Design and application of group-specific oligonucleotide probes for detecting and monitoring mouse clostridia

Clostridia dominate the rodent intestinal bacterial community and play an important role in physiological functions of the host. However, their ecology and diversity are still unclear. In our previous report, we showed that phylogenetically novel groups of clostridia inhabit the mouse intestine and contribute to the normalization of germfree mice. In this study, five new oligonucleotide probes were designed and applied to detect these clostridial groups that are essential for the normalization of germfree mice. Faecal microbiota of conventional mouse strains and specific pathogen-free mice from different breeding colonies were analysed by fluorescence in situ hybridization using these five probes. Our results showed that the composition of clostridia differed among mouse strains and also among mouse groups of the same inbred strain from different breeding colonies. These five new probes for mouse clostridia were able to detect the difference in clostridial diversity in each mouse group. In addition to Clostridium, we also analysed Bacteroides and Lactobacillus using previously described probes and the number or the frequency of occurrence of Bacteroides was shown to be different among mouse groups analysed. The oligonucleotide probe set including our newly developed and previously described probes used in this study can be applied to monitoring of significant groups of mouse intestinal microbiota.     

Effect of intestinal flora on megaenteron of mice.

CF#1 germfree (GF) and conventional (CV) mice as well as offspring of conventionalized GF (GF-CV) mice were orally inoculated with Escherichia coli 0115a, c: K(B), a causative agent of megaenteron in mice. Although CV and GF mice showed no clinical signs and survived, all of the GF-CV mice died with diarrhea by day 14 after inoculation. Thickened wall of the large intestine, microscopically showing proliferation of crypt type cells, was seen in GF and GF-CV mice but not in CV mice. In addition, in GF-CV mice, hemorrhage and severe erosion with marked inflammatory reactions were observed. While the inoculated E. coli could not colonize in CV mice, a level of 10⁸ cells/g feces was maintained in GF mice from day 1 after inoculation to the end of examination (on day 28) and in GF-CV mice from day 5 to the time of death. Newly prepared germfree (GF-CV-GF) mice obtained hysterectomy from GF-CV mice showed a low sensitivity as comparable to that in GF mice. On the other hand, ex-germfree mice produced by oral administration of feces of GF-CV mice showed severe infection as comparable to that seen in GF-CV mice. These results suggest that the intestinal flora may play roles both on protecting from the infection of pathogenic E. coli and on enhancing the infection.     

Bacterial populations colonizing and degrading rice straw in anoxic paddy soil

Rice straw is a major substrate for the production of methane, a greenhouse gas, in flooded rice fields. The bacterial community degrading rice straw under anoxic conditions was investigated with molecular methods. Rice straw was incubated in paddy soil anaerobically for 71 days. Denaturing gradient gel electrophoresis (DGGE) of the amplified bacterial 16S rRNA genes showed that the composition of the bacterial community changed during the first 15 days but then was stable until the end of incubation. Fifteen DGGE bands with different signal intensities were excised, cloned, and sequenced. In addition, DNA was extracted from straw incubated for 1 and 29 days and the bacterial 16S rRNA genes were amplified and cloned. From these clone libraries 16 clones with different electrophoretic mobilities on a DGGE gel were sequenced. From a total of 31 clones, 20 belonged to different phylogenetic clusters of the clostridia, i.e., clostridial clusters I (14 clones), III (1 clone), IV (1 clone), and XIVa (4 clones). One clone fell also within the clostridia but could not be affiliated to one of the clostridial clusters. Ten clones grouped closely with the genera Bacillus (3 clones), Nitrosospira (1 clone), Fluoribacter (1 clones), and Acidobacterium (2 clones) and with clone sequences previously obtained from rice field soil (3 clones). The relative abundances of various phylogenetic groups in the rice straw-colonizing community were determined by fluorescence in situ hybridization (FISH). Bacteria were detached from the incubated rice straw with an efficiency of about 80 to 90%, as determined by dot blot hybridization of 16S rRNA in extract and residue. The number of active (i.e., a sufficient number of ribosomes) Bacteria detected with a general eubacterial probe (Eub338) after 8 days of incubation was 61% of the total cell counts. This percentage decreased to 17% after 29 days of incubation. Most (55%) of the active cells on day 8 belonged to the genus Clostridium, mainly to clostridial clusters I (24%), III (6%), and XIVa (24%). An additional 5% belonged to the Cytophaga-Flavobacterium cluster of the Cytophaga-Flavobacterium-Bacteroides phylum, 4% belonged to the alpha, beta, and gamma Proteobacteria, and 1.3% belonged to the Bacillus subbranch of the gram-positive bacteria with a low G+C content. The results show that the bacterial community colonizing and decomposing rice straw developed during the first 15 days of incubation and was dominated by members of different clostridial clusters, especially clusters I, III, and XIVa.     

Increased intracellular content of enteroglucagon in proximal colon is related to intestinal adaptation to germ-free status

Changes in the frequency of endocrine cells are evidence of intestinal adaptation to germ-free (GF) status. Not only the distribution of these cells along the intestine, but also the differences in intracellular content of these regulatory peptides may be explored to explain functional and structural aspects of GF intestinal adaptation. Focusing on the endocrine L-cells, we analyzed the intracellular content of enteroglucagon (EG) and peptide YY (PYY) throughout the intestine of the 14 GF and 14 conventional (CV) mice by using immunohistochemistry and the supra-optimal dilution technique. The percentage of EG-immunoreactive cells, but not of PYY-immunoreactive cells stained at supra-optimal dilution was significantly higher in the proximal colon of GF mice than in the CV counterparts (P < 0.05). Since the content of co-stored PYY did not differ between GF and CV mice, the higher content of EG was compatible with a selective cellular response. Moreover, in the cecum of GF mice, the density of EG-immunoreactive cells was significantly higher than that of PYY-immunoreactive cells (P < 0.05). These results are consistent with preferential production of EG by L-cells at the expense of PYY in the proximal colon and in the enlarged cecum of GF mice. In addition, they may reflect the dynamics of the GF intestinal epithelium and/or be correlated with the higher serum levels of these peptides. The role of endocrine cells needs to be better studied in human and other experimental adaptative conditions in order to elucidate the regulatory mechanisms of intestinal functions.     

三种微生物学级别的SSB小鼠生长和繁殖性能的比较

本文对无菌，清洁和普通三种微生物学级别的近交系ＳＳＢ小鼠核心群的生长，繁殖性能的观察结果进行比较分析。在哺乳期风无菌小鼠生长优于清洁级和普通级小鼠；而在离乳后育成初期体重增长又慢于者。计算机统计分析表明：三种级别种鼠的各胎次胎间隔无显著差异。清洁和无菌种鼠的产仔数和离乳率均无显著差异；而普通种鼠的繁殖性能不如清洁和无菌种鼠。     

Coprophagy in the germfree mouse

Coprophagy was observed in germfree (GF) ICR mice of both sexes, and the results were compared with those of conventional mice. Frequency of coprophagy per animal per day in GF mice was 5.1 in males and 5.8 in females. In conventional (CV) mice, the frequencies were 6.2 in males and 5.3 in females (data from Zoological Science 2:249-255, 1985), with no significant differences compared with GF mice. Coprophagy in CV mice was frequently observed during 6-8 hr after lighting, whereas such close time relationships tended to weaken in GF animals. In a comparison of levels of constituents per unit weight between feces and diet, fecal crude protein and crude fat exhibited lower values than those in the diet. Levels of fecal crude ash and crude fiber were higher than those in the diet, and nitrogen-free extract was almost equal to that in the diet. No essential difference in these tendencies was found compared with CV mice. Levels of fecal vitamin B1, B2, B12 and folic acid were lower than those in the diet. In CV mice, except for vitamin B1, these vitamins exhibited either almost equal or much higher levels compared with those in the diet (data from Experimental Animals 35: 381-386, 1986). From the fact that coprophagy was observed in GF mice, it is suggested that the behavior is inherent in the mouse.     

Bifidobacterium bifidum monoassociation of gnotobiotic mice: Effect on enterocyte brush-border enzymes

The effect of intestinal colonization withBifidobacterium bifidum (Gram-positive anaerobic bacterium colonizing the intestine of healthy new-born mammals, exhibiting a probiotic effect, protecting the intestinal mucosa against colonization by pathogenic microflora) on enterocyte brush-border enzymes was examined in weaned 23-d- and in 2-month-old gnotobiotic inbred mice and compared with that in corresponding germ-free (GF) and conventional (CV) controls. The two groups of GF mice were associated with humanB. bifidum 11 d before the end of the experiment. Specific activity of enterocyte brush-border enzymes—lactase, alkaline phosphatase and γ-glutamyltranspeptidase was significantly higher in both age groups of GF mice in comparison with CV ones; on the other hand, sucrase and glucoamylase activities were higher in CV mice. Monoassociation withB. bifidum accelerates biochemical maturation of enterocytes resulting in a shift of specific activities of brush-border enzymes between the values found for GF and CV mice. This effect ofB. bifidum supplementation was less pronounced for alkaline phosphatase, sucrase, glucoamylase and dipeptidyl peptidase IV in immature gut of weaned mice than of 2-month-old ones.     

Bacterial Populations Colonizing and Degrading Rice Straw in Anoxic Paddy Soil

Rice straw is a major substrate for the production of methane, a greenhouse gas, in flooded rice fields. The bacterial community degrading rice straw under anoxic conditions was investigated with molecular methods. Rice straw was incubated in paddy soil anaerobically for 71 days. Denaturing gradient gel electrophoresis (DGGE) of the amplified bacterial 16S rRNA genes showed that the composition of the bacterial community changed during the first 15 days but then was stable until the end of incubation. Fifteen DGGE bands with different signal intensities were excised, cloned, and sequenced. In addition, DNA was extracted from straw incubated for 1 and 29 days and the bacterial 16S rRNA genes were amplified and cloned. From these clone libraries 16 clones with different electrophoretic mobilities on a DGGE gel were sequenced. From a total of 31 clones, 20 belonged to different phylogenetic clusters of the clostridia, i.e., clostridial clusters I (14 clones), III (1 clone), IV (1 clone), and XIVa (4 clones). One clone fell also within the clostridia but could not be affiliated to one of the clostridial clusters. Ten clones grouped closely with the genera Bacillus (3 clones), Nitrosospira (1 clone), Fluoribacter (1 clones), and Acidobacterium (2 clones) and with clone sequences previously obtained from rice field soil (3 clones). The relative abundances of various phylogenetic groups in the rice straw-colonizing community were determined by fluorescence in situ hybridization (FISH). Bacteria were detached from the incubated rice straw with an efficiency of about 80 to 90%, as determined by dot blot hybridization of 16S rRNA in extract and residue. The number of active (i.e., a sufficient number of ribosomes) Bacteria detected with a general eubacterial probe (Eub338) after 8 days of incubation was 61% of the total cell counts. This percentage decreased to 17% after 29 days of incubation. Most (55%) of the active cells on day 8 belonged to the genus Clostridium, mainly to clostridial clusters I (24%), III (6%), and XIVa (24%). An additional 5% belonged to the Cytophaga-Flavobacterium cluster of the Cytophaga-Flavobacterium-Bacteroides phylum, 4% belonged to the α, β, and γ Proteobacteria, and 1.3% belonged to the Bacillus subbranch of the gram-positive bacteria with a low G+C content. The results show that the bacterial community colonizing and decomposing rice straw developed during the first 15 days of incubation and was dominated by members of different clostridial clusters, especially clusters I, III, and XIVa.     

Differential effect of Bacillus firmus on immune response and enterocyte brush-border enzyme levels in BALB/c and B10.BR mice

A nonpathogenic bacterium of external environment possessing remarkable immunomodulatory activity, Bacillus firmus (BF) inactivated with formaldehyde, was given intragastrically to two genetically different mouse strains BALB/c (H-2d) and B10.BR/SnPh (B10.BR, H-2k) reared in conventional (CV) and B10.BR strain also in germ-free (GF) conditions. Repeated intragastric administration of BF (500 micrograms every other day over two weeks, starting at the age of 3 months) significantly enhanced intestinal IgA levels in CV BALB/c mice but did not affect intestinal IgA in CV B10.BR mice. In GF B10.BR mice, IgG levels in sera and intestinal washings increased after BF administration compared to CV B10.BR mice. In CV BALB/c mice, specific activity of enterocyte brush-border enzymes (lactase, gamma-glutamyltransferase, alkaline phosphatase) decreased after BF treatment; sucrase (sucrose alpha-glucosidase) activity was not affected. On the other hand, in B10.BR mice, specific activity of gamma-glutamyltransferase and dipeptidyl peptidase IV were higher after administration of BF in both CV and GF groups relative to untreated controls. The activities of lactase and glucoamylase (glucan 1,4-alpha-glucosidase) were significantly stimulated only in the group of GF B10.BR mice treated with formolized BF. The stimulation of immunoglobulin production after BF treatment was accompanied by changes in the levels of enterocyte brush-border enzymes; this responsiveness to BF treatment was genetically regulated.     

Restricted distribution of the butyrate kinase pathway among butyrate-producing bacteria from the human colon

The final steps in butyrate synthesis by anaerobic bacteria can occur via butyrate kinase and phosphotransbutyrylase or via butyryl-coenzyme A (CoA):acetate CoA-transferase. Degenerate PCR and enzymatic assays were used to assess the presence of butyrate kinase among 38 anaerobic butyrate-producing bacterial isolates from human feces that represent three different clostridial clusters (IV, XIVa, and XVI). Only four strains were found to possess detectable butyrate kinase activity. These were also the only strains to give PCR products (verifiable by sequencing) with degenerate primer pairs designed within the butyrate kinase gene or between the linked butyrate kinase/phosphotransbutyrylase genes. Further analysis of the butyrate kinase/phosphotransbutyrylase genes of one isolate, L2-50, revealed similar organization to that described previously from different groups of clostridia, along with differences in flanking sequences and phylogenetic relationships. Butyryl-CoA:acetate CoA-transferase activity was detected in all 38 strains examined, suggesting that it, rather than butyrate kinase, provides the dominant route for butyrate formation in the human colonic ecosystem that contains a constantly high concentration of acetate.     

Comparative studies of some functional and morphological parameters in the livers of germfree, conventional and ex-germfree mice.

Some parameters of hepatic function and morphology were studied to compare germfree (GF) and conventional (CV) BALB/c mice. The levels of lipid peroxide (LPO) and aniline-hydroxylase (AH) activity in the livers and the serum total cholesterol (TC), triglyceride (TG) and phospholipid (PL) were significantly lower in GF than in CV 8-week-old mice. There were no significant differences in the histology and lectin-histochemistry of the livers in the GF and CV mice. On the other hand, in ex-GF mice which were induced by housing 4-week-old GF mice together with age-matched CV mice, the levels of LPO and AH activity in the liver and the serum TC, TG and PL contents increased rapidly within the first week and then approached values almost identical to those in CV mice 4 weeks later (i.e. at 8 weeks of age). The histologic picture of the liver was similar among the GF, CV and ex-GF mice.     

COMPARISON OF DNA RENEWAL IN GERM-FREE AND CONVENTIONAL MICE USING [125I]IODODEOXYURIDINE AND [3H]THYMIDINE

Germ-free (GF) and conventional (CV) C3H mice received a single injection of 1 μCi [³H]thymidine and 3 μCi [¹²⁵I]iododeoxyuridine to provide simultaneous labeling of DNA with the two precursors. Thymus, spleen, mesenteric lymph nodes, bone marrow (femora), small intestine, colon and skin were examined for total organ activity and rate of DNA renewal 1–8 days after injection. Precursor incorporation, assayed on day 1, was lower in the thymus, mesenteric lymph nodes and femora (and, to a lesser extent, in the spleen and colon) of GF mice as compared to CV animals. The opposite was observed in the small intestine and skin, i.e. total organ activity was higher in GF animals. Differences in precursor incorporation were partly due to differences in organ weights between the two groups of mice. In comparison to CV animals, DNA renewal rates were diminished in the mesenteric lymph nodes, bone marrow, colon (following a 3-day plateau) and spleen of GF mice. Little, if any, difference was observed between the two groups with respect to the rate of DNA turnover in the thymus and skin. Radioactivity of the small intestine remained constant for 2 days. Thereafter intestinal activity in GF mice declined at an initial slow rate between days 2 and 5 followed by a rapid decrease between days 5 and 8. In CV mice the first phase of activity loss was short with the rapid decline in intestinal activity beginning on day 3. From the slopes of the regression lines, the percentage thymidine reutilization was estimated. Reutilization varied from 0 to 63% in the various organs examined, with the greatest difference between GF and CV mice occurring in the mesenteric lymph nodes.     

Characteristic faecal flora of NC mice

The composition of faecal flora of NC mice was compared with that of CF #1 mice. NC- and CF #1-germfree (GF) mice were cage-mated with NC- or CF #1-conventional (CV) mice in an isolator. The faecal flora of these ex-GF mice was dependent on the recipient mouse strain modifying colonization by the donor mouse bacteria. Although NC- and CF #1-pups removed by hysterectomy were fostered to different strains, almost all these mice at 8 weeks old had a strain characteristic pattern of faecal flora regardless of the foster strains. In GF mice mono-associated with a Lactobacillus strain or a Bifidobacterium strain isolated from faeces of CV mice, the numbers of these bacteria in the stomach and small intestine of NC mice were lower than those of CF #1 mice. In GF mice associated with chloroform-treated faeces of CV mice, and a Lactobacillus strain or a Bifidobacterium strain, the numbers of these bacteria in the stomach and all parts of the intestine of NC mice were considerably lower than those of CF #1 mice. These results suggested that the composition of faecal flora of NC mice were characteristic, i.e. the fact that the numbers of lactobacilli were low compared with CF #1 mice with ordinary faecal flora and the colonization of bifidobacteria, peptococcaceae and eubacteria on ES agar in NC mice intestine differed, was due to genetic factors.     

Production of germfree mice by embryo transfer.

We applied the embryo transfer technique to germfree (GF) mouse production. Embryos harvested from superovulated mice were transferred aseptically, in a sterile environment, to the uterus of GF recipient females which had been mated with vasectomized GF males. One of the recipients became pregnant and delivered offspring. Sterility tests confirmed that the vasectomized males, newborns, recipient female mice, embryo-containing culture media, and the inside of the vinyl film isolator were germfree. These results suggest that the embryo transfer technique can be successfully applied to the production of GF mice.     

Novel endophytic nitrogen-fixing clostridia from the grass Miscanthus sinensis as revealed by terminal restriction fragment length polymorphism analysis

Anaerobic nitrogen-fixing consortia consisting of N2-fixing clostridia and diverse nondiazotrophic bacteria were previously isolated from various gramineous plants (K. Minamisawa, K. Nishioka, T. Miyaki, B. Ye, T. Miyamoto, M. You, A. Saito, M. Saito, W. Barraquio, N. Teaumroong, T. Sein, and T. Tadashi, Appl. Environ. Microbiol. 70:3096-3102, 2004). For this work, clostridial populations and their phylogenetic structures in a stand of the grass Miscanthus sinensis in Japan were assessed by a 16S rRNA gene-targeted terminal restriction fragment length polymorphism (TRFLP) analysis combined with most-probable-number (MPN) counts. PCR primers and restriction enzymes were optimized for analyses of the plant clostridia. Clostridia were detected in strongly surface-sterilized leaves, stems, and roots of the plants at approximately 10(4) to 10(5) cells/g of fresh weight; they made up a large proportion of N2-fixing bacterial populations, as determined by MPN counts associated with an acetylene reduction assay. Phylogenetic grouping by MPN-TRFLP analysis revealed that the clostridial populations belonged to group II of cluster XIVa and groups IV and V of cluster I; this result was supported by a culture-independent TRFLP analysis using direct DNA extraction from plants. When phylogenetic populations from M. sinensis and the soil around the plants were compared, group II clostridia were found to exist exclusively in M. sinensis.     

Colonization resistance against Pseudomonas aeruginosa in gnotobiotic mice

Gnotobiotic (GB) mice were colonized with various groups of intestinal bacteria to determine which members of the indigenous flora would exert colonization resistance against Pseudomonas aeruginosa. P. aeruginosa was cultured from the faeces at levels of 10(3)-10(4) cells/g in GB mice inoculated with either the combination of bacteroides and clostridia obtained from conventional (CV) mice or the combination of bacteroides, lactobacilli and clostridia obtained from limited flora mice. The combination of lactobacilli and clostridia from CV mice also did not eliminate P. aeruginosa from GB mice. However, P. aeruginosa was not detected in the faeces of GB mice by 14 days after inoculation with the combination of bacteroides, lactobacilli and clostridia obtained from CV mice. Thus, a complex indigenous flora consisting of bacteroides, lactobacilli and certain clostridia obtained from CV mice but not clostridia obtained from limited flora mice is required to exert complete colonization resistance against P. aeruginosa in GB mice.     

Ammonia in the intestinal contents from germfree rats

The values of pH and the concentrations of nitrogen (N) and ammonia in the middle part of the small intestinal and cecal contents of germfree (GF) and conventionalized (CVZ) seven-week-old rats were compared. The pH of the small intestinal and cecal contents of GF rats was higher than that of CVZ rats. There was no difference in total N per fresh weight in contents from the middle part of the small intestine between GF and CVZ, whereas total N per fresh weight of the cecal contents was higher in CVZ than in GF rats. The ammonia concentrations per fresh weight or per total N in the intestinal and cecal contents of CVZ rats were higher than those of GF rats.     

Distribution of estrone sulfatase activity in the intestine of germfree and conventional rats

The intestinal content, the mucosa and the rest of the intestinal wall of germfree (GF) and conventional ( CVL ) rats were tested for in vitro hydrolysis of [3H]estrone sulfate. In homogenates from GF rat intestine some estrone sulfate hydrolysis was detected in those from the proximal small intestine (PSI) (4.2 +/- 0.1% hydrolyzed after 4 h), but not in those from the distal small intestine (DSI) and the caecum. Estrone sulfate was also hydrolyzed by the homogenates of the mucosa and the rest of the intestinal wall from each of the segments tested (PSI: 12.8 +/- 0.4% (mucosa) and 21.5 +/- 2.1 (wall); DSI: 8.2 +/- 0.9% (mucosa) and 17.3 +/- 1.7% (wall); caecum: 8.8 +/- 1.6% (mucosa) and 17.3 +/- 0.5% (wall) ). In the homogenates of CVL rat intestine, the estrone sulfatase activity in the rest of the intestinal wall did not differ considerably from the values for GF rats, when expressed per mg protein of the homogenate. The mucosa of the CVL rats, however, showed higher rates of hydrolysis than the mucosa of the GF rats. The microbial estrone sulfatase activity in the intestinal content of CVL rats, tested by anaerobic incubation, was high in the caecum (91.7 +/- 6.6% after 4 h), but very low in the PSI (2.2 +/- 0.7%) and DSI (1.3 +/- 0.5%). Serial dilutions of the caecal content also showed higher viable numbers of estrone sulfate hydrolyzing bacteria. These results add further weight to the suggestion that estrone sulfate may be absorbed from the small intestine, but has to be hydrolyzed in the caecum by the gut microflora prior to absorption.     

Plasmodium berghei and Eperythrozoon coccoides: Antibody and immunoglobulin synthesis in germfree and and conventional mice simultaneously infected

The immune response to Eperythrozoon coccoides and the malaria parasite Plasmodium berghei was evaluated in germfree (GF) and conventionally reared (CV) mice infected with both parasites. Following infection, the mice showed significant changes in the levels of the immunoglobulins IgM, 7Sγin1, 7Sγ2a, and 7Sγ2b, but no detectable changes in IgA. Increases in immunoglobulin levels were first observed in GF mice, but by the twelth day both GF and CV mice had comparable levels. 7Sγ2a globulin had a bimodal distribution in both groups of mice which probably was due to heterogeneity in the allotype of this immunoglobulin. IgM levels closely paralleled the antibody responses to P. berghei suggesting that most of the antibody to this parasite was IgM. Relatively low levels of antibody to both parasites, in comparison to the large immunoglobulin response, were detected in GF and CV mice. The possible causes for the low titers are discussed.     

Novel Endophytic Nitrogen-Fixing Clostridia from the Grass Miscanthus sinensis as Revealed by Terminal Restriction Fragment Length Polymorphism Analysis

Anaerobic nitrogen-fixing consortia consisting of N₂-fixing clostridia and diverse nondiazotrophic bacteria were previously isolated from various gramineous plants (K. Minamisawa, K. Nishioka, T. Miyaki, B. Ye, T. Miyamoto, M. You, A. Saito, M. Saito, W. Barraquio, N. Teaumroong, T. Sein, and T. Tadashi, Appl. Environ. Microbiol. 70:3096-3102, 2004). For this work, clostridial populations and their phylogenetic structures in a stand of the grass Miscanthus sinensis in Japan were assessed by a 16S rRNA gene-targeted terminal restriction fragment length polymorphism (TRFLP) analysis combined with most-probable-number (MPN) counts. PCR primers and restriction enzymes were optimized for analyses of the plant clostridia. Clostridia were detected in strongly surface-sterilized leaves, stems, and roots of the plants at approximately 10⁴ to 10⁵ cells/g of fresh weight; they made up a large proportion of N₂-fixing bacterial populations, as determined by MPN counts associated with an acetylene reduction assay. Phylogenetic grouping by MPN-TRFLP analysis revealed that the clostridial populations belonged to group II of cluster XIVa and groups IV and V of cluster I; this result was supported by a culture-independent TRFLP analysis using direct DNA extraction from plants. When phylogenetic populations from M. sinensis and the soil around the plants were compared, group II clostridia were found to exist exclusively in M. sinensis.