RESEARCH ARTICLE
The cellulose-degrading microbial community of the human gut
varies according to the presence or absence of methanogens
Christophe Chassard, Eve Delmas, Céline Robert & Annick Bernalier-Donadille
Unité de Microbiologie UR454, INRA, Centre de Recherches de Clermont-Ferrand/Theix, Saint Genès-Champanelle, France
Correspondence: Annick BernalierDonadille, Unité de Microbiologie UR454,
INRA, Centre de Recherches de ClermontFerrand/Theix, 63 122 Saint GenèsChampanelle, France. Tel./fax: 133 4 73 62
46 42; e-mail: [email protected]
Received 14 January 2010; revised 11 June
2010; accepted 16 June 2010.
Final version published online 19 July 2010.
DOI:10.1111/j.1574-6941.2010.00941.x
MICROBIOLOGY ECOLOGY
Editor: Julian Marchesi
Keywords
human intestinal microbiota; cellulose
degradation; bacterial community structure;
metabolic activity; methanogenesis.
Abstract
Cellulose-degrading microorganisms involved in the breakdown of plant cell wall
material in the human gut remain rather unexplored despite their role in intestinal
fermentation. Microcrystalline cellulose-degrading bacteria were previously identified
in faeces of methane-excreting individuals, whereas these microorganisms were
undetectable in faecal samples from non-methane excretors. This suggested that the
structure and activity of the cellulose-degrading community differ in methane- and
non-methane-excreting individuals. The purpose of this study was to characterize in
depth this cellulose-degrading community in individuals of both CH4 statuses using
both culture-dependent and molecular methods. A new real-time PCR analysis was
developed to enumerate microcrystalline cellulose-degrading ruminococci and used
to confirm the predominance of these hydrolytic ruminococci in methane excretors.
Culture-dependent methods using cell wall spinach (CWS) residue revealed the
presence of CWS-degrading microorganisms in all individuals. Characterization of
CWS-degrading isolates further showed that the main cellulose-degrading bacteria
belong essentially to Bacteroidetes in non-methane-excreting subjects, while they are
predominantly represented by Firmicutes in methane-excreting individuals. This
taxonomic diversity was associated with functional diversity: the ability to degrade
different types of cellulose and to produce H2 from fermentation differed depending
on the species. The structure of the cellulolytic community was shown to vary
depending on the presence of methanogens in the human gut.
Introduction
Dietary fibres, i.e. indigestible plant residues, are important
in maintaining healthy gastrointestinal functions. These
fibres enter the human colon, where they are fermented by
the gut microbial community (Flint et al., 2007). Microbial
fermentation of dietary fibres is now recognized as providing human health benefits, for example regulation of
digestive transit and a preventive role in cardiovascular
diseases and cancers through the production of short-chain
fatty acids and release of plant polymer-associated bioactive
compounds with antioxidant activity (Schweizer & Edwards, 1991). Plant cell wall polysaccharides, mainly cellulose, hemicelluloses and pectin, found in cereals, vegetables
and fruits, represent one of the most important fractions of
dietary fibres. The daily consumption of plant cell wall
polysaccharides varies from 10 to 25 g in European adults
or about 30% of the total dietary fibre ingested (Lairon et al.,
FEMS Microbiol Ecol 74 (2010) 205–213
2003, 2005). Despite the potential health benefits associated
with microbial digestion of these substrates, studies on gut
bacterial communities involved in the breakdown and
fermentation of these insoluble polysaccharides are lacking.
Attempts to elucidate the cellulose-degrading microbial
community have only been partially successful as only a
restricted number of individuals appear to harbour such
cellulose-degrading organisms (Bétian et al., 1977; Montgomery, 1988; Wedekind et al., 1988). We demonstrated
previously that the presence or absence of microcrystalline
cellulose-degrading bacteria was related to the methane
status of volunteers (Robert & Bernalier-Donadille, 2003).
This population was found in the faecal samples of the
majority of methane-excreting subjects (CH1
4 ) studied,
while it was not detected in specimens from non-methaneexcreting individuals (CH
4 ). The dominant microcrystalline
cellulose-degrading species isolated from CH1
4 subjects
belong to new Ruminococcus sp. and some Enterococcus sp.
2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
206
closely related to Enterococcus faecalis (Robert & BernalierDonadille, 2003). These microcrystalline cellulose-degrading species were shown to produce high amounts of H2 from
cellulose fermentation.
The absence of cultivable microcrystalline cellulose-degrading organisms in faecal samples from CH
4 individuals
does not reflect the total absence of cellulose-degrading
bacteria in this non-methane-producing ecosystem as both
cellulase and xylanase activities could be detected in this
faecal microbial community (Robert & Bernalier-Donadille,
2003). A unique isolate described as Bacteroides cellulosilyticus (Robert et al., 2007), a new Bacteroides sp. able to grow
with pure celluloses, was isolated recently from a CH
4
subject. We therefore hypothesize that the structure and
activity of the cellulolytic community from CH
4 individuals
may differ from that of CH1
4 individuals. The objective of
the present study was to characterize in depth the cellulose1
degrading community from the gut of CH
4 and CH4
individuals using both culture-dependent and molecular
approaches. Populations of microcrystalline cellulose-degrading ruminococci and methanogens were first investigated using real-time PCR analysis in a large number of
subjects (n = 49). Culture-dependent studies were further
performed using different types of pure celluloses and cell
wall spinach (CWS) residue in order to identify cellulosedegrading organisms, especially in CH
4 individuals.
Materials and methods
Human subjects
Forty-nine healthy subjects, aged 20–54 years, were recruited
for the present study. All consumed a diversified Western
diet including 10–15 g of fibre per day. No vegetarians were
included in the study. All volunteers had a normal body
mass index (BMI between 18 and 25 kg m2). None had
received antibiotics during the 3 months preceding the study
and none had any previously diagnosed gastrointestinal
disease. All participants gave informed consent to the
protocol, which was approved by the local ethics committee.
Breath methane concentration was determined using a
Microlyser-SC gas chromatograph (Quintron Instrument,
Biotech, France). Twenty-nine volunteers, who did not expire
any CH4, were selected as non-methane excretors (CH
4 ). The
other 20 subjects, selected as methane excretors (CH1
4 ),
excreted CH4 at a concentration in excess of at least 1 p.p.m.
of the ambient atmosphere (Bond et al., 1971).
Collection and preparation of faecal samples for
analyses
Freshly voided faeces were obtained from all subjects, stored
at 4 1C under anaerobic conditions and processed within a
few hours. Each sample was mixed, and five aliquots were
2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
C. Chassard et al.
stored at 80 1C for DNA extraction and quantitative realtime PCR analysis. One gram of faecal sample was collected
1
from 18 selected volunteers (nine CH
4 and nine CH4 ) and
was 10-fold (wet w/v) diluted in an anaerobic mineral
solution. Serial 10-fold dilutions to 1011 were then carried
out for culture-dependent analysis.
Quantitative real-time PCR analysis
Population levels of microcrystalline cellulose-degrading
ruminococci and methanogens were evaluated in faecal
samples from CH1
4 (n = 20) and CH4 individuals (n = 29)
TM
system, Roche Moleusing real-time PCR (LightCycler
cular Biochemicals Mannheim, Germany).
Primers specifically targeting the 16S rRNA gene of the new
cellulose-degrading Ruminococcus phylum (Robert & Bernalier-Donadille, 2003) were designed and validated (RumcelF:
GTTCAGTCCTTCGGGACACAA, RumcelR: TAGCAATTCC
GACTTCAT, amplicon = 336 pb). Primers were initially selected based on the sequence alignment of cellulose-degrading
human strains, Ruminococcus sp. from clostridial cluster IV
and XIVa, while the specificity of the primers was checked
with the Probe match function of the Ribosome Database
Project II. Primer specificity was verified using a panel of
4 20 strains belonging to the major bacterial genera found in
humans and two major cellulose-degrading Ruminococcus sp.
(Ruminococcus albus and Ruminococcus flavefaciens) from the
rumen ecosystem (Table 1). Amplification was performed in a
total volume of 25 mL with a primer concentration of
0.5 mmol L1 each according to the following protocol: one
cycle of denaturation (95 1C for 10 min), 35 cycles of amplification (95 1C for 15 s, 60 1C for 10 s and 72 1C for 30 s), one
melting curve programme (65–95 1C with a heating rate of
0.1 1C s1) and a cooling step to 45 1C. Standard template
DNA was prepared from the 16S rRNA gene of Ruminococcus
sp. strain 18P13 (DSMZ 18848) by amplification with F8 (5 0 AGAGTTTGATCMTGGCTC-3 0 ) and 1492R (5 0 -GNTACCTT
GTTACGACTT-3 0 ) primers and purification as described
previously (Mosoni et al., 2007). Standard curves were prepared with five standard concentrations from 10-fold serial
dilutions in water, ranging from 108 to 104 copies. Methanogens were quantified using the primers described previously
and following the conditions of analysis defined by OheneAdjei et al. (2007).
Media and enumeration procedures
All liquid and solid media were prepared, dispensed and
inoculated using strictly anaerobic techniques (Hungate,
1969), with 100% O2-free CO2 gas. Total anaerobes,
cellulose and CWS-degrading bacteria, and methanogens
were enumerated using the most probable number (MPN)
estimation (Clarke & Owens, 1983). Three tubes were
FEMS Microbiol Ecol 74 (2010) 205–213
207
Cellulolytic microbiota and CH4 production in the human gut
Table 1. Validation of 16S rRNA gene primers for real-time PCR analysis
of human cellulolytic ruminococci
Bacterial strains
Clostridial cluster IV
Ruminococcus sp. strain 18P13z
Ruminococcus sp. strain 9SE51z
Ruminococcus sp. strain 25F8
Ruminococcus sp. strain 7L75z
Ruminococcus callidus (ATCC 27760)
Ruminococcus albus (ATCC 27211)
Ruminococcus flavefaciens (ATCC 19208)
Clostridial cluster XIVa
Ruminococcus obeum (ATTC 29174)
Blautia producta (DSM 2950)
Blautia hydrogenotrophica (DSM 10507)
Roseburia intestinalis (strain XB6B4)‰
Roseburia faecis (strain 11SE39)z
Others
Bacteroides cellulosilyticus (DSM 14838)k
Bacteroides xylanisolvens (DSM 18836)
Bacteroides vulgatus (DSM 1447)
Fibrobacter succinogenes (ATCC 19169)
Enterococcus faecalis (strain 7L76)z
Bifidobacterium adolescentis (DSM 20083)
Atopobium parvulum (DSM 20469)
Specific PCR
detection
with Rumcel
Cellulosedegrading
bacteriaw
1
1
1
1
1
1
1
1
1
1
1
1
1
1, species detected; , species not detected.
w
1, reported as cellulose-degrading species (Chesson & Forsberg, 1988;
Robert & Bernalier-Donadille, 2003; Robert et al., 2007); , not
reported as cellulose-degrading species.
z
Robert & Bernalier-Donadille (2003).
‰
Chassard et al. (2007).
z
This study.
k
Robert et al. (2007).
inoculated per dilution for each cellulosic substrate and for
total anaerobe enumeration.
Total anaerobes were enumerated in a complex medium
containing clarified rumen fluid (Leedle & Hespell, 1980) and
bacterial growth was recorded in culture tubes of each dilution
inoculated (107–1011) after 48 h of incubation at 37 1C.
Cellulose- and CWS-degrading bacteria were enumerated
in basal cellulolytic (BC) medium as described previously
(Robert & Bernalier-Donadille, 2003). Two pure celluloses
(Whatman no. 1 filter paper cellulose and Sigmacells 101
cellulose) and a CWS extract were used as substrates. CWS
residue was obtained by pebble milling fresh spinach in
liquid nitrogen as described previously (Robert et al., 2007).
A strip of 50 mg Whatman no. 1 filter paper cellulose was
added to each culture tube before the addition of the
prereduced BC medium (10 mL per tube). After incubation
at 37 1C for 21 days, cellulose utilization in the cultures
obtained from each dilution inoculated (103–109) was
detected visually by noting the degradation of the piece of
FEMS Microbiol Ecol 74 (2010) 205–213
filter paper. Sigmacells 101 cellulose and cell wall spinach
were rehydrated in distilled water and solutions of each
substrate (2%) were pebble milled for one night before
addition to the BC medium at a final concentration
of 0.4%. After incubation at 37 1C for 15 days, the presence
of endoglucanase activity, the carboxymethyl cellulase
(CMCase), was examined in each culture tube inoculated
with faecal dilutions (from 103 to 109). CMCase activity
was detected on agar plates containing the following mixture: 50 mM sodium phosphate buffer (pH 6.8), 0.3%
carboxymethylcellulose (Sigma Chemicals Co., St. Louis)
and 1.5% bacteriologic agar type A (Biokar Diagnostics).
Supernatant fluids (0.2 mL) of each culture were deposited
in a hole formed in the agar mixture and plates were
incubated for 10 h at 37 1C. The CMCase activity was
visualized with Congo red after extensive washing of the
agar plates with 1M NaCl (Forano et al., 1994).
H2-utilizing methanogens were enumerated as described by
Doré et al. (1995). Briefly, faecal sample dilutions from 102 to
109 were inoculated into broth culture medium. After inoculation, culture tubes were pressurized with H2/CO2 (80/20 v/v,
202 kPa). Vials with CH4 levels above 0.1% after 15 days of
incubation at 37 1C were considered positive.
Isolation of CWS-degrading bacteria from faecal
samples
Enrichments of CWS-degrading bacteria were obtained
from the highest dilution tubes in MPN estimation in
CWS–BC medium, showing CMCase activity. Two to three
weekly subcultures were performed in CWS-BC broth
medium before transferring the enriched cultures into roll
tubes containing the same medium with 2% Bacto agar and
0.7% hydrated CWS residue. Colonies developing in agar
medium were transferred from roll tubes into a CWS–BC
broth medium. The CMCase activity of the culture was then
checked on the agar mixture as described above. Two to five
successive subcultures on roll tubes and broth BC medium
were carried out to obtain a pure culture of isolates showing
CMCase activity. Isolates were then examined for purity,
morphology and Gram staining by bright-field microscopy,
in CWS- and glucose (2 g L1)-grown cultures.
Degradation of CWS and pure celluloses by
isolates
CWS-degrading isolates were cultivated in a liquid BC
medium (10 mL per tube) containing 100 mg of Avicels
or Sigmacells 101 cellulose or 100 mg of CWS as the
sole energy source. Inocula were composed of 0.3 mL of
5-day-old cultures on a CWS–BC medium. Three culture
tubes were inoculated for each incubation time and each
strain studied. Substrate degradation by the isolates was
determined by measuring the dry weight of the substrate
2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
208
C. Chassard et al.
remaining in the culture tubes after 10 days of incubation at
37 1C (Chassard et al., 2005).
Fermentation end-products of each isolate were determined
in the 10-day-old cultures obtained with the different pure
celluloses or CWS as the sole energy source. Headspace gases
were analysed by gas-phase chromatography (Chassard
et al., 2005), and short-chain fatty acid production was
quantified by 1D 1H NMR (Chassard et al., 2008).
Number of 16S rRNA gene
copies of cellulolytic
ruminococci (log N g–1)
Fermentation of pure celluloses and CWS by
isolates
CH4+ subjects
CH4– subjects
9
8
7
3
Sequencing of the 16S rRNA gene and
phylogenetic placement of CWS-degrading
isolates
Sequencing of the 16S rRNA gene and phylogenetic assignment of isolates were performed as described previously
(Chassard et al., 2007). The 16S rRNA gene (c. 1500 bp) was
amplified using the universal primers F8 and 1492R. Two
PCR reactions were performed per strain. PCR products
were purified using the Stratapreps PCR Purification kit
(Ozyme, St Quentin en Yvelines, France) according to the
manufacturer’s instructions. The concentration of amplified
DNA was estimated in ethidium bromide-stained agarose
gels by densitometry and c. 50 ng of each product was
included in a 20-mL sequencing reaction. Amplicons were
sequenced to full length in both directions using an ABI
Prism BigDie terminator cycle sequencing kit (Perkin Elmer
Corporation, Norwalk, CT) with appropriate primers by the
ABI Prism 310 genetic analyzer (Perkin Elmer Corporation).
Data and statistical analysis
The results of bacterial counts are reported as log10 microorganisms per gram wet weight of faeces. Data were expressed as mean SE (n = number of faecal samples
studied). Statistical analyses were performed with INSTAT 3.0
GRAPHPAD software using Student’s t-test. All tests were twotailed and the level used to establish significance was
P o 0.05. The statistical correlation between bacterial population levels was studied using the linear Pearson correlation.
Results
Quantification of population levels of human
microcrystalline cellulose-degrading
ruminococci and methanogens in faecal samples
from CH1
4 and CH4 individuals
Newly designed primers for human cellulose-degrading ruminococci enumeration were validated successfully for specificity (Table 1). A broad range of dominant species including
Ruminococcus sp. with a human origin and major cellulose2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
4
5
6
7
8
9
Number of 16S rRNA gene copies of
methanogens (log N g–1)
Fig. 1. Correlation between human cellulolytic ruminococci and methanogen populations in faecal samples from methane-producing (CH1
4)
and non-methane-producing (CH
4 ) faecal samples (n = 49), estimated by
quantitative real-time PCR.
degrading Ruminococcus sp. from the rumen ecosystem
(R. albus and R. flavefaciens) were tested. All microcrystalline
cellulose-degrading Ruminococcus strains previously isolated
from the human gut (Robert & Bernalier-Donadille, 2003)
were enumerated using RumcelF and RumcelR primers.
Amplification was observed neither with other human intestinal bacteria nor with cellulose-degrading Ruminococcus
sp. of ruminal origin (Table 1). As expected, the population
levels of methanogens were significantly higher in CH1
4
individuals than in CH
4 ones (log 7.8 1.2 vs. log 3.2 0.8 g
faeces, respectively). The population of human microcrystalline cellulose-degrading ruminococci was detected in the
majority of subjects regardless of the methane excretion status
(Fig. 1), and this population level ranged from log 6.1 to
log 9.2 g faeces depending on the subject considered. The level
of this ruminococci population was, however, significantly
higher (P o 0.0001) in CH1
4 individuals (log 8.1 0.6 g
faeces) than in CH
4 individuals (log 7.2 0.5 g faeces). A
correlation (r2 = 0.463) was observed between the population
levels of human microcrystalline cellulose-degrading rumi
nococci and methanogens in faecal samples of CH1
4 and CH4
1
subjects (Fig. 1). In CH4 subjects, a high population level of
methanogens was determined to correlate significantly
(P = 0.008) with a high population level of human cellulosedegrading ruminococci.
Evaluation of the cellulose- and CWS-degrading
population in faecal samples from CH1
4 and CH4
individuals
The population levels of cultivable cellulose- and CWSdegrading communities were estimated in faecal samples
using different types of substrates. The results of these
FEMS Microbiol Ecol 74 (2010) 205–213
209
Cellulolytic microbiota and CH4 production in the human gut
bacterial counts are shown in Table 2. As reported previously
(Robert & Bernalier-Donadille, 2003), microcrystalline (filter paper) cellulose-degrading bacteria could not be cultivated from CH
4 faecal samples while the level of this
population reached 106–107 g1 faeces in CH1
4 faecal specimens. The use of Sigmacells101 cellulose allowed the
detection of cellulose-degrading organisms in all CH1
4 faecal
samples as well as in six of the nine CH
4 individuals studied.
The Sigmacells101 cellulose-degrading community was
detected at similar levels in CH1
4 and in the six CH4 individuals, with the average number of these bacteria nearly
reaching 107 g1 faeces. The Sigmacells101 cellulose-degrading population, however, remained undetectable ( o 103 g1
faeces) in three CH
4 faecal samples. By contrast, CWSdegrading bacteria were found in all faecal samples from both
CH1
4 and CH4 subjects. The average number of the CWSdegrading population was not significantly different between
CH1
4 and CH4 individuals, with the population levels of
this community varying from 106 to 109 bacteria g1 faeces,
depending on the faecal sample considered. The CWS-degrading population was detected at higher levels than either the
filter paper- or the Sigmacells 101-degrading communities, in
1
both CH
4 and CH4 individuals.
Isolation of CWS-degrading bacteria from faecal
samples
Enrichments of CWS-degrading bacteria were obtained from
seven faecal samples of subjects harbouring high levels
(107–108) of CWS-degrading bacteria. The highest dilution
cultures from MPN estimation exhibiting carboxymethylcellulase activity were used for the isolation of CWS-degrading
organisms (Table 3). These cultures were transferred two to
three times on CWS–BC medium. A total of 14 CWSdegrading strains were then obtained after three to four
successive isolations on agar roll tubes with CWS as a substrate.
The seven isolates obtained from three CH
4 faecal
samples were Gram-negative coccobacillus (Table 3) and
Table 2. Distribution of the cellulose- and cell wall spinach-degrading populations in faecal samples from methane-excreting (CH1
4 ) and non-methaneexcreting (CH
4 ) individuals (n = 18)
Microbial counts (log10 N g1 faeces)
Subjects
Total
anaerobes
Filter paper-degrading
bacteria
Sigmacells 101-degrading
bacteria
Cell wall spinach-degrading
bacteria
Methanogenic
archaea
CH1
4
n=9
CH
4
n=9
P
11.3 0.9
[9.9–12.2]
10.7 0.6
[9.9–11.9]
NS
6.3 0.9
[5.7–7.7]
o 3.0
7.7 1.6
[5.3–9.5]
6.6 2.9
[3.0–9.9]
NS
8.6 0.8
[7.9–9.5]
7.8 0.9
[6.2–9.2]
NS
8.8 1.1
[7.0–9.9]
4.5 1.8
[3.0–7.6]
0.001
0.001
NS, not statistically different (P 4 0.05).
Table 3. Origin and phylogenic analysis of CWS-degrading strains isolated from seven human faecal samples according to the methane status
Subjects
Methane
status
CWS-degrading
population level
E
G
CH
4
CH
4
8.5
6.9
I
CH
4
7.6
K
M
CH1
4
CH1
4
8.1
9.5
N
O
CH1
4
CH1
4
7.9
9.5
Strain
Bacterial species (%16S rRNA gene
sequence similarity)
CRE27
35AE31
35AE37
35AE34
35AE35
31S15
31S18
7SE20
8SE23
8SE26
9SE51
11SE37
11SE38
11SE39
Bacteroides cellulosilyticusw (100%)
Bacteroides cellulosilyticus (99%)
Bacteroides cellulosilyticus (98%)
Bacteroides cellulosilyticus (98%)
Bacteroides cellulosilyticus (98%)
Bacteroides cellulosilyticus (98%)
Bacteroides cellulosilyticus (98%)
Enterococcus faecalis (99%)
Enterococcus faecalis (99%)
Enterococcus faecalis (99%)
Ruminococcus sp. nov.z (98%)
Roseburia faecis (99%)
Roseburia faecis (99%)
Roseburia faecis (99%)
The population level is expressed as log N g1 faeces.
w
Robert et al. (2007).
Robert & Bernalier-Donadille (2003).
z
FEMS Microbiol Ecol 74 (2010) 205–213
2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
210
C. Chassard et al.
expressed CMCase activity in CWS-broth cultures.
Seven different strains were obtained from CH1
4 faecal
samples and all exhibited positive or variable Gram staining
(Table 3). Three strains were identified as facultative anaerobes (strains 7SE20, 8SE23, 8SE26), while only one strain
was identified as a Gram-positive coccus (strain 9SE51).
Three other strains (11SE37, 11SE38 and 11SE39) were
identified as Gram-variable curved rods. All isolates expressed CMCase activity in CWS broth medium.
Sequencing of the 16S rRNA gene of CWSdegrading isolates
A nearly complete sequence of the 16S rRNA gene (around
1400 nucleotides) was determined for the 14 CWS-degrading isolates. Sequence similarity analysis showed that all
Gram-negative coccobacilli isolated from CH
4 faecal samples belonged to the Bacteroides taxon (Eckburg et al., 2005)
(Table 3). Phylogenetic analysis further demonstrated that
these isolates could be classified into only one subgroup that
was closely related (at least 98% of 16S rRNA gene sequence
similarity) to B. cellulosilyticus, the cellulose-degrading Bacteroides sp. described previously by Robert et al. (2007). The
16S rRNA gene sequences of these Bacteroides isolates were
deposited in the EMBL database (AJ583244, AJ583245,
AJ583246, AJ583247, AJ583248 and AJ583249).
16S rRNA gene sequence analysis of the three facultative
anaerobes (strains 7SE20, 8SE23, and 8SE26) demonstrated
that these isolates could be assigned to Enterococcus sp., as
their sequence showed 99% similarity to E. faecalis (Table 3).
Strain 9SE51 was assigned to the Ruminococcus genus (cluster
IV of Clostridiacaea), with the closest relative of this isolate
corresponding to the microcrystalline cellulose-degrading
Ruminococcus sp. described previously by Robert & BernalierDonadille (2003). The 16S rRNA gene sequence of this
Ruminococcus isolate was deposited in the EMBL database
(FM954974). The three identified Gram-variable curved rods
(strains 11SE37, 11SE38 and 11SE39) were assigned to the
genus Roseburia (Table 3), with their sequences (EMBL
accession numbers FM954975, FM954976 and FM954977)
showing 99% similarity to the type strain of Roseburia faecis.
Ability of different CWS-degrading isolates to
degrade and ferment pure celluloses and CWS
Four CWS-degrading species, all representatives of the
various phylogenetic groups of isolates described by Robert
& Bernalier-Donadille (2003) and identified in this study,
were analysed for their ability to degrade and ferment both
celluloses and CWS (Fig. 2). Bacteroides isolates degraded
different types of pure cellulose such as Avicels pH-101 in
addition to CWS, but were unable to degrade microcrystalline (filter paper) cellulose. In contrast, Roseburia isolates
degraded CWS, but not the pure celluloses tested (Avicels
pH-101, Sigmacells 101 and filter paper). Enterococcus and
Ruminococcus isolates degraded all different types of celluloses including filter paper cellulose. Furthermore, these
Gram-positive species were also particularly efficient in
degrading CWS (Fig. 2).
Metabolite production and ratios varied depending on
the strain considered. Bacteroides sp. produced mainly
acetate, propionate and succinate during cellulose metabolism, while Enterococcus and Ruminococcus isolates synthesized mostly acetate and succinate and Roseburia sp. formed
butyrate (data not shown). The production of hydrogen also
varied considerably depending on the species (Fig. 2).
Bacteroides sp. produced only small amounts of hydrogen
in cellulose-grown cultures, while Enterococcus, Ruminococcus and Roseburia sp. produced large amounts of this gas
from cellulose and/or CWS fermentation.
Discussion
The present study demonstrates that cellulose-degrading
organisms are present in the gut microbiota of all individuals, but that the structure of this cellulose-degrading
community differs largely depending on the CH4 status of
Bacteroides sp. Ruminococcus sp. Enterococcus sp. Roseburia sp.
strain CRE21
strain 18P13
strain 7L76
strain 11SE39
Substrate
degradation
(expressed as mg
of dry matter
disappeared)
Hydrogen
production
(expressed in
µmol)
60
50
40
30
20
10
0
10
20
30
40
50
60
70
80
CWS
2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
Fig. 2. Cellulose or CWS disappearance and
production of H2 by Bacteroides sp., Ruminococcus sp., Enterococcus sp. and Roseburia sp.
after 10 days of incubation at 37 1C. Substrate
degradation is expressed as mg of dry matter
disappeared and H2 production in mmol.
FEMS Microbiol Ecol 74 (2010) 205–213
211
Cellulolytic microbiota and CH4 production in the human gut
the subject. The dominant cellulose degraders isolated from
non-methane-excreting subjects belong mainly to Bacteroidetes, while they are predominantly composed of Firmicutes
in methane-excreting individuals.
Enumeration of cellulose-degrading organisms in human
faecal samples appears to be dependent on the type of cellulose
used in the culture medium. This hypothesis stems from our
previous results (Robert & Bernalier-Donadille, 2003), demonstrating that microcrystalline cellulose-degrading bacteria
could only be detected in methane-excreting subjects. The
relationship between a detectable microcrystalline cellulosedegrading population and the CH4 status of the volunteers was
confirmed in the present study. We further demonstrate here
that the population level of the main microcrystalline cellulose-degrading Ruminococcus sp. identified previously (Robert
& Bernalier-Donadille, 2003) correlates with the population
levels of methanogens in the gut. While this Ruminococcus
group was detected in all faecal samples, this bacterial group
was determined to represent a subdominant community in
CH
4 individuals. However, no microcrystalline cellulose-degrading Ruminococcus sp. could be cultivated from CH
4
subjects, suggesting that some members of this phylogenetic
subgroup could be non-cellulose-degraders such as Ruminococcus bromii, another member of the cluster IV subgroup
capable of degrading resistant starch, but not cellulose.
The use of crystalline Sigmacells 101 cellulose was not as
successful in detecting the cellulose-degrading community
in faeces from both CH1
4 and CH4 subjects. The presence of
s
the Sigmacell 101 cellulose-degrading population was not
related to the CH4 status of volunteers in contrast to the
microcrystalline cellulose-degrading population. However,
Sigmacells 101 cellulose-degrading organisms could not be
found in three of the nine CH
4 faecal samples examined. In
contrast, CWS-degrading populations were detected in all
subjects, regardless of the methane excretion status. CWS
enrichments were therefore selected for further isolation of
CWS-degrading organisms. CWS was determined to be an
effective substrate for the exploration of the bacterial
diversity of cellulose-degrading microorganisms in humans.
All bacterial species isolated previously from CH1
4 subjects
using microcrystalline (filter paper) cellulose (Robert &
Bernalier-Donadille, 2003) were also identified using CWS
as a substrate. Several more species belonging to the
Bacteroides and Roseburia genera were also identified using
CWS as a substrate. All seven CWS-degrading strains
isolated from CH
4 individuals were determined to be closely
related to the cellulose-degrading Bacteroides sp., B. cellulosilyticus (Robert et al., 2007). These different Bacteroides
isolates were determined as belonging to the same Bacteroides phylotype despite being isolated from three different
subjects.
Bacteroides represents a predominant taxon in the colon
(Eckburg et al., 2005; Chassard et al., 2008). Several species
FEMS Microbiol Ecol 74 (2010) 205–213
within this taxon have been shown to express xylanolytic
activity (Bacteroides ovatus, Bacteroides fragilis, Bacteroides
xylanisolvens) (Hespell & Whitehead, 1990; Chassard et al.,
2007); however, very few species exhibited clear cellulosedegrading activity. Only B. cellulolysiticus (Robert et al., 2007)
and one other Bacteroides-like strain isolated from human
faeces by Bétian et al. (1977) have been reported to be capable
of degrading pure cellulose. Identification of the B. cellulosilyticus subgroup through our enrichment cultures suggests that
this subgroup may represent an important cellulose-degrading
Bacteroides phylotype of the human gut. This finding further
supports the hypothesis that the predominant Bacteroidetes
contribute to plant cell wall degradation in the human gut
through hydrolysis of main polysaccharides such as xylan and
cellulose (Macfarlane & Macfarlane, 2006). Our attempt to
specifically quantify the B. cellulosilyticus subgroup by realtime PCR in faecal samples remains elusive. Primer sets
designed for this Bacteroides subgroup were capable of detecting the targeted cellulose-degrading species; however, other
non-cellulose-degrading Bacteroides sp. were also quantified.
This is likely attributed to the close phylogenetic relationship
between B. cellulosilyticus and some predominant Bacteroides
sp. found in the human gut such as Bacteroides intestinalis
(Robert et al., 2007).
While CWS-degrading isolates from CH
4 faecal samples
were determined as belonging to the Bacteroidetes phylum,
CWS-degrading isolates obtained from faecal samples of
CH1
4 subjects were identified as Firmicutes and could be
assigned to different bacterial genera including Ruminococcus, Enterococcus and Roseburia. This result supports and
extends the previous hypothesis (Robert & Bernalier-Donadille, 2003) of Gram-positive cellulose-degrading organisms
dominating in CH1
4 individuals. Although Ruminococcus
and Enterococcus isolates were mainly obtained using microcrystalline cellulose as a substrate (Robert & BernalierDonadille, 2003), CWS represented a more favourable
substrate for the growth and isolation of Enterococcus sp.
and Roseburia sp. The distribution of these two bacterial
subgroups in CH1
4 and CH4 subjects was not evaluated in
this study. Roseburia sp. are an important group of butyrateproducing bacteria that predominate in the gut of all healthy
human individuals (Aminov et al., 2006; Duncan et al.,
2006), while Enterococcus sp. are also indigenous, but subdominant bacterial species. The role of Enterococcus sp. and
Roseburia sp. in cellulose breakdown in the gut has not been
explored extensively in contrast to the Ruminococcus sp.,
which are predominant cellulose degraders in the rumen
(Flint et al., 2008).
The taxonomic diversity of the cellulose-degrading com
munity found in CH1
4 and CH4 individuals strongly correlated with functional diversity. The ability to degrade and
ferment different types of cellulose varied considerably
depending on the bacterial species considered. Bacteroides
2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
212
isolates degraded pure celluloses, but not microcrystalline
filter paper, while Ruminococcus sp. and Enterococcus sp.
efficiently hydrolysed all types of substrates (Robert & Bernalier-Donadille, 2003). Roseburia sp. expressed CMCase activity
when grown on CWS, but were unable to grow on pure
cellulose. Roseburia sp. were also shown to express high
xylanase activity (Chassard et al., 2007), suggesting that this
bacterial species was selected on CWS medium because of its
ability to use other complex polysaccharides present in CWS.
Roseburia sp. could thus potentially interact with cellulosedegrading microorganisms to ensure efficient degradation of
complex substrates such as spinach in the human gut.
While the different Bacteroides isolates did not produce H2
from cellulose fermentation, large amounts of this gas were
released by the dominant cellulose-degrading organisms isolated from CH1
4 individuals (Robert & Bernalier-Donadille,
2003; Chassard et al., 2005). Thus, different cellulose-degrading species may contribute differently to H2 production
during fibre fermentation in the human gut. The presence of
H2-producing cellulolytic species in CH1
4 subjects, but not in
CH
4 individuals, further implies that the structure of this
fibrolytic community may vary according to the presence or
absence of methanogens in the gut ecosystem.
Cross-feeding between H2-producing and H2-consuming
microorganisms is a crucial mechanism, essential for the
maintenance of efficient degradation and fermentation of
organic matter. Hydrogen produced at high levels by several
hydrolytic and saccharolytic species in CH1
4 subjects should
be used by methanogens, to ensure an optimal fermentation
of substrates (Robert et al., 2001). This bacterial equilibrium
could, however, be impaired in certain situations, leading to
lower or higher energy salvage of the gut microbiota from
the diet (Samuel & Gordon, 2006; Turnbaugh et al., 2006).
The disturbance of this equilibrium could further impact
the ability of the host to harvest and store calories. Further
studies are needed to elucidate the potential role of bacterial
populations and/or interactions in diseases such as obesity.
In this context, the role of the different H2-producing and
non-H2-producing cellulose-degrading species and their
interactions with hydrogenotrophic microorganisms on gut
health should be explored.
In conclusion, the composition and activity of the cellulosedegrading population were shown to vary considerably according to the presence or absence of methanogens in the human
gut, further suggesting that microbial factors might influence
the CH4 status and affect host energy balance in humans.
Acknowledgements
C.C. and C.R. were supported by a fellowship from the
French Ministère de la Recherche et de l’Enseignement
Supérieur. The authors thank Dr Marcus Yaffee and Dr
Amanda Payne for critically reading the manuscript.
2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
C. Chassard et al.
References
Aminov RI, Walker AW, Ducan SH, Harmsen HJ, Welling GW &
Flint HJ (2006) 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 Microb 72: 6371–6376.
Bétian HG, Linehan BA, Bryant MP & Holdeman LV (1977)
Isolation of cellulolytic Bacteroides sp. from human feces. Appl
Environ Microb 33: 1009–1010.
Bond JH, Engel RR & Levitt MD (1971) Factors influencing
pulmonary methane excretion in man. J Exp Med 133:
572–588.
Chassard C, Gaillard-Martinie B & Bernalier-Donadille A (2005)
Interaction between H2-producing and non-H2-producing
cellulolytic bacteria from the human colon. FEMS Microbiol
Lett 242: 339–344.
Chassard C, Goumy V, Leclerc M, Del’homme C & BernalierDonadille A (2007) Characterization of the xylan-degrading
microbial community from human faeces. FEMS Microbiol
Ecol 61: 121–131.
Chassard C, Scott KP, Marquet P, Martin JC, Del’homme C,
Dapoigny M, Flint HJ & Bernalier-Donadille A (2008)
Assessment of metabolic diversity within the intestinal
microbiota from healthy humans using combined molecular
and cultural approaches. FEMS Microbiol Ecol 66: 496–504.
Chesson A & Forsberg CW (1988) Polysaccharide degradation by
rumen microorganisms. The Rumen Microbial Ecosystem
(Hobson PN, ed), pp. 251–284. Elsevier Applied Science,
London.
Clarke KR & Owens JP (1983) A simple and versatile
microcomputer for the determination of most probable
number. J Microbiol Meth 1: 133–137.
Doré J, Morvan B, Rieu-Lesme F, Goderel I, Gouet P & Pochart P
(1995) Most probable number enumeration of H2-utilizing
acetogenic bacteria from the digestive tract of animals and
man. FEMS Microbiol Lett 15: 7–12.
Duncan SH, Aminov RI, Scott KP, Louis P, Stanton TB & Flint HJ
(2006) Proposal of Roseburia faecis sp. nov., Roseburia hominis
sp. nov. and Roseburia inulinivorans sp. nov., based on isolates
from human faeces. Int J Syst Evol Micr 56: 2437–2441.
Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen L,
Sargent M, Gill SR, Nelson KE & Relman DA (2005) Diversity
of the human intestinal microbial flora. Science 308:
1635–1638.
Flint HJ, Duncan SH, Scott KP & Louis P (2007) Interactions and
competition within the microbial community of the human
colon: links between diet and health. Environ Microbiol 9:
1101–1111.
Flint HJ, Bayer EA, Rincon MT & White BA (2008)
Polysaccharide utilization by gut bacteria: potential for new
insights from genomic analysis. Nat Rev Microbiol 6: 121–131.
Forano E, Broussolle V, Gaudet G & Bryant JA (1994) Molecular
cloning, expression and characterization of a new
endoglucanase gene from Fibrobacter succinogenes S85. Curr
Microbiol 28: 7–14.
FEMS Microbiol Ecol 74 (2010) 205–213
213
Cellulolytic microbiota and CH4 production in the human gut
Hespell RB & Whitehead TR (1990) Physiology and genetics of
xylan degradation by gastrointestinal tract bacteria. J Dairy Sci
73: 3013–3022.
Hungate RE (1969) A roll tube method for the cultivation of strict
anaerobes. Methods in Microbiology, Vol. 3B (Norris JR &
Ribbons DW, eds), pp. 117–132. Academic Press, London.
Lairon D, Bertrais S, Vincent S, Arnault N, Galan P, Boutron MC &
Hercberg S (2003) Dietary fibre intake and clinical indices in the
french supplementation en vitamines et minéraux antioxydants
(SU.VI.MAX) adult cohort. P Nutr Soc 62: 11–15.
Lairon D, Arnault N, Bertrais S, Planells R, Clero E, Hercberg S &
Boutron-Ruault MC (2005) Dietary fiber intake and risk
factors for cardiovascular disease in french adults. Am J Clin
Nutr 82: 1185–1194.
Leedle JA & Hespell RB (1980) Differential carbohydrate media
and anaerobic replica plating techniques in delineating
carbohydrate utilizing subgroups in rumen bacterial
populations. Appl Environ Microb 39: 709–719.
Macfarlane S & Macfarlane GT (2006) Composition and metabolic
activities of bacterial biofilms colonizing food residues in the
human gut. Appl Environ Microb 72: 6204–6211.
Montgomery L (1988) Isolation of human colonic fibrolytic
bacteria. Lett Appl Microbiol 6: 55–57.
Mosoni P, Chaucheyras-Durand F, Béra-Maillet C & Forano E
(2007) Quantification by real-time PCR of cellulolytic bacteria
in the rumen of sheep after supplementation of a forage diet
with readily fermentable carbohydrates: effect of a yeast
additive. J Appl Microbiol 103: 2676–2685.
FEMS Microbiol Ecol 74 (2010) 205–213
Ohene-Adjei S, Teather RM, Ivan M & Forster RJ (2007)
Postinoculation protozoan establishment and association
patterns of methanogenic archaea in the ovine rumen. Appl
Environ Microb 73: 4609–4618.
Robert C & Bernalier-Donadille A (2003) The cellulolytic
microflora of the human colon: evidence of microcrystalline
cellulose-degrading bacteria in methane-excreting subjects.
FEMS Microbiol Ecol 46: 81–89.
Robert C, Del’homme C & Bernalier-Donadille A (2001)
Interspecies H2-transfer in cellulose degradation between
fibrolytic bacteria and H2-utilising microorganisms from the
human colon. FEMS Microbiol Lett 205: 209–214.
Robert C, Chassard C, Lawson PA & Bernalier-Donadille A
(2007) Bacteroides cellulosilyticus sp. nov., a cellulolytic
bacterium from the human gut microbial community. Int J
Syst Evol Micr 57: 1516–1520.
Samuel BS & Gordon JI (2006) A humanized gnotobiotic mouse
model of host–archaeal–bacterial mutualism. P Natl Acad Sci
USA 103: 10011–10016.
Schweizer TF & Edwards CA (1991) Dietary fibre – a component
of food. Nutritional function in health and disease. SpringerVerlag, London.
Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER &
Gordon JI (2006) An obesity-associated gut microbiome with
increased capacity for energy harvest. Nature 444: 1027–1031.
Wedekind KJ, Mansfield HR & Mongomery YL (1988)
Enumeration and isolation of cellulolytic and hemicellulolytic
bacteria from human feces. Appl Environ Microb 54:
1530–1535.
2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c