Vol. 40, No.6APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Dec. 1980, p. 1017-10220099-2240/80/12-1017/06$02.00/0

Biochemical Characterization of Cholesterol-ReducingEubacterium

GLEN E. MOTT,* ALLEN W. BRINKLEY, AND CYNTHIA L. MERSINGERDepartment ofPathology, The University of Texas Health Science Center, San Antonio, Texas 78284

We characterized two isolates of cholesterol-reducing Eubacterium by con-ducting conventional biochemical tests and by testing various sterols and glycer-olipids as potential growth factors. In media containing cholesterol and plasmen-ylethanolamine, the tests for nitrate reduction, indole production, and gelatin andstarch hydrolyses were negative, and no acid was produced from any of 22carbohydrates. Both isolates hydrolyzed esculin to esculetin, indicating ,B-glycos-idase activity. In addition to plasmenylethanolamine, five other lipids whichcontain an alkenyl ether residue supported growth of Eubacterium strain 403 ina lecithin-cholesterol base medium. Of six steroids tested, cholesterol, cholest-4-en-3-one, cholest-4-en-3,8-ol (allocholesterol), and androst-5-en-3,B-ol-17-one sup-ported growth of Eubacterium strain 403. All four steroids were reduced to the3,8-ol, 5fl-H products. The A5 steroids cholest-5-en-3a-ol (epicholesterol) and22,23-bisnor-5-cholenic acid-3,B-ol were not reduced and did not support growthof the Eubacterium strain.

Organisms which reduce cholesterol to co-prostanol have been isolated from the rat cecum(4) and from human (12) and baboon (10) feces.These small, strictly anaerobic, gram-positivebacilli have been tentatively assigned to thegenus Eubacterium, but few of their metaboliccharacteristics are known. In earlier work, theorganisms grew only in a complex medium con-taining brain, and therefore assessment of bio-chemical characteristics was difficult. In addi-tion to cholesterol, we have identified plasmen-ylethanolamine (PLE) (10) as the principalgrowth factor supplied by brain medium, and wehave developed a simplified medium which canbe used for classical biochemical tests and forevaluation of other growth factors. This reportdescribes the results of a variety of biochemicaltests with two cholesterol-reducing isolates ofthe genus Eubacterium, ATCC 21408 and strain403, an organism that we isolated from baboonfeces. We also have evaluated several glycerylether lipids and steroids for their capacity tosupport growth of Eubacterium strain 403.

MATERIALS AND METHODSOrganisms and culture conditions. Two strains

of Eubacterium, ATCC 21408 (4) and strain 403 (10),were characterized in this study. Both organisms re-duce cholesterol to coprostanol in vitro. The cultureswere maintained and all experimental studies wereperformed in an anaerobic chamber as described pre-viously (10). Standard brain medium (10) was used forroutine maintenance of these bacteria and as a positivecontrol medium. Since the medium was opaque withsuspended lipids, growth could not be assessed by

turbidity. Therefore, we relied on a clotted appearanceand a Gram stain for evidence of growth.

Biochemical characterization. The base mediumused for biochemical characterization contained thefollowing components (in grams per liter): Casitone(Difco Laboratories, Detroit, Mich.), 10; yeast extract,10; lecithin (type II-S; Sigma Chemical Co., St. Louis,Mo.), 1.0; sodium thioglycolate, 0.5; PLE (Supelco,Inc., Beliefonte, Pa.), 0.25; and cholesterol, 2; the pHwas 7.2. This formula approximated that of PYG (8)plus cholesterol, lecithin, and PLE, and it was pre-pared in the same manner as described previously forlecithin-cholesterol medium (LCM) (10). PLE was 40to 70% pure in various lots, as determined by theiodine uptake method of Gottfried and Rapport (7).Each batch of PLE was evaporated to dryness underN2, and the residue was dissolved in carbon disulfide(redistilled) and stored at -20'C, which minimizeddecomposition.The two Eubacterium isolates were tested for re-

duction of nitrate, production of indole, hydrolysis ofstarch, gelatin, and esculin, and fermentation of amyg-dalin, arabinose, cellobiose, erythritol, fructose, glu-cose, glycogen, inositol, lactose, maltose, mannitol,mannose, melezitose, melibiose, raffinose, rhamnose,ribose, salicin, sorbitol, sucrose, trehalose, and xylose.We added sterile solutions of substrate to sterile basemedium in the concentrations recommended (8). Thetest media were inoculated with 0.05 ml of cultures ofthe two test organisms grown in base medium. Allcultures were incubated at 35°C until the mediumclotted, indicating growth. The final pH of each me-dium containing carbohydrate was determined andcompared with the pH of uninoculated controls andcultures in base medium without carbohydrate. Ni-trate reduction, indole production, and starch, gelatin,and esculin hydrolyses were tested as indicated byHoldeman and Moore (8).

D17

on March 11, 2020 by guest

http://aem.asm

.org/D

ownloaded from

1018 MOTT, BRINKLEY, AND MERSINGER

The short-chain volatile and nonvolatile carboxylicacids from cultures containing glucose were analyzedby gas-liquid chromatography (GLC) on a ShimadzuGC-3BF gas chromatograph (American InstrumentCo., Inc. [Trevenol Laboratories, Inc.], Silver Spring,Md.). The extraction and methylation procedureswere carried out as described by Holdeman and Moore(8). Two microliters of each sample was injected ontoa glass column (3-mm inside diameter by 2 m) packedwith 10% SP-1000/1% H3PO4 on 100/120 ChromosorbW AW (Supelco, Inc.).We repeated the test for ,B-glycosidase activity of

Eubacterium strain 403 in 2 ml of base medium con-taining 5 mg of esculin per ml to allow isolation anddefinitive identification ofthe product(s) of hydrolysis.The culture was incubated at 350C for 7 days and thenextracted with chloroform-methanol (2:1) (5). Twoportions of the lipid extract were spotted on a 250-,um-thick silica gel thin-layer chromatography (TLC) platewhich was developed in chloroform-methanol-water(65:25:4). One side of the plate containing a portion ofthe sample was sprayed heavily with 1% aqueous ferricammonium citrate solution, which produces a gray-black color with esculetin, the hydrolysis product ofesculin. The band from the corresponding unsprayedarea of the TLC plate was scraped and eluted withchloroform-methanol-water (86:14:1) through a frit-ted-glass filter. The solvents were removed by flashevaporation, and the residue was silylated with 0.1 mlofTRI-SIL/BSA (Pierce Chemical Co., Rockford, Ill.)for 45 min at 70°C. The sample was analyzed by GLCon a 1.2-m glass column packed with 3% OV-17 (Ap-plied Science Laboratories Inc., State College, Pa.) at170°C. The principal GLC peak was identified byGLC-mass spectrometry with a Hewlett-Packard5980A system (Hewlett-Packard, Avondale, Pa.).We assayed for arginine dihydrolase activity in sep-

arate cultures of the two Eubacterium isolates grownfor 14 days in 2 ml of base medium containing 1% L-arginine. The clotted cultures were centrifuged at 1,000x g for 15 min, and 2 pl of the supernatant was spottedon a precoated cellulose TLC plate (0.1-mm celluloseMN 300; Brinkmann Instruments Inc., Westbury,N.Y.) as recommended by Zolg and Ottow (17). Stand-ards (2 pl of 0.01 M aqueous solutions) of arginine,citrulline, ornithine, agmatine, and putrescine werealso applied to the TLC plate. The plate was developedin n-butanol-acetone-40% methylamine-water (10:10:2:5), air dried, sprayed with 1% ninhydrin (in metha-nol), and heated at 90°C for 1 min.

Preparation of ether lipids. We used the methodof Snyder et al. (14) for the preparation of 1-monoal-kenylglycerol from 5.1 mg of PLE by hydrogenolysiswith Vitride T (70% in toluene; Eastman OrganicChemicals, Rochester, N.Y.). Products were extractedwith ethyl acetate and fractionated by TLC on SilicaGel G plates (Merck & Co., Inc., Rahway, N.J.) indiethyl ether-30% ammonium hydroxide (100:0.25).The band which migrated with approximately thesame Rf as 1-palmitoylglycerol (Supelco, Inc.) wasscraped and eluted with chloroform-methanol (2:1)through a fritted-glass filter. The presence of the al-kenyl ether residue was shown by iodine uptake (7).By analysis of the trimethylsilyl ethers by GLC-massspectrometry, we identified a mixture of 1-monoalken-

ylglycerols. The mixture contained approximately 40%C16-0 and 60% C180 alkenyl ether residues.The preparation of 2-acyl-1-alkenylglycerol was ac-

complished by hydrolysis of PLE with phospholipaseC from Clostridium perfringens (P-L Biochemicals,Inc., Milwaukee, Wis.). Approximately 3 mg of PLEwas sonicated in 0.1 M phosphate buffer, pH 7.0.' Thereaction mixture, containing 0.2 ml of 0.04 M calciumchloride, 2 to 4 U of phospholipase C, and the PLEsuspension, was incubated overnight at room temper-ature with gentle mixing. The hydrolysate was ex-tracted by the method of Folch et al. (5), and the lowerlipid phase was chromatographed on silica gel TLCplates in chloroform-methanol (96:4). The band thatmigrated with the same Rf (0.65) as 1,2-dipalmitoylgly-cerol (Applied Science Laboratories Inc.) was elutedwith chloroform-methanol (2:1). Iodine uptake (7) bythis band indicated that the alkenyl ether residue wasintact.

Plasmenic acid (2-acyl-1-alkenyl-sn-glycerol-3-phosphate) was prepared from PLE by hydrolysis withphospholipase D from cabbage (Boehringer-Mann-heim Corp., Indianapolis, Ind.) with the followingmodifications of the procedure of Dawson and Hem-ington (2). A suspension of PLE was prepared bysonication of 3.5 mg of PLE in a solution containing0.3 ml of 0.05 M acetate buffer (pH 5.6) and 0.5 ml of18 mM sodium dodecyl sulfate at 0°C for 10 min. Thereaction was initiated by the addition of 0.3 ml of 120mM calcium chloride and 2 U of phospholipase D.The suspension was incubated for 7 h at 37°C withmixing. The reaction was stopped by extracting themixture with chloroform-methanol (2:1) (5). The ex-tract was fractionated by TLC on silica gel in chloro-form-methanol-7 N ammonium hydroxide (60:35:5).The band that migrated with the same Rf (0.12) asphosphatidic acid (Supelco, Inc.) was eluted from thesilica gel with chloroform-methanol (2:1).The 2-lysoplasmenylethanolamine was prepared by

mild alkaline hydrolysis of PLE with the procedure ofDawson et al. (3) and was purified by TLC in chloro-form-methanol-water (65:25:4). The band correspond-ing to a standard of lysophosphatidylethanolaminewas eluted with chloroform-methanol (2:1).We obtained plasmenylcholine from P-L Biochem-

icals, Inc. Its purity was 42%, as determined by iodineuptake. Impurities observed by TLC were predomi-nately aldehyde and lysoplasmenylcholine.

Lysoplasmanylcholine, 1-alkyl-sn-glycero-3-phos-phorylcholine, supplied by Donald J. Hanahan, Uni-versity of Texas Health Science Center, San Antonio,Tex., was exposed to HCI fumes for 10 min to assurethe absence of alkenyl ether groups. The fumes wereremoved under N2, and the residue was dissolved inethanol.

Testing of ether lipids as possible growth fac-tors. The lipids were tested for their ability to supportgrowth of Eubacterium strain 403 in modified LCMwhich contained the following components, per liter:Casamino Acids, 10 g; yeast extract, 10 g; lecithin, 1 g;cholesterol, 2 g; K2HPO4, 5 g; sodium thioglycolate, 0.5g; and resazurin, 1.0 mg. The medium was prepared aspreviously described (10). Each lipid was dissolved inchloroform-methanol (2:1) and added at a concentra-tion of approximately 1 mg per ml of modified LCM.

APPL. ENVIRON. MICROBIOL.

on March 11, 2020 by guest

http://aem.asm

.org/D

ownloaded from

CHOLESTEROL-REDUCING EUBACTERIUM 1019

The medium was lyophilized, autoclaved, and inocu-lated with 0.1 ml of a 2 x 10' dilution of a starterculture. The cultures were incubated for 12 to 14 daysat 37°C and then analyzed for coprostanol by GLC forevidence of growth as previously described (10).

Preparation of cultures containing test ste-roids. Five steroids were tested for their capacity tosupport growth of Eubacterium strain 403 in the pres-

ence and absence of cholesterol. Epicholesterol (cho-lest-5-en-3a-ol), cholest-4-en-3-one, allocholesterol(cholest-4-en-3,B-ol), androst-5-en-3,B-ol-17-one, and22,23-bisnor-5-cholenic acid-3,1-ol were obtained fromSteraloids, Wilton, N.H. Bisnorcholenic acid-3,B-ol wasdissolved in dimethyl formamide with heating, and 4mg was added to each of two tubes. The other steroidswere dissolved in chloroform, and 4 mg in 0.5 ml ofchloroform was added to each of the duplicate culturetubes. Two milliliters of modified LCM (see above)was added to one set of the tubes containing the teststerols. Each of these tubes contained 2 mg of choles-terol and 2 mg of the test sterol per ml of medium. A2-ml amount of the same medium without cholesterolwas added to the other set of tubes containing 2 mg ofthe test sterols. All tubes contained 0.5 mg of PLE per

ml of medium. The components were mixed thor-oughly and lyophilized to remove the solvents. Theresidues were suspended in 2 ml of distilled water,sterilized by autoclaving, and reduced in an anaerobicchamber as described previously (10). The inoculumwas a 0.1-ml sample of a 10-' dilution of a 3-dayculture of Eubacterium strain 403 grown in brainmedium (10). The cultures were incubated anaerobi-cally for 14 days at 35°C. Since the tubes with 2 mg ofcholest-4-en-3-one per ml did not appear to be growingat 7 days, we also inoculated tubes containing 0.2 mgof cholest-4-en-3-one per ml of medium with and with-out cholesterol.

Extraction, fractionation, and identification ofsteroid reaction products. The cultures, except forthose containing androst-5-en-3,B-ol-17-one, were sa-

ponified in 5 ml of 1.25 N KOH in ethanol at 55°C for1 h. A total of 3 ml of water was added, and the neutralsteroids were extracted with three 10-ml portions ofpetroleum ether. The petroleum ether extracts were

evaporated to dryness on a rotary evaporator anddissolved in 5 ml of chloroform.

After the petroleum ether extraction, the aqueouslayers of only the cultures containing bisnorcholenicacid-3,8-ol were acidified to a pH of 1 with 12 N HCIand extracted with three 10-ml portions of diethylether to remove the acidic lipids, including the bisnor-cholenic acid-3ft-ol. The diethyl ether extracts were

evaporated to dryness and dissolved in 5 ml of meth-anol, and 300-g1 portions of these extracts were evap-orated to dryness in screw-capped tubes (13 by 100mm). The residues were methylated with a solutioncontaining 0.9 ml of ethereal diazomethane and 0.1 mlof absolute methanol, and the solution remained atroom temperature for 10 min. The solvents were re-

moved under N2.The methylated residues of the bisnorcholenic acid-

318-ol cultures and 300-pl samples of each of the petro-leum ether extracts of the saponified cultures contain-ing bisnorcholenic acid-3,B-ol or cholest-4-en-3-onewith and without cholesterol were evaporated to dry-

ness, and trimethylsilyl derivatives were producedwith 100 pl of TRI-SIL/BSA by heating at 65°C for 45min. The steroid derivatives were separated by GLCon a 1.5-m glass column packed with 3% OV-17 (Ap-plied Science Laboratories Inc.) at 250°C.

Portions of the petroleum ether extract from cul-tures which contained epicholesterol and allocholes-terol were fractionated on a model 201 high-perform-ance liquid chromatograph (Waters and Associates,Milford, Mass.) equipped with a Radial Pak B silicacolumn (Waters and Associates). The samples wereeluted with isooctane-isopropanol (995:5) at a flowrate of 2.0 mil/min. The substrates and products weredetected with a refractive index detector.The two cultures containing androst-5-en-3,B-ol-17-

one were extracted by the procedure of Folch et al.(5). The extracts were evaporated to dryness andresuspended in chloroform-methanol (2:1). Samples(300-j1 portions) of each extract were applied to eachhalf of prescored silica gel 60 TLC plates (20 by 20 cm;Brnkmann Instruments Inc.). Standards of choles-terol, coprostanol, 5fi-androstan-3,f (and 3a)-ol-17-one, and androst-5-en-3,B-ol-17-one were also spotted.The plate was developed in isooctane-ethyl acetate-acetic acid (15:75:0.6) and separated along the pre-scored lines, and the half of the plate with the stand-ards and one set of sample extracts was sprayed withanisaldehyde reagent (16) and heated at 1000C for 10to 15 min to visualize the bands. The androstanederivatives gave bright violet colors; cholesterol ap-peared brown-violet, and coprostanol appeared red-violet. The Rf values of 5.8-androstan-3,t-ol-17-one, 5fi-androstan-3a-ol-17-one, and 5,B-androst-5-en-3,B-ol-17-one were 0.50, 0.44, and 0.40, respectively; cholesteroland coprostanol overlapped at approximately 0.71.Three bands with R, values of 0.39 to 0.50, 0.52 to 0.62,and 0.62 to 0.87, which corresponded to spots in thesamples, were scraped from the adjacent unsprayedhalf of the plate. Each scraping was eluted with 3 x 4ml of chloroform-methanol (2:1). The extracts of thebands were evaporated to dryness and treated with100 id of methoxylamine hydrochloride (MOX; PierceChemical Co.) at 65°C for 30 min to produce meth-oxime derivatives of ketones (6). The mixtures weresilylated with TRI-SIL/BSA at 65°C for 45 min andseparated by GLC at 250°C as described above. Thetest steroids and their metabolites were identified bycomparison of GLC retention times with authenticstandards and by GLC-mass spectrometry with a Hew-lett-Packard 5980A combined GLC-mass spectrome-try system.

RESULTSBiochemical characterization. The two

isolates of cholesterol-reducing Eubacteriumwere tested for 27 biochemical reactions as rec-ommended by Holdeman and Moore (8) foridentification as to species within the genus Eu-bacterium. The two isolates were identical intheir reactions. No acid was produced from anyof 22 carbohydrates, as determined by change inpH. No short-chain acids were produced fromglucose, as determined by GLC analysis. Indoleproduction, nitrate reduction, and gelatin and

VOL. 40, 1980

on March 11, 2020 by guest

http://aem.asm

.org/D

ownloaded from

1020 MOTT, BRINKLEY, AND MERSINGER

starch hydrolyses were all negative. Both cul-tures hydrolyzed esculin and therefore have ,8-glucosidase activity. An extract of an esculinculture of Eubacterium strain 403 also was frac-tionated by TLC. A gray-black spot appeared atan Rf of 0.5 when the plate was sprayed withdilute ferrous ammonium citrate. A second por-

tion of sample which was not sprayed was elutedfrom the TLC scrapings and analyzed by GLCas the trimethylsilyl ether derivative. A singlemajor peak with a retention time of 8 min wasobserved. GLC-mass spectrometry produced aparent mass ion at M/z 322, which is consistentwith the molecular weight of the ditrimethylsilylderivative of esculetin (6,7-dihydroxycoumarin).The two organisms were tested also for the

presence of the arginine dihydrolase pathway.No products of arginine were detected by TLC,nor was there stimulation of growth by arginine.Ether lipids as growth factors. We have

previously reported that the plasmalogen PLEsupports growth of cholesterol-reducing Eubac-terium in LCM (10). All of the following alkenyl-containing compounds also supported growth inLCM, as evidenced by clotting of the mediumand by coprostanol formation: 2-lysoplasmen-ylethanolamine, plasmenic acid, 2-acyl-1-alken-ylglycerol, 1-monoalkenylglycerol, and plasmen-ylcholine. Lysoplasmanylcholine, a 1-alkyl gly-cerophospholipid which does not contain an al-kenyl residue, did not support growth.Steroids as growth factors. All cultures

that contained both cholesterol and the teststeroid and the inoculated LCM - PLE controlcultures clotted or produced a heavy precipitatewithin 4 days. Clotting was not observed inmedium without cholesterol, but the physicalappearance of the medium changed if growthoccurred (Table 1).The metabolites of the test steroids and the

approximate percentages produced are given inTable 2. Coprostanol was the major metaboliteof all cultures containing cholesterol. Two of thetest steroids, allocholesterol and cholest-4-en-3-one, were reduced to coprostanol. The copros-tanol reported in Table 2 was derived only fromthe test steroids and not from cholesterol. Wecalculated the amount of coprostanol derivedfrom the test steroids by subtracting the sum ofthe peak areas of coprostanol and cholesterol inthe control culture from the sum of cholesteroland coprostanol in the test cultures. The differ-ence was assumed to be the amount of copros-tanol derived from the test steroid. Also, sincecoprostanone and cholest-4-en-3-one are onlyminor metabolites of cholesterol, the presence ofeither of these steroids was attributed to the teststeroid. The amount of coprostanol derived from

APPL. ENVIRON. MICROBIOL.

TABLE 1. Appearance of cultures during incubation

Test steroids Description of culture atindicated time

Control .................. NC'

Allocholesterol ........... Grainy (6 days);precipitate (14 days)

Cholest-4-en-3-one Clear medium;(2 mg/ml) ............. precipitate in bottom

(14 days)

Cholest-4-en-3-one Clear medium (14(0.2 mg/ml) ............ days)

22,23-Bisnor-5-cholenicacid-3,8-ol .............. NCa

Androst-5-en-3,8-ol-17-one Grainy (6 days)a NC, No change in appearance.

TABLE 2. Metabolism of test steroids byEubacterium strain 403

Choles- Metabolites (% of testterola steroids)

Epicholesterol + None- None

Allocholesterol + Coprostanol (37)- Coprostanol (66)

Cholest-4-en-3-one + Coprostanone (45),(2 mg/ml) coprostanol (33)

- Coprostanone (25)

Cholest-4-en-3-one + Coprostanolb,(0.2 mg/ml) coprostanoneb

- Coprostanol (42)

22,23-Bisnor-5- + Nonecholenic acid-3,8- - Noneol

Androst-5-en-3/8-ol- + 5,8-Androstane-3,8-17-one ol-17-one (64)

- 5fl-Androstane-3,8-ol-17-one (14)

a + Culture with cholesterol; -, culture withoutcholesterol.

b Could not be accurately quantitated.

cholest-4-en-3-one in the culture with the lowconcentration of cholest-4-en-3-one and choles-terol could not be readily estimated. The prod-uct of cholest-4-en-3-one metabolism was co-prostanone when the substrate concentrationwas 2 mg/ml; however, the major product wascoprostanol when the substrate concentrationwas 0.2 mg/ml. The identity ofthese metaboliteswas confirmed by GLC-mass spectrometry anal-

on March 11, 2020 by guest

http://aem.asm

.org/D

ownloaded from

CHOLESTEROL-REDUCING EUBACTERIUM 1021

ysis. The product eluted from GLC at 6.5 minhad the same retention time as a trimethylsilylderivative of coprostanol and gave a massspectrum identical to the mass spectrum of tri-methylsilyl-coprostanol; the base peak was atM/z 370, with fragments at M/z 355, 215, 257,and 403 and a small parent peak at M/z 460.The GLC peak at 11.3 min had the same reten-tion time as coprostanone and a mass spectrumidentical to that of coprostanone; the base peakwas at M/z 386, with fragments atM/z 231, 316,353, and 371. These results confirmed the iden-tity ofcoprostanol and coprostanone as productsof cholest-4-en-3-one. Two of the test steroids,epicholesterol and bisnorcholenic acid-3,B-ol,were not metabolized in the presence or absenceof cholesterol. Neither of these steroids sup-ported growth. The double bond of androst-5-en-3,B-ol-17-one was reduced to an androstanederivative. A 3,8-ol, 5,8-H product could not bedifferentiated with certainty from the 3a-ol, 5a-H isomer by TLC, GLC, or GLC-mass spectrom-etry. However, by analogy to the reduction ofcholesterol to the 3,B-ol, 5,B-H isomer (copros-tanol) by Eubacterium strains, the androstanederivative probably is 5,f-androstan-3fi-ol-17-one. The mass spectrum of the trimethylsilyl-MOX derivative of the product was identical tothat produced by a derivatized standard of 5fl-androstan-3f8-ol-17-one. Scanning from M/z 100to 500, the base peak was at M/z 270, with otherfragments greater than M/z 200 atM/z 360, 300,376, and 391 (the molecular ion).

DISCUSSIONThe strains of cholesterol-reducing Eubacte-

rium are unique in their requirement for choles-terol or certain other steroids for growth (4).The metabolism of steroids by Eubacteriumstrain 403 was similar to that reported for Eu-bacterium ATCC 21408 (4). Cholesterol, allo-cholesterol, cholest-4-en-3-one, and androst-5-en-3,8-ol-17-one supported growth and were re-duced by both organisms. We have also recentlyobserved metabolism of the A4-3-one steroidsdeoxycorticosterone and progesterone by Eu-bacterium strain 403 (unpublished data). How-ever, 22,23-bisnor-5-cholenic acid-3,8-ol, with aring structure which is identical to cholesteroland an ionic side chain (carboxylic), did notsupport growth of Eubacterium strain 403, norwas it metabolized in the presence of cholesterol.Eyssen et al. (4) reported that coprostanone, apossible intermediate in coprostanol formation,also would support growth. The results indicatethat these Eubacterium strains have little spec-ificity for the length of the steroid side chain butmay be inhibited by an ionic side chain. Also,

reduction of the A4 or A5 double bond or the 3-ketone or both appears to be required forgrowth. The wide range of steroids which sup-port growth of these organisms probably rulesthem out as structural membrane componentsbut suggests that they might have a vital meta-bolic role, possibly as electron acceptors, as pro-posed by Eyssen et al. (4).The mechanism of microbial coprostanol for-

mation from cholesterol is probably an intra-molecular hydrogen transfer, wi-th cholest-4-en-3-one and coprostanone as intermediates (1, 11,13). Our cultures containing cholest-4-en-3-oneresulted primarily in the production of copros-tanone or coprostanol. At a high concentration(2.0 mg/ml) of cholest-4-en-3-one, a greater per-centage of the substrate was converted to co-prostanone than to coprostanol, but at a lowconcentration (0.2 mg/ml), cholest-4-en-3-onewas converted principally to coprostanol. Wehave subsequently repeated this experiment andhave not observed significant amounts of co-prostanone or coprostanol, which may be due tolimited growth. In medium containing 2 mg ofcholest-4-en-3-one per ml, the cell numbers in-creased from 1 x 102 to 6.5 x 105 organisms perml of medium during 6 days of growth. Cellnumbers were determined by the plate counttechnique with an agar medium that we haverecently developed (la). A similar experimentwith cholesterol showed that the cell densityreached 106 to 107 cells per ml of medium beforecoprostanol could be detected by our methods.Therefore, some steroids, such as cholest-4-en-3-one, which support growth may not alwaysresult in sufficient cell numbers to produce de-tectable steroid metabolites.

In addition to the requirement for steroids,these organisms apparently require an alkenylether lipid. We reported that PLE or its 2-lysoderivative would serve as a growth factor inLCM and that none of 21 other lipids supportedgrowth (10). We found that four derivatives ofPLE and plasmenylcholine, all with an intact 1-alkenyl ether linkage, would support growth.Cultures with the 1-alkyl glycerophospholipid 2-lysoplasmanylcholine did not produce copros-tanol.Members of the genus Eubacterium are non-

sporeforming, gram-positive, anaerobic bacillithat generally produce energy by fermentingsugars to short-chain acids (8, 9). However, likeEubacterium lentum, the cholesterol-reducingorganisms do not produce acids from carbohy-drate-containing media. Unlike E. lentum (15),cholesterol-reducing strains of Eubacterium donot derive energy from the arginine dihydrolasepathway. They have a unique requirement for

VOL. 40, 1980

on March 11, 2020 by guest

http://aem.asm

.org/D

ownloaded from

1022 MOTT, BRINKLEY, AND MERSINGER

cholesterol or certain other steroids and alkenylether lipids, which distinguishes them from anypreviously recognized species (8, 9).

ACKNOWLEDGMENTS

The technical assistance ofAndrew Gottesman, Don Smith,and Evelyn Jackson is gratefully acknowledged. Susan Wein-traub performed the mass spectral analyses.

This work was supported by Public Health Service grantHL-19362 from the National Heart, Lung, and Blood Institute.

LITERATURE CITED1. Bjorkhem, I., J.-A. Gustafsson, and 0. Wrange. 1973.

Microbial transformation of cholesterol into copros-tanol: properties of a 3-oxo-A4-steroid-5,8-reductase.Eur. J. Biochem. 37:143-147.

la.Brinkley, A. W., A. R. Gottesman, and G. E. Mott.1980. Growth of cholesterol-reducing Eubacterium oncholesterol-brain agar. Appl. Environ. Microbiol. 40:1130-1132.

2. Dawson, R. M. C., and N. Hemington. 1967. Someproperties of purified phospholipase D and especiallythe effect of amphipathic substances. Biochem. J. 102:76-6.

3. Dawson, R. M. C., N. Hemington, and J. B. Daven-port. 1962. Improvement in the method of determiningindividual phospholipids in a complex mixture by suc-cessive chemical hydrolyses. Biochem. J. 84:497-501.

4. Eyssen, H. J., G. G. Parmentier, F. C. Compernoile,G. DePauw, and M. Piessens-Denef. 1973. Biohydro-genation of sterols by Eubacterium ATCC 21,408-Novaspecies. Eur. J. Biochem. 36:411-421.

5. Folch, J., M. Lees, and G. H. Sloane-Stanley. 1957. Asimple method for the isolation and purification of totallipids from animal tissues. J. Biol. Chem. 226:497-509.

6. Gardiner, W. L, and E. C. Horning. 1966. Gas-liquidchromatographic separation of C,g and C2, human uri-nary steroids by a new procedure. Biochim. Biophys.Acta 115:524-526.

7. Gottfried, E. L., and M. M. Rapport. 1963. The bio-chemistry of plasmalogens. III. Concentrations in tis-sues of the rat as a function of age. Biochemistry 2:646-648.

8. Holdeman, L. V., and W. E. C. Moore (ed.). 1973.Anaerobe laboratory manual. Anaerobe Laboratory,Virginia Polytechnic Institute and State University,Blacksburg.

9. Holdeman, L V., and W. E. C. Moore. 1974. Genus II.Eubacterium, p. 641-657. In R. E. Buchanan and N. E.Gibbons (ed.), Bergey's manual of determinative bac-teriology, 8th ed. The Williams & Wilkins Co., Balti-more.

10. Mott, G. E., and A. W. Brinkley. 1979. Plasmenylethan-olamine: growth factor for cholesterol-reducing Eubac-terium. J. Bacteriol. 139:755-760.

11. Parmentier, G., and H. Eyssen. 1974. Mechanism ofbiohydrogenation of cholesterol to coprostanol by Eu-bacterium ATCC 21408. Biochim. Biophys. Acta 348:279-284.

12. Sadzikowski, M. R., J. F. Sperry, and T. D. Wilkins.1977. Cholesterol-reducing bacterium from humanfeces. Appl. Environ. Microbiol. 34:355-362.

13. Schoenheimer, R., D. Rittenberg, and M. Graff. 1935.Deuterium as an indicator in the study of intermediarymetabolism. IV. The mechanism of coprosterol forma-tion. J. Biol. Chem. 111:183-192.

14. Snyder, F., M. L Blank, and R. L. Wykle. 1971. Theenzymic synthesis of ethanolamine plasmalogens. J.Biol. Chem. 246:3639-3645.

15. Sperry, J. F., and T. D. Wilkins. 1976. Arginine, agrowth-limiting factor for Eubacterium lentum. J. Bac-teriol. 127:780-784.

16. Tung, J., and R. Ostwald. 1969. Effect of dietary cho-lesterol on bile acid composition of gall bladder bilefrom guinea pigs. Lipids 4:216-223.

17. Zolg, W., and J. C. G. Ottow. 1973. Improved thin-layertechnique for detection of arginine dihydrolase amongthe Pseudomonas species. Appl. Microbiol. 26:1001-1003.

APPL. ENVIRON. MICROBIOL.

on March 11, 2020 by guest

http://aem.asm

.org/D

ownloaded from