Within the phylum Firmicutes, Clostridiales was the most abundant bacterial class (754 clones), representing 56 OTUs. These OTUs were affiliated with six different Clostridium clusters, of which clusters I and XIVa were predominant. Clones affiliated with Clostridium cluster I were identified in all intestinal segments, with the highest number of clones identified in the colon. This finding differs from studies in humans that reported that very few clones obtained from the colon were affiliated with Clostridium cluster I (Hold et al., 2002; Wang et al., 2003; Delgado et al., 2006). Clostridium cluster XIVa was the most diverse cluster (20 OTUs) in the feline intestinal tract, with clones isolated predominantly in the distal intestine (i.e. colon). Similar to our results, the majority of isolated clones from the colon of humans and horses were also affiliated with Clostridium cluster XIVa (Daly et al., 2001; Wang et al., 2003). Clones affiliated with Clostridium cluster IV could only be obtained from the colon and were not isolated in any other segment of the feline intestinal tract.
Lactobacillales were another major constituent of the phylum Firmicutes and were predominantly isolated from the jejunum and colon. This is similar to a previous study that reported Lactobacillus spp. to be a predominant bacterial group in the jejunum and colon of cats when analyzed by culture-based methods (Osbaldiston & Stowe, 1971).
Clones classified in the phylum Proteobacteria were more commonly isolated in the small intestine compared with the large intestine, and 32% of all clones obtained from the duodenum were classified as Epsilonproteobacteria. In a previous study using culture-based techniques, Escherichia spp. were isolated from the colon of all six cats evaluated (Osbaldiston & Stowe, 1971). In contrast, only one clone classified as an E. coli-like organism was isolated in our study, and only from the ileum. This absence of Proteobacteria from the colon is consistent with human studies where it has been reported that these bacteria are only found in low numbers in the human colon and feces using 16S rRNA gene clone libraries (Suau et al., 1999; Hold et al., 2002; Wang et al., 2003; Eckburg et al., 2005; Delgado et al., 2006).
Clones classified in the phylum Actinobacteria were predominantly isolated from the ileum and colon. No Bifidobacterium spp. were detected in our clone libraries. In contrast, previous studies using culture-based techniques have reported Bifidobacterium spp. to be present in feline feces. In one study, Bifidobacterium spp. were cultured from feces in 60% of healthy cats (Itoh et al., 1984). Also, another study utilizing FISH analysis of fecal extracts reported the detection of Bifidobacterium spp. in 90% of healthy cats and 64% of cats with IBD (Inness et al., 2007). One potential explanation for these differences between studies could be the differences in the methodologies used. For example, it has been reported that Bifidobacterium spp. are uncommonly detected in 16S rRNA gene libraries, probably due to bias in the commonly used universal primers or PCR protocols (Wilson & Blitchington, 1996; Hold et al., 2002; Wang et al., 2003; Suchodolski et al., 2008). Therefore, a 16S rRNA gene approach may underestimate the presence of Bifidobacterium spp. In contrast, traditional bacterial culture techniques may have led to an overestimation of the diversity of Bifidobacterium spp. due to insufficient selectivity of culture media (Greetham et al., 2002). Future studies using genus-specific primers will be useful for the accurate characterization of Bifidobacterium spp. in the feline intestinal tract.
In this study, we analyzed samples obtained from the stomach, duodenum, jejunum, ileum, colon, and rectum of an SPF cat. Characterization of the bacterial microbiota in these samples revealed differences when compared with the intestinal microbiota of four conventionally raised healthy colony cats. Clones obtained from the SPF cat were predominantly classified within the phylum Firmicutes (98%), and this proportion was markedly higher compared with the other cats (67%). Further classification showed that clones obtained from the SPF cat belonged exclusively to the class Clostridiales. In contrast, clones belonging to the classes Clostridiales, Bacilli, and Mollicutes were identified in the healthy non-SPF cats. Additionally, 93% of clones obtained from the SPF cat were affiliated with Clostridium cluster I, and all clones obtained from the stomach and duodenum were affiliated with this cluster. In contrast, only 58% of clones obtained from the four healthy non-SPF cats were affiliated with Clostridium cluster I. In addition, 27% of clones obtained from the healthy non-SPF cats were affiliated with Clostridium cluster XIVa, compared with 3% of clones obtained from the SPF cat. Unfortunately, only one SPF cat was analyzed in this study, which makes it difficult to conclude whether the intestinal microbiota of SPF cats is generally different from conventionally raised cats.
To our knowledge, only one culture-based study has compared the intestinal microbiota between conventionally raised and SPF cats (Itoh et al., 1984). Clostridium spp. were significantly higher in cesarian section-derived, barrier-maintained cats compared with conventionally raised cats. The SPF cats displayed a microbial community diversity that was generally similar to the conventionally raised cats, with seven bacterial classes identified (Itoh et al., 1984). This is in contrast to the present study, where only four bacterial classes were identified in the SPF cat, with 98% of all clones classified as Clostridiales. Differences in the methods used (i.e. bacterial culture vs. 16S rRNA gene analysis), diets, or environmental differences in the housing facility may explain the different results of both studies. Given the fact that SPF cats are commonly used in immunological studies, and the microbial communities are believed to play an important role in the development of the host immune system, the potential differences in the intestinal microbial communities between SPF cats and conventionally raised cats warrant further characterization (Brown et al., 1991; Falk et al., 1998).
In conclusion, the molecular approach described in this study facilitated the next step in understanding the complex phylogenetic diversity of the microbial communities in the feline intestinal tract, and allowed the identification of several previously uncharacterized bacterial phylotypes. Further studies are warranted to characterize the intestinal microbiota in diseased cats using 16S rRNA gene clone libraries.