APPLIED AND ENVIRONMENTAL MICROBIOLOGY, June 1976, p. 907-912 Copyright © 1976 American Society for Microbiology
Vol. 31, No. 6 Printed in U.S.A.
Anaerobic Bacteria from the Large Intestine of Mice1 MARTHA A. HARRIS, C. ADINARAYANA REDDY,* AND GORDON R. CARTER Department of Microbiology and Public Health, Michigan State University, East Lansing, Michigan 48824 Received for publication 3 February 1976
Anaerobic bacteria from the colon of laboratory mice were enumerated and isolated using strict anaerobic techniques. Direct microscopic counts revealed 4.4 x 1010 organisms in each gram (wet weight) of colon contents. Actual cultural counts averaged 3.2 x 1010 organisms, which was 73% of the direct microscopic count. The tentatively identified genera were Bacteroides, Eubacterium, Fusobacterium, Lactobacillus, Peptostreptococcus, and Propionibacterium. Strains of Fusobacterium, Lactobacillus, Peptostreptococcus, and Propionibacterium were biochemically homogeneous. Strains of Bacteroides and Eubacterium, on the other hand, were biochemically heterogeneous and were subdiv;'led into several distinct groups. The data indicate that many of the isolates are different from previously described species of the respective genera and may belong to new species.
There have been several studies conducted on the isolation of anaerobic bacteria from the gastrointestinal tract of mice. Spears and Freter (13) isolated the anaerobic flora in the ceca of mice, by the agar plate-anaerobic jar procedure and the Hungate roll tube method, but did not identify the isolates. Aranki et al. (1) reported anaerobic, gram-positive bacilli, gram-negative cocco-bacilli, and rods with tapered ends in the ceca of mice but did not identify the isolates. On the basis of morphological studies, several investigators suggested that bacteria belonging to the genera Bacteroides, Clostridium, Eubacterium, Fusobacterium, and Lactobacillus were present in the gastrointestinal tract of mice (5, 8, 9, 12), but the data were inconclusive. The current investigation is therefore concerned with the enumeration and identification of the predominant anaerobic genera and with the description of subgroups within a given genus. The significance of this investigation is related to the fact that anaerobic bacteria are the predominant microbes in the gastrointestinal tract of mammals and they are frequently involved in clinical infections in humans. The incidence of nonsporing anaerobes in clinical infections of other animals has not been well documented. A knowledge of the predominant genera and species in the large intestine of healthy mice is essential to better understand the microbial ecology of this habitat and the specific organisms that may be involved in various pathological processes.
MATERIALS AND METHODS Animals and diets. Male, Swiss Webster mice were housed in individual steel cages (with wood shaving for bedding) in an animal room. The mice were fed Wayne Mouse Breeder Blox, a complete diet, and were provided with drinking water from individual glass bottles. Culture methods. The anaerobic roll tube techniques used in this study were those of Hungate (7) as modified by Moore (10) except where indicated otherwise. An oxygen-free gas mixture containing 85% nitrogen, 12% carbon dioxide, and 3% hydrogen was used to exclude air in preparing media, and during various cultural manipulations. The anaerobic culture system described by Holdeman and Moore (6) was used in this investigation. Culture media. Prereduced anaerobically sterilized media, autoclaved, dispensed, and inoculated in the complete absence of oxygen were used throughout this study. The medium used for agar roll tubes and maintenance slants was similar in composition to medium 10 of Caldwell and Bryant (3, 6), except that 0.0002% menadione was added and the volatile fatty acid mixture was excluded. Commercial prereduced anaerobically sterilized media purchased from the Robbin Laboratories division of Scott Laboratories were used for all biochemical tests. Isolation procedures. Adult male mice were sacrificed by the use of anhydrous ether and pinned on a dissecting board. The colon was removed aseptically and its contents were squeezed out and immediately weighed. Complete anaerobiosis was maintained after this step. The weighed colon contents were blended anaerobically under the oxygen-free gas mixture in a Waring blender containing a known volume of the anaerobic salts solution (6) for approximately 1 min. One milliliter of the blended intestinal contents was transferred aseptically and an' Article no. 7560, Michigan Agricultural Experiment aerobically to a sterile rubber stoppered tube (18 by Station. 907
908
APPL. ENVIRON. MICROBIOL.
HARRIS, REDDY, AND CARTER
150 mm) containing 9 ml of sterile anaerobic salts solution. Serial 10-fold dilutions were made to 1011. Gram-stained smears were made from the 10' dilution of the colon contents of each mouse and all morphotypes were recorded. Wet mounts were made from the 101 dilution and examined for motility with a phase-contrast microscope. Direct microscopic counts were made from the 103 dilution as described by Holdeman and Moore (6). Four roll tubes were made of each dilution from 107 to 1011 and the tubes were incubated at 37 C for 7 days. Colony counts were made in roll tubes containing 30 to 300 colonies. The procedures for analysis of the cultural counts were the same as those given by Holdeman and Moore (6). Isolation and characterization of bacteria. A platinum or stainless-steel transfer needle was used to pick representative colonies from the 109, 1010, and 1011 dilution roll tubes. Isolated colonies were stabbed into slants (3) of maintenance medium. After a 24-h incubation period, the organisms in the water of syneresis were examined for morphology, culture purity, and Gram reaction. A few of these cultures found to be mixed on microscopic examination were purified by streaking brain heart infusion roll tubes. Discrete colonies were picked and stabbed into slants of maintenance medium. Each isolate was examined for aerotolerance by streaking one-fourth of a blood agar plate and incubating it at 37 C aerobically for 24 h. For analysis of metabolic end products produced from carbohydrates, bacterial isolates were grown in peptone-yeast extract-glucose broth (PYG) for 48 h. The volatile acid end products were extracted with ether and the nonvolatile acids were methylated and extracted into chloroform as described by Holdeman and Moore (6). Both fractions were analyzed by a Dohrmann gas chromatograph using a Resoflex column. Helium was used as the carrier gas for the system with a flow rate of 120 cc/min. The column and thermal conductivity detector were maintained at approximately 118 to 120 C, whereas the injection port was maintained at 145 C. Procedures described by Holdeman and Moore (6) were used to perform biochemical tests. Four drops of an overnight culture in supplemented brain heart infusion broth, inoculated from the pure culture slants, was used as inoculum for differential biochemical media. The final pH in carbohydrate media was determined by a Beckman zeromatic pH meter. The final pH in carbohydrate broth above 6.0, between 6.0 to 5.5, or below 5.5 was read as no acid, weak acid, and strong acid production, respectively. All isolates were tested for their ability to grow in the presence of bile, hydrolyze esculin, liquify gelatin, produce indole, reduce nitrates, and produce acid, curd, and/or digestion of milk. All gram-positive rods were subjected to a heat test by inoculating a tube of starch broth with the test organism, heating at 80 C for 10 min, and incubating at 37 C for 24 h. Lack of growth in the starch broth after the heat test is taken to imply that the bacterial isolate is not a clostridial species. Young cultures, 12 h old, grown in peptone-yeast broth were Gram stained to determine the Gram reaction. A phase-contrast micro-
scope was used to determine motility from wet mounts prepared from peptone-yeast broth cultures.
RESULTS AND DISCUSSION Direct microscopic examination of the intestinal contents. The microbial morphotypes observed on direct microscopic examination of Gram-stained smears and wet mounts of the mouse colon contents (diluted 101) were similar to the microbial morphotypes observed by earlier investigators (5, 9). Gram-negative rods were the predominant morphotypes observed, although small numbers of gram-positive rods and cocci were also present. Gram-negative spiral-shaped organisms were observed in microscopic smears but none of these motile organisms was isolated. It is possible that: (i) the isolation medium did not provide the fastidious nutritional requirements of the spiral-shaped organisms; (ii) very small numbers of these organisms were present in the higher dilutions; or more importantly, (iii) special procedures (6) designed for the selective isolation of spirochetes would be necessary. Bacterial counts. In previous studies, 1010 to 1011 organisms/g of mouse intestinal contents were reported (11, 12). In the present investigation, a direct microscopic clump count revealed 4.4 x 10'° organisms/g (wet weight) of colon contents which was in close agreement with previous results. The mean cultural count per gram (wet weight) of colon contents was 3.2 x 1010 organisms (Table 1). This constituted approximately a 73% recovery of the organisms seen in the direct clump count, indicating the relative efficiency of the cultural methods used in this investigation. There were no noticeable differences in the number or types of anaerobes recovered from roll tubes inoculated with the colon contents from the three mice used in this particular study. Direct microscopic counts were generally higher than actual culture counts probably because: (i) some of the flora of the mouse such as spiral-shaped organisms are seen in smears but are rarely cultured (5, 12); (ii) some of the bacteria seen in clump counts TABLE 1. Enumeration of the cultural counts of anaerobic bacteria in the intestinal contents of mice" Animal no.
Wt of con- Avg counts Bacteria/g (wet tents (g) tion weight)
Mouse 1 0.22 32 1.3 x 10"' Mouse 11 0.31 53 1.6 x 10"' Mouse 111 0.52 78 6.7 x 10"' " Procedures in the Anaerobe Laboratory Manual (6) were used for enumerating the isolates.
VOL. 31, 1976
ANAEROBES FROM THE LARGE INTESTINE OF MICE
may be nonviable bacteria; and (iii) the medium may not be providing all the factors required for the growth, especially fatty acids known to be required for growth of several anaerobes from a number of anaerobic microbial habitats (2, 3, 6). Isolation and identification of intestinal anaerobes. Predominant bacterial groups in the mouse intestine were classified into genera based on Gram reaction, morphology, motility, oxygen tolerance, and acid metabolic end products. From the results presented in Table 2, gram-negative rods were assigned to the genera Bacteroides and Fusobacterium, gram-positive rods were placed in the genera Eubacterium, Lactobacillus, Propionibacterium, and gram-positive cocci in the genus Peptostreptococcus. The results were in close agreement with those of previous investigators with regard to the predominant anaerobic genera present in the intestinal tract of mice (5, 8, 9, 12) except that Clostridium species were not isolated in the present investigation. Savage et al. (12) observed the presence of Clostridium species in frozen sections of the large bowel of mice. Gordon and Dubos (5) and Dubos et al. (4) reported Clostridium species in the 107 and 10W dilutions of mouse cecal contents and fecal flora, respectively. In the present study, all the organisms isolated were from the 10' to 1011 dilution roll tubes and this may account for the fact that Clostridium species were not isolated by us. Bacteroides. Bacteroides isolates were from 1010 to 1011 dilutions and were obligately anaerobic, gram negative, non-sporulating rods with rounded or tapered ends. They produced varying amounts of acetate, lactate, propionate, and succinate and were nonmotile. None of the isolates produced black pigment on anaerobically incubated blood agar plates. The addition of hemin to the PYG broth, greatly enhanced the
909
growth of most isolates in comparison to the same medium without hemin. The results of this investigation are in agreement with the findings of Gordon and Dubos (5) and Lee et al. (8) who also reported the presence of Bacteroides species in the intestine of mice. Based on biochemical characteristics presented in Table 3, the Bacteroides isolates were divided into six distinct groups. As discussed below, none of the six groups are exactly similar to the presently recognized species of Bacteroides, suggesting that these isolates may belong to Bacteroides species that have not yet been described. Bacteroides group no. 1 isolates were nonmotile and fermented a number of carbohydrates including fructose, lactose, and maltose but not melezitose. End products produced by these strains included major amounts of acetic and small amounts of succinic acids. No other acid end products were detectable. Based on these characteristics they appeared similar toB. clostridiiformis subsp. clostridiiformis (3, 6), but differ from previously described strains of this species in not fermenting sucrose and in not hydrolyzing esculin (2, 6). Bacteroides group no. 2 isolates were nonmotile, nonfermentative rods that produced small amounts of acetic and succinic acids and traces of lactate as end products. These isolates were similar to B. coagulans but differed from typical strains of this species in hydrolyzing esculin, not growing in the presence ofbile and even more important in not being able to curdle milk. B. coagulans by definition (2) curdles milk. Thus this group of isolates appear to be a new Bacteroides species. Bacteroides group no. 3 isolates were nonmotile, did not produce black pigment on blood agar, produced succinic and acetic acids as major end products, fermented fructose and rhamnose, produced indole, gave growth in 20% bile, and were stimulated by the addition of heme to
TABLE 2. Some features used for identifying major groups of bacteria from the
mouse
large intestine
Fermenta-
Morphological group
b
Oxygen products from eranceatol- tion glucose'
Gram-negative rods ALSP A Rounded to tapered rods ABL A Tapered rods Gram-positive rods ALS A Short rods LA AT Short rods AP A Short to medium-sized rods AT ALPS Gram-positive cocci A, Anaerobic; AT, aerotolerant anaerobe. A, Acetic; B, butyric; L, lactic; P, propionic; S, succinic.
Population levels
109-10"
109-10X}
1O9-1OlI 109 109
1O
Tentative identification
Bacteroides Fusobacterium Eubacterium Lactobacillus
Propionibacterium Peptostreptococcus
910
APPL. ENVIRON. MICROBIOL.
HARRIS, REDDY, AND CARTER
TABLE 3. Biochemical reactions of predominant anaerobic bacteria isolated form the mouse intestine Group
a o ~
8O ~
|E0 B
B
B
A
0
a
to
IV
Bacteroides a A A A a a a A A A 2 Abl Group no. 1 c -++ 2 - _+.W - -_ + + Group no. 2 A A -a a a- -- A w c -+ + + 3 - a Group no. 3 2 - a a A -a A a Group no. 4 + 3 - -- + _ Group no. 5 __ a A A a a a A A cd + 2 - a Group no. 6 Eubacterium 1 - - + A A A A Group no. 1 c +-. 2 A A + A A AA - A a A AA - A A - c ++ - Group no. 2 1 - - A A A A .- A - + c - + + + Group no. 3 2 - a + - A -- - a - a a a A A w c -- + Fusobacterium + - + + 5 - - - A A -- - a -- A a Lactobacillus + c + -_ 6 - A + A A A A - A - A a A - A Peptostreptococcus a + cd - + + + 1 - Propionibacterium a _ AIAIAIA.I a The number of strains studied was not proportional to the actual numbers grown from samples as colonies were not picked in a random fashion. b Symbols: A, Acid; a, weak acid; +, positive; -, negative; w, weak reaction; c, curd; d, digestion. -
the medium. Based on these characteristics they key to B. fragilis group (2, 6). However, these strains are significantly different from typical B. fragilis strains in not fermenting lactose, maltose, and sucrose. Therefore, further work needs to be done to determine the exact taxonomic status of this group of isolates. Bacteroides group no. 4 isolates produced major amounts of lactic acid and trace amounts of acetic and succinic acids. All strains produced acid from glucose and weakly fermented fructose, rhamnose, and mannose. These strains appeared somewhat similar to B. furcosus but differed from previously described strains of this species in failing to grow in the presence of bile, in fermenting raffinose, and in not hydrolyzing esculin (2, 6). These data suggest that this group of isolates may belong to a new Bacteroides species. Bacteroides group no. 5 isolates are nonmotile, did not produce indole, did not hydrolyze gelatin or esculin, did not reduce nitrate, and did not ferment carbohydrates. Only trace amounts of acetic acid is produced in PYG medium. Based on the above-mentioned characteristics, this group of isolates may be considered similar to B. pneumosintes (2, 6). However, B. pneumosintes, by definition, is a small coccobacillus, while the group no. 5 isolates were medium-sized rods. Therefore, Bacteroides group no. 5 isolates probably belong to a new Bacteroides species which is somewhat related to but different from B. pneumosintes (2, 6). Strains designated Bacteroides group no. 6
isolates were gram-negative, anaerobic rods and produced major amounts of acetate and lactate, fermented a moderate number of carbohydrates, and grew in the presence of bile. On the basis of characteristics studied, these isolates appeared to be different from any of the Bacteroides species described to date (2). Fusobacterium. Savage et al. (12) reported the presence of Fusobacterium species in the mouse intestine. Gordon and Dubos (5) isolated organisms similar to F. russii in the 1010 to 1011 dilutions of the cecal contents of laboratory mice. In agreement with these earlier results, Fusobacterium species were isolated from the mouse intestinal contents in this investigation. These organisms were anaerobic, nonmotile, gram-negative, nonsporulating rods with tapered ends and produced butyric acid as the major acid end product in PYG broth. The biochemical characteristics of the Fusobacterium species presented in Table 3 show that these isolates are different from any of the previously described Fusobacterium species (2, 6). Eubacterium. In agreement with the results of the present investigation, Gordon and Dubos (5) reported that Eubacterium species were present at levels of 1010 to 1011 organisms/g of cecal contents. Savage et al. (12) also demonstrated, by electron microscopy, organisms similar to Eubacterium in frozen sections of the large bowel of mice. In agreement with these earlier findings, Eubacterium strains were isolated from 1010 to 1011 dilution of the mouse colon contents in the present study. All the isolates were subjected to heat tests and found
VOL. 31, 1976
ANAEROBES FROM THE LARGE INTESTINE OF MICE
-to be negative. The negative heat tests indicated that these gram-positive bacilli were not Clostridium species. All the isolates were gram-positive nonsporeforming, bacilli that occur in chains, and clumps (6). The biochemical characteristics of these Eubacterium isolates are presented in Table 3. Eubacterium group no. 1 isolate produced acetic, lactic, and succinic acids, but no butyric acid. Among the carbohydrates tested, only fructose, glucose, lactose, and maltose were fermented. This strain reduced nitrate, hydrolyzed esculin, and coagulated milk. Based on these criteria, this isolate does not fit with the description for any of the presently described species of Eubacterium and appears to be a new species. Intestinal isolates classified as Eubacterium group no. 2 fermented a variety of carbohydrates and produced acetic, formic, and small amounts of lactic and succinic acids. These strains appeared similar to E. contortum in not producing butyric acid or indole and in producing acid from fructose, maltose and rhamnose (2, 6). However, they differed from conventional strains of E. contortum in not fermenting arabinose and in reducing nitrates (6). Strain designated Eubacterium group no. 3 produced acid from fructose, glucose, lactose, and maltose, produced indole, hydrolyzed gelatin, and grew in 20% bile. This strain was stimulated by the addition of heme to the PYG medium and produced acetic, lactic, and succinic acids but no butyric acid. Based on these characteristics, Eubacterium group no. 3 appears to be a new species. Lactobacillus. The genus Lactobacillus includes facultative, gram-positive bacilli that occur singly, in pairs, or in chains. All lactobacilli isolated in this study produced lactic acid as the major acid end product. Small amounts of acetic acid were produced by some strains. These isolates became aero-tolerant after repeated subculture. Lactobacillus strains isolated in the present study, based on biochemical characteristics presented in Table 3, appear to be different from Lactobacillus species described to date and probably belong to a new species of Lactobacillus. Lactobacillus species have been demonstrated in the gastrointestinal tract of mice by earlier workers (4, 11, 6). Dubos et al. (4) and Lee and Dubos (8) reported the presence of lactobacilli in mouse cecal flora and fecal contents, respectively. Moore et al. (11) reported lactobacilli at a concentration of 106 to 108/g of small intestinal contents. In the present study, about 109 lactobacilli were present per g of colon contents. Propionibacterium. Genus Propionibacte-
911
rium contains species that are gram-positive, nonsporeforming rods that produce acetic and propionic acids as the major end products. A Propionibacterium strain has been isolated from mouse colon contents in this study. To the best of our knowledge, this is the first reported isolation of propionibacteria from the intestinal tract of mice. The biochemical reactions of this new isolate are presented in Table 3. The strain appeared similar to P. acnes in fermenting glucose and fructose, in failing to ferment mannose and hydrolyze esculin, and in being able to reduce nitrates, hydrolyze gelatin, and produce indole (2, 6). However, this isolate fermented lactose, maltose, and sucrose, which is not characteristic of typical P. acnes, although a few strains of this species have been known to ferment maltose and sucrose (6). Our Propionibacterium isolate differs from P. granulosum in not fermenting mannose, fermenting lactose, in producing indole and in digesting gelatin. Peptostreptococcus. The genus Peptostreptococcus includes anaerobic to aerotolerant, gram-positive cocci which do not require fermentable carbohydrate for growth (6). Peptostreptococcus strains isolated in the present study were anaerobic when first isolated but became aerotolerant on repeated subculture. All the isolates could grow in the absence of a fermentable carbohydrate. Acetic, succinic, and propionic acids and variable amounts of lactic acids were produced in PYG broth. Based on biochemical reactions presented in Table 3, these strains appear similar to previously described strains of P. intermedius (2, 6). The results indicate that organisms isolated from mouse colon contents represent several genera of nonsporeforming anaerobes and that in most cases the isolates could not be unequivocally speciated. The data strongly suggest that many of the isolates are different from the previously described species of the respective genera and may belong to new species. It should be pointed out that all the isolates were obtained from the lumen contents of the colon of mice and do not necessarily reflect the types and numbers of bacteria which may be colonizing the epithelium in colon. ACKNOWLEDGMENTS We thank J. A. Breznak and J. M. Tiedje for reviewing and suggesting improvements in the manuscript. LITERATURE CITED 1. Aranki, A., S. Syed, E. Kenney, and R. Freter. 1969. Isolation of anaerobic bacteria from human gingiva and mouse cecum by means of a simplified glove box procedure. Appl. Microbiol. 17:568-576. 2. Buchanan, R., and N. Gibbons (ed.). 1974. Bergey's
912
3. 4. 5.
6. 7.
8.
APPL. ENVIRON. MICROBIOL.
HARRIS, REDDY, AND CARTER
Manual of determinative bacteriology. The Williams and Wilkins Co., Baltimore. Caldwell, D., and M. Bryant. 1966. Medium without rumen fluid for non-selective enumeration and isolation of rumen bacteria. Appl. Microbiol. 14:794-801. Dubos, R., R. Schaedler, R. Costello, and P. Hoet. 1965. Indigenous, normal and autochthonous flora of the gastrointestinal tract. J. Exp. Med. 122:67-75. Gordon, J., and R. Dubos. 1970. The anaerobic bacterial flora of the mouse cecum. J. Exp. Med. 132:251260. Holdeman, L., and W. Moore (ed.). 1972. Anaerobe laboratory manual. Virginia Polytechnic Institute Anaerobe Laboratory, Blacksburg, Va. Hungate, R. 1950. The anaerobic mesophilic cellulolytic bacteria. Bacteriol. Rev. 14:1-49. Lee, A., J. Gordon, and R. Dubos. 1968. Enumeration of oxygen sensitive bacteria usually present in the in-
9. 10. 11.
12. 13.
testine of healthy mice. Nature (London) 220:11371139. Lee, A., J. Gordon, C. Lee, and R. Dubos. 1971. The mouse intestinal microflora with emphasis on the strict anaerobes. J. Exp. Med. 133:339-352. Moore, W. 1966. Techniques for routine culture of fastidious anaerobes. Int. J. Syst. Bacteriol. 16:173-190. Moore, W., Et Cato, and L. Holdeman. 1969. Anaerobic bacteria of the gastrointestinal flora and their occurrences in clinical infections. J. Infect. Dis. 119:641649. Savage, D., J. McAllister, and C. Davis. 1971. Anaerobic bacteria on the mucosal epithelium of the murine large bowel. Infect. Immun. 4:492-502. Spears, R., and R. Freter. 1967. Improved isolation of anaerobic bacteria from the mouse cecum by maintaining strict anaerobiosis. Proc. Soc. Exp. Biol. Med. 124:903-908.
Vol. 31, No. 6 Printed in U.S.A.
Anaerobic Bacteria from the Large Intestine of Mice1 MARTHA A. HARRIS, C. ADINARAYANA REDDY,* AND GORDON R. CARTER Department of Microbiology and Public Health, Michigan State University, East Lansing, Michigan 48824 Received for publication 3 February 1976
Anaerobic bacteria from the colon of laboratory mice were enumerated and isolated using strict anaerobic techniques. Direct microscopic counts revealed 4.4 x 1010 organisms in each gram (wet weight) of colon contents. Actual cultural counts averaged 3.2 x 1010 organisms, which was 73% of the direct microscopic count. The tentatively identified genera were Bacteroides, Eubacterium, Fusobacterium, Lactobacillus, Peptostreptococcus, and Propionibacterium. Strains of Fusobacterium, Lactobacillus, Peptostreptococcus, and Propionibacterium were biochemically homogeneous. Strains of Bacteroides and Eubacterium, on the other hand, were biochemically heterogeneous and were subdiv;'led into several distinct groups. The data indicate that many of the isolates are different from previously described species of the respective genera and may belong to new species.
There have been several studies conducted on the isolation of anaerobic bacteria from the gastrointestinal tract of mice. Spears and Freter (13) isolated the anaerobic flora in the ceca of mice, by the agar plate-anaerobic jar procedure and the Hungate roll tube method, but did not identify the isolates. Aranki et al. (1) reported anaerobic, gram-positive bacilli, gram-negative cocco-bacilli, and rods with tapered ends in the ceca of mice but did not identify the isolates. On the basis of morphological studies, several investigators suggested that bacteria belonging to the genera Bacteroides, Clostridium, Eubacterium, Fusobacterium, and Lactobacillus were present in the gastrointestinal tract of mice (5, 8, 9, 12), but the data were inconclusive. The current investigation is therefore concerned with the enumeration and identification of the predominant anaerobic genera and with the description of subgroups within a given genus. The significance of this investigation is related to the fact that anaerobic bacteria are the predominant microbes in the gastrointestinal tract of mammals and they are frequently involved in clinical infections in humans. The incidence of nonsporing anaerobes in clinical infections of other animals has not been well documented. A knowledge of the predominant genera and species in the large intestine of healthy mice is essential to better understand the microbial ecology of this habitat and the specific organisms that may be involved in various pathological processes.
MATERIALS AND METHODS Animals and diets. Male, Swiss Webster mice were housed in individual steel cages (with wood shaving for bedding) in an animal room. The mice were fed Wayne Mouse Breeder Blox, a complete diet, and were provided with drinking water from individual glass bottles. Culture methods. The anaerobic roll tube techniques used in this study were those of Hungate (7) as modified by Moore (10) except where indicated otherwise. An oxygen-free gas mixture containing 85% nitrogen, 12% carbon dioxide, and 3% hydrogen was used to exclude air in preparing media, and during various cultural manipulations. The anaerobic culture system described by Holdeman and Moore (6) was used in this investigation. Culture media. Prereduced anaerobically sterilized media, autoclaved, dispensed, and inoculated in the complete absence of oxygen were used throughout this study. The medium used for agar roll tubes and maintenance slants was similar in composition to medium 10 of Caldwell and Bryant (3, 6), except that 0.0002% menadione was added and the volatile fatty acid mixture was excluded. Commercial prereduced anaerobically sterilized media purchased from the Robbin Laboratories division of Scott Laboratories were used for all biochemical tests. Isolation procedures. Adult male mice were sacrificed by the use of anhydrous ether and pinned on a dissecting board. The colon was removed aseptically and its contents were squeezed out and immediately weighed. Complete anaerobiosis was maintained after this step. The weighed colon contents were blended anaerobically under the oxygen-free gas mixture in a Waring blender containing a known volume of the anaerobic salts solution (6) for approximately 1 min. One milliliter of the blended intestinal contents was transferred aseptically and an' Article no. 7560, Michigan Agricultural Experiment aerobically to a sterile rubber stoppered tube (18 by Station. 907
908
APPL. ENVIRON. MICROBIOL.
HARRIS, REDDY, AND CARTER
150 mm) containing 9 ml of sterile anaerobic salts solution. Serial 10-fold dilutions were made to 1011. Gram-stained smears were made from the 10' dilution of the colon contents of each mouse and all morphotypes were recorded. Wet mounts were made from the 101 dilution and examined for motility with a phase-contrast microscope. Direct microscopic counts were made from the 103 dilution as described by Holdeman and Moore (6). Four roll tubes were made of each dilution from 107 to 1011 and the tubes were incubated at 37 C for 7 days. Colony counts were made in roll tubes containing 30 to 300 colonies. The procedures for analysis of the cultural counts were the same as those given by Holdeman and Moore (6). Isolation and characterization of bacteria. A platinum or stainless-steel transfer needle was used to pick representative colonies from the 109, 1010, and 1011 dilution roll tubes. Isolated colonies were stabbed into slants (3) of maintenance medium. After a 24-h incubation period, the organisms in the water of syneresis were examined for morphology, culture purity, and Gram reaction. A few of these cultures found to be mixed on microscopic examination were purified by streaking brain heart infusion roll tubes. Discrete colonies were picked and stabbed into slants of maintenance medium. Each isolate was examined for aerotolerance by streaking one-fourth of a blood agar plate and incubating it at 37 C aerobically for 24 h. For analysis of metabolic end products produced from carbohydrates, bacterial isolates were grown in peptone-yeast extract-glucose broth (PYG) for 48 h. The volatile acid end products were extracted with ether and the nonvolatile acids were methylated and extracted into chloroform as described by Holdeman and Moore (6). Both fractions were analyzed by a Dohrmann gas chromatograph using a Resoflex column. Helium was used as the carrier gas for the system with a flow rate of 120 cc/min. The column and thermal conductivity detector were maintained at approximately 118 to 120 C, whereas the injection port was maintained at 145 C. Procedures described by Holdeman and Moore (6) were used to perform biochemical tests. Four drops of an overnight culture in supplemented brain heart infusion broth, inoculated from the pure culture slants, was used as inoculum for differential biochemical media. The final pH in carbohydrate media was determined by a Beckman zeromatic pH meter. The final pH in carbohydrate broth above 6.0, between 6.0 to 5.5, or below 5.5 was read as no acid, weak acid, and strong acid production, respectively. All isolates were tested for their ability to grow in the presence of bile, hydrolyze esculin, liquify gelatin, produce indole, reduce nitrates, and produce acid, curd, and/or digestion of milk. All gram-positive rods were subjected to a heat test by inoculating a tube of starch broth with the test organism, heating at 80 C for 10 min, and incubating at 37 C for 24 h. Lack of growth in the starch broth after the heat test is taken to imply that the bacterial isolate is not a clostridial species. Young cultures, 12 h old, grown in peptone-yeast broth were Gram stained to determine the Gram reaction. A phase-contrast micro-
scope was used to determine motility from wet mounts prepared from peptone-yeast broth cultures.
RESULTS AND DISCUSSION Direct microscopic examination of the intestinal contents. The microbial morphotypes observed on direct microscopic examination of Gram-stained smears and wet mounts of the mouse colon contents (diluted 101) were similar to the microbial morphotypes observed by earlier investigators (5, 9). Gram-negative rods were the predominant morphotypes observed, although small numbers of gram-positive rods and cocci were also present. Gram-negative spiral-shaped organisms were observed in microscopic smears but none of these motile organisms was isolated. It is possible that: (i) the isolation medium did not provide the fastidious nutritional requirements of the spiral-shaped organisms; (ii) very small numbers of these organisms were present in the higher dilutions; or more importantly, (iii) special procedures (6) designed for the selective isolation of spirochetes would be necessary. Bacterial counts. In previous studies, 1010 to 1011 organisms/g of mouse intestinal contents were reported (11, 12). In the present investigation, a direct microscopic clump count revealed 4.4 x 10'° organisms/g (wet weight) of colon contents which was in close agreement with previous results. The mean cultural count per gram (wet weight) of colon contents was 3.2 x 1010 organisms (Table 1). This constituted approximately a 73% recovery of the organisms seen in the direct clump count, indicating the relative efficiency of the cultural methods used in this investigation. There were no noticeable differences in the number or types of anaerobes recovered from roll tubes inoculated with the colon contents from the three mice used in this particular study. Direct microscopic counts were generally higher than actual culture counts probably because: (i) some of the flora of the mouse such as spiral-shaped organisms are seen in smears but are rarely cultured (5, 12); (ii) some of the bacteria seen in clump counts TABLE 1. Enumeration of the cultural counts of anaerobic bacteria in the intestinal contents of mice" Animal no.
Wt of con- Avg counts Bacteria/g (wet tents (g) tion weight)
Mouse 1 0.22 32 1.3 x 10"' Mouse 11 0.31 53 1.6 x 10"' Mouse 111 0.52 78 6.7 x 10"' " Procedures in the Anaerobe Laboratory Manual (6) were used for enumerating the isolates.
VOL. 31, 1976
ANAEROBES FROM THE LARGE INTESTINE OF MICE
may be nonviable bacteria; and (iii) the medium may not be providing all the factors required for the growth, especially fatty acids known to be required for growth of several anaerobes from a number of anaerobic microbial habitats (2, 3, 6). Isolation and identification of intestinal anaerobes. Predominant bacterial groups in the mouse intestine were classified into genera based on Gram reaction, morphology, motility, oxygen tolerance, and acid metabolic end products. From the results presented in Table 2, gram-negative rods were assigned to the genera Bacteroides and Fusobacterium, gram-positive rods were placed in the genera Eubacterium, Lactobacillus, Propionibacterium, and gram-positive cocci in the genus Peptostreptococcus. The results were in close agreement with those of previous investigators with regard to the predominant anaerobic genera present in the intestinal tract of mice (5, 8, 9, 12) except that Clostridium species were not isolated in the present investigation. Savage et al. (12) observed the presence of Clostridium species in frozen sections of the large bowel of mice. Gordon and Dubos (5) and Dubos et al. (4) reported Clostridium species in the 107 and 10W dilutions of mouse cecal contents and fecal flora, respectively. In the present study, all the organisms isolated were from the 10' to 1011 dilution roll tubes and this may account for the fact that Clostridium species were not isolated by us. Bacteroides. Bacteroides isolates were from 1010 to 1011 dilutions and were obligately anaerobic, gram negative, non-sporulating rods with rounded or tapered ends. They produced varying amounts of acetate, lactate, propionate, and succinate and were nonmotile. None of the isolates produced black pigment on anaerobically incubated blood agar plates. The addition of hemin to the PYG broth, greatly enhanced the
909
growth of most isolates in comparison to the same medium without hemin. The results of this investigation are in agreement with the findings of Gordon and Dubos (5) and Lee et al. (8) who also reported the presence of Bacteroides species in the intestine of mice. Based on biochemical characteristics presented in Table 3, the Bacteroides isolates were divided into six distinct groups. As discussed below, none of the six groups are exactly similar to the presently recognized species of Bacteroides, suggesting that these isolates may belong to Bacteroides species that have not yet been described. Bacteroides group no. 1 isolates were nonmotile and fermented a number of carbohydrates including fructose, lactose, and maltose but not melezitose. End products produced by these strains included major amounts of acetic and small amounts of succinic acids. No other acid end products were detectable. Based on these characteristics they appeared similar toB. clostridiiformis subsp. clostridiiformis (3, 6), but differ from previously described strains of this species in not fermenting sucrose and in not hydrolyzing esculin (2, 6). Bacteroides group no. 2 isolates were nonmotile, nonfermentative rods that produced small amounts of acetic and succinic acids and traces of lactate as end products. These isolates were similar to B. coagulans but differed from typical strains of this species in hydrolyzing esculin, not growing in the presence ofbile and even more important in not being able to curdle milk. B. coagulans by definition (2) curdles milk. Thus this group of isolates appear to be a new Bacteroides species. Bacteroides group no. 3 isolates were nonmotile, did not produce black pigment on blood agar, produced succinic and acetic acids as major end products, fermented fructose and rhamnose, produced indole, gave growth in 20% bile, and were stimulated by the addition of heme to
TABLE 2. Some features used for identifying major groups of bacteria from the
mouse
large intestine
Fermenta-
Morphological group
b
Oxygen products from eranceatol- tion glucose'
Gram-negative rods ALSP A Rounded to tapered rods ABL A Tapered rods Gram-positive rods ALS A Short rods LA AT Short rods AP A Short to medium-sized rods AT ALPS Gram-positive cocci A, Anaerobic; AT, aerotolerant anaerobe. A, Acetic; B, butyric; L, lactic; P, propionic; S, succinic.
Population levels
109-10"
109-10X}
1O9-1OlI 109 109
1O
Tentative identification
Bacteroides Fusobacterium Eubacterium Lactobacillus
Propionibacterium Peptostreptococcus
910
APPL. ENVIRON. MICROBIOL.
HARRIS, REDDY, AND CARTER
TABLE 3. Biochemical reactions of predominant anaerobic bacteria isolated form the mouse intestine Group
a o ~
8O ~
|E0 B
B
B
A
0
a
to
IV
Bacteroides a A A A a a a A A A 2 Abl Group no. 1 c -++ 2 - _+.W - -_ + + Group no. 2 A A -a a a- -- A w c -+ + + 3 - a Group no. 3 2 - a a A -a A a Group no. 4 + 3 - -- + _ Group no. 5 __ a A A a a a A A cd + 2 - a Group no. 6 Eubacterium 1 - - + A A A A Group no. 1 c +-. 2 A A + A A AA - A a A AA - A A - c ++ - Group no. 2 1 - - A A A A .- A - + c - + + + Group no. 3 2 - a + - A -- - a - a a a A A w c -- + Fusobacterium + - + + 5 - - - A A -- - a -- A a Lactobacillus + c + -_ 6 - A + A A A A - A - A a A - A Peptostreptococcus a + cd - + + + 1 - Propionibacterium a _ AIAIAIA.I a The number of strains studied was not proportional to the actual numbers grown from samples as colonies were not picked in a random fashion. b Symbols: A, Acid; a, weak acid; +, positive; -, negative; w, weak reaction; c, curd; d, digestion. -
the medium. Based on these characteristics they key to B. fragilis group (2, 6). However, these strains are significantly different from typical B. fragilis strains in not fermenting lactose, maltose, and sucrose. Therefore, further work needs to be done to determine the exact taxonomic status of this group of isolates. Bacteroides group no. 4 isolates produced major amounts of lactic acid and trace amounts of acetic and succinic acids. All strains produced acid from glucose and weakly fermented fructose, rhamnose, and mannose. These strains appeared somewhat similar to B. furcosus but differed from previously described strains of this species in failing to grow in the presence of bile, in fermenting raffinose, and in not hydrolyzing esculin (2, 6). These data suggest that this group of isolates may belong to a new Bacteroides species. Bacteroides group no. 5 isolates are nonmotile, did not produce indole, did not hydrolyze gelatin or esculin, did not reduce nitrate, and did not ferment carbohydrates. Only trace amounts of acetic acid is produced in PYG medium. Based on the above-mentioned characteristics, this group of isolates may be considered similar to B. pneumosintes (2, 6). However, B. pneumosintes, by definition, is a small coccobacillus, while the group no. 5 isolates were medium-sized rods. Therefore, Bacteroides group no. 5 isolates probably belong to a new Bacteroides species which is somewhat related to but different from B. pneumosintes (2, 6). Strains designated Bacteroides group no. 6
isolates were gram-negative, anaerobic rods and produced major amounts of acetate and lactate, fermented a moderate number of carbohydrates, and grew in the presence of bile. On the basis of characteristics studied, these isolates appeared to be different from any of the Bacteroides species described to date (2). Fusobacterium. Savage et al. (12) reported the presence of Fusobacterium species in the mouse intestine. Gordon and Dubos (5) isolated organisms similar to F. russii in the 1010 to 1011 dilutions of the cecal contents of laboratory mice. In agreement with these earlier results, Fusobacterium species were isolated from the mouse intestinal contents in this investigation. These organisms were anaerobic, nonmotile, gram-negative, nonsporulating rods with tapered ends and produced butyric acid as the major acid end product in PYG broth. The biochemical characteristics of the Fusobacterium species presented in Table 3 show that these isolates are different from any of the previously described Fusobacterium species (2, 6). Eubacterium. In agreement with the results of the present investigation, Gordon and Dubos (5) reported that Eubacterium species were present at levels of 1010 to 1011 organisms/g of cecal contents. Savage et al. (12) also demonstrated, by electron microscopy, organisms similar to Eubacterium in frozen sections of the large bowel of mice. In agreement with these earlier findings, Eubacterium strains were isolated from 1010 to 1011 dilution of the mouse colon contents in the present study. All the isolates were subjected to heat tests and found
VOL. 31, 1976
ANAEROBES FROM THE LARGE INTESTINE OF MICE
-to be negative. The negative heat tests indicated that these gram-positive bacilli were not Clostridium species. All the isolates were gram-positive nonsporeforming, bacilli that occur in chains, and clumps (6). The biochemical characteristics of these Eubacterium isolates are presented in Table 3. Eubacterium group no. 1 isolate produced acetic, lactic, and succinic acids, but no butyric acid. Among the carbohydrates tested, only fructose, glucose, lactose, and maltose were fermented. This strain reduced nitrate, hydrolyzed esculin, and coagulated milk. Based on these criteria, this isolate does not fit with the description for any of the presently described species of Eubacterium and appears to be a new species. Intestinal isolates classified as Eubacterium group no. 2 fermented a variety of carbohydrates and produced acetic, formic, and small amounts of lactic and succinic acids. These strains appeared similar to E. contortum in not producing butyric acid or indole and in producing acid from fructose, maltose and rhamnose (2, 6). However, they differed from conventional strains of E. contortum in not fermenting arabinose and in reducing nitrates (6). Strain designated Eubacterium group no. 3 produced acid from fructose, glucose, lactose, and maltose, produced indole, hydrolyzed gelatin, and grew in 20% bile. This strain was stimulated by the addition of heme to the PYG medium and produced acetic, lactic, and succinic acids but no butyric acid. Based on these characteristics, Eubacterium group no. 3 appears to be a new species. Lactobacillus. The genus Lactobacillus includes facultative, gram-positive bacilli that occur singly, in pairs, or in chains. All lactobacilli isolated in this study produced lactic acid as the major acid end product. Small amounts of acetic acid were produced by some strains. These isolates became aero-tolerant after repeated subculture. Lactobacillus strains isolated in the present study, based on biochemical characteristics presented in Table 3, appear to be different from Lactobacillus species described to date and probably belong to a new species of Lactobacillus. Lactobacillus species have been demonstrated in the gastrointestinal tract of mice by earlier workers (4, 11, 6). Dubos et al. (4) and Lee and Dubos (8) reported the presence of lactobacilli in mouse cecal flora and fecal contents, respectively. Moore et al. (11) reported lactobacilli at a concentration of 106 to 108/g of small intestinal contents. In the present study, about 109 lactobacilli were present per g of colon contents. Propionibacterium. Genus Propionibacte-
911
rium contains species that are gram-positive, nonsporeforming rods that produce acetic and propionic acids as the major end products. A Propionibacterium strain has been isolated from mouse colon contents in this study. To the best of our knowledge, this is the first reported isolation of propionibacteria from the intestinal tract of mice. The biochemical reactions of this new isolate are presented in Table 3. The strain appeared similar to P. acnes in fermenting glucose and fructose, in failing to ferment mannose and hydrolyze esculin, and in being able to reduce nitrates, hydrolyze gelatin, and produce indole (2, 6). However, this isolate fermented lactose, maltose, and sucrose, which is not characteristic of typical P. acnes, although a few strains of this species have been known to ferment maltose and sucrose (6). Our Propionibacterium isolate differs from P. granulosum in not fermenting mannose, fermenting lactose, in producing indole and in digesting gelatin. Peptostreptococcus. The genus Peptostreptococcus includes anaerobic to aerotolerant, gram-positive cocci which do not require fermentable carbohydrate for growth (6). Peptostreptococcus strains isolated in the present study were anaerobic when first isolated but became aerotolerant on repeated subculture. All the isolates could grow in the absence of a fermentable carbohydrate. Acetic, succinic, and propionic acids and variable amounts of lactic acids were produced in PYG broth. Based on biochemical reactions presented in Table 3, these strains appear similar to previously described strains of P. intermedius (2, 6). The results indicate that organisms isolated from mouse colon contents represent several genera of nonsporeforming anaerobes and that in most cases the isolates could not be unequivocally speciated. The data strongly suggest that many of the isolates are different from the previously described species of the respective genera and may belong to new species. It should be pointed out that all the isolates were obtained from the lumen contents of the colon of mice and do not necessarily reflect the types and numbers of bacteria which may be colonizing the epithelium in colon. ACKNOWLEDGMENTS We thank J. A. Breznak and J. M. Tiedje for reviewing and suggesting improvements in the manuscript. LITERATURE CITED 1. Aranki, A., S. Syed, E. Kenney, and R. Freter. 1969. Isolation of anaerobic bacteria from human gingiva and mouse cecum by means of a simplified glove box procedure. Appl. Microbiol. 17:568-576. 2. Buchanan, R., and N. Gibbons (ed.). 1974. Bergey's
912
3. 4. 5.
6. 7.
8.
APPL. ENVIRON. MICROBIOL.
HARRIS, REDDY, AND CARTER
Manual of determinative bacteriology. The Williams and Wilkins Co., Baltimore. Caldwell, D., and M. Bryant. 1966. Medium without rumen fluid for non-selective enumeration and isolation of rumen bacteria. Appl. Microbiol. 14:794-801. Dubos, R., R. Schaedler, R. Costello, and P. Hoet. 1965. Indigenous, normal and autochthonous flora of the gastrointestinal tract. J. Exp. Med. 122:67-75. Gordon, J., and R. Dubos. 1970. The anaerobic bacterial flora of the mouse cecum. J. Exp. Med. 132:251260. Holdeman, L., and W. Moore (ed.). 1972. Anaerobe laboratory manual. Virginia Polytechnic Institute Anaerobe Laboratory, Blacksburg, Va. Hungate, R. 1950. The anaerobic mesophilic cellulolytic bacteria. Bacteriol. Rev. 14:1-49. Lee, A., J. Gordon, and R. Dubos. 1968. Enumeration of oxygen sensitive bacteria usually present in the in-
9. 10. 11.
12. 13.
testine of healthy mice. Nature (London) 220:11371139. Lee, A., J. Gordon, C. Lee, and R. Dubos. 1971. The mouse intestinal microflora with emphasis on the strict anaerobes. J. Exp. Med. 133:339-352. Moore, W. 1966. Techniques for routine culture of fastidious anaerobes. Int. J. Syst. Bacteriol. 16:173-190. Moore, W., Et Cato, and L. Holdeman. 1969. Anaerobic bacteria of the gastrointestinal flora and their occurrences in clinical infections. J. Infect. Dis. 119:641649. Savage, D., J. McAllister, and C. Davis. 1971. Anaerobic bacteria on the mucosal epithelium of the murine large bowel. Infect. Immun. 4:492-502. Spears, R., and R. Freter. 1967. Improved isolation of anaerobic bacteria from the mouse cecum by maintaining strict anaerobiosis. Proc. Soc. Exp. Biol. Med. 124:903-908.
