APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 1986, p. 978-984 Vol. 51, No. 50099-2240/86/050978-04$02.00/0Copyright © 1986, American Society for Microbiology

Comparison of Proteins Involved in Chondroitin Sulfate Utilizationby Three Colonic Bacteroides Species

LAUREN LIPESKI, ELLEN P. GUTHRIE, MILDRED O'BRIEN, SUSAN F. KOTARSKI, ANDABIGAIL A. SALYERS*

Department of Microbiology, University of Illinois, Urbana, Illinois 61801

Received 12 November 1985/Accepted 26 February 1986

Three species of colonic bacteria can ferment the mucopolysaccharide chondroitin sulfate: Bacteroidesovatus, Bacteroides sp. strain 3452A (an unnamed DNA homology group), and B. thetaiotaomicron. Proteinsassociated with the utilization of chondroitin sulfate by B. thetaiotaomicron have been characterized previously.In this report we compare chondroitin lyases and chondroitin sulfate-associated outer membrane polypeptidesof B. ovatus and Bacteroides sp. strain 3452A with those of B. thetaiotaomicron. All three species produce twosoluble cell-associated chondroitin lyases, chondroitin lyase I and II. Purified enzymes from the three specieshave similar pH optima, Km values, and molecular weights. However, peptide mapping experiments show thatthe chondroitin lyases from B. ovatus and Bacteroides sp. strain 3452A are not identical to those of B.thetaiotaomicron. A cloned gene that codes for the chondroitin lyase II from B. thetaiotaomicron hybridized ona Southern blot with DNA from B. ovatus or Bacteroides sp. strain 3452A only when low-stringency conditionswere used. Antibody to chondroitin lyase II from B. thetaiotaomicron did not cross-react with chondroitin lyaseII from B. ovatus or Bacteroides sp. strain 3452A. Chondroitin lyase activity in all three species was inducibleby chondroitin sulfate. B. ovatus and Bacteroides sp. strain 3452A, like B. thetaiotaomicron, have outermembrane polypeptides that appear to be regulated by chondroitin sulfate, but the chondroitin sulfate-associated outer membrane polypeptides differ in molecular weight. Despite these differences, the ability ofintact bacteria to utilize chondroitin sulfate, as indicated by growth yields in carbohydrate-limited continuousculture and the rate at which the chondroitin lyases were induced, was the same for all three species.

Bacteroides thetaiotaomicron, B. ovatus, and Bacteroidessp. strain 3452A (an unnamed DNA homology group) have aDNA-DNA cross-homology of 30 to 40% with each other,whereas their cross-homology with other colonic Bacte-roides species is much lower (4). These three species alsoresemble each other in that they can utilize some polysac-charides, such as chondroitin sulfate, that are not utilized byother Bacteroides species (16). If polysaccharides such aschondroitin sulfate are important sources of carbon andenergy for these species in the colon, the genes that code forkey proteins in the breakdown pathway might have beenmore conserved than other genes. Alternatively, these spe-cies could have evolved completely different strategies forpolysaccharide breakdown. If so, such differences mightindicate that the organisms occupy different ecologicalniches. To gain some insight into how different the polysac-charide-degrading systems of these three organisms mightbe, we compared the proteins that are involved in the firststeps of chondroitin sulfate breakdown.

Chondroitin sulfate breakdown by B. thetaiotaomicronhas been studied in some detail. B. thetaiotaomicron pro-duces two chondroitin sulfate-degrading enzymes, chondroi-tin lyase I and II (9). These enzymes are inducible bychondroitin sulfate (14), are cell-associated (15), and arevery similar with respect to molecular weight and kineticproperties (9). In addition to the two chondroitin lyases,there are chondroitin sulfate-associated outer membranepolypeptides (6). No function has yet been proven for theseouter membrane polypeptides, but they may be responsiblefor bringing chondroitin sulfate into contact with the peri-plasmic chondroitin lyases (6, 15). No previous work has

* Corresponding author.

been done on the breakdown of chondroitin sulfate by B.ovatus or Bacteroides sp. strain 3452A. In the present study,we have characterized the chondroitin lyases and chon-droitin sulfate-associated outer membrane polypeptides ofB. ovatus and Bacteroides sp. strain 3452A and have com-pared them with the chondroitin lyases and chondroitinsulfate-associated outer membrane polypeptides of B. theta-iotaomicron.

MATERIALS AND METHODS

Bacterial strains and growth conditions. B. ovatus 0038,Bacteroides sp. strain 3452A, and B. thetaiotaomicron 5482were obtained from the culture collection of the VirginiaPolytechnic Institute Anaerobe Laboratory, Blacksburg, Va.Bacteria were cultivated in a defined medium whichcontained chondroitin sulfate, glucuronic acid, or glucose asthe sole carbohydrate source (5, 15). For batch cultures, theatmosphere was a mixture of oxygen-free carbon dioxide(20%) and nitrogen (80%), and the concentration ofcarbohydrate was 5 mg/ml. Chondroitin sulfate III, chon-droitin sulfate A, chondroitin sulfate C (Sigma Chemical Co.,St. Louis, Mo.), and glucose were autoclaved in the medium.Glucuronic acid was filter sterilized and added to theautoclaved medium. For the continuous cultures that wereused in the induction experiments, the atmosphere wasoxygen-free carbon dioxide and the concentration of glucosewas 2 mg/ml. In continuous-culture experiments to determinegrowth yield on chondroitin sulfate, the limiting carbohydratewas chondroitin sulfate III (2 mg/ml). Continuous culturetechniques have been described previously (5).The rate at which chondroitin lyase activity increased

after bacteria had been exposed to chondroitin sulfate wasmeasured as described previously (14). Briefly, bacteria

978

on July 14, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

CHONDROITIN SULFATE UTILIZATION BY BACTEROIDES SPP.

were equilibrated in glucose-limited continuous culture at ageneration time of 10 h. Bacteria were then removed fromthe continuous-culture vessel and mixed with chondroitinsulfate type A (Sigma). Anaerobic conditions were main-tained throughout. At intervals, samples were removed andthe specific activity of chondroitin lyase was determined.

Localization, purification, and characterization ofchondroitin lyases. The location of chondroitin lyase activitywas determined as described previously (15). The purifica-tion of chondroitin lyases from Bacteroides sp. strain 3452Aand B. ovatus was essentially the same as described previ-ously (9), except that the phosphate buffer concentrationwas 20 mM and the heparin agarose affinity column waseluted with a linear gradient of 0. 15 to 0.45 M NaCl in 20 mMphosphate buffer (pH 7.6). Enzyme assays and kinetic stud-ies were carried out as described by Linn et al. (9).Molecular weight determinations. Native molecular

weights of the enzymes were estimated by gel filtrationchromatography on a Sephacryl S-300 column (1.5 by 60cm). The column was equilibrated and eluted with 0.15 MNaCl in 20 mM phosphate buffer (pH 7.6). Fractions (1.0 ml)were collected and analyzed for chondroitin lyase activity orprotein concentration (A280). Molecular weight standardswere thyroglobulin (670,000), 3-amylase (225,000),ovalbumin (45,000), and myoglobin (18,000). Denatured mo-lecular weights of the enzymes were determined by sodiumdodecyl sulfate (SDS)-polyacrylamide gel electrophoresis(PAGE) as described by Laemmli (8). The separating gel wasan 8 to 15% acrylamide gradient.

Renaturation studies. To ascertain that a particular poly-peptide on the SDS-PAGE gel was actually the chondroitinlyase, a modification of the renaturation procedure of Bur-gess and Hager (1) was used to obtain active enzyme fromSDS-PAGE gels. Chondroitin lyase samples (1.0 total en-zyme unit) in SDS sample buffer (5.0%,-mercaptoethanol,20% glycerol, 3.0% SDS, 0.063 M Tris, pH 6.8) were appliedto cylindrical SDS-PAGE gels (0.9 by 10 cm; 8.75%acrylamide separating gel). After electrophoresis the cylin-drical gels were not stained for protein, but were cut into1.0-cm slices and placed in 15-ml glass tubes which had beentreated with Sigmacote (Sigma). Elution buffer (1.0 ml ofelution buffer per 1.0-cm gel slice) was added, and theprotein was eluted for 9.0 h at room temperature withoccasional stirring (1). The eluted protein was then treated asdescribed by Burgess and Hager (1). Molecular weightmarkers and the preparation containing chondroitin lyaseprotein were run on parallel cylindrical gels, and protein wasvisualized by Coomassie blue staining. Chondroitin lyaseactivity in the eluant from the gel slices was assayed asdescribed previously (9). The products which resulted fromthe action of renatured chondroitin lyase on chondroitinsulfate A were resolved by descending paper chromatogra-phy in glacial acetic acid-n-butanol-1 M NH40H (3:2:1) asdescribed previously and compared with authentic standards(15).

Peptide mapping. The protein that was used for peptidedigests was run on a 10% polyacrylamide SDS-PAGE gel(0.8 mM). Protein in the SDS-PAGE gel was first visualizedby staining and destaining the gel with Coomassie blue R250in 40% methanol-10% acetic acid and 10% methanol-10%acetic acid, respectively, and the appropriate band was thencut from the gel. Partial digestion of the protein in the gelslices with N-chlorosuccinamide was performed by theprocedure of Lischwe and Ochs, and the gel slices wereloaded onto a 15% acrylamide (0.9 mM) SDS-PAGE gel toseparate the peptide fragments (10). After electrophoresis,

the gel was silver stained by the method of Burk andEschenbruch (2). Peptide digests of different preparations ofthe same chondroitin lyase gave reproducible patterns.

Southern blot analysis. The gene that codes for B. theta-iotaomicron chondroitin lyase II (about 3 kilobases) hasbeen cloned on a 7.8-kilobase EcoRI fragment in pBR328 (3).The recombinant plasmid that contains the clonedchondroitin lyase II gene (pA818) was labeled with 32P bynick translation (12) and used to probe EcoRI digests ofBacteroides sp. strain 3452A, B. ovatus, and B. thetaiota-omicron chromosomal DNA. Southern blot analysis, plas-mid isolation, and restriction endonuclease digestions wereperformed as described previously (3, 11). ChromosomalDNA was isolated by the method of Saito and Miura (13).High-stringency conditions were as follows: hybridization at42°C (50% formamide) followed by two washes with 0.2%SDS in 2x SSC (lx SSC is 0.15 M NaCl plus 0.015 Msodium citrate) at 60°C and two washes with 0.2% SDS in0.5x SSC at 50°C. Low-stringency conditions were as fol-lows: hybridization at 42°C (no formamide) followed by twowashes with 0.2% SDS in 2x SSC at 25°C and two washeswith 0.2% SDS in 0.5x SSC at 50°C.

Western blot analysis. Antiserum was obtained by inocu-lating a New Zealand White rabbit with purified B. thetaio-taomicron chondroitin lyase II. To avoid possible cross-contamination with chondroitin lyase I, chondroitin lyase IIwas purified from an E. coli clone containing pA818, aplasmid that carries the cloned chondroitin lyase II genefrom B. thetaiotaomicron (3). Chondroitin lyase I and IIfrom the three Bacteroides species were resolved on a 10%SDS-polyacrylamide gel and then blotted as described byKranz and Gennis (7). Hybridizing antibody was detectedwith 125I-labeled protein A from Staphylococcus aureus (7).Outer membrane polypeptides. Outer membranes of Bac-

teroides sp. strain 3452A and B. ovatus were obtained asdescribed by Kotarski and Salyers (6). Inner membranecontamination was less than 10%, as determined by thesuccinic acid dehydrogenase assay (6). SDS-PAGE gel anal-ysis of outer membrane polypeptides was performed on an 8to 20% polyacrylamide gradient gel as described previously(6).

RESULTSLocation and purification of chondroitin lyases. In both B.

ovatus and Bacteroides sp. strain 3452A, chondroitin lyaseactivity was cell associated and partitioned with solubleprotein rather than with membranes. The effect of variouspurification steps on chondroitin lyase specific activity isshown in Table 1. Chromatography of the DEAE effluent ona heparin-agarose column resolved the chondroitin lyaseactivity of each species into two separate peaks. For both B.ovatius and Bacteroides sp. strain 3452A, one peak(chondroitin lyase I) eluted from the column at ca. 0.3 MNaCl and a second peak (chondroitin lyase II) eluted fromthe column at ca. 0.4 M NaCl (Fig. 1). Chondroitin lyase IIof both species generally appeared as a single band on aSDS-PAGE gel (Mr, ca. 107,000) after purification throughheparin-agarose, although minor contaminants were presentin some preparations. Chondroitin lyase I preparations fromboth species after purification through heparin-agarose con-tained a number of minor contaminants. The predominantprotein in chondroitin lyase I preparations had a molecularweight of ca. 113,000. Further purification of chondroitinlyase I could be achieved by chromatography on cellulosephosphate (9), but yields were low and the preparations stillcontained more than one protein.

VOL. 51, 1986 979

on July 14, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

TABLE 1. Purification of chondroitin lyase activity fromBacteroides sp. strain 3452A and B. ovatus

Total Total S cPurification step activity protein Spact

(U)a (mg) (U/mg)

Bacteroides sp. strain 3452ACrude extract 471 1,302 0.4Streptomycin sulfate supernatant 459 928 0.5DEAE-Sephacel eluant 362 197 1.8Heparin-agarose pooled fractionsPeak I 21 5.1 4.0Peak II 52 1.5 35.0

B. ovatusCrude extract 518 1,629 0.3Streptomycin sulfate supernatant 510 1,303 0.4DEAE-Sephacel eluant 397 592 0.8Heparin-agarose pooled fractionsPeak I 28 11.5 2.4Peak II 190 8.4 23.0

a 1 = 1 p.mol of disaccharide released per min when the enzyme wasincubated with chondroitin sulfate A at 37°C.

Renaturation studies. To ascertain that the prominentproteins of M, 107,000 and 113,000 were in fact chondroitinlyases, concentrated pooled fractions from the heparin-agarose column were applied to an SDS-PAGE gel. Afterelectrophoresis and renaturation, the location of thechondroitin lyase activity in the gel was determined. Typicalresults for the chondroitin lyase I and II from Bacteroidessp. strain 3452A are shown in Fig. 2. Similar results wereobtained with chondroitin lyase preparations from B. ovatus.Chondroitin lyase activity was detected only in the gel slicesthat corresponded to the migration distance of the prominentpolypeptide that was visible on a parallel Coomassie blue-

.1.1-1-1."'Z.jIqIq(b

-_jI

IZZ-Z-Q)

lz.(Zi

i.3

20 40 60 80 100 120

Eluant Volume (ml)

K)-

0:Z

FIG. 1. Elution of chondroitin lyase activity (solid line) from a

heparin-agarose column. The NaCl concentration is indicated by thedashed line. (A) Profile for B. ovatus; (B) profile for Bacteroides sp.

strain 3452A.

A

t 2.0

ID~ 1.i..

2

B

5 Q25

- 17.6

I

NV

_ t%

- 5.2

i_q*k- *11.%I %s.k. '..,

0.2. 04 0.6Mobility

0.8 I£.

X 2 t

0.2 0.4 as 0.8, 1.0

Mobility

FIG. 2. Chondroitin lyase activity eluted from fractions of anSDS-PAGE gel. (A) Results obtained when Bacteroides sp. strain3452A chondroitin lyase I was applied to the SDS-PAGE gel; (B)results obtained when Bacteroides sp. strain 3452A chondroitinlyase II was applied to the SDS-PAGE gel. The inset of each figureis a photograph of a parallel gel that was stained with Coomassieblue to locate protein. Arrows indicate points at which the gel wassectioned. Chondroitin lyase activity is indicated by the solid line.The migration distance of various molecular weight standards isindicated by solid circles and the dashed line.

stained gel. When renatured enzyme was incubated withchondroitin sulfate A, the products generated comigratedwith authentic disaccharide standards on paper chro-matograms, whereas incubation of negative renaturationsamples with chondroitin sulfate A did not result in substratebreakdown. Approximately 3% of the chondroitin lyaseactivity applied to the gel was recovered. The recovery-limiting step of the procedure was SDS-PAGE of the sample.At this step there was an 81 to 90% loss of the activitypresent in a sample processed through the other four steps ofthe procedure. Electrophoretic elution of protein from gelslices did not increase the recovery of chondroitinase activ-ity. These results confirmed that the bands on the SDS-PAGE gel that had been tentatively identified as chondroitinlyase I and II in B. ovatus and Bacteroides sp. strain 3452Awere in fact the chondroitin lyases.

Molecular weight and peptide mapping. The denaturedmolecular weights of the chondroitin lyases (I and II) from B.

980 LIPESKI ET AL.

t,q --..

1%.

),3 ,,zq). - c,--... 1.. tz IIZI

1.CZ

on July 14, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

CHONDROITIN SULFATE UTILIZATION BY BACTEROIDES SPP.

1 2 3 4 5 6

-116K4-97.4K

- 67K

-- 45K

~ 29K

'-cr 19K

FIG. 3. Comparison of migration distances of the variouschondroitin lyases on an 8 to 15% SDS-PAGE gradient gel:chondroitin lyase I from B. thetaiotaomicron (lane 1), Bacteroidessp. strain 3452A (lane 2), B. ovatus (lane 3); chondroitin lyase II

from B. ovatus (lane 4), Bacteroides sp. strain 3452A (lane 5), and B.thetaiotaomicron (lane 6). Migration distances of molecular weightmarkers (K) are indicated by arrows and numbers at the right side ofthe gel.

ovatus, Bacteroides sp. strain 3452A, and B. thetaiotaomi-cron, as determined by SDS-PAGE, were similar and fellwithin a molecular weight range of 106,000 to 115,000 (Fig.3). The molecular weights that were estimated from migra-tion distances in the SDS-PAGE gel are compared in Table2. Undenatured molecular weights, estimated by chromatog-raphy on Sephacryl S-300, are also compared in Table 2. Ingeneral, the molecular weight of chondroitin lyase I from thethree species was about 6,000 greater than the molecularweight of the corresponding chondroitin lyase II (Fig. 3).Molecular weight differences of the chondroitin lyases werereproducible when different enzyme preparations were used.

Typical partial peptide digests of chondroitin lyases (I andII) from the three species are shown in Fig. 4. The peptidedigest data show that the chondroitin lyase I proteins of B.ovatus, Bacteroides sp. strain 3452A, and B. thetaiotaomi-cron are similar but not identical. The chondroitin lyase II

proteins of the three species are also similar but not identi-

cal. The differences in the digest patterns were reproduciblewhen repeated with different enzyme preparations.

Kinetic studies. The properties of chondroitin lyases fromB. ovatus and Bacteroides sp. strain 3452A are shown inTable 2. Properties of the enzymes from B. thetaiotaomi-cron, taken from a previously published report (9), are alsogiven for comparison. Pooled heparin-agarose fractions wereused as a source of chondroitin lyase II. Chondroitin lyase Iwas further purified by cellulose phosphate chromatography(9) before being used to generate the results in Table 2. Theenzymes were similar with respect to pH optima, relativeactivity on chondroitin sulfates A or C, and Km values.

Southern blot analysis. Small differences in protein struc-ture can produce large differences in the peptide digests.Because the molecular weights and kinetic properties of thechondroitin lyases were so similar, we wanted to determinewhether the proteins from the three species were as differentas they appeared to be by peptide digest analysis. Accord-ingly, we hybridized a recombinant plasmid that containedthe gene for chondroitin lyase II from B. thetaiotaomicron(3) with EcoRI digests of chromosomal DNA from B. theta-iotaomicron, B. ovatus, and Bacteroides sp. strain 3452A.When high-stringency hybridization conditions were used,there was no detectable hybridization between the probe andDNA from B. ovatus or Bacteroides sp. strain 3452A. Theonly detectable hybridization with DNA from B. thetaiota-omicron was to a 7.8-kilobase EcoRI fragment which corre-sponds to the insert size of the probe (data not shown).When low-stringency hybridization conditions were used,some cross-reactive sequences were detected in Bacteroidessp. strain 3452A and B. ovatus DNA, but the cross-reactiveEcoRI fragments differed in size from that of B. thetaiota-omicron (Fig. 5).Western blot analysis. Antiserum to chondroitin lyase II

from B. thetaiotaomicron did not cross-react with chon-droitin lyase I from B. thetaiotaomicron or with chondroitinlyase I and II from B. ovatus or Bacteroides sp. strain 3452A(Fig. 6).

Regulation of chondroitin lyase activity. The chondroitinlyases of B. ovatus and Bacteroides sp. strain 3452A wereregulated by chondroitin sulfate. After bacteria that hadbeen grown in glucose-limited continuous culture were ex-posed to chondroitin sulfate, the chondroitin lyase specificactivity increased 10- to 30-fold within 90 min. The rate atwhich the specific activity increased was similar for all threespecies (data not shown).

TABLE 2. Comparison of the properties of chondroitin lyase I and II from Bacteroides sp. strain 3452A, B. ovatus,and B. thetaiotaomicron

Mol wta (103) pH optimum for Chondroitin sulfate Ca Km (mg/ml)a forStrain and lyase type chondroitin sulfate' activity (t activity on chondroitin sulfate

SDS-PAGE Sephacryl S-300 A c chondroitin sulfate A) A c

Bacteroides sp. strain 3452AI 111 93 t 10 7.4 7.6 95 0.12 0.03II 107 115 ± 10 7.6 7.6 90 0.07 0.09

B. ovatusI 114 93 ± 10 7.4 7.6 90 0.10 0.07II 106 115 ± 10 7.6 7.6 85 0.04 0.06

B. thetaiotaomicronbI 115 115 ± 10 7.2 7.2 120 0.06 0.05II 109 128 ± 10 7.6 7.6 70 0.05 0.05

a Values represent the average of at least duplicate determinations, made with different enzyme preparations. Variations among the values was less then 20%.b Values, except for molecular weight data, are from a previously published report (9).

VOL. 51, 1986 981

on July 14, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

982 LIPESKI ET AL.

1 2 3 4 5 6-<-45K

fi.-- 29K

I 1I8K

I1 14-

FIG. 4. Comparison of N-chlorosuccinamide-generated peptidedigests of the various chondroitin lyases. Digest of chondroitin lyaseI from B. ovatus (lane 1), Bacteroides sp. strain 3452A (lane 3), andB. thetaiotaomicron (lane 5). Digest of chondroitin lyase II from B.ovatus (lane 2), Bacteroides sp. strain 3452A (lane 4), and B.thetaiotaomicron (lane 6).

Outer membrane polypeptides. We previously found evi-dence of differences in outer membrane polypeptide compo-

sition when B. thetaiotaomicron was grown on chondroitinsulfate rather than on glucuronic acid, a component ofchondroitin sulfate (5) (Fig. 7). In one-dimensional gels ofouter membranes from B. ovatus cells, there were threepolypeptides (Mr, ca. 105,000, 35,000, and 28,000) thatappeared to le produced at higher levels when bacteria weregrown on chdndroitin sulfate. Also, a number of polypeptide

-6.7

--*404

--a2D

FIG. 5. Autoradiogram of a Southern blot in which EcoRI-digested chromosomal DNA from B. ovatus (lane b), Bacteroidessp. strain 3452A (lane c), and B. thetaiotaomicron (lane a) was

probed with 32P-labeled pA818, which carries the gene for B.thetaiotaomicron chondroitin lyase II. Low-stringency hybridiza-tion conditions were used (see Materials and Methods). Arrows atthe right side of the gel show the migration distance of molecularweight standards (in kilobases).

FIG. 6. Autoradiogram of a Western blot in which antiserum tochondroitin lyase II from B. thetaiotaomicron was used as theprobe. Lanes contain 2 to 3 ,ug of purified chondroitin lyase I fromB. thetaiotaomicron (lane a), B. ovatus (lane c), or Bacteroides sp.strain 3452A (lane e), or 2 to 3 ,ug of purified chondroitin lyase IIfrom B. thetaiotaomicron (lane b), B. ovatus (lane d), or Bacteroidessp. strain 3452A (lane f).

bands decreased in intensity. In outer membranes fromBacteroides sp. strain 3452A, three polypeptides (Mr ca.60,000, 44,000, and 40,000) were consistently seen as moreintensely staining bands when bacteria were grown onchondroitin sulfate, and other polypeptide bands decreasedin intensity. In one-dimensional gels of outer membranesfrom B. thetaiotaomicron, five chondroitin sulfate-enhancedpolypeptides were visible (Mr, ca. 105,000, 63,000, 54,000,28,000, and 26,000). The 105,000- and 28,000-molecular-weight polypeptides of B. thetaiotaomicron were probablynot the same as those of B. ovatus, because antiserum toouter membranes of B. thetaiotaomicron, which detectedthe 105,000- and 28,000-molecular-weight polypeptides ofthat species, did not cross-react with the 105,000- and28,000-molecular-weight polypeptides of B. ovatus (data notshown). As with the other two species, some polypeptidebands also decreased in intensity, and the polypeptide profile

-616K_ 97K

_- 19K

FIG. 7. Comparison of outer membrane polypeptides of bacteriagrown on glucuronic acid, a component of chondroitin sulfate (lanes1, 3, and 5), with outer membrane polypeptides of bacteria grown onchondroitin sulfate (lanes 2, 4, and 6): B. thetaiotaomicron (lanes 1and 2), B. ovatus (lanes 3 and 4), and Bacteroides sp. strain 3452A(lanes 5 and 6).

APPL. ENVIRON. MICROBIOL.

on July 14, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

CHONDROITIN SULFATE UTILIZATION BY BACTEROIDES SPP.

was less complex when bacteria were grown on chondroitinsulfate.Growth in continuous culture. The ability of intact orga-

nisms to utilize chondroitin sulfate, as indicated by growth inchondroitin sulfate-limited continuous cultures at dilutionrates of 0.07 to 0.08 h-', was similar for all three species.Optical densities at 650 nm were 0.8 to 0.9 in all cases. Theconcentrations of cell dry weight (average of results of twoseparate chemostat experiments) were 0.41 mg/ml (B. theta-iotaomicron), 0.39 mg/ml (B. ovatus), and 0.38 mg/ml (Bac-teroides sp. strain 3452A). These differences were not sig-nificant. Chondroitin lyase specific activities of bacteriagrowing in continuous culture (average of results from twoseparate experiments) also did not differ significantly: 0.58U/mg of protein (B. thetaiotaomicron), 0.58 U/mg of protein(B. ovatus), and 0.54 U/mg of protein (Bacteroides sp. strain3452A). Growth rates in batch culture were identical (datanot shown).

DISCUSSIONThe results of this study show that the chondroitin lyases

of B. thetaiotaomicron, B. ovatus, and Bacteroides sp. strain3452A are quite distinct at the molecular level. This can beseen from differences in the peptide digest patterns of thechondroitin lyases from the different species (Fig. 4). Also, acloned fragment containing the gene that codes for thechondroitin lyase II from B. thetaiotaomicron cross-hybridized with DNA from the other two Bacteroides spe-cies only when low-stringency conditions were used. Fi-nally, antiserum raised against purified B. thetaiotaomicronchondroitin lyase II did not cross-react with chondroitinlyase II from B. ovatus or from Bacteroides sp. strain 3452A.The failure of the anti-chondroitin lyase II antiserum to

cross-react indicates that the amino acid sequence of the B.thetaiotaomicron chondroitin lyase II is quite different fromthat of chondroitin lyase II from B. ovatus or Bacteroides sp.strain 3452A. However, it is also possible that this antiserumonly detects a few epitopes on the protein. If so, the lack ofcross-reactivity could be the result of a relatively smallnumber of changes. The fact that there is detectable cross-hybridization between a clone which contains the gene forchondroitin lyase II and DNA from B. ovatus andBacteriodes sp. strain 3452A when low-stringency hybrid-ization conditions were used indicates that there may besome homology between the DNA sequences of thechondroitin lyase II genes from the three species. However,sequences within the cloned fragment which lie outside thechondroitin lyase II gene could be responsible for some ofthe observed homology.

Despite differences at the molecular level, the chondroitinlyases from the three species were very similar at thefunctional level, i.e., they had similar pH optima, specificactivities, fractionation properties, and Km values (Table 2).

All three species have outer membrane polypeptides thatappear to be regulated by polypeptides. Some polypeptidebands decreased in intensity, and others were enhancedwhen bacteria were grown on chondroitin sulfate rather thanon glucuronic acid. However, the molecular weights of thechondroitin sulfate-associated outer membrane polypeptidesof the three species differed. Since the function of thesepolypeptides is now known, it was not possible to assesstheir functional similarity directly. Nonetheless, it is inter-esting to note that the outer membrane of all three speciesresponded to growth on chondroitin sulfate in a similar way.

Despite differences at the molecular level among thechondroitin lyases and chondroitin sulfate-associated pro-

teins of the different species, the general features ofchondroitin sulfate utilization, such as the cellular locationof the enzymes and regulation of the enzymes and outermembrane polypeptides by chondroitin sulfate, were thesame in all three species. Moreover, there was no detectabledifference in the efficiency of chondroitin sulfate utilization,as measured by growth yields and chondroitin lyase specificactivities of bacteria growing in continuous culture, growthrates in batch culture, and rate of chondroitin lyase induc-tion. Apparently, considerable changes can occur at themolecular level without significantly affecting the overallfunctioning of the system. There may be more subtle effectsthat cannot be detected by these techniques, and such effectscould be significant for organisms in the colon. However, atthe level of the intact cell, the three species appear to bevirtually identical with respect to chondroitin sulfate utiliza-tion.Our findings support the hypothesis that all three Bacte-

roides species originally had the same set of genes coding forproteins involved in utilization of chondroitin sulfate, butthat divergence due to mutation has occurred. Since thedifferences observed are consistent with the low DNA-DNAcross-homologies of these species (30 to 40%), it appearsthat there is no strong selection for conservation of thesequences of genes associated with chondroitin sulfate utili-zation. The fact that the general features of the chondroitinsulfate utilization system have been preserved indicates thatsuch features as cellular location and regulation of theenzymes, production of two similar chondroitin lyases, andthe possible involvement of outer membrane polypeptidesmay be essential for the utilization of chondroitin sulfate.Recently, in apparent contradiction to this, we found that amutant of B. thetaiotaomicron that produces chondroitinlyase I but not active chondroitin lyase II can still grow inbatch culture on chondroitin sulfate (E. P. Guthrie and A. A.Salyers, submitted for publication). However, it is stillpossible that both enzymes are necessary for efficient utili-zation of chondroitin sulfate under conditions encounteredin the colon.

ACKNOWLEDGMENTThis work was supported by Public Health Service grant AI-17876

from the National Institute of Allergy and Infectious Disease.

LITERATURE CITED

1. Burgess, R., and D. Hager. 1980. Elution of protein from sodiumdodecyl sulfate-polyacrylamide gels, removal of sodium dode-cyl sulfate, and renaturation of enzymatic activity: results withsigma subunit of Escherichia coli RNA polymerase, wheat germDNA topoisomerase, and other enzymes. Anal. Biochem.108:76-86.

2. Burk, R., and M. Eschenbruch. 1982. Experimentally improvedreliability of ultrasensitive silver staining of protein in polyacryl-amide gels. Anal. Biochem. 125:96-99.

3. Guthrie, E. P., N. B. Shoemaker, and A. A. Salyers. 1985.Cloning and expression in Escherichia coli of a gene coding fora chondroitin lyase from Bacteroides thetaiotaomicron. J. Bac-teriol. 164:510-515.

4. Johnson, J. L. 1978. Taxonomy of the Bacteroides. int. J. Syst.Bacteriol. 28:245-256.

5. Kotarski, S. F., and A. A. Salyers. 1981. Effect of long genera-tion times on growth of Bacteroides thetaiotaomicron in carbo-hydrate-limited continuous culture. J. Bacteriol. 146:853-860.

6. Kotarski, S. F., and A. A. Salyers. 1984. Isolation and charac-terization of outer membranes of Bacteroides thetaiotaomicrongrown on different carbohydrates. J. Bacteriol. 158:102-109.

7. Kranz, R. G., and R. B. Gennis. 1982. A quantitative radioim-

VOL. 51, 1986 983

on July 14, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

984 LIPESKI ET AL. APPL. ENVIRON. MICROBIOL.

munochemical screening method for specific gene products.Anal. Biochem. 127:247-257.

8. Laemmli, U. K. 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

9. Linn, S., T. Chan, L. Lipeski, and A. A. Salyers. 1983. Isolationand characterization of two chondroitin lyases from Bacteroidesthetaiotaomicron. J. Bacteriol. 156:859-866.

10. Lischwe, M., and D. Ochs. 1982. A new method for partialpeptide mapping using N-chlorosuccinimide/urea and peptidesilver staining in sodium dodecyl sulfate-polyacrylamide gels.Anal. Biochem. 127:453-457.

11. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y. 31:533-542.

12. Rigby, P. W. J., M. Dieckmann, C. Rhodes, and P. Berg. 1977.

Labeling deoxyribonucleic acid to high specific activity in vitroby nick translation with DNA polymerase I. J. Mol. Biol.113:237-251.

13. Saito, H., and K. Miura. 1963. Preparation of transformingdeoxyribonucleic acid by phenol treatment. Biochim. Biophys.Acta 72:619-629.

14. Salyers, A. A., and S. F. Kotarski. 1980. Induction ofchondroitin sulfate lyase activity in Bacteroides thetaiotaomi-cron. J. Bacteriol. 143:781-788.

15. Salyers, A. A., and M. O'Brien. 1980. Cellular location ofenzymes involved in chondroitin sulfate breakdown by Bacte-roides thetaiotaomicron. J. Bacteriol. 143:772-780.

16. Salyers, A. A., J. R. Vecellotti, S. H. West, and T. D. Wilkins.1977. Fermentation of mucin and plant polysaccharides bystrains of Bacteroides from the human colon. Appl. Environ.Microbiol. 33:319-322.

on July 14, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from