Isolation And Characterization Of Thirteen Intestinal

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Mar. 1981, p. 737-745

0099-2240/81/030737-09$02.00/0

Vol. 41, No. 3

Isolation and Characterization of Thirteen Intestinal Microorganisms Capable of 7a-Dehydroxylating Bile Acids SEIJU HIRANO,* RYOSEI NAKAMA, MICHIHIRO TAMAKI, NORIYUKI MASUDA, AND HIROSHI ODA Department of Bacteriology, Faculty ofMedicine, Kagoshima University, Kagoshima 890, Japan

Thirteen anaerobic bacteria capable of performing the 7a-dehydroxylation of both cholic acid and chenodeoxycholic acid were isolated from human feces and also from sewage. Ten organisms from heat-treated samples were species of Clostridium identical or closely related to the Clostridium bifermentans-C. sordellii group and consisted of four strains elaborating 7a-dehydroxylase alone and six strains capable of catalyzing both 7a-dehydrogenation and 7a-dehydroxylation. The remaining three organisms, recovered from fresh human feces, were gram-positive, nonflagellated, nonsporeforming, anaerobic rods and comprised two distinct species. Strain HD-17, still unidentified, had both activities, but was unique in that it exclusively 7a-dehydroxylated cholic acid while biotransforming chenodeoxycholic acid, preferably through 7a-dehydrogenation. Two unclassified strains, b-8 and c-25, metabolized both acids through 7a-dehydroxylation and 7adehydrogenation. Except for strains b-8 and c-25, all of the 7a-dehydroxylating bacteria split the conjugated bile acid series, and hydrolases were detected in cellfree filtrates of early stationary-phase broth cultures.

The removal of the 7a-hydroxy substituent on the steroid nucleus (7a-dehydroxylation) is a unique reaction confined to microbial action. The reductive conversion through this reaction of the primary bile acids, cholic acid (CA) and chenodeoxycholic acid (CDCA), is the most important of the bile acid transformations taking place in the intestinal tract of humans and other animals. It gives rise to the formation of deoxycholic acid (DCA) and lithocholic acid (LCA), respectively, both of which play a physiologically significant role after entering the enterohepatic circulation. The reaction is carried out extensively in the lower part of the bowel and virtually all of the bile acids excreted in the feces are products of this reaction (5). Available information concerning the microorganisms accounting for this phenomenon, however, is incomplete and conflicting. Although one group of investigators asserted that this property is widely distributed among the intestinal microorganisms (1, 2), it is generally accepted that it is difficult to demonstrate this activity in pure bacterial cultures (6, 24), and only a few strains of intestinal anaerobes have so far been shown to be capable of performing this reaction (3, 8, 11, 1316, 29). This study is aimed at evaluating the prevalence and variety of 7a-dehydroxylating organisms among the human intestinal microflora. When many bacterial strains isolated from human feces and municipal sewage were screened,

the enzymatic activity was found to be not as rare as had been anticipated. The isolation and the characterization of 13 dehydroxylating organisms thus obtained, 10 sporeforming and 3 nonsporing gram-positive anaerobic bacteria, are reported in this paper. MATERLALS AND METHODS Sources and isolation of microorganisms. Bacteria were isolated from freshly voided feces from healthy adults and from influx fluids at a municipal sewage disposal plant. A 1:10 suspension of feces in sterile phosphate buffer (0.02 M, pH 7.0, containing 0.1% sodium thioglycolate), or in buffered peptoneyeast extract broth (see below), or raw sewage was used as the inoculum after removal of coarse debris by centrifugation at low speed. In the first three experiments, the specimens were heated at 80 to 90°C for 20 min to destroy vegetative forms of bacteria, and 0.1-ml portions of appropriate dilutions were spread on GAM agar medium (Nissui Pharmaceutical Co.). In the fourth experiment, dilutions of a pool of fecal suspensions from three persons were plated on GAM agar medium supplemented with 150 yg of neomycin sulfate per ml to facilitate the isolation of Bacteroides spp. In the fifth experiment, two fecal samples were inoculated individually in bile acid broth (see below), loopfuls from which were streaked on plain GAM agar plates after 24 and 48 h of anaerobic incubation. In the sixth to eighth experiments, serial dilutions from each of three fecal samples were directly plated on GAM agar for a viable cell count. All seeded plates were incubated for 48 h in an

737

738

APPL. ENVIRON. MICROBIOL.

HIRANO ET AL.

anaerobic jar under a gas mixture of 90% N2 and 10% CO2 (purified previously by passing over heated copper gauze). Plates containing 10 to 100 colonies were examined, and representative colonies (virtually all colonies in experiments VI to VIII) were picked and stabbed into tubes of semisolid GAM agar medium; the tubes were then incubated aerobically for 3 days. After purification by reisolation of colonies, the cultures were tested for aerobic growth on the same GAM plate, and only obligately anaerobic organisms were preserved in semisolid GAM tubes for further examination. In experiments VI to VIII, all of the isolated cultures, including both facultative and obligate anaerobes, were studied. All manipulations except anaerobic incubation were carried out under atmospheric conditions, with precautions taken to minimize the exposure of samples to air. Transformation of bile acid by isolated bacterial strains. The basal medium used was a peptoneyeast extract broth prepared essentially according to the formula given in the Virginia Polytechnic Institute manual (19) but modified by using 2% peptone instead of 1% peptone and by dissolving the ingredients in 0.02 M phosphate buffer at pH 7.5 to maintain the pH above neutral during the entire period of incubation (2, 25). No carbohydrate was included, since the presence of a fermentable carbohydrate seemed to suppress the dehydroxylation (presumably due to acidification of the medium), although it stimulated bacterial growth. Bile acid, free or conjugated, was added to this medium to a final concentration of 200 ,uM from a stock solution of 10 mg of bile acid per ml of ethanol. The medium was then dispensed in 4-ml quantities in 12- by 105-mm test tubes (for incubation for a definite period of time) or in 30-ml amounts in 100-ml flasks (for successive sampling). The medium was prepared on the day of use and inoculated immediately after autoclaving and cooling. The inoculum was a 0.1-ml portion from an overnight culture of the test strain in peptone-yeast extract-glucose broth. The inoculated tubes were assayed for bile acids after 4 days of incubation in an anaerobic jar (under

an

atmosphere of

pure

N2 to

avoid acidification of the culture medium by C02). The medium in the flasks was inoculated under flushing with N2 through the liquid phase, and 4-ml samples were removed for assay at various time intervals during incubation under N2 for up to 7 days. After incubation, bacterial growth was assessed by measuring the transmittance of light at 660 nm on a photoelectric colorimeter, and the final pH of the culture medium was checked by a pH meter. Preparation for GLC. The spent culture medium was extracted three times with an equal volume of ethyl acetate after acidification to pH 2.0 with 6 N HCI. The combined solvent fractions were washed repeatedly with distilled water until the washing fluid became neutral in reaction. After evaporation to dryness under a stream of air at 45°C, the bile acid residue was reconstituted with 2 ml of methanol and 1 drop of concentrated H2SO4, and the solution was allowed to stand at room temperature overnight (methylation). After addition of 4 ml of distilled water and adjustment to pH 10 with saturated NaHCO3, the methylated bile acids were extracted thoroughly with ethyl acetate,

and the extract was washed and evaporated in the manner described above. The final residue was dissolved in 0.5 ml of chloroform, and 3-,ul portions were injected into a gas-liquid chromatography (GLC) column (3% QF-1). After analysis, the chloroform was evaporated and the methylated bile acid residue was exposed to silylating reagents to convert all free hydroxy groups on the steroid nucleus to trimethylsilyl ethers, as recommended by Makita and Wells (22). Three-microliter portions of the reaction mixture were directly injected into a 3% Hi Eff-8B column. Analysis of bile acids by GLC. A Hitachi 163 gas-liquid chromatograph equipped with a flame ionization detector was used isothermally. A U-shaped glass tube (3-mm inside diameter by 2 mj packed with 3% QF-1 coated on Chromosorb WAW DMCS, 80 to 100 mesh, was used for separation of methyl cholanoates, and a similar tube packed with 3% Hi Eff-8B coated on 80- to 100-mesh Gas-Chrom Q was used for analysis of the silyl ethers of methylated bile acids. The column oven was held at 260°C for QF-1 and at 230°C for Hi Eff-8B. The injector and detector were held at 50°C higher than the respective column temperatures. N2 was used as the carrier gas at a flow rate of 30 to 40 ml/min. H2 and air were at 1.1 and 1.3 kg/ cm 2 , respectively. Individual GLC peaks were identified by retention time relative to that of methyl deoxycholate or its silyl ether depending upon the derivation of the test sample (= relative retention time). A standard pool of authentic bile acids was analyzed in parallel, and the relative retention times of the sample and the standard acids were compared. For bile acids such as 3-keto- and 12ketocholanoates, for which we have no authentic comparatives, the published relative retention times (9, 10) were consulted for comparison. Quantitative estimation was performed by measuring the peak area with an electronic integrator (Shimadzu Chromatopac EIA) and relating it to the peak area of a known amount of the corresponding derivative of deoxycholate. Quantities of individual bile acids were expressed as percent composition, after it had been confirmed that the recovery of the total bile acids in each sample amounted to 80 to 90% of the substrate bile acid added. Blank tubes containing bile acid in sterile broth showed the same rate of recovery, and tubes inoculated without substrate showed no bile acid peaks. Combined gas chromatography-mass spectrometry. The silylated compounds separated on Hi Eff-8B were analyzed by mass spectrometry with a Hitachi M-60 gas chromatograph-mass spectrometer. The column was kept at 240°C, the injector was held at 280°C, and the separator was kept at 320°C. The helium inlet pressure was 1.2 kg/cm2. The ionizing voltage was kept at 20 eV, and the accelerating voltage was held at 3.2 kV. Recording was done from mle 0 to 700 in 4 s. Background spectra, mainly from the stationary phase, were recorded and subtracted from the sample spectra. Examination of bacteriological properties. Colonial morphology and the lecithinase reaction were observed on GAM agar plates supplemented with 10% egg yolk after anaerobic incubation for 2 days. Urease and catalase tests were applied to the bacterial growth

VOL. 41, 1981

INTESTINAL 7a-DEHYDROXYLATING BACTERIA

739

the same plate according to the techniques of from Gaschro-Kogyo Co. Sutter et al. (31). Other biochemical and fermentative RESULTS reactions were examined as described by Holdeman and Moore (19), using ordinary peptone-yeast extract Screening of intestinal isolates for bile as the basal medium. Final acidity of the carbohydrate broth was determined with a pH meter on day 3 of acid transformation. A total of 347 strains anaerobic incubation. Volatile and non-volatile fatty isolated by three procedures were examined for acids produced in 2-day cultures in peptone-yeast ex- their ability to transform CDCA in broth cultract or peptone-yeast extract-glucose were analyzed tures. The overall results are summarized in by GLC in an ethylene glycol adipate column under Table 1. The 7a-dehydroxylating conversion of temperature programming, as recommended by Hir- CDCA into LCA was shown by 10 of the 77 akawa (17). For electron microscopy, cells grown on presumed clostridial strains isolated from heatGAM agar plates were suspended in a small amount treated samples (experiments I to III) and 3 of of sterile distilled water. A drop of the suspension was the 240 isolates obtained by plating without mounted on a Formvar-carbon-coated grid. After negative staining with 2% potassium phosphotungstate, selective treatment (experiments V to VIII). the specimen was examined in a Hitachi H-300 elec- None of the 30 presumed bacteroides strains recovered from plates containing neomycin (extron microscope. Bile acids. Bile acids are referred to in the text by periment IV) showed the activity. The 7a-dethe following abbreviations for their trivial names: hydrogenation giving rise to the formation of CDCA, chenodeoxycholic acid (3a,7a-dihydroxy-5,f- 7KL was found in 19 (25%) of the clostridial, 13 cholanoic acid); CA, cholic acid (3a,7a,12a-trihydroxy- (43%) of the bacteroides, and 100 (42%) of the 5,B-cholanoic acid); DCA, deoxycholic acid (3a,12a-di- unselected isolates. Furthernore, 31 (40%) of the hydroxy-5,8-cholanoic acid); 7KD, 7-ketodeoxycho- clostridial strains and 8 (3%) of the unselected lic acid (3a,12a-dihydroxy-7-keto-5fl-cholanoic acid); 7KL, 7-ketolithocholic acid (3a-hydroxy-7-keto-5,8- isolates were found to epimerize the 3a-hydroxy cholanoic acid); and LCA, lithocholic acid (3a-hy- group of CDCA into a fl-configuration (3a-epimerization). droxy-5fl-cholanoic acid). Transformation of CDCA and CA by 7aThe free bile acids used as substrates were: CDCA (99.9% pure), a gift from Tokyo Tanabe Pharmaceu- dehydroxylating bacteria. The 13 strains tical Co.; CA (99.1% pure), purchased from Sigma found to dehydroxylate CDCA in the above test Chemical Co.; and 7KL (lot no. 322), obtained from were compared for their activities against CDCA Gaschro-Kogyo Co., all of which proved to be pure and CA, the two major components of human when tested by GLC. Glycine and taurine conjugates bile (Table 2). Clostridium sordellii 4709, a of bile acids (sodium salt, A grade) were obtained from Calbiochem. The conjugates were contaminated with stock culture originally obtained from Kanazawa small amounts of free bile acids; therefore, parallel University (N. Nishida), was also capable of 7acultures containing each conjugate (but uninoculated) dehydroxylating CDCA and CA and was included in the study for comparison. Table 2 were set up and served as controls. The methyl esters of bile acids used as GLC references were all obtained shows the percentages of substrate converted on

TABLE 1. Distribution of bile acid-transforrning activities among bacterial strains isolated from human feces and municipal sewage Isolation' Viable countb

Expt Source

I Feces

Treatment

Heat

Sewage Heat

No. of isolates producing:c No. of 7a-dehyDesignation ofstrains isolates LCA 7KL droxylating exam333a CDCA 7KL anld 3/s- CC and ind ined C CDCA 7KL 6 2 12 2 I-55, I-102 4 4 2 15 I-B6, II-27 4 3 3 18 N-9-2, N-9-4, N-9-10 9 1 2 15 HU-2, HU-6 8 1 3 S-il 17

Heat 8 x 104 Heat 5 X 104 Sewage Heat 13 30 IV Feces Neomycin 2 x 109 HD-17 24 8 59 1 None V Feces 20 57d 5 x 109 None VI Feces 1 23 b-8 2 x 1010 64d None VII Feces c-25 22 1 2 x 10'0 6od None VIII Feces a For bacterial isolation, see text. b Units per gram (wet weight) of feces (or 1 ml of sewage). c Four-day cultures originally containing CDCA were assayed by GLC for metabolic bile acids. d Preliminary screening of mixed cultures (five strains each); only strains from positive cultures were tested II Feces III Feces

individually.

740

APPL. ENVIRON. MICROBIOL.

HIRANO ET AL.

TABLE 2. Comparative activity of 7a-dehydroxylating bacteria against CDCA and CA % of CDCA converted to:b % of CA converted to:b Bacterial strain' LCA

I-55 I-102 I-B6 II-27

42,56, 15,32 36,53,21, 18 22,56,20,16 42,42,29,26

7KL

0,0,0,0 0,0,0,0 0,0,0,0 0,0,0,0

N-9-2 N-9-4 N-9-10 HU-2 HU-6 S-11 4709

1, 2, 1, 1 2,3, 1, 1 1, 2 3,1 2,1 2,8 2, 9, 3, 3

27, 30, 24, 38 24,30,27,27 26, 27 32, 32 32, 30 15,39 44, 60, 29,41

HD-17

2, 2, 3, 1

25, 44, 36, 52

DCA 29, 10, 7 8,6,6

13,7,5 9,12,4

12, 4 14, 3 4 5, 5 8, 6

6,9 13,17,10,17 100, 65, 76

7KD

0,0,0 0,0,0 0,0,0 0,0,0

18, 30 20, 29 24 25, 32 28, 31 14, 19 20, 18, 30, 24 0, 11, 6

b-8 6,8,4 5,6,21 13,11,20 7,8,20 c-25 2,10,4 17, 6, 20 11, 7,18 26, 21, 14 For sources of strains, see Table 1. Strain 4709 is a reference culture of C. sordellii. b Four-day cultures containing CDCA or CA were assayed for bile acids, and percentages of each substrate converted into LCA or DCA through 7a-dehydroxylation and into 7KL or 7KD by 7a-dehydrogenation were calculated. Values from repeated tests are presented.

into the respective major metabolites on day 4 and 7a-hydroxy and 7-keto acids were used inof anaerobic incubation, when the metabolites terchangeably, producing the same metabolites had reached practically their maximum concen- (Fig. 2). trations, as shown in Fig. 1, in which the changes Besides the principal metabolites produced in the composition of bile acids during the course through 7a-dehydroxylation and 7a-dehydroof incubation for up to 7 to 10 days are illus- genation, small amounts of a 3-keto derivative trated. The values in Table 2 represent the of CDCA were frequently produced by certain results of tests successively performed during strains (Fig. 1). It can be seen in Fig. 1 that the in vitro transfers of the organisms covering strain HD-17, which extensively dehydroxylates a period of >1 year. CA, is also capable of metabolizing the resulting All of the strains consistently dehydroxylated DCA into 12-keto and 3-keto derivatives. both CDCA and CA, yielding LCA and DCA, These bile acid-transforming activities have respectively (Table 2). The 7a-dehydroxylation been maintained almost unchanged from the took place early in the incubation and ceased time of isolation onwards. after 24 h, when bacterial growth had reached a Identification of 7a-dehydroxylation maximum level. The concentration of LCA or products. In the preceding experiments, bile DCA remained undiminished upon prolonged acid metabolites were separated as methyl esters incubation, suggesting an end product nature on QF-1 and tentatively identified on the basis (Fig. 1). The relative activity against both 7a- of relative retention times. To confirm their hydroxy bile acids varied considerably with dif- identity, the methylated samples were rechroferent bacterial strains: the 1-55 group (I-55, matographed after trimethylsilylation, and the I-102, I-B6, and II-27) constantly dehydroxy- separated peaks were analyzed by mass speclated CDCA to a greater extent than CA, trometry. The suspected 7a-dehydroxylation whereas HD-17 was unique for an extensive products, LCA and DCA, showed the same GLC dehydroxylation of CA and restricted activity mobilities and mass spectra as those of the same against CDCA (Table 2; Fig. 1). derivatives of the corresponding authentic acids Most of the dehydroxylating organisms cata- simultaneously processed for comparison. The lyzed 7a-dehydrogenation also, giving rise to the mass spectra were also compatible with the specformation of a 7-keto acid, 7KL or 7KD, with tra published by Sjovall et al. (27) for known the exception of the I-55 group, which showed references. no indication of the presence of an oxidative Deconjugation by 7a-dehydroxylating process at C-7 (Table 2). In contrast to dehy- bacteria. Except for b-8 and c-25, the 7a-dehydroxylation, dehydrogenation was reversible, droxylating strains were capable of deconjugat-

CDCA 1001

CA 100'

I -102

................1-102

C

502

DCA 4

1

100'.

C

7

4709 CA

500

,^______

P.-

z

---

2

rA z

12 3

100-

100,-

c-25

0

7KD_

DCA

i

7

C-25

0

CD0CA ^ 50

50-

/-

-^~~~ r ~S~~~ 1

/3k7a

[

I

le

4

5S.

--

-

-

CP-

- :*

7

1

4

7

INCUBATION (DAYS) FIG. 1. Changes in percent composition of metabolites during the course of incubation. The test strains (I-102, 4709, c-25, and HD-1 7) were grown in broth containing CDCA or CA for 7 to 10 days, and samples were removed for assay at the indicated times. The percent composition of individual bile acids was calculated by GLC of the methylated samples on QF-1. 3k7a, 3-Keto derivative of CDCA; 3k12a and 3al2k, 3-keto and 12keto derivatives of DCA. 741

742

APPL. ENVIRON. MICROBIOL.

HIRANO ET AL.

100

4709

100 [

z

CDCA

0

Z o

HU-2

CDCA

I

50

50

7KL CA

1

2

3

5

LCA

-

7

1

2

3

5

7

INCUBATION (DAYS) FIG. 2. Changes in percent composition of metabolites formed from 7KL by strains 4709 and HU-2 during 7 days of anaerobic incubation. Cultural and analytical procedures are the same as for Fig. 1.

ing the series of glycine and taurine conjugates of bile acids in growing cultures. Strong hydrolytic activities were also detected in cell-free supernatant fractions from early stationaryphase cultures in GAM broth containing no bile salt, indicating that the hydrolases were constitutively formed in cultures and excreted extracellularly in an early phase of bacterial growth. Bacteriological characterization of 7adehydroxylating bacteria. The main characteristics of 7a-dehydroxylating bacteria are summarized in Table 3. All of the organisms are gram-positive, obligately anaerobic rods; the 10 organisms isolated from heated specimens formed abundant spores in in vitro cultures and were therefore assigned to the Clostridium genus. These clostridial strains resemble each other in the bacteriological features tested so far and are very similar, if not identical, to the C. bifermentans-C. sordellii group, with the possible exception of strain S-11, which should be placed in a separate species because of its failure to liquefy gelatin. The remaining three organisms, isolated from fresh feces, are nonflagellated and nonsporing and comprise two distinct species. Strain HD-17 assumed large curved forms, 0.8 by 1.5 to 4.0 pm (Fig. 3), and fermented glucose and other carbohydrates. Abundant growth was attained in GAM broth, and colonies on GAM plates were 5 mm or more in diameter, irregular, umbilicate, glistening, and gray. On the other hand, strains b-8 and c-25, which are identical in all criteria examined, are much more slender than HD-17, measuring 0.5 by 1.0 to 1.5 ,um, and occur in chains of two or three elements with a conspicuous cellular appearance similar to that presented for Eubacterium len-

tum by Bokkenheuser et al. (4) (Fig. 3). The organisms lacked any activity against carbohy-

drates and proteins, and growth in broth was poor and was not improved by the addition of 2% (wt/vol) arginine. Colonies on GAM agar plates were less than 1 mm in diameter, circular, entire, convex, and translucent. DISCUSSION Distribution of 7a-dehydroxylase activity among intestinal bacteria. In this study, three categories of human intestinal bacteria were screened for bile acid-transforming activity: 77 heat-resistant presumed clostridial strains, 30 neomycin-tolerant presumed bacteroides strains, and 240 unselected strains. 7a-Dehydroxylase activity was demonstrated most frequently in the first group. Of the 77 clostridial strains, 10 were found to be capable of catalyzing this reaction, restricted to certain species related, if not identical, to C. bifermentans-C. sordellii. Some of these active strains were also recovered from municipal sewage. The results agree with those of Hayakawa and Hattori (16), Hayakawa (15), and Ferrari and colleagues (11, 12), who reported that certain strains of C. sordellii and C. bifermentans possess this activity. It is certain that 7a-dehydroxylation is not a rare property among specified species of Clostridium, but this cannot explain the extensive occurrence of this reaction in vivo since it does not seem likely that clostridia of this kind would be indigenous to the human gut in any appreciable amount. It may be only an indication of the diversity of metabolic activity in this genus. Not a single strain of C. perfringens, which is indigenous to the normal intestine in significant num-

VOL. 41, 1981

INTESTINAL 7a-DEHYDROXYLATING BACTERIA

743

TABLE 3. Bacteriological characteristics of 7a-dehydroxylating strains Strains

Characteristic

I-55, I-B6, I-102, 11-27 +, rod + +

N-9-2 N-9-4 N-9-10 'HU-2, HU-6 +, rod +, rod + + + + + + + + + +

4709

5-11

HD-17

b-85 b-85

Gram stain and form +, rod +, rod +, rod +, rod Flagella + + Spores + + Indole production + + Gelatin liquefaction + + Glucose, acid + + + + Lactose, acid Sucrose, acid + Fructose, acid + + + + + Maltose, acid + + + + + Mannose, acid + + _ + Esculin hydrolyzed + + + + + Starch hydrolyzed Lecithinase + + + + + Catalase Urease + + + Fatty acids' in: PY a,p,b,ib,iv,ic a,p,ib,iv,ic a,p,ib,iv,ic a,p,ib,iv,ic a,p,ib,iv a,p,b,iv a PYG a, ic a, ic a, ic a, ic a, iv a, ib, iv a a Fatty acids produced by 2-day cultures: a, acetic; p, propionic; b, butyric; ib, isobutyric; iv, isovaleric; ic, isocaproic. PY, Peptone-yeast extract; PYG, peptone-yeast extract-glucose.

I

C.-

FIG 3 Electron micrographs of representative 7a-dehydroxylating bacteri. (A) 155 (B) HD 17 (C) HU6; (D) c 25. Bar 1 ,km.

bers and accounted for the majority of the clostridial isolates studied, showed this dehydroxylation activity.

The second group, which consisted of presumed bacteroides, included no dehydroxylatingr organisms. Similar negative results were ob-

744

HIRANO ET AL.

tained in our previous experiment with 64 taxonomically defined Bacteroides fragilis strains from human intestinal contents (18). Stellwag and Hylemon (29) were also unable to find any active strains among 10 reference and 70 laboratory strains of intestinal Bacteroides. Although several organisms ascribed to the Bacteroides genus have been reported to be capable of 7a-dehydroxylating cholic acid (3, 8, 14), whether this predominant intestinal species plays any primary role in in vivo dehydroxylation seems open to question. An isolate from human feces, which was previously designated Bacteroides 28S (14), has been reclassified as C. leptum (29, 30). From the third group, consisting of nonselectively isolated strains, three dehydroxylating organisms were obtained: strain HD-17 from the 59 anaerobes isolated indirectly from broth cultures of fresh feces and strains b-8 and c-25 from the 181 colonies isolated on anaerobic plates for viable counts. All three of these organisms are gram-positive, nonsporeforming, anaerobic rods. Of the organisms about which data have been published, the lactobacilli isolated by Gustafsson et al. (13) and described in detail by Midtvedt and Norman (23, 25) are the sole bacteria belonging to this category. As for bile acid-transforming activity, however, the lactobacilli failed to split conjugated bile acids, in contrast to the strong hydrolysis by HD-17, and the former bacilli dehydroxylated CA and CDCA to a similar extent whereas the 7a-dehydroxylation by HD-17 showed an obvious preference for CA over CDCA. Both equally elaborated 3a-, 7a-, and 12a-dehydrogenases in addition to 7a-dehydroxylase. Strains b-8 and c-25 share some negative characteristics with E. lentum bacteria (26), but the cultures failed to respond to arginine (28), and their accurate taxonomy, particularly the relation to E. lentum, remains to be determined. Concerning the metabolism of bile acids by Eubacterium, Midtvedt and Norman (24) reported 3a- and 7a-dehydrogenation by three strains of Eubacterium species, including one E. lentum strain, whereas Macdonald et al. (20, 21) demonstrated 3a- and 12a-dehydrogenases but no 7a-dehydrogenation in many strains of E. lentum. Bokkenheuser et al. (4) revealed the presence of a corticoid 21-dehydroxylase in E. lentum, but no 7a-dehydroxylation of bile acids has ever been reported with bacteria belonging to this genus. Strains b-8 and c-25 in this study apparently possess not only 3a- and 7a-dehydrogenases, but also 7a-dehydroxylase activity. Comparative activity against CA and CDCA. The 7a-dehydroxylating organisms in this study consistently dehydroxylated both CA

APPL. ENVIRON. MICROBIOL.

and CDCA, and most of these strains metabolized the two acids to similar extents, as reported by Midtvedt and Norman (25) for culture of Lactobacillus strain II and by Aries and Hill (2) with cell-free preparations of various bacterial species. In this regard, strain HD-17 was exceptional in that it dehydroxylated CA almost exclusively, with little action on CDCA. A similar result was also reported by Ferrari and Beretta (12) with a cell-free extract from C. bifermentans. Dehydroxylation and dehydrogenation of the 7a-hydroxy group. Except for four strains of the 1-55 group, the 7a-dehydroxylating organisms, including nonsporeforming ones, were also capable of oxidizing the same hydroxy group in CA and CDCA, like all the previously reported bacteria (2, 3, 12, 16, 25). C. leptum, although reported by Stellwag and Hylemon (29) to lack the oxidative process, was capable of converting CA into 7KD at the time of isolation (14). It should be noted in this connection that 7a-dehydrogenation, in contrast to reductive dehydroxylation, is an oxidative process. Consequently, the relative rates of these two reactions may vary according to the degree of anaerobiosis of the environment. The anaerobic conditions in this study, in which oxidative keto formation was invariably demonstrated, occasionally even to a greater extent than dehydroxylation, do not seem to have reproduced the strictly anaerobic circumstances in the colon, where virtually all the bile acid metabolites are in the reduced form. Drasar and Hill (7) have emphasized that there should be little keto formation under ideal conditions for 7a-dehydroxylation. In this regard, the I-55 group is unique and unprecedented. The strains have never demonstrated any detectable 7-keto formation since the time of isolation. ACKNOWLEDGMENT We thank N. Akimori, Naka Works, Hitachi Ltd., Katsuta, Japan, for excellent assistance with the mass spectrometry. LITERATURE CITED 1. Aries, V., J. S. Crowther, B. S. Drasar, and M. J. Hill. 1969. Degradation of bile salts by human intestinal bacteria. Gut 10:575-576. 2. Aries, V., and M. J. Hill. 1970. Degradation of steroids by intestinal bacteria. II. Enzymes catalyzing the oxidoreduction of the 3a-, 7a- and 12a-hydroxyl groups in cholic acid, and the dehydroxylation of the 7a-hydroxyl group. Biochim. Biophys. Acta 202:535-543. 3. Bokkenheuser, V., T. Hoshita, and E. H. Mosbach. 1969. Bacterial 7-dehydroxylation of cholic acid and allocholic acid. J. Lipid Res. 10:421-426. 4. Bokkenheuser, V. D., J. Winter, S. M. Finegold, V. L. Sutter, A. E. Ritchie, M. E. C. Moore, and L. V. Holdeman. 1979. New markers for Eubacterium lentum. Appl. Environ. Microbiol. 37:1001-1006. 5. Carey, J. B., Jr. 1973. Bile salt metabolism in man, p.

VOL. 41, 1981

INTESTINAL 7a-DEHYDROXYLATING BACTERIA

62-63. In P. P. Nair and D. Kritchevsky (ed.), The bile acids, vol. 2. Plenum Publishing Corp., New York. 6. Dickinson, A. B., B. E. Gustafsson, and A. Norman. 1971. Determination of bile acid conversion potencies of intestinal bacteria by screening in vitro and subsequent establishment in germfree rats. Acta Pathol. Microbiol. Scand. Sect. B 79:691-698. 7. Drasar, B. S., and M. J. Hill. 1974. Human intestinal flora, p. 118. Academic Press, Inc., London. 8. Edenharder, R., and J. Slemrova. 1976. Die Bedeutung des bakteriellen Steroidabbau fur die Atiologie des Dickdarmkrebses. IV. Spaltung von Glykocholsaure, Oxydation und Reduktion von Cholsaure durch saccharolytische Bacteroides-Arten. Zentralbl. Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt. 1 Orig. Reihe B 162:350-373. 9. Elliott, W. H., L. B. Walsh, M. M. Mui, M. A. Thorne, and C. M. Siegfried. 1969. Bile acids. XXVIII. Gas chromatography of new bile acids and their derivatives. J. Chromatogr. 44:452-464. 10. Eneroth, P., and J. Sjovall. 1969. Methods of analysis in the biochemistry of bile acids. Methods Enzymol. 15: 237-280. 11. Ferrari, A., and F. Aragozzini. 1972. Attivita di un ceppo di Clostridium bifermentans su alcuni acidi biliari. Ann. Microbiol. (Milan) 22:131-136. 12. Ferrari, A., and L. Beretta. 1977. Activity on bile acids of a Clostridium bifermentans cell-free extract. FEBS Lett. 75:163-165. 13. Gustafsson, B. E., T. Midtvedt, and A. Norman. 1966. Isolated fecal microorganisms capable of 7a-dehydroxylating bile acids. J. Exp. Med. 123:413-432. 14. Hattori, T., and S. Hayakawa. 1969. Isolation and characterization of a bacterium capable of 7a-dehydroxylating cholic acid from human faeces. Microbios 3:287294. 15. Hayakawa, S. 1973. Microbiological transformations of bile acids. Adv. Lipid Res. 11:143-192. 16. Hayakawa, S., and T. Hattori. 1970. 7a-Dehydroxylation of cholic acid by Clostridium bifermentans strain ATCC 9714 and Clostridium sordellii strain NCIB 6929. FEBS Lett. 6:131-133. 17. Hirakawa, K. 1979. Resting cell studies on fermentation of sugar and peptone by clostridia. Acta Med. Univ. Kagoshima. 21:117-130. 18. Hirano, S., N. Masuda, H. Mukai, K. Hirakawa, and T. Imamura. 1979. Transformation of bile acids by Bacteroides fragilis strains isolated from the human intestine. Jpn. J. Bacteriol. 34:403-411. (In Japanese.)

745

19. Holdeman, L. V., and W. E. C. Moore. 1973. Anaerobe laboratory manual, 2nd ed., p. 108-110. Virginia Polytechnic Institute and State University, Blacksburg. 20. Macdonald, I. A., J. F. Jellett, D. E. Mahony, and L. V. Holdeman. 1979. Bile salt 3a- and 12a-hydroxysteroid dehydrogenases from Eubacterium lentum and related organisms. Appl. Environ. Microbiol. 37:9921000. 21. Macdonald, I. A., D. E. Mahony, J. F. Jellett, and C. E. Meier. 1977. NAD-dependent 3a- and 12a-hydroxysteroid dehydrogenase activities from Eubacterium lentum ATCC No. 25559. Biochim. Biophys. Acta 489: 466-476. 22. Makita, M., and W. W. Wells. 1963. Quantitative analysis of fecal bile acids by gas-liquid chromatography. Anal. Biochem. 5:523-530. 23. Midtvedt, T. 1967. Properties of anaerobic gram-positive rods capable of 7a-dehydroxylating bile acids. Acta Pathol. Microbiol. Scand. 71:147-160. 24. Midtvedt, T., and A. Norman. 1967. Bile acid transformations by microbial strains belonging to genera found in intestinal contents. Acta Pathol. Microbiol. Scand. 71:629-638. 25. Midtvedt, T., and A. Norman. 1968. Parameters in 7adehydroxylation of bile acids by anaerobic lactobacilli. Acta Pathol. Microbiol. Scand. 72:313-329. 26. Moore, W. E. C., E. P. Cato, and L. V. Holdeman. 1971. Eubacterium lentum (Eggerth) Prevot 1938: emendation of description and designation of the neotype strain. Int. J. Syst. Bacteriol. 21:299-303. 27. Sjovall, J., P. Eneroth, and R. Ryhage. 1971. Mass spectra of bile acids, p. 209-248. In P. P. Nair and D. Kritchevsky (ed.), The bile acids, vol. 1. Plenum Publishing Corp., New York. 28. Sperry, J. F., and T. D. Wilkins. 1976. Arginine, a growth-limiting factor for Eubacterium lentum. J. Bacteriol. 127:780-784. 29. Stellwag, E. J., and P. B. Hylemon. 1978. Characterization of 7a-dehydroxylase in Clostridium leptum. Am. J. Clin. Nutr. 31:243-247. 30. Stellwag, E. J., and P. B. Hylemon. 1979. 7a-Dehydroxylation of cholic acid and chenodeoxycholic acid by Clostridium leptum. J. Lipid Res. 20:325-333. 31. Sutter, V. L., V. L. Vargo, and S. M. Finegold. 1975. Wadsworth anaerobic bacteriology manual, 2nd ed., p. 85-89. Anaerobic Bacteriology Laboratory, Wadsworth Hospital Center, Veterans Administration, and Department of Medicine, UCLA School of Medicine, Los Angeles.