Short Communication

International Journal of Obesity (2008) 32, 1720–1724; doi:10.1038/ijo.2008.155; published online 9 September 2008

Human colonic microbiota associated with diet, obesity and weight loss

S H Duncan1, G E Lobley1, G Holtrop2, J Ince1, A M Johnstone1, P Louis1 and H J Flint1

  1. 1Rowett Institute of Nutrition and Health, University of Aberdeen, Bucksburn, Aberdeen, UK
  2. 2Biomathematics and Statistics Scotland, Bucksburn, Aberdeen, UK

Correspondence: Professor HJ Flint, Microbial Ecology Group, Rowett Institute of Nutrition and Health, University of Aberdeen, Greenburn Road, Bucksburn, Aberdeen AB21 9SB, UK. E-mail:

Received 27 May 2008; Revised 4 August 2008; Accepted 10 August 2008; Published online 9 September 2008.





It has been proposed that the development of obesity in humans is influenced by the relative proportions of the two major phyla of bacteria (Bacteroidetes and Firmicutes) present in the large intestine.



To examine the relationships between body mass index, weight loss and the major bacterial groups detected in fecal samples.



Major groups of fecal bacteria were monitored using fluorescent in situ hybridization (FISH) in obese and non-obese subjects under conditions of weight maintenance, and in obese male volunteers undergoing weight loss on two different reduced carbohydrate weight-loss diets given successively for 4 weeks each.



We detected no difference between obese and non-obese individuals in the proportion of Bacteroidetes measured in fecal samples, and no significant change in the percentage of Bacteroidetes in feces from obese subjects on weight loss diets. Significant diet-dependent reductions in a group of butyrate-producing Firmicutes were, however, detected in fecal samples from obese subjects on weight loss diets.



Diets designed to achieve weight loss in obese subjects can significantly alter the species composition of the gut microbiota, but we find no evidence that the proportions of Bacteroidetes and Firmicutes among fecal bacteria have a function in human obesity.


weight loss, dietary carbohydrate, gut bacteria, bacteroidetes, firmicutes

There has been much recent interest in the relationship between the types of bacteria that colonize the gut and the development of obesity. Many dietary components that are resistant to initial digestion in the small intestine are subsequently fermented by the microbial community of the large intestine, producing short-chain fatty acids that are absorbed across the colonic mucosa. These short-chain fatty acids have been estimated to provide 10% of total dietary energy supply in humans1 and, in addition to increasing the amount of dietary energy available to the host, may also stimulate lipogenesis through the regulation of fasting-induced adipose factor.2 On the other hand, the energy that is derived by the host from a dietary-resistant (or ‘non-digestible’) carbohydrate through microbial conversion to short-chain fatty acids is less than from the equivalent amount of sugar that is absorbed directly in the small intestine.1 Non-digestible carbohydrates may also increase satiety.3 Furthermore, changes in microbial metabolism and in gut microbiota composition that result from carbohydrate fermentation have several potentially beneficial consequences for the host. These include the supply of butyrate to the colonic mucosa and the stimulation of bifidobacteria.4, 5

It has been reported, on the basis of evidence from 16S rRNA sequencing, that the proportions of the two major bacterial phyla found in the human fecal microbiota, the Firmicutes and the Bacteroidetes, differ between obese and lean humans,6 with the suggestion that the colonic microbiota found in obese subjects is more efficient than that of lean subjects at recovering energy from resistant dietary components.7 In support of this, transfer of the colonic microbiota from ob/ob mice to germ-free animals led to increased fat gain, equivalent to an extra 2% energy retention of the calories consumed, compared with transfer from control mice,7 although there was also a non-significant difference in food intake. The ob/ob mice harbored a greater proportion of Firmicutes and less Bacteroidetes among their gut microbiota compared with lean mice.7, 8 In obese human subjects undergoing weight loss over a 12-month period, the proportion of Bacteroidetes was reported to increase from approximately 2% of total fecal bacteria to >20%.6


Relationship of percentage of Bacteroides in human fecal samples to body mass index

In recent studies with 18 obese human male volunteers, we used 16S rRNA-based quantitative fluorescent in situ hybridization (FISH) to study the composition of fecal microbiota.9 These data have now been extended to include an additional 15 obese male subjects and 14 non-obese subjects. These studies received prior approval from the North of Scotland Research Ethics Service, and written informed consent was obtained from all subjects. Counts for a particular bacterial group were expressed relative to the total number of bacteria present in the feces, which was estimated using a broad probe (Eub338) that quantifies approximately 73% of bacteria detectable by direct staining using 4,6-diamido-2-phenylindole.10 Total bacterial numbers did not differ between obese (body mass index (BMI) >30kg/m2) and non-obese (BMI <30kg/m2) subjects when on weight maintenance (M) diets (5.52 and 5.59 × 1010/g feces, respectively, P=0.85, standard error of the difference (SED)=6.68 × 109). To determine the numbers of Bacteroides, the Bac303 probe was used, as this detects the great majority of Gram-negative human gut bacteria that belong to this phylum.11 The percentage of total fecal bacteria present as Bacteroide did not differ between obese subjects on the M diet and non-obese controls (27.2 vs 21.9%, SED=2.96, P=0.084) (Figure 1). These findings contrast with a previous report6 but are in good agreement with other estimates from human feces,12, 13, 14, 15 although considerable inter-individual variation is widely observed (see Figure 1). Furthermore, there was no significant relationship between BMI (range: 20–44kg/m2) and the proportion of Bacteroides (correlation: 0.13, P=0.40, Figure 1).

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Relationship between body mass index (BMI) and the percentage of Bacteroides among fecal bacteria. Bacteroides cells present in fecal samples were enumerated by fluorescent in situ hybridization (FISH) using the Bac303 probe, and expressed as a percentage of bacteria detected by the Eub338 probe. Methods for sample processing, FISH and enumeration, were described earlier.9 Each data point represents a sample from a different volunteer. Obese or overweight males (triangles) provided a sample after 3–7 days on a controlled weight maintenance diet in which 55% of calories came from carbohydrates.9 Also shown are non-obese female (squares) and male (diamonds) volunteers, consuming their habitual western style diets. The majority of the volunteers lived in the vicinity of Aberdeen, UK. Best fit linear regression is: % Bacteroidetes=19.2+0.19 × BMI (R2=0.13, P=0.40).

Full figure and legend (12K)


Changes in gut bacterial populations in volunteers on reduced carbohydrate weight-loss diets

A total of 23 obese subjects shown in Figure 1 also underwent weight loss regimes (energy intake <8.5MJday−1) for a period of 8 weeks, comprising 4 weeks on each of two diets, either high-protein low carbohydrate, ketogenic (LC) or high-protein moderate-carbohydrate, non-ketogenic (MC), offered in a balanced cross-over design. The composition of each meal in terms of energy, macronutrients and fiber content was calculated from food composition tables and was carefully controlled.16, 17 The proportion of total fecal bacteria detected as Bacteroides (by the Bac303 probe) was not affected either by diet or by diet order, thus within subjects even substantial changes in the amount of carbohydrate offered had no impact (Figure 2). This was despite a reduction in the total number of bacteria, detected by the Eub338 probe, between the maintenance (M) and weight loss (MC and LC) diets (4.92 vs 3.32 and 3.09 × 1010, SED=0.46 × 1010, P<0.001). There was no significant relationship between changes in the percentage of Bacteroides in the feces and the amount (kg) of weight lost (R2: 0.08, P=0.11), although there was a marginally significant, weak positive correlation between the change in the percentage of Bacteroides and weight loss expressed as percentage of initial body weight (R2: 0.14, P=0.043, data not shown). This contrasts with a R2 of 0.8 for the relationship between the percentage of Bacteroidetes and the percentage of weight loss reported elsewhere for obese subjects on a low carbohydrate diet.6 In the studies reported here, however, almost as many subjects showed decreased (10) as opposed to increased (13) fecal percentage of Bacteroides after 8 weeks on reduced carbohydrate diets.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Impact of diets of different carbohydrate (CHO) and NSP content on the proportion of bacteria in feces. Obese male volunteers (23) were offered a maintenance diet for 3 days (M; 399g CHO and 24g NSP per day) and then 4 weeks on two weight loss diets containing either moderate CHO (MC; 164g CHO and 12.2g NSP per day) or low CHO (LC; 24g CHO and 6g NSP per day), offered in a randomized sequence (11 subjects M-LC-MC, 12 subjects M-MC-LC).9 Diets: M (filled bars), MC (hatched bars) and LC (open bars). Bacterial populations, estimated by FISH microscopy on stool samples collected at the end of each period, and expressed as percentages of the total Eub338 count are shown for: Bacteroides (detected with the Bac303 probe); Roseburia+Eubacterium rectale (‘Rrec’, detected with the Rrec584 probe); and other Clostridium coccoides-related bacteria (‘Erec-Rrec’, calculated from the difference between the counts obtained with the Erec482 and Rrec584 probes). Error bars reflect the standard error of the mean based on between subject variations. Overall mean values at the start of the study were 26.79% Bacteroides, 11.49% Rrec and 12.52% Erec-Rrec, and the mean initial body weight was 113.76kg.

Full figure and legend (47K)

Previous reports also indicate an increase in the proportion of Firmicutes present in the feces or colon of obese human subjects and ob/ob mice,7, 8 as compared with lean controls. In the current study, a panel of probes (Prop853, Fprau645, Erec 482, Rfla729 and Rbro730)9 was used that account for the majority of bacteria in the Firmicutes phylum. In total, these showed no change in the percentage of Firmicutes with the various weight loss diets (M 53.9%, LC 49.8%, MC 55.2%, SED=2.61, P=0.32). Within this phylum, however, various bacteria were sensitive to the dietary interventions. In particular, the Roseburia+Eubacterium rectale group (detected with the Rrec482 probe) that has a key function in the formation of butyrate,4, 9, 10 and probably has a beneficial impact on colonic health, showed a striking reduction (P<0.001) in the proportion present in the feces as the estimated total carbohydrate (399, 164 and 24gday−1), starch (187, 95 and 3gday−1) and non-starch polysaccharide (NSP; 28, 12.2 and 6gday−1) intakes17 decreased between the M, MC and LC diets, respectively. This was a response to diet composition, as demonstrated by the fact that the proportions of Roseburia+E. rectale decreased with diet order MC to LC, but increased with diet order LC to MC (Figure 2; order effect P=0.02), whereas weight loss continued over time with either diet order (Figure 2).

Roseburia+E. rectale form a part of a larger group, Clostridium coccoides, that is detected with the Erec482 probe. Interestingly, as noted earlier,9 other bacteria within the C. coccoides cluster tended to increase as the Roseburia+E. rectale group declined (Figure 2, P<0.03). These data show that short-term weight loss per se neither drives, nor is driven by, changes in these groups of bacteria. Rather, the changes in these bacterial populations are probably in response to diet, in particular the amount and type of carbohydrate present. Similarly, although weight-stable obese and non-obese subjects had similar percentages of Bifidobacteria (3.87 vs 4.34, SED=1.36, P=0.73), a significant reduction was also detected for this group in obese subjects following the 4-week weight loss period on both the LC and MC diet (1.87 and 2.09%, SED=0.88, P<0.037 vs M).9 This is also likely to be a response to the supply of fermentable carbohydrate.

The major differences between our findings and those of Ley et al.,6 particularly with respect to the Bacteroidetes, might relate to diet, the cohort of volunteers, methods of sample preparation and storage or the methods of detection used. The latter possibility was tested with fresh feces taken from six of the obese subjects used in Figure 1. One portion was prepared for FISH enumeration, whereas another, following homogenization, was used for DNA extraction followed by real-time PCR. Primers used to selectively amplify Bacteroides 16S rRNA genes were Bac303F18 (GAAGGTCCCCCACATTG) and Bfr-Fmrev (CGCKACTTGGCTGGTTCAG) (modified from19). The mean number of Bacteroides cells detected by FISH against Eub338 tended to be lower than the estimate for Bacteroides relative to total bacterial 16S rRNA gene copies obtained by PCR (21.2 vs 28.1%, SED=3.21, P=0.086) using a Bio-Rad MyiQ cycler and amplification conditions, as described earlier.20 These findings appear to exclude the possibility that FISH systematically overestimates the percentage of Bacteroides compared with a method based on PCR amplification when fresh fecal samples are compared. Problems with the loss of microbial diversity upon frozen storage of fecal samples have been reported earlier,21 and this issue is under further investigation. It must be emphasized, however, that the FISH procedure used in this study routinely involves preservation of fresh fecal samples in paraformaldehyde prior to frozen storage.



In conclusion, the data reported here show that the FISH technique is capable of detecting diet-related changes for particular bacterial groups, and the magnitude of response reported for Bacteroidetes during weight loss in other studies6 should have been easily detected in the current experiments. Nevertheless, we could find no relationship in humans between BMI or absolute weight loss and the relative populations of the major groups of human colonic bacteria, including the Bacteroidetes, in stool samples from both obese and non-obese subjects. Our findings therefore do not support the hypothesis6, 7 that the proportions of Bacteroidetes and Firmicutes, at least at the phylum level, have a major function in determining human obesity. We do not rule out the possibility that more detailed analysis of the community will reveal changes in species composition between obese and lean subjects, which might result from differences either in eating habits or host physiology. The extent of inter-individual differences at the level of bacterial species14, 15 and of inter-sample variation12 makes such analyses challenging. Nevertheless, we were able to confirm our earlier observations9 that reduced carbohydrate weight-loss diets result in a reduction in the proportions of at least one group of Gram-positive bacteria that is responsible for butyrate production, together with bifidobacteria. The consequences of these diet-dependent changes on the gut microbiota will undoubtedly have impacts on aspects of both colonic health and function, as well as affecting host metabolism, thus warranting further investigation.



  1. McNeil NI. The contribution of the large intestine to energy supplies in man. Am J Clin Nutr 1984; 39: 338–342. | PubMed | ChemPort |
  2. Backhed F, Ding H, Wang T, Hooper LV, Koh GY, Nagy A et al. The gut microbiota as an environmental factor that regulates fat storage. Proc Nat Acad Sci USA 2004; 101: 15718–15723. | Article | PubMed | ChemPort |
  3. Nilsson AC, Ostman EM, Holst JJ, Bjorck IME. Including indigestible carbohydrates in the evening meal of healthy subjects improves glucose tolerance, lowers inflammatory markers, and increases satiety after a subsequent standardized breakfast. J Nutr 2008; 138: 732–739. | PubMed | ChemPort |
  4. Pryde SE, Duncan SH, Hold GL, Stewart CS, Flint HJ. The microbiology of butyrate formation in the human colon. FEMS Microbiol Lett 2002; 217: 133–139. | Article | PubMed | ISI | ChemPort |
  5. Cani PD, Neyrick AM, Fava F, Knaut C, Burcelin RG, Tuohy KM et al. Selective increases of bifidobacteria in gut microflora improve high fat diet induced diabetes in mice through a mechanism associated with endotoxemia. Diabetologia 2007; 50: 2374–2383. | Article | PubMed | ChemPort |
  6. Ley RE, Turnbaugh PJ, Klein S, Gordon JI. Microbial ecology—human gut microbes associated with obesity. Nature 2006; 444: 1022–1023. | Article | PubMed | ISI | ChemPort |
  7. Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER, Gordon JI. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature 2006; 444: 1027–1031. | Article | PubMed | ISI |
  8. Turnbaugh PJ, Backhed F, Fulton L, Gordon JI. Diet-induced obesity is linked to marked but reversible alterations in the mouse distal gut microbiome. Cell Host Microbe 2008; 3: 213–223. | Article | PubMed | ChemPort |
  9. Duncan SH, Belenguer A, Holtrop G, Johnstone AM, Flint HJ, Lobley GE. Reduced dietary intake of carbohydrate, by obese subjects, results in decreased concentrations of butyrate and butyrate-producing bacteria in feces. Appl Environ Microbiol 2007; 73: 1073–1078. | Article | PubMed | ChemPort |
  10. Aminov I, Walker AW, Duncan SH, Harmsen HJM, Welling GW, Flint HJ. Molecular diversity, cultivation, and improved detection by fluorescent in situ hybridization of a dominant group of human gut bacteria related to Roseburia spp. or Eubacterium rectale. Appl Environ Microbiol 2006; 72: 6371–6376. | Article | PubMed | ChemPort |
  11. Manz W, Amann R, Ludwig W, Vancanneyt M, Schleifer KH. Application of a suite of 16S rRNA-specific oligonucleotide probes designed to investigate bacteria of the phylum Cytophaga–Flavobacter–Bacteroides in the natural environment. Microbiology 1996; 142: 1097–1106. | PubMed | ISI | ChemPort |
  12. Franks AH, Harmsen HJM, Raangs GC, Jansen GJ, Schut F, Welling GW. Variations of bacterial populations in human feces measured by fluorescent in situ hybridization with group-specific 16S rRNA-targeted oligonucleotide probes. Appl Environ Microbiol 1998; 64: 3336–3345. | PubMed | ChemPort |
  13. Suau A, Bonnet R, Sutren M, Godon JJ, Gibson GR, Collins MD et al. Direct analysis of genes encoding 16S rRNA from complex communities reveals many novel molecular species within the human gut. Appl Environ Microbiol 1999; 65: 4799–4807. | PubMed | ISI | ChemPort |
  14. Hold GL, Pryde SE, Russell VJ, Furrie E, Flint HJ. Assessment of microbial diversity in human colonic samples by 16S rDNA sequence analysis. FEMS Microbiol Ecol 2002; 39: 33–39. | Article | ChemPort |
  15. Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen L, Sargent M et al. Diversity of the human intestinal microbial flora. Science 2005; 308: 1635–1638. | Article | PubMed | ISI |
  16. Johnstone AM, Horgan GW, Murison SD, Bremner DM, Lobley GE. Effects of high protein weight loss diets on hunger, appetite, and weight loss in obese men feeding ad libitum. Am J Clin Nutr 2008; 87: 44–55. | PubMed | ChemPort |
  17. Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate DAT. McCance and Widdowson's The Composition of Foods. The Royal Society of Chemistry: Cambridge, UK, 1991.
  18. Bartosch S, Fite A, Macfarlane GT, McMurdo MET. Characterization of bacterial communities in feces from healthy elderly volunteers and hospitalized elderly patients by using real-time PCR and effects of antibiotic treatment on the fecal microbiota. Appl Environ Microbiol 2004; 70: 3575–3581. | Article | PubMed | ISI | ChemPort |
  19. Liu C, Song Y, McTeague M, Vu AW, Wexler H, Finegold SM. Rapid identification of the species of the Bacteroides fragilis group by multiplex PCR assays using group- and species-specific primers. FEMS Microbiol Lett 2003; 222: 9–16. | Article | PubMed | ChemPort |
  20. Fuller Z, Louis P, Mihajlovski A, Rungapamestry V, Ratcliffe B, Duncan AJ. Influence of cabbage processing methods and prebiotic manipulation of colonic microflora on glucosinolate breakdown in man. Br J Nutr 2007; 98: 364–372. | Article | PubMed | ChemPort |
  21. Ott SJ, Musfeldt M, Timmis KN, Hampe J, Wenderoth DF, Schreiber S. In vitro alterations of intestinal bacterial microbiota in fecal samples during storage. Diagnostic Microbiol Infect Dis 2004; 50: 237–245. | Article | ChemPort |


The Rowett Institute of Nutrition and Health, and Biomathematics and Statistics Scotland receive support from the Scottish Government Rural Environment Research and Analysis Directorate. We also acknowledge support for this work from the World Cancer Research Fund.



Extra navigation