Correspondence: Alexander G. Haslberger, Department of Nutritional Sciences, University of Vienna, UZAII, Althanstrasse 2, A-1090 Vienna, Austria. Tel.: +43 699 122 11212; fax: +43 1 879 5896; e-mail: email@example.com
The gastrointestinal microbiota produces short-chain fatty acids, especially butyrate, which affect colonic health, immune function and epigenetic regulation. To assess the effects of nutrition and aging on the production of butyrate, the butyryl-CoA:acetate CoA-transferase gene and population shifts of Clostridium clusters lV and XlVa, the main butyrate producers, were analysed. Faecal samples of young healthy omnivores (24 ± 2.5 years), vegetarians (26 ± 5 years) and elderly (86 ± 8 years) omnivores were evaluated. Diet and lifestyle were assessed in questionnaire-based interviews. The elderly had significantly fewer copies of the butyryl-CoA:acetate CoA-transferase gene than young omnivores (P=0.014), while vegetarians showed the highest number of copies (P=0.048). The thermal denaturation of the butyryl-CoA:acetate CoA-transferase gene variant melting curve related to Roseburia/Eubacterium rectale spp. was significantly more variable in the vegetarians than in the elderly. The Clostridium cluster XIVa was more abundant in vegetarians (P=0.049) and in omnivores (P<0.01) than in the elderly group. Gastrointestinal microbiota of the elderly is characterized by decreased butyrate production capacity, reflecting increased risk of degenerative diseases. These results suggest that the butyryl-CoA:acetate CoA-transferase gene is a valuable marker for gastrointestinal microbiota function.
Recent evidence suggests that 1000–1150 different species are capable of living in the gut ecosystem. An individual harbours at least 160 species (Qin et al., 2010), with high interindividual variations in species diversity and evenness. It has been reported that the microbiota composition is influenced by diet (Larsen et al., 2010) and age (Mariat et al., 2009), as well as genetic factors (Khachatryan et al., 2008). The gastrointestinal microbiota produces short-chain fatty acids (SCFAs). Butyrate is of particular interest due to its anticarcinogenic and anti-inflammatory potential (Maslowski et al., 2009), its effects on the intestinal barrier (Peng et al., 2007), satiety (Cani et al., 2009) and epigenetic regulation (Rada-Iglesias et al., 2007). Two of the most important groups of butyrate producers are Faecalibacterium prausnitzii from the Clostridium cluster IV, and the Eubacterium rectale/Roseburia spp. from the Clostridium cluster XIVa (Walker et al., 2010). Both clusters (now also known as Ruminococcaceae and Lachnospiraceae) consist of producers and nonproducers of butyrate (Pryde et al., 2002). Isolated dietary compounds have been shown to promote growth of butyrate producers (Hernot et al., 2009). For example, the consumption of inulin significantly stimulated growth of F. prausnitzii (Louis & Flint, 2009). In colonic in vitro model systems, resistant starch stimulated the growth of E. rectale (Leitch et al., 2007).
Butyrate is easily taken up by the gut mucosa and faecal butyric acid levels give little information about the butyrate-producing capacity of the gut microbiota. Therefore, a function-based approach was suggested for the enumeration of butyrate-producing bacteria (Louis & Flint, 2007) targeting the butyryl-CoA:acetate CoA-transferase gene. Furthermore, the butyryl-CoA:acetate CoA-transferase route, using acetate as a cosubstrate, is suggested to be the most important route for butyrate production in the gut ecosystem (Duncan et al., 2004). Alternative routes are via butyrate kinase and phosphotransbutyrylase, which are found in a minority of bacteria (Louis et al., 2004) in the human gastrointestinal tract.
To evaluate the relevance of the butyryl-CoA:acetate CoA-transferase gene in butyrate-producing communities, we compared the amount of butyryl-CoA:acetate CoA-transferase gene in young omnivores, vegetarians and the elderly by quantitative PCR (qPCR) with a 16S rRNA gene quantification of Clostridium clusters IV and XIVa considering the effects of aging and nutrition.
Materials and methods
Study groups consisted of 15 institutionalized, mobile elderly persons [age: 86 ± 8 years, body mass index (BMI) 21.75 ± 5.08], 15 young healthy vegetarians (age: 26 ± 5 years, BMI 21.02 ± 2.71) and 17 young healthy omnivores (age: 24 ± 2.5 years, BMI 22.68 ± 3.4) consuming a central European diet. All subjects were interviewed using a questionnaire assessing age, gender, body height, weight, individual health status, lifestyle and dietary habits. Approval was obtained from the Viennese Human Ethics committee (EK 07-153VK).
Faeces was collected from each participant individually and stored at −18 °C until processed. DNA was extracted using the DNA stool Mini Kit (Qiagen) following the manufacturer's protocol with minor modifications and immediately stored at −20 °C. Efficiency and quality of extraction was controlled by photometry (Nanodrop) and gel electrophoresis.
Quantification of specific metabolic genes and 16S rRNA genes by qPCR
The butyryl-CoA:acetate CoA-transferase genes were amplified with degenerated primers BCoATscrF/R as listed in Table 1 (Louis & Flint, 2007) on a Rotor-Gene 3000A (Qiagen) using the SensiMix SYBR Kit (Quantace). Amplification of one of the faecal samples with BCoATscrF/R leads to a highly concentrated butyryl-CoA:acetate CoA-transferase gene mix. This purified PCR product was used for quantification of samples. Amplification with primer pair BcoATscr resulted in clearly distinguishable and assignable melt peaks. Total bacteria (Yu & Morrison, 2004) and Clostridium clusters IV and XIVa were quantified (Meier et al., 1999; Matsuki et al., 2004) on a Rotor-Gene 3000A (Qiagen) using the SensiMix Probe Kit (Quantace). The primers and probes used in this study are listed in Table 1. Samples were quantified using standards derived from one clone [clone library CleptF/R; Promega Vector System, specificity in library confirmed (Liszt et al., 2009) with known concentration in the case of Clostridium cluster IV and from a Blautia coccoidesT pure culture for cluster XIVa].
Table 1. Primers and probes used for quantification of 16S rRNA genes and butyryl CoA:acetate CoA-transferase genes with qPCR
Thermal denaturation of butyryl-CoA:acetate CoA-transferase amplicons
Melt curves from amplified PCR products were divided into three areas (Fig. 1b), as described by Louis & Flint (2009). These peaks were assigned to represent bacteria related to Eubacterium hallii and Anaerostipes coli (82.5–85.0 °C), Roseburia/E. rectale spp. (85.5–89.0 °C) and F. prausnitzii (89.5–92.5 °C) as illustrated in Fig. 1a and b.
All statistical analysis (Spearman's rank, Kolmogorov–Smirnov, F- Kruskal–Wallis- and t-tests) was done using originpro 8G (http://www.originlab.com).
Analysis of the dietary habits indicated similar consumptions of fruit and milk products in the individual groups. Vegetarians stated significantly more frequent consumption of vegetables (χ2; P<0.03) but similar patterns of consumption of carbohydrates, including whole-grain products. The elderly stated significantly more frequent consumption of meat and similar vegetable consumption (χ2 test; P<0.04) compared with omnivores. The exercise levels of vegetarians and omnivores were comparable.
Vegetarians had 12 ± 62% more and the elderly had 31 ± 21% less 16S rRNA gene relative to absolute quantified genes compared with omnivores (Fig. 2a). Many SCFA-synthesizing bacteria belong to the Clostridium clusters lV and XlVa. The Clostridium cluster lV (Fig. 2b) was significantly more abundant in omnivores (36.3 ± 11.2%) than in the elderly (27 ± 11.7%, P=0.04) quantified relative to total bacterial 16S rRNA genes. Vegetarians harboured 31.86 ± 17.00% of Clostridium cluster IV. The Clostridium cluster XlVa (Fig. 2c) was significantly more abundant in omnivores (19.01 ± 6.7%, P>0.01) and vegetarians (14.52 ± 5.6%, P=0.049) than in the elderly (9.89 ± 6.64%).
Butyryl-CoA:acetate CoA-transferase genes in faeces by qPCR
The elderly had significantly fewer copies (1.52 × 1011± 1.36 × 1010 copies g−1 faeces) of the butyryl-CoA:acetate CoA-transferase gene compared with the omnivores (4.96 × 1011± 3.22 × 1010 copies g−1 faeces, P=0.01) and the vegetarians (1.37 × 1012± 1.47 × 1011 copies g−1 faeces, P=0.048) (Fig. 2d). The amount of the butyryl-CoA:acetate CoA-transferase gene did not correlate significantly with the amount of total bacteria.
Melt curve analysis
The E. hallii/A. coli melt peaks tend to be higher in vegetarians (P=0.08) and omnivores (P=0.09) than in the elderly. The abundance of E. rectale/Roseburia spp. melt peak differed significantly between vegetarians and the elderly (P=0.04). Melt peak attributed to F. prausnitzii was significant lower in the elderly than in omnivores (P=0.049) (Fig. 1a).
Spearman's rank showed no significant correlation between the amount of the butyryl-CoA:acetate CoA-transferase gene and that of Clostridium clusters IV and XIVa at an individual level.
Analysis of the overall abundance of bacterial 16S rRNA genes reveals that the vegetarians harboured more bacteria than the omnivores. The low numbers of bacteria in the elderly individuals (Fig. 2a) may reflect physiological alterations such as prolonged colonic transit time, reduced dietary energy requirement and food uptake (Morley, 2007). Figure 2b illustrates the significantly higher abundance of Clostridium cluster IV in omnivores. Mueller et al. (2006) detected the highest levels of the Clostridium cluster IV in a Swedish study population, whose dietary habits were characterized by a high consumption of fish and meat (Mueller et al., 2006). Despite high meat consumption in the elderly, the generally smaller capacity for energy harvest from food may decrease the abundance of Clostridium cluster IV (Li et al., 2008). The elderly gut microbiota is also characterized by a significantly lower relative contribution of Clostridium cluster XIVa compared with young study participants (Fig. 2c). A decrease of Clostridium cluster XIVa has already been described as an effect of the ageing process (Bartosch et al., 2004). Elderly study participants had lower physical activity levels and did not consume whole-grain products, whereas the other groups stated regular consumption of fibre-rich products.
Vegetarians and omnivores have significantly more copies of the butyryl-CoA:acetate CoA-transferase genes compared with the elderly (Fig. 2d). Although no clear correlation with Clostridium cluster IV and XIVa levels were found, the elderly tended to harbour fewer butyrate producers than did young individuals. Melt curve analysis showed that the butyryl-CoA:acetate CoA-transferase gene variant related to E. rectale/Roseburia spp. is significantly more variable in vegetarians than in the elderly (Fig. 1a). Correspondingly, Clostridium cluster XIVa seems to be more abundant in vegetarians. Biagi et al. (2010) found lower quantities of Roseburia intestinalis, E. hallii and E. rectale in the elderly (>75 years) than in young adults using the HITchip method. The abundance of the Clostridium cluster XIVa does not show significant correlations with the abundance of the butyrate gene variant as determined by melting curve analysis related to Roseburia/E. rectale spp., as this cluster also contains many nonbutyrigenic bacteria. As illustrated in Fig. 1a, the level of the melt peak attributed to F. prausnitzii was significantly lower in the elderly. This is of particular interest as this species has been reported to influence gut inflammation processes by exerting a butyrate-independent anti-inflammatory effect (Sokol et al., 2009).
The vegetarian diet may have favoured growth of the Roseburia/E. rectale spp. that carries the butyryl-CoA:acetate CoA-transferase gene, without causing an increase in the abundance of butyrate producers in the entire Clostridium cluster XIVa. Omnivores and vegetarians had a similar potential to harbour butyryl-CoA:acetate CoA-transferase genes and members of Clostridium clusters IV and XIVa, possibly caused by a similar intake of fruit and carbohydrates.
These results suggest that the elderly group in this study harbours less total bacteria and an even lower abundance of Clostridium clusters IV and XIVa. Together with a lower abundance of the butyryl-CoA:acetate CoA-transferase gene, the results indicate that in the elderly, microbiota may be characterized by a low butyrate production capacity. In respect of the important nutritive, anti-inflammatory and anticancerogenic potential of butyrate in the human colon, these findings demonstrate that these microbiota specificities may contribute to the development of degenerative diseases (Guigoz et al., 2008) and anorexia in advanced age (Donini et al., 2010).
Consideration of the results of the analysis must take into account methodological limitations. Despite the extraction controls discussed in Materials and methods, DNA extraction can be influenced by diet and the consistency of faeces. Previous studies reported a higher abundance of Clostridium cluster XIVa than cluster IV, which is not seen in our results (Fig. 2b and c). This points to the need to compare different standards for these clusters. Amplification of the 16S rRNA gene can be particularly biased due possible multiple operons for this gene. The use of degenerated primers carries a certain risk for unspecific amplification of nontarget DNA. To estimate the accuracy of our amplification, we checked every PCR product in 2% agarose gels where all PCR products gave bands of the expected size. Our melt curve analysis assumes that the intensity of individual peaks represents the initial proportion of the different butyryl-CoA:acetate CoA-transferase gene variants.
In conclusion, the quantification of the butyryl-CoA:acetate CoA-transferase gene may be a suitable biomarker for butyrate production for an individualized assessment of gastrointestinal health and microbiota function in addition to analysis of gastrointestinal microbiota.
We thank all the study participants. We thank Dr Guadalupe Pinar and Dr Katja Sterflinger for giving us access to DNA quantification machinery. The Austrian Science Fund (FWF) funded this study.