Characterization of Several Bovine Rumen Bacteria

Transcription

1 JOURNAL OF BACTERIOLOGY, May, American Society for Microbiology Vol. 91, No. 5 Printed in U.S.A. Characterization of Several Bovine Rumen Bacteria Isolated with a Xylan Medium1 B. A. DEHORITY Department of Animal Science, Ohio Agricultural Research and Development Center, Received for publication 4 March 1966 Wooster, Ohio ABSTRACT DEHORITY, B. A. (Ohio Agricultural Research and Development Center, Wooster). Characterization of several bovine rumen bacteria isolated with a xylan medium. J. Bacteriol. 91: Studies were conducted to characterize eight strains of bacteria isolated from bovine rumen contents, by use of a medium containing xylan as the only added carbohydrate source. Based on morphology, biochemical reactions, nutritional requirements, and fermentation products, five of the eight strains were identified as Butyrivibrio fibrisolvens. Many properties of the remaining three strains resembled Bacteroides ruminicola; however, propionic acid was consistently found as a fermentation product. When the type strains for B. ruminicola subsp. ruminicola and B. ruminicola subsp. brevis were compared with the present isolates, it was found that propionic acid was a normal fermentation product for the type strain B. ruminicola subsp. ruminicola when grown in a 40% rumen fluid-0.5 % glucose broth. Production of propionic acid was markedly reduced for all strains when grown in a 20% rumen fluid-i % glucose broth. The three remaining strains were thus placed in the species B. ruminicola, and further classified into the subspecies ruminicola (one strain) and brevis (two strains) on the basis of their requirement for hemin. Although the type strain of B. ruminicola subsp. brevis did not produce propionic acid, both of the present isolates classified as this subspecies produced substantial amounts. One strain of B. ruminicola subsp. brevis had an absolute requirement for volatile fatty acids. Either isobutyric or DL-2-methylbutyric acid would satisfy this requirement, whereas isovaleric acid was ineffective. It is of interest that xylan-fermenting bacteria isolated from 107 and 108 dilutions of rumen contents by use of a xylan medium are similar to the xylan fermenters isolated at the same dilutions with a nonselective medium. Fermentation of the fiber portion of forages, probably one of the more important functions of the rumen bacterial population, has primarily been studied in pure culture, with regard only to those bacteria which can' attack the cellulosic fraction. The ability of mixed cultures of rumen bacteria and several species isolated in pure culture to digest hemicellulose has been reviewed in an earlier publication (14). Bryant (3, 4) has listed a number of rumen bacterial species which are known to ferment xylan, a major portion of the hemicellulose fraction of forages. However, most of these bacteria were isolated with a nonselective medium. Butterworth, Bell, and Gar- vock (11), using wheat flour pentosan as a substrate, isolated a xylan-fermenting bacterium from the bovine rumen and tentatively assigned it to the genus Butyrivibrio. A sheep rumen bacterium was isolated by Walker (22) with the use of a medium containing wheaten hay hemicellulose as the only added carbohydrate. This organism was tentatively placed in the genus Bacteroides on the basis of fermentation products and biochemical characteristics. Hobson and Purdom (16) isolated two types of xylan-fermenting sheep rumen bacteria in the wheat flour pentosan medium previously described by Butterworth et al. (11). One organism was identified as a species of Butyrivibrio, and the second organism, a gram- 1 Approved for publication as Journal Article No. variable coccobacillus, was studied in some detail by the Associate Director of the Ohio Agricultural Research and Development Center, Wooster, isolation and characterization of a number of but not named. The present paper describes the Ohio. strains of bovine rumen bacteria isolated in a me- 1724

2 VOL. 91, 1966 XYLAN-FERMENTING RUMEN BACTERIA 1725 dium with purified xylan as the only added carbohydrate source. MATERIALS AND METHODS Rumen ingesta was obtained through a permanent rumen fistula, from a steer maintained on mixed hay. The anaerobic cultural techniques used throughout the study were similar to those described by Hungate (17). Carbon dioxide and nitrogen were freed from oxygen by passage through a column of hot reduced copper turnings. For isolation of the bacteria, rumen ingesta was diluted 1OX with the anaerobic dilution solution (ADS) of Bryant and Burkey (5), agitated for 3 min in a Waring Blendor under a stream of CO2, and further diluted in ADS. Roll tubes were inoculated with amounts to give a final concentration of 10-7, and 10-8 g of rumen ingesta per tube. The roll-tube medium was identical to the rumen fluid-glucose-cellobioseagar described by Bryant and Burkey (5), except that 0.5% purified xylan (Calbiochem) was substituted for the soluble sugars. The finely ground xylan became almost completely soluble at this concentration during autoclaving; however, some precipitation of the substrate was observed in the solidified medium. Thus, a somewhat lower xylan concentration was used in all subsequent media. Colonies were picked from the roll tubes after incubation for 24 to 48 hr at 38 C and stabbed into rumen fluid-xylan-agar (RXA) slants. Xylan concentration in the slant medium was 0.3%. From a total of 25 isolates, eight strains were chosen on the basis of morphology and the Gram stain for further study. Culture purity was checked by culturing appropriate dilutions in RXA roll tubes. Four colonies were picked from the tube with the highest dilution showing growth. If the bacteria from all four colonies were similar in morphology and the Gram stain, the culture was considered to be pure; one of these four colonies was then maintained as the stock culture. All cultures were carried in RXA slants. The methods used for characterization of these bacteria have been reported previously (6, 8, 9, 13), except as outlined below. Aside from tests on the fermentation of various carbon sources, glucose was used as a substrate throughout. Cellulose digestion was determined quantitatively by the method of Crampton and Maynard (12). Composition of the defined medium plus "vitamin-free" casein hydrolysate (complete medium) was identical to that reported by Scott and Dehority (21). Determination of minimal ph for strains requiring CO2 was made in poorly buffered glucose broth (13); with the remaining strains, Na2CO3 was deleted from the medium and nitrogen served as the gaseous phase. The complete medium of Scott and Dehority (21) with all volatile fatty acids (VFA) omitted was used as a basal medium to study the VFA requirement of Bacteroides ruminicola H2b. Nine different media were used, with the following additions made to the VFA free basal: (i) acetic acid; (ii) valeric acid; (iii) isobutyric acid; (iv) DL-2-methylbutyric acid; (v) isovaleric acid; (vi) ii plus iii; (vii) ii plus iv; (viii) ii plus v; (ix) all acids. Whether the acids were added singly or in various combinations, the following concentrations, in milligrams per 100 ml of medium, were used: acetic, 133; isobutyric, 6.6; isovaleric, 8.0; DL-2-methylbutyric, 8.0; and valeric, 8.0. Fermentation products were determined in cultures incubated for 7 days in 40% rumen fluid-0.5% glucose broth and 40% rumen fluid-0.3% xylan broth by use of silica gel chromatography (E. G. Linke, M.S. Thesis, The Ohio State University, Columbus, 1952). Gas chromatography was used to determine qualitatively ethyl alcohol production as well as to check the ratios of volatile fatty acids determined with the silica gel column. RESULTS Morphology of the eight strains was studied with Gram-stained smears from the water of syneresis of 16- to 20-hr RXA slant cultures. Strains H4a, H1Ob, H13b, H16a, and H17c were gram-negative, slightly curved rods, varying in size from 0.4 to 0.6,u by 0.7 to 2.0 A. Curving was more pronounced in strains H1Ob, H13b, and H16a. Strains H4a, H13b, and H17c tended to occur in pairs joined end to end. Strains H2b, H8a, and H15a were coccoid to oval in shape, gram-negative, and varied in size from 0.5 to 0.6,u. Although some rod forms were seen, the proportion was quite low, and cells usually occurred as singles or pairs, with some occasional short chains being present. All eight strains were strict anaerobes and were consistent in their ability to liquify gelatin and digest casein. None of the strains produced indole or reduced nitrate. Substrates fermented by all strains were: glucose, xylose, cellobiose, arabinose, fructose, sucrose, lactose, and galactose. Table 1 presents the data on those physiological characteristics which differed among the strains. The fermentation products produced by these strains after 1 week of incubation in 40% rumen fluid-0.5% glucose broth and 40% rumen fluid- 0.3% xylan broth are shown in Table 2. Strains H4a, HlOb, H13b, H16a, and H17c can readily be classified as belonging to the species Butyrivibrio fibrisolvens (8) on the basis of the following characteristics: morphology and Gram stain, motility, lack of a requirement for CO2 and rumen fluid (except H13b), gas production, cellulose digestion (except H4a), and fermentation products. Although none of the strains were identical to the type strain Dl (8), their physiological characteristics were within the limits of previously isolated strains of this species (8, 11, 15-19). The only exception appears to be the extremely large requirement of strain H13b for rumen fluid. Motility could not be demonstrated for strain H17c with the test employed (migration from stab inoculation in 0.5% agar-0.1 % glucose medium); however, subsequent observa-

3 1726 DEHORITY J. BACTERIOL. TABLE 1. Physiological characteristics which differed among eight strains of rumen bacteria isolated by use of a xylan medium Characteristic Strain H4a HiOb H13b H16a H17c H2b H8a Hl5a Motility _ H2S production Cellulose digestion, %, a Growth in Trypticase-yeast extract medium i C02 requirement Starch digestion ± Final ph, glucose Gas production b Acid from Maltose ± + Inulin Dextrin Growth on complete mediumc, Plus 10% CRFd Minus VFAe... ± f Minus casein hydrolysate... + i - i Minus NH4... i a Cellulose substrate added at 0.75%; values less than 1% were considered as no digestion. b Refers to very slight gas production. c Defined medium plus acid hydrolysate of casein (21). d Clarified rumen fluid. e Volatile fatty acids. Concentrations used, in milligrams per 100 ml of medium, were: acetic, 133; isobutyric, 6.6; isovaleric, 8.0; and valeric, 8.0. f These strains required rumen fluid and thus could not be studied. TABLE 2. Fermentation products produced from the fermentation of glucose and xylan by eight strains of rumen bacteria isolated by use of xylan medium Product Strain H4a HiOb H13b H16a H17c H2b H8a Hl5a In 40% rumen fluid-0.5% glucose medium Butyric acid * Propionic acid Acetic acid Formic acid Succinic acid Lactic acid Ethyl alcohol... + ±4-t + ±t In 40% rumen fluid-0.3% xylan medium Butyric acid Propionic acid Acetic acid Formic acid Succinic acid Lactic acid * Results expressed as millimoles of product per 100 ml of medium. t Present in only trace amounts.

4 VOL. 91, 1966 XYLAN-FERMENTING RUMEN BACTERIA 1727 TABLE 3. Growth response of strains H2b, H8a, and HISa to hemin and an enzymatic hydrolysate of casein Medium Strain H2b H8a HlSa Complete medium* Plus hemin (0.0002%) Minus volatile fatty acids Minus casein hydrolysate Minus NH4+ (CHA)t ±... i 4± Minus NH4+ (CHE) * Defined medium plus acid hydrolysate of casein (21). t Ammonia-free acid hydrolysate of casein as sole nitrogen source. t Ammonia-free enzymatic hydrolysate of casein as sole nitrogen source. tions of wet mounts by phase-contrast microscopy revealed that this strain was motile. Identification of the remaining three strains (H2b, H8a, and H15a) was somewhat more complicated. Although they differed in certain physiological characteristics, their similarity in morphology, lack of motility and gas production, requirement for CO2, and fermentation products suggested that they might all belong to the same species. On the basis of morphology and physiological characteristics, these three strains appeared to be closely related to Bacteroides ruminicola (9); however, the consistent production of propionic acid differs markedly from the fermentation products reported. Studies by Bryant and Robinson (7) and Pittman and Bryant (20) indicated that the rumen fluid requirement of B. ruminicola subsp. ruminicola could be replaced by hemin, and that, although slight growth of both subspecies occurred with an acid hydrolysate of casein as the sole nitrogen source, maximal growth was obtained with an enzymatic hydrolysate of casein. Evaluation of these nutritional requirements (Table 3) substantiated classification of these strains as being very closely related to B. ruminicola. The type strains of B. ruminicola subsp. brevis (strain GA 33) and ruminicola (strain 23) were obtained from M. P. Bryant, Department of Dairy Science, University of Illinois, Urbana. Fermentation products of the type strains and the three present isolates were compared in the 40%7, rumen fluid-0.50% glucose medium used earlier in this study and in the 20% rumen fluid-0.5 %. Trypticase-1 % glucose medium used by Bryant et al. (9) in their classification studies. Table 4 presents the fermentation products of these bacteria on the two different media. The type strain for B. ruminicola subsp. ruminicola (strain 23) produced propionic acid in amounts approximately equal to that of the present isolates, when the organisms were grown in a 40% rumen fluid-0.5% glucose broth. However, propionic acid producduction was almost negligible when the organisms were grown in the 200% rumen fluid-0.5 %, Trypticase-i % glucose medium previously used by TABLE 4. Fermentation acids produced by Bacteroides ruminicola subsp. ruminicola (strain 23), subsp. brevis (strain GA33), and strains H2b, H8a, and H15a on two different media Total acid meq % of total acid produced* Medium Strain produced (meq/100 ml) Butyric Propionic Acetic Formic Succinic Lactic 40C/% rumen fluid-0.5% glucose GA H2b H8a HlSa Blank values 23, GA H2b, H8a, H15a % rumen fluid-0.5% trypticase-1% GA glucose H2b H8a H15a Blank values All strains * Values of 3% or less are considered of doubtful significance.

5 1728 DEHORITY J. BACTERIOL. Bryant et al. (9). The type strain for subspecies brevis (strain GA33) did not produce propionate in either medium. Based on the above results, it appeared that the three strains H2b, H8a, and H15a should be classified as belong to the species B. ruminicola. A division of the species into two subspecies on the basis of requirement for hemin (9) would place strains H2b and H8a into the subspecies brevis and strain H15a into the subspecies ruminicola. Strain H2b was studied in detail to determine which acids would satisfy this organism's requirement for volatile fatty acids. Growth occurred only in those tubes which contained isobutyric or DL-2-methylbutyric acid. Six additional strains of the original 25 isolated were examined in an attempt to gain further information on the production of propionic acid by B. ruminicola. Although these strains were not reisolated, as was the case with the first eight, they were extensively examined microscopically and appeared to be pure cultures. Based on morphology, nutritional requirements, and biochemical reactions, these six strains were presumptively identified as B. ruminicola subsp. ruminicola (three strains), B. ruminicola subsp. brevis, B. fibrisolvens, and Eubacterium ruminantium (2). These presumptively identified strains of B. ruminicola produced from 17.8 to 19.5 meq% of propionic acid as a fermentation product when grown in 40% rumen fluid-0.5% glucose medium. DISCUSSION In a review of the groups of anaerobic bacteria active in the rumen, Bryant (4) has listed some important characteristics of the predominant species. Included in this list is their ability to use xylan as an energy source. Although most of these bacteria were isolated by use of a nonselective medium, those species which were capable of digesting xylan and occurred in fairly high numbers were identical to those isolated in the present study. The main exception to this would be the cellulolytic ruminococci, which were not found in this series of isolations with a selective xylan medium. These observations support the hypothesis that, based on numbers and reactions, a good deal of information can be obtained about the rumen bacterial population by the isolation of strains with a nonselective medium. The possibility exists, with both cellulose- and xylan-digesting bacteria, that important species may exist in the rumen which are only capable of using the oligo- or polysaccharide forms as sources of energy. Most of the evidence accumulated to date, however, including the present work, has been to the contrary. In the three previous studies (11, 16, 22) where rumen bacteria were isolated in a selective xylan medium, species of Butyrivibrio were consistently isolated, except in the study of Walker (22). Even in this instance, however, a number of the characteristics reported, as well as the inability of a number of workers to detect motility in Butyrivibrio, do not exclude the possibility that his organism was a species of Butyrivibrio. Combined with the present results, the evidence for the importance of B. fibrisolvens in the digestion of forage hemicellulose becomes fairly clear, and strongly supports the proposal to this effect made originally by Bryant and Small (8). Propionic acid was not found as a fermentation end product by Bryant et al. (9) in their original report on the characterization of Bacteroides ruminicola as a new species. However, in later studies, Bryant et al. (10) and Bladen, Bryant, and Doetsch (1) found several strains of bacteria very closely related to B. ruminicola, which normally produced some propionic acid in a 20% rumen fluid-0.5% Trypticase-1l% glucose medium. On the other hand, the positively identified strains of B. ruminicola did not produce propionic acid as a fermentation product in this medium, whereas all but one strain (GA33) did produce this acid in a 40% rumen fluid-0.5% glucose medium. The discrepancy in end products observed between the two media suggests the possibility that such factors as ph and the level of acids originally present in the medium may be responsible. Gill and King (15) and Lee and Moore (18), working with several strains of Butyrivibrio fibrisolvens, have demonstrated the effect of these factors on the ratios of fermentation acids produced. Based on the present findings, it is concluded that many strains of Bacteroides ruminicola subsp. ruminicola and brevis possess the ability to decarboxylate succinic acid to propionic acid. The extent to which this occurs is governed by certain conditions of the fermentation which have not yet been elaborated. Although the type strain of subspecies brevis did not produce propionic acid, the strains classified as subspecies brevis in the present study produced considerable quantities. On the basis of the data obtained with the present strains of B. ruminicola and type strain 23 of subspecies ruminicola, it is suggested that the description of the species B. ruminicola subsp. ruminicola and brevis (9) be amended to include propionic acid as a fermentation product. All previously isolated strains of B. ruminicola subsp. brevis will grow in a medium in which rumen fluid is replaced by Trypticase and yeast extract (1, 7, 9). Although VFA plus acetate were stimulatory to the growth of some strains and could replace the casein hydrolysate require-

6 VOL. 91, 1966 XYLAN-FERMENTING RUMEN BACTERIA 1729 ment of others, these VFA were not essential. Strain H2b exhibits an absolute requirement for VFA, which can be satisfied by either isobutyric or DL-2-methylbutyric acid alone. Acetate was included in all media, so the possible effect of this acid is not yet known. ACKNOWLEDGMENTS I wish to acknowledge the technical assistance of H. W. Scott in determining the fermentation products. I am also indebted to M. P. Bryant for supplying the type strains of B. ruminicola subsp. ruminicola and brevis, and for his valuable criticism in the preparation of this manuscript. This investigation was supported by funds allocated to the Ohio Agricultural Research and Development Center on the recommendation of the North Central Technical Committee on Biochemistry of Forage Utilization (NC-63). LITERATURE CITED 1. BLADEN, H. A., M. P. BRYANT, AND R. N. DOETSCH A study of bacterial species from the rumen which produce ammonia from protein hydrolyzate. Appl. Microbiol. 9: BRYANT, M. P Bacterial species of the rumen. Bacteriol. Rev. 23: BRYANT, M. P Some aspects of ruminal metabolism revealed by pure culture studies of the indigenous flora. Proc. Intern. Grassland Congr. Eighth, p BRYANT, M. P Symposium on microbial digestion in ruminants: identification of groups of anaerobic bacteria active in the rumen. J. Animal Sci. 22: BRYANT, M. P., AND L. A. BURKEY Cultural methods and some characteristics of some of the more numerous groups of bacteria in the bovine rumen. J. Dairy Sci. 36: BRYANT, M. P., AND R. N. DoETsCH A study of actively cellulolytic rod-shaped bacteria of the bovine rumen. J. Dairy Sci. 37: BRYANT, M. P., AND I. M. ROBINSON Some nutritional characteristics of predominant culturable ruminal bacteria. J. Bacteriol. 84: BRYANT, M. P., AND N. SMALL The anaerobic monotrichous butyric acid-producing curved rod-shaped bacteria of the rumen. J. Bacteriol. 72: BRYANT, M. P., N. SMALL, C. BouMA, AND H. CHU Bacteroides ruminicola n. sp. and Succinimonas amylolytica the new genus and species. Species of succinic acid-producing anaerobic bacteria of the bovine rumen. J. Bacteriol. 76: BRYANT, M. P., N. SMALL, C. BouMA, AND I. ROBINSON Studies on the composition of the ruminal flora and fauna of young calves. J. Dairy Sci. 41: BUTTERWORTH, J. P., S. E. BELL, AND M. G. GARVOCK Isolation and properties of the xylan-fermenting bacterium 11. Biochem. J. 74: CRAMPTON, E. W., AND L. A. MAYNARD The relation of cellulose and lignin content to the nutritive value of animal feeds. J. Nutr. 15: DEHORrry, B. A Isolation and characterization of several cellulolytic bacteria from in vitro rumen fermentations. J. Dairy Sci. 46: DEHORITY, B. A Degradation and utilization of isolated hemicellulose by pure cultures of cellulolytic rumen bacteria. J. Bacteriol. 89: GILL, J. W., AND K. W. KING Nutritional characteristics of a Butyrivibrio. J. Bacteriol. 75: HOBSON, P. N., AND M. R. PURDOM Two types of xylan fermenting bacteria from the sheep rumen. J. Appl. Bacteriol. 24: HUNGATE, R. E The anaerobic mesophilic cellulolytic bacteria. Bacteriol. Rev. 14: LEE, H. C., AND W. E. C. MOORE Isolation and fermentation characteristics of strains of Butyrivibrio from ruminal ingesta. J. Bacteriol. 77: MARGHERITA, S. S., AND R. E. HUNGATE Serological analysis of Butyrivibrio from the bovine rumen. J. Bacteriol. 86: PITIMAN, K. A., AND M. P. BRYANT Peptides and other nitrogen sources for growth of Bacteroides ruminicola. J. Bacteriol. 88: ScoTT, H. W., AND B. A. DEHoRrrY Vitamin requirements of several cellulolytic rumen bacteria. J. Bacteriol. 89: WALKER, D. J Isolation and characterization of a hemicellulose fermenting bacterium from the sheep rumen. Australian J. Agr. Res. 12: