Dietary format alters fecal bacterial populations in the domestic cat (Felis catus)

The effects of short-term (5-week) exposure to wet or dry diets on fecal bacterial populations in the cat were investigated. Sixteen mixed-sex, neutered, domestic short-haired cats (mean age = 6 years; mean bodyweight = 3.4 kg) were randomly allocated to wet or dry diets in a crossover design. Fecal bacterial DNA was isolated and bacterial 16S rRNA gene amplicons generated and analyzed by 454 Titanium pyrosequencing. Cats fed dry diets had higher abundances (P < 0.05) of Actinobacteria (16.5% vs. 0.1%) and lower abundances of Fusobacteria (0.3% vs. 23.1%) and Proteobacteria (0.4% vs. 1.1%) compared with cats fed the wet diet. Of the 46 genera identified, 30 were affected (P < 0.05) by diet, with higher abundances of Lactobacillus (31.8% vs. 0.1%), Megasphaera (23.0% vs. 0.0%), and Olsenella (16.4% vs. 0.0%), and lower abundances of Bacteroides (0.6% vs. 5.7%) and Blautia (0.3% vs. 2.3%) in cats fed the dry diet compared with cats fed the wet diet. These results demonstrate that short-term dietary exposure to diet leads to large shifts in fecal bacterial populations that have the potential to affect the ability of the cat to process macronutrients in the diet.


Introduction
The domestic cat is an obligate carnivore, evolving on high-protein, low-carbohydrate (CHO) diets (Zoran 2002;Ritchie et al. 2010). Obesity levels in domestic cats are between 25% and 40% (Scarlett et al. 1994) and are increasing (German 2006). Intestinal microbiota have been implicated in the development of obesity in humans and rodent models Turnbaugh et al. 2006Turnbaugh et al. , 2008Turnbaugh et al. , 2009). This may be due, in part, to shifts in the intestinal microbiota, which may result in alterations in energy metabolism (Turnbaugh et al. 2008). Therefore, there is increasing interest in the effects of diet on the intestinal microbiota profiles of the domestic cat, in order to ascertain any effects of metabolic disturbances.
Other than two studies performed recently (Barry et al. 2012;Hooda et al. 2012), most of the studies that have investigated the effects of diet composition on fecal microbiota in adult cats have focused on specific bacterial species using gel-or PCR-based methods. Several of these studies have evaluated the effects of prebiotics or dietary fiber (Barry et al. 2010;Kanakupt et al. 2011), while others have focused on protein : CHO ratios Vester et al. 2009;Hooda et al. 2012). Lubbs et al. (2009), for example, reported that high-protein diets increased Clostridium perfringens and Fusobacterium and decreased Bifidobacterium populations in the feces of adult cats. In weaned kittens fed the same diets, high dietary protein decreased fecal Escherichia coli, Bifidobacterium, and Lactobacillus populations ).
Changes in dietary fiber also affect bacterial populations in the domestic cat, including decreased Fusobacterium species (Bueno et al. 2000). Pectin supplementation in adult cats increased the concentrations of C. perfringens, E. coli, and Lactobacillus spp., while fructooligosaccharide supplementation increased Bifidobacterium and decreased E. coli concentrations (Barry et al. 2012).
To our knowledge, all the studies investigating the effects of diet on bacterial composition utilizing nextgeneration sequencing in cats and dogs have examined the effects of dry diets (Barry et al. 2012;Hooda et al. 2012;Tun et al. 2012), with only preliminary investigations into the differences in bacterial composition between raw (meat) and kibbled diets fed to dogs reported (Beloshapka et al. 2011). Recent investigations focused on the composition and function of the intestinal microbiota of cats have shown differences at the phylum level between cats maintained in a research colony and cats living in a domestic setting (Tun et al. 2012). Typically, domestic cats are fed standard wet or dry diets that may greatly differ in moisture, CHO, protein, and fat content. Therefore, we conducted this study to understand the effects of two common conventional diet formats (i.e., wet or dry) on the bacterial composition in the gastrointestinal tract of cats. Because these diets contained many ingredient and nutrient differences, bacterial shifts may not be attributed to any one aspect of the diet, but the diets as a whole. Preliminary data using denaturing gradient gel electrophoresis and DNA sequencing indicated that cats that were changed from a wet to a dry diet showed an 86% change in the fecal microbiota composition, with major changes in Fusobacteriaceae and Comamonadaceae observed (Bermingham et al. 2011). The use of next-generation sequencing may provide a much more detailed view of the dietary effects on the bacterial communities of domestic cats, and allow the identification of minor changes that were not able to be identified on a taxonomic level by means of using denaturing gradient gel electrophoresis alone.
The hypothesis was that short-term dietary changes would alter the bacterial populations within the digestive tract of the domestic cat. The aim of this study was to investigate the effects of a short-term (5-week) dietary exposure to wet and dry diets on the fecal bacterial population of domestic cats using next-generation sequencing.

Animals and diets
The protocol for this study was approved by the Massey University Animal Ethics Committee (MUAEC # 09/103).
All cats were housed at the Centre for Feline Nutrition (Massey University, Palmerston North, New Zealand) according to the Animal Welfare (Companion Cats) Code of Welfare (2007). Prior to the study, all cats were maintained on wet diets as part of standard feeding practices at the Centre for Feline Nutrition. In order to ensure that the cats were clinically and physiologically healthy prior to the study commencing, a complete blood count and thyroid assessment was carried out on each cat (data not shown).
Sixteen mixed-sex, neutered, domestic short-hair cats averaging 6 years of age (range = 1-10 years) and 3.4 kg bodyweight at the start of the study were used in a crossover design to determine the effects of short-term (5-week) exposure to wet (canned) or dry (kibbled) diets on the fecal bacterial communities. The cats were housed in two dietary treatments in adjacent colony cages (1400 9 2400 9 4400 cm). Cats were offered food ad libitum, receiving either a commercially available Association of American Feed Control Officials (AAFCO)-tested, wet diet or a dry diet (Table 1) for 5 weeks. At week 5, cats were placed in individual cages (80 9 80 9 110 cm) for 5 days to determine individual feed intake and fecal output. Fresh fecal samples were collected, mixed for homogeneity, subsampled, and stored at À85°C until bacterial community analysis. Cats were then changed to the alternative diet and the experimental procedure repeated.
Apparent total intestinal digestibility of dietary energy and macronutrients (crude protein, crude fat, crude fiber, and ash; nitrogen-free extracts by difference) were deter- mined for each diet. Individual food intake and refusals and fecal output were recorded daily. Total feces were collected over the 5-day collection period and frozen (À20°C), freeze dried, and ground to a fine powder using an electric grinder (Model CG-2; Breville, Oldham, UK), before analysis. Diet and fecal samples were analyzed for moisture using a convection oven at 105°C (AOAC 930.15, 925.10) and ash using a furnace at 550°C (AOAC 942.05). Crude protein and crude fat were determined using the Leco total combustion method (AOAC 968.06) and acid hydrolysis/Mojonnier extraction (AOAC 954.02), respectively. Gross energy (kJ/g) was determined using bomb calorimetry. Crude fiber was determined using the gravimetric method (AOAC 978.10) and nitrogen-free extracts by difference (Table 1).
Nutrient digestibility (Wichert et al. 2009) and metabolizable energy intake (NRC 2006) were determined. Briefly, the digestibility of macronutrients was determined using % digestibility = [(content in diet À content in feces)/content in diet] 9 100. Metabolizable energy intake was calculated by correcting gross energy (determined via bomb calorimetry) content of the diet by energy digestibility and crude protein content (Bermingham et al. 2012).

Bar-coded amplification of bacterial 16S rRNA genes and amplicon pooling
Primers Ba9F (5′-GAG TTT GAT CMT GGC TCA G-3′) (Weisburg et al. 1991) and Ba515Rmod1 (5′-CCG CGG CKG CTG GCA C-3′) modified from Lane et al. (1985) for PCR amplification of bacterial 16S rRNA genes were synthesized by Integrated DNA Technologies (Coralville, IA). Primers containing the Roche GS FLX adaptors A (5′-CCA TCT CAT CCC TGC GTG TCT CCG ACT CAG-3′) or B (5′-CCT ATC CCC TGT GTG CCT TGG CAG TCT CAG-3′) were used for Titanium sequencing (Rius et al. 2012). A two-base linker sequence between the bar code and the bacteria-specific primer, and a unique 12-base error-correcting bar code was attached to adaptor A for sample identification (Fierer et al. 2008). Each PCR reaction contained 40 lL of Taq PCR MasterMix (Qiagen, Hilden, Germany), 28 lL non-bar-coded primer (0.6 lmol/L), and 8 lL of bar coded primer (2 lmol/L). Before the addition of template DNA, a 19-lL aliquot was transferred into a sterile tube to serve as no-template negative control. The remaining 57 lL were spiked with 3 lL of DNA at a concentration between 20 and 40 ng/lL and divided into three aliquots of 20 lL. Amplification was performed as follows on a Mastercycler proS (Eppendorf, Hamburg, Germany): initial denaturation at 95°C for 2 min, 30 cycles of denaturing (95°C, 20 sec), annealing (52°C, 20 sec), and elongation (72°C, 1 min), and a final 7-min extension at 72°C. Triplicate PCR products were pooled, and correct sizes of PCR products and signal absence from the negative controls were verified by agarose gel electrophoresis. Subsequently, amplicons derived from the samples were purified using a High Pure PCR product purification kit (Roche Diagnostics, Mannheim, Germany), quantified using the Quant-iT dsDNA BR assay kit (Invitrogen, Carlsbad, CA) on a Qubit fluorometer (Invitrogen) and pooled in equimolar ratio into a single pool. The amplicon pool was sent to Macrogen (Seoul, Korea) for Titanium pyrosequencing on a 454 Life Sciences Genome Sequencer FLX machine (454 Life Sciences, Branford, CT).

Bionumerics and statistics
Sequences were analyzed using the Quantitative Insights into Microbial Ecology (QIIME) version 1.3 pipeline using default parameters (Caporaso et al. 2010). Sequences passing quality control metrics were assigned to samples according to their 12-bp bar codes. Sequences sharing a minimum pair-wise similarity of 97% were binned into operational taxonomic units. Representative sequences from each operational taxonomic unit were aligned using PyNAST and a phylogenetic tree constructed using FastTree. Taxonomy for each operational taxonomic unit was assigned using the Ribosomal Database Project classifier with a support threshold of 80% (Wang et al. 2007). Beta diversity between samples was compared by Principal Coordinate analysis of weighted UniFrac distances. Bacterial abundance was tested using the Kruskal-Wallis rank sum test with diet as the main effect. Bodyweight was analyzed using REML variance components analysis (GenStat v12). Results are reported as mean and associated standard error of the mean (SEM) and were considered significant at P < 0.05 and a trend between P > 0.05 and P < 0.10.

Metabolizable energy intake and nutrient digestibility
Metabolizable energy intake was different (P < 0.001) between diets (73.6 kcal/kg vs. 130.7 kcal/kg bodyweight/ day in cats maintained on the dry and wet diets, respectively). The digestibility of crude fat and energy were not different between diets, whereas dry matter (DM) and crude protein digestibility were lower in cats fed dry diets (Table 2). Bodyweight was not significantly affected by diet (3.5 kg vs. 3.5 kg; SEM 0.12) in dry versus wet cats, respectively.

Fecal bacterial profiles
Pyrosequencing of bacterial 16S rRNA gene amplicons resulted in a total of 147,703 sequences, with an average of 4616 (range = 2175-8223) sequences per sample. The number of operational taxonomic units identified was 3924. Rarefaction measures (CHAO) indicated that the diversity of the bacterial population in cats fed dry diets were lower compared with cats fed wet diets (Fig. 1).
The effects of diet format on fecal bacterial composition at the phylum level are shown in Table 3. The three most abundant phyla in cats fed the dry diet were Firmicutes, Actinobacteria, and Bacteriodetes, whereas in cats fed the wet diet, Firmicutes and Fusobacteria were the predominant phyla. Cats fed the dry diet had higher proportions of Actinobacteria (P < 0.05) and lower proportions of Fusobacteria, Proteobacteria, and unclassified bacteria (P < 0.05) compared with cats exposed to the wet diet.
A total of 28 bacterial families were identified in this study. Of these, 17 were affected by diet format (data not shown). Major shifts associated with feeding dry diets included higher proportions (as a percentage of total reads) of Lactobacillaceae ( Forty-six bacterial genera were identified in this study. Short-term exposure to dietary format significantly affected the proportions of 30 of these. There were 15 genera identified in cats fed the wet diet that were not present in the cats fed the dry diet. In cats fed the dry diet, the top five genera identified were Lactobacillus, Megasphaera, Olsenella, Prevotella, and Streptococcus. In the cats fed the wet diet,  1. The effects of short-term exposure to a wet or dry diet on faecal microbial diversity. The rarefaction curve indicates the faecal microbiota CHAO1 diversity index (Chao1 index at 97% sequence identity cut-off) observed over the number of sequences sampled between cats fed wet (-) and dry (-) diets. Data are reported as means AE SEM (n = 16 cats per treatment).  Of the classified genera affected, the dry diet increased the levels of Lactobacillus, Megasphaera, and Olsenella. Cats fed the wet diet had higher levels of Blautia, Bacteroides, Faecalibacterium, Sutterella, and Sporacetigenium (Table 4).
Principal coordinate analysis of Unifrac distances showed that community profiles were clearly separated by diet (Fig. 2).

Discussion
This is the first report that details the effects of dietary format on fecal bacterial populations in the domestic cat using next-generation sequencing. This study identified a dramatic shift in fecal bacterial communities induced by a 5-week exposure to a dry or wet diet with greater Actinobacteria and lower Fusobacteria and Proteobacteria observed in cats exposed to the dry diet. Five bacterial phyla were identified in the feces of cats, consistent with other studies investigating the intestinal microbiota in healthy cats (Ritchie et al. 2008(Ritchie et al. , 2010Desai et al. 2009;Handl et al. 2011). The phyla identified in cats fed the wet diet in this study were similar to those fed the dry diet, but the proportion of phyla changed between dietary formats. Tun et al. (2012) showed that in client-owned cats fed dry diets, Bacteroidetes (68%), Firmicutes (13%), and Proteobacteria (6%) were the predominant phyla, whereas data from laboratory cats (including this study) had higher Firmicutes (57-78%) and lower Bacteroidetes (0.2-16%) populations (Ritchie et al. 2010;Hooda et al. 2012). It is unknown whether this reflects a difference in the laboratory cats per se compared with client-owned cats, or whether differences are due to environmental differences in microbial exposure (e.g., grooming), diet, or methodological differences (e.g., primer design).
High protein : CHO diets (approximately 50% DM crude protein) decreased Actinobacteria and increased Fusobacteria levels in the growing kitten (Hooda et al. 2012), similar to the effects observed in the wet diet (high protein : CHO; 42:5% DM) in this study. While changes in phyla reported in Hooda et al. (2012) were similar to those observed in the adult cats fed wet diets in this study, differences were noted in bacterial genera between the two studies. For example, Hooda et al. (2012) observed increased Dialister, Acidaminococcus, Bifidobacteria, Megasphaera, and Mitsuokella with moderate protein : CHO diets (approximately 34% DM crude protein), which were not observed in this study. These differences may reflect differences in age (growing vs. adult cats in this study), diet (dry vs. wet format), or living conditions. It is possible that different bacterial species may have similar functions; however, functional data were not described for either Hooda et al. (2012) or this study, thereby emphasizing the increasing importance of understanding functional differences underlying the changes in bacterial composition.
A major genus of interest in the phylum Actinobacteria is the Bifidobacterium; species within this genus are thought to play a role in intestinal health. In contrast to some recent next-generation sequencing studies, bifidobacteria were not observed in either dietary group in this study. Bifidobacteria, which are carbolytic bacteria (Ritchie et al. 2010) capable of starch digestion (Suchodolski 2011), have been shown to be present in the intestine of the cat in other studies (Handl et al. 2011;Jia et al. 2011a;Hooda et al. 2012). It is possible that the absence of bifidobacteria observed in this study is due to primer bias (Palmer et al. 2007;Sim et al. 2012). In this study, the changes in the phylum Actinobacteria were mainly due to increased proportions of Coriobacteriaceae, namely the Olsenella genus, in cats fed the dry diet. Increases in Coriobacteriaceae abundance may have important consequences for health, as it has been associated with decreased blood glucose levels in mice (Claus et al. 2011) and increased blood non-high-density lipoprotein plasma concentrations and cholesterol absorption in hamsters (Martinez et al. 2009). High Coriobacteriaceae populations have also been observed in geriatric cats (8-14 years) fed dry diets (Jia et al. 2011b), although blood lipids and glucose were not measured in Jia et al. (2011a). The levels of Bacteriodetes reported in the literature vary from very low levels (0.2%) in kittens fed moderateor high-protein diets (Hooda et al. 2012) to 68% in client-owned cats (Tun et al. 2012). Although the overall abundance of the Bacteroidetes phylum remained unchanged between diets in this study, the composition within this phylum varied with diet on a lower taxonomic level (genus). For example, the large downward shift in the genus Bacteroides in cats fed the wet diet was largely compensated for by the upward shift in unclassified Prevotellaceae. This finding suggests that the Bacteriodetes phylum consists of diverse bacterial taxa capable of degrading both protein and CHO sources . Further research with more in-depth sequencing is needed to identify the specific genera and species responding to diet and to identify any health implications it may have.
Within the Firmicutes phylum, differences between the Clostridia and Bacilli were observed, including members of the Lactobacillaceae (Lactobacillus), Peptostreptococcaceae, and Veillonellaceae (Megasphaera) families. In this study, fecal Lactobacillus populations were greater in cats fed the dry diet. Lactobacillus spp. are typically regarded as a beneficial group of microbes. Previous studies have reported higher fecal Lactobacillus in cats fed dry, moderate-protein, moderate-CHO diets compared with those fed high-protein, low-CHO diets (Hooda et al. 2012). In contrast, fecal Lactobacillus populations were lower in dogs fed raw meat versus dry, extruded diets (Beloshapka et al. 2011). These results from previous reports are in agreement with what was observed in cats of this study; this presumably reflects the saccharolytic nature of Lactobacillus and the CHO content of the dry diets used in these studies. In this study, Clostridium spp. increased in cats fed the wet diet, which concurs with previous studies that have shown elevated proportions of Clostridium in cats fed high-protein diets Ritchie et al. 2010). These results suggest that members of the Clostridium genus in cats may be well adapted for proteolytic activity. Similarly, Peptostreptococcus spp., which have been associated with amino acid fermentation and breakdown (Suchodolski 2011), were also elevated in cats fed the wet diet. The large fecal bacterial shifts from CHO-utilizing to protein-utilizing bacteria observed in this study appear to reflect the differences in macronutrient profiles of the two diets. Proportions of Megasphaera (a member of the Veillonellaceae family), a major butyrate producer, were increased in cats fed the dry diet, which is in agreement with the data reported by Hooda et al. (2012), where cats fed a moderate-protein, moderate-CHO diet showed similar increases compared with those fed a high-protein, low-CHO diet.
Fusobacteria species contribute to amino acid fermentation and breakdown (Loesche and Gibbons 1968;Potrykus et al. 2008), which may explain the increased levels of this phylum in cats fed the wet diet in this study. Previous studies have observed low Fusobacteria populations, although diet was not detailed in these studies (Ritchie et al. 2010;Handl et al. 2011). However, higher Fusobacteria levels have been observed more recently by Hooda et al. (2012); however, primer design differed to that of this study. While Fusobacteria have been associated with a number of diseases in humans (Hooda et al. 2012), the cats in this study appeared to be healthy despite the high abundance of bacteria belonging to this phylum.
Changes in bacterial phylogeny observed in this study do not necessarily mean changes in function. Therefore, developing more robust data sets, including the analysis of functional changes (metagenomics and metatranscriptomics) is vital for understanding the effects of diet composition on the development of obesity phenotypes in domestic cats.