"I - American Society for Microbiology

0 downloads 0 Views 3MB Size Report
Biotechnologies,1 and Station de Microscopie Electronique, Institut Pasteur, 75724 Paris Cedex 15, France. Received 14 December 1993/Accepted 28 February ...
Vol. 176, No. 10

JOURNAL OF BAcrERIOLOGY, May 1994, p. 2828-2834

0021-9193/94/$04.00+0

Copyright © 1994, American Society for Microbiology

Subcellular Localization of Clostridium thermocellum ORF3p, a Protein Carrying a Receptor for the Docking Sequence Borne by the Catalytic Components of the Cellulosome SYLVIE SALAMITOU,1 MARC LEMAIRE,1 T. FUJINO,'t HELENE OHAYON 2 PIERRE GOUNON 2 PIERRE BEGUIN,l* AND JEAN-PAUL AUBERT1

Unite de Physiologie Cellulaire and URA 1300 Centre National de la Recherche Scientifique, Departement des Biotechnologies,1 and Station de Microscopie Electronique, Institut Pasteur, 75724 Paris Cedex 15, France Received 14 December 1993/Accepted 28 February 1994

The ORF3 gene of Clostridium thermocelum encodes a polypeptide (ORF3p) which contains a receptor domain for the docking sequence borne by the catalytic subunits of the cellulosome and a triplicated domain related to some bacterial cell surface proteins. It was thus surmised that ORF3p is a surface protein. In this study, this hypothesis was confirmed. Subcellular fractionation, Western blotting (immunoblotting), and electron microscopy of immunocytochemically labeled cells indicated that ORF3p produced by C. thermocellum was located in the outer surface layer of the bacterium. This layer appeared to consist of a soft matrix shedding of particulate fragments. Nonsedimenting ORF3p derived from sonicated cells was associated with highmolecular-mass fractions (>20 MDa), probably corresponding to fragments of the outer cell layer. The same high-molecular-mass fractions also contained the cellulosomal marker CipA. Contrary to CipA, however, ORF3p was not associated with 2- to 4-MDa fractions corresponding to individual cellulosomes, and a significant fraction of ORF3p failed to bind to cellulose. It is proposed that ORF3 and ORF3p be renamed olpA and OlpA, respectively (for outer layer protein).

coli clone and used to prepare ORF3p-specific antibodies. These antibodies served to detect by Western blotting (immunoblotting) the presence of ORF3p in cell fractions of C. thernocellum. The location of ORF3p was also investigated by electron microscopy of immunochemically labeled cells. In addition, the behavior of ORF3p-containing fractions was investigated by gel filtration and cellulose affinity chromatography.

The cellulolytic system of Clostridium thermocellum consists of a high-molecular-weight multienzyme complex termed the cellulosome (17, 18). Electron micrographs of C. thermocellum cells labeled with cellulosome-specific antibodies showed that cellulosomes are distributed in the protuberances formed by the outermost layer of the bacterial cell surface. This layer is heavily stained by cationized ferritin (2). Adherence of the cells to cellulose is correlated with the presence of cellulosome-containing protuberances (3, 20). The catalytic subunits of the cellulosome contain a duplicated segment which serves as a docking sequence mediating attachment of the catalytic subunits to a large scaffolding component termed CipA (24). A CipA molecule contains nine receptor domains for the docking sequence (6, 8). Sequencing of the DNA downstream from cipA revealed three new open reading frames. One of these, termed ORF3, encodes a protein, called ORF3p, which possesses an NH2terminal domain similar to the receptor domains carried by CipA and a COOH-terminal domain homologous to segments present in some bacterial cell surface proteins. On the basis of these data, it was suggested that ORF3p is a surface layer protein (7). In the accompanying article (21a), it was shown that the duplicated segment of CipA did not bind to the receptor domain of ORF3p. However, the duplicated segment of CelD has about the same affinity for the receptor domain of ORF3p as for one of the receptor domains carried by CipA. In this study, the subcellular location of ORF3p was investigated. ORF3p was purified from a recombinant Escherichia

MATERIALS AND METHODS Bacterial strains, plasmids, and culture conditions. E. coli TG1 (9) carrying pCT1700 (Fig. 1) was grown in Luria-Bartani E

"I

I

N/P Sp H

(S) P/N I

Sp pTZl 8R

IacZ' orf3 0,4kb

E

G/P/T/S-rich

S-layer-like

segment

repeat

Receptor domair

FIG. 1. Structure of pCIT1700. The DNA of the vector is indicated by a thin line. C. thermocellum DNA lying outside ORF3 is indicated by a thick line. Segments encoding the various regions identified in ORF3p (7) are shown as boxes of different patterns. M.C.S., multiple cloning site; E, EcoRI; (S), SalI site deleted by mung bean nuclease treatment; P/N and N/P, fusions between PstI and NsiI and vice versa; H, Hindlll; Sp, SphI. S-layer, surface layer.

* Corresponding author. Mailing address: Unite de Physiologie Cellulaire, Departement des Biotechnologies, Institut Pasteur, 28 rue du Dr. Roux, 75724 Paris Cedex 15, France. Phone: 33 1 45 68 88 19. Fax: 33 1 45 68 87 90. t Permanent address: Nagoya Seiraku Co. Ltd., Nagoya 468, Japan.

2828

LOCALIZATION OF C. THERMOCELLUM ORF3p PROTEIN

VOL. 176, 1994

A

B 12

3

4

A 5

6

1

7

2

B 3

4

1 2

2829

C 3

4

1

2

3

4

97 -

97 97

66-

-

6666-

4

.4E4

tS

45 -

3

454531 -

21-

FIG. 2. Monitoring of the different steps of ORF3p purification. (A) SDS-PAGE analysis of the various fractions. (B) Probing of proteins for the binding of "2I-labeled CelD. About 40 ,ug of protein was loaded in lane 1, 10 Fig was loaded in lane 2, and 5 pug was loaded in lanes 3 to 7. Lanes: 1, suspension of broken cells; 2, pellet of aggregated proteins before extraction; 3, dialyzed guanidine hydrochloride extract; 4, dialyzed and cleared guanidine hydrochloride extract; 5 and 6, pools from the first and second pH 7.5 DEAE columns; 7, pool from the pH 8.5 DEAE column. Numbers at left are in kilodaltons.

medium supplemented with carbenicillin (100 ,ug/ml) (1). C. thermocellum NCIB 10682 was grown anaerobically with gentle stirring at 60°C in complete CM3-3 medium (23) containing either 10 g of cellulose MN300 (Macherey, Nagel and Co., Duren, Federal Republic of Germany) or 5 g of cellobiose (Fluka AG, Buchs, Switzerland) per liter. DNA manipulations. Restriction enzymes were used as recommended by the suppliers. DNA manipulations were performed as described by Maniatis et al. (19). Subcloning and expression of OR". A 1.65-kb NsiI fragment containing ORF3, with the exception of the first 63 nucleotides, was cloned at the PstI site of pTZ18R. In-frame fusion with the lacZ' gene of pTZ18R was achieved by digestion of the recombinant plasmid with SalI, trimming of single-stranded ends with mung bean nuclease, and ligation (Fig. 1). In the resulting construct, termed pCT1700, the first 21 codons encoding the putative signal peptide of ORF3p were deleted. SDS-PAGE and blotting techniques. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (15). Electrophoretic transfer of separated proteins to nitrocellulose was performed as described previously (1) or with a Bio-Rad semidry blotting cell according to the manufacturer's directions. Blots were probed either with anti-ORF3p antiserum and 1251-labeled protein A or with 125I-labeled 68-kDa endoglucanase CelD (24). Unless otherwise stated, the immune serum was saturated with a crude extract from E. coli TG1(pTZ19R) (10). Purification of ORF3p. All operations were performed at 4°C. Protein concentrations were assessed by use of the Bradford reagent (Bio-Rad) as directed by the manufacturer, with bovine serum albumin as a standard. E. coli TG1 cells harboring pCT1700 were cultivated at 42°C and harvested in stationary phase. Under these conditions,

FIG. 3. Identification and localization of ORF3p in C. thernocellum fractions. (A) Identification of ORF3p in envelopes from C. thermocellum cells grown in the presence of cellobiose. Blots were incubated either with anti-ORF3p immune serum saturated with a crude extract of TG1(pTZ19) (lanes 1 and 2) or with immune serum incubated with a crude extract of TG1(pCT1700) (lanes 3 and 4). Lanes: 1 and 3, ORF3p purified from E. coli (0.5 kug); 2 and 4, envelope fractions of cells harvested from 330 ,ul of a cellobiose-grown culture at an OD6w of 1.1. (B and C) Western blotting of fractions from cellobiose-grown (B) or cellulose-grown (C) cells with anti-ORF3p immune serum. Aliquots corresponding to the same fraction of the initial volume of the cellobiose- or cellulose-grown culture were loaded. Samples from the two cultures were normalized relative to each other by loading of the same quantity of protein (15 pLg) from the supernatant of sonicated cells, corresponding to 88 pl of the cellobiose-grown culture and 150 ,lI of the cellulose-grown culture. In both panels, culture supernatant was loaded in lane 1, sonicated cell supernatant (15 pg of protein) was loaded in lane 2, and envelope fraction was loaded in lane 3. The second wash of the envelope fraction was loaded in lane 4 of panel B, and the SDS extract from the sonicated, cellulose-bound fraction was loaded in lane 4 of panel C. Numbers at left of panels A and B are in kilodaltons.

OR,F3p aggregated and formed inclusion bodies. Cells from a

1-liter culture were suspended in 70 ml of 50 mM PC buffer (4) containing 1 mM phenylmethylsulfonyl fluoride and 5 mM EDTA, disrupted in an Aminco French pressure cell at 100 MPa, and centrifuged for 30 min at 27,000 x g. The amount of protein in the crude extract was 828 mg in 69 ml. The pellet was washed twice with 0.154 M NaCl (saline), taken up in 10 ml of 50 mM Tris-HCl buffer (pH 7.5) (buffer 1), and extracted in 40 ml of 5 M guanidine hydrochloride for 1 h. The extract was centrifuged at 27,000 x g for 30 min, and the supernatant was dialyzed against three changes of buffer 1, yielding 119 mg of protein in 41 ml. Precipitated material was removed from the dialyzed sample by centrifugation for 30 min at 27,000 x g. The cleared supernatant (108 mg of protein in 40 ml) was loaded on a DEAE-Trisacryl column (8 by 2.6 cm; IBF) equilibrated with buffer 1. Elution was performed with a linear NaCl gradient (0 to 0.3 M in 150 ml of buffer 1). The pooled ORF3p fractions were dialyzed against buffer 1, yielding 38 mg of total protein in 14 ml, and chromatographed a second time under the same conditions. After the second chromatography, the preparation (30 mg of protein in 16 ml) was dialyzed against 50 mM Tris-HCl buffer (pH 8.5) and loaded on a DEAE-Trisacryl column equilibrated with the same buffer. Elution was achieved with a linear NaCl gradient (0 to 0.3 M in 150 ml of the same buffer). The pooled ORF3p fractions (16 mg of protein in 10 ml) were dialyzed against buffer 1 and concentrated to 1.5 ml with an Amicon ultrafiltration cell (membrane PM10). Aliquots were kept at - 80°C.

2830

SALAMITOU ET AL.

J. BAC-rERIOL.

FIG. 4. Thin sections (A, B, and E) and whole cells (C and D) of C. thermocellum. Thin sections in panels A and E were stained with

anti-ORF3p antibodies. Cells in panel B were stained with cationized ferritin prior to sectioning. Cells in panels C and D were stained with anti-ORF3p antibodies. Bars, 200 nm.

For gel filtration chromatography experiments, 2 mg of purified ORF3p was applied to an Ultrogel AcA 44 column (80 by 1.6 cm; LKB) equilibrated with buffer 1. Elution was performed at a flow rate of 12 ml/h. Preparation of ORF3p-specific antibodies. Rabbit antiORF3p antiserum was prepared as described previously (24) with purified ORF3p as an immunogen. ORF3p-specific antibodies were further purified by immunoaffinity chromatography. Purified ORF3p (300 ,ug) was electrophoresed on a 10% acrylamide gel and transferred to a nitrocellulose membrane. The ORF3p band was stained with Ponceau S (1) and cut out.

The nitrocellulose membrane was incubated for 1 h at room temperature in phosphate-buffered saline (PBS) (10) containing 5% dried milk and 0.1% Tween 20 to block nonspecific adsorption to the membrane. The membrane was then incubated for 4 h at room temperature on a rocking platform with 4 ml of antiserum, previously saturated for 4 h at 37°C with 4 ml of a TG1(pTZ19) crude extract (12 mg of protein per ml). After three washes for 15 min each in PBS containing 0.1% Tween 20 and one wash in PBS, the nitrocellulose was cut into pieces and antibodies were eluted by treatment with 0.7 ml of 0.2 M HCl-glycine buffer (pH 2.2) at room temperature for 15

VOL. 176, 1994

LOCALIZATION OF C. THERMOCELLUM ORF3p PROTEIN

2831

FIG. 4-Continued.

min. The pH of the eluate was brought to 7 with 0.3 ml of 1 M K2HPO4, and the preparation was dialyzed overnight at 4°C against PBS. Fractionation of C. thermocelum cells. Fractionation was done as described in the accompanying article (21a). Size exclusion chromatography. Samples from the supernatant of sonicated cells were analyzed by gel filtration chromatography with Sepharose CL-4B (Pharmacia). The column dimensions were 88 by 2.5 cm. The column was equilibrated with 50 mM Tris-HCI (pH 7.5). Elution was performed at a flow rate of 12 ml/h. Fractions of 2.2 ml were collected. Cellulose binding assay. Microgranular cellulose (400 mg; Sigma Chemical Co., St. Louis, Mo.) was incubated with gentle stirring for 3 h at room temperature with 160 p.g of protein in 8 ml of 20 mM Tris-HCI (pH 7.7). The cellulose was then packed in a 1-ml syringe. Nonadsorbed material was collected in the effluent. The packed material was washed with 3 ml of the same buffer. Cellulose-bound proteins were eluted with 3.5 ml of 1% triethylamine. The pH of the eluted fractions was brought to 7 with 6 N HCl. Electron microscopy. Cells grown in cellobiose-supple-

mented CM3-3 medium were harvested at an optical density at 600 nm (OD600) of 1.2. For immunocytochemical labeling of whole cells, cells from 1 ml of culture were centrifuged, washed

with saline, and suspended gently in 0.7 ml of saline, and 0.3 ml of Karnovsky's fixative (4% paraformaldehyde and 5% glutaraldehyde in 0.1 M phosphate buffer [pH 7.4]) was added. The samples were fixed for 5 min at room temperature, centrifuged, washed in PBS, and suspended in 1 ml of the same buffer. Cells were applied to Formvar-carbon-coated nickel grids previously made hydrophilic by glow discharge. Grids were treated with drops of the following reagents: PBS-50 mM NH4Cl (10 min), PBS-1% bovine serum albumin (5 min), purified specific antibodies diluted in PBS-1% bovine serum albumin (1/2 to 1/100) (1 h), PBS (three washes, 2 to 5 min each), anti-rabbit immunoglobulin G (heavy and light chains) antibody-gold conjugates (10 nm; BioCell Research Laboratories, Cardiff, United Kingdom) diluted 1/25 in PBS-0.01% fish skin gelatin (Sigma G7765) (30 to 45 min), PBS (one wash, 1 min), and distilled water (three washes, 1 min each). Samples on the grids were then fixed with 1% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) (2 min), rinsed with distilled water, and negatively stained with 1% ammonium molybdate in water. For immunolabeling of thin sections, cells from 50 ml of culture were centrifuged, washed twice in saline, and taken up in 500 jil of saline. Cacodylate buffer (500 [lI, 0.1 M, pH 7.4) and 300 pAl of Karnovsky's fixative in 0.1 M cacodylate buffer were added. Cells were fixed for 1 h at 4°C and rinsed with 0.1

2832

SALAMITOU ET AL.

J. BAcrERIOL. Thyroglobuulin

Void volurne

-200 - 116 -

-

ORF3p _

I

97 66

A

i

- 45

CipA -_-

F 0 -200

if W;+S*f: -

116 -97

-66 0!B

Y

ORF3p -_

w 4_ 5 i->0;0-

Fraction number

FIG. 5. Chromatography on a Sepharose CL -4B column of the supernatant of sonicated cells of a cellobiose-grow in culture. A sample containing 7 mg of protein was applied to the colur of the selected fractions were analyzed by We elution volumes of blue dextran (void volume) anc (669 kDa) are indicated. Blots were probed with anti-OlRF3p antiserum (A) or with 1"I-labeled CelD (B). Horizontal arrows ir idicate the positions of ORF3p and CipA. Numbers at right are in kilc)daltons.

smtn Ablqott.s (25Th) sthyroglobulin

M cacodylate buffer (pH 7.4). Postfixation waas performed with 0.5% uranyl acetate in water for 1 h at 4°C. . Thereafter, cells were rinsed for 30 min in distilled water and suspended in 2% agarose in PBS. The agar-cell pellet was dehyydrated in ascending grades of alcohol (30 to 100%) in an; automatic freezesubstitution system (Leica AFS), which progYressively lowered the temperature from 4 to 50°C. Embeddiing was performed with Lowicryl HM20 resin, and polymerizat ion proceeded at 50°C for 48 h (5). Thin sections were cut on an LKB Nova ultramicrotome, collected on Formvar-coatecd nickel grids, and immunolabeled following the protocol outllined above. Sections were stained for 10 min with 2% urany '1 acetate in water and for 1 min with Millonig's lead citrate (1 1, 12). The same protocol was used for the con itrol experiments, except that anti-ORF3p antibodies was omit tted. Labeling of whole cells with cationized fe-rritin and visualization of the cell samples as thin sections wrere performed as described previously (3). Specimens were examined with a Philips CM12 or JEOL 1010 electron microscope. -

-

RESULTS Purification of ORF3p. As no enzymatic activity could be assigned to ORF3p, the purification of the p)rotein was moni-

tored at each step by SDS-PAGE and by probing of proteins for the binding of 125I-labeled CelD. Figure 2 shows the monitoring of the purification of ORF3p, a 57-kDa protein displaying the ability to bind to 125I-CelD. The elution volume for ORF3p on an Ultrogel AcA 44 gel filtration column corresponded to the exclusion volume of a 60-kDa globular protein, suggesting that ORF3p is monomeric. Identification of ORF3p in C. thermocellum. Antibodies raised against ORF3p purified from E. coli TG1(pCT1700) were used to detect the corresponding protein in C. thermocellum. The lower-molecular-mass band recognized by the antiserum in Fig. 3A, lane 1, probably resulted from proteolytic degradation of ORF3p purified as displayed in Fig. 2A and B, lane 7. Western blot experiments performed with envelope fractions of C. thermocellum revealed a radioactive band of 59 kDa (Fig. 3A, lane 2). No band was observed with preimmune serum (data not shown). When immune serum was saturated with a crude extract of TG1(pCT1700), labeling of the 59-kDa band disappeared (Fig. 3A, lane 4). The 59-kDa component was able to bind 251I-labeled 68-kDa CelD (Fig. 3B), providing further evidence of its identity with ORF3p. Distribution of ORF3p in different fractions of C. thermnocellum. Western blotting performed with various fractions from a cellobiose-grown culture is shown in Fig. 3B. Similar amounts of ORF3p were detected in the culture supernatant fraction, in the supernatant of sonicated cells, and in the envelope fraction. Little if any ORF3p could be washed off from the envelope fraction by 1 M NaCl (Fig. 3B, lane 4). A similar distribution of ORF3p was observed in a cellulosegrown culture (Fig. 3C), except that almost no ORF3p was found in the culture supernatant fraction. This result might be explained by the presence of ORF3p in the cellulose-bound fraction. Indeed, when the sonicated cellulose-bound fraction was boiled in the presence of SDS, ORF3p could be extracted in quantities comparable to those detected in the culture supernatant of the cellobiose-grown culture (Fig. 3C, lane 4). However, the fractionation 'procedure does not allow the determination of which proportion of ORF3p present in the sonicated cellulose-bound fraction derives from cosedimenting cells and cell debris. Samples from cellulose- and cellobiose-grown cultures were normalized so that the same amount of protein was loaded for the supernatant of sonicated cells and, for each culture, the same proportion of each fraction was loaded on the blot. Under these conditions, there was no obvious difference in the total amount of ORF3p produced in the presence of cellobiose or cellulose. In either case, ORF3p was present in amounts too small to be reliably identified by Coomassie blue staining of proteins extracted from the cell envelope fraction and separated by SDS-PAGE. In particular, ORF3p did not correspond to either of the two major envelope proteins, whose Mrs were 50,000 and 130,000, respectively (data not shown). The distribution of ORF3p among the various fractions of the cultures could be interpreted in several ways. ORF3p could be either a cytoplasmic protein which partially bound to envelopes and cosedimented with them or a cell surface component which was released from the envelope fraction upon sonication. An example of the second possibility is provided by the cell-bound cellulosome, which is present at the surface of the cells but appears in the soluble fraction after sonication (3, 18). Likewise, it was not clear whether the presence of ORF3p in the culture medium resulted from true secretion, from shedding of cell envelope material, or from lysis of the cells. These questions were investigated by electron microscopy of immunocytochemically labeled cells.

LOCALIZATION OF C. THERMOCELLUM ORF3p PROTEIN

VOL. 176, 1994

A 12

fraction that was bound could not be washed off with the washing buffer.

B

3

2

1

3 Z

~ ~

~

~

2833

-200

~

E

-116

-97

-

66

-45

FIG. 6. Cellulose affinity chromatography. Void volume fractions from the Sepharose CL-4B column of Fig. 5 were pooled and subjected to cellulose affinity chromatography as described in Materials and Methods. The same proportion of each fraction was analyzed by Western blotting. Blots were probed with anti-ORF3p antiserum (A) or with "2I-labeled CelD (B). Lanes 1, nonadsorbed fraction; 2, Tris-HCl wash; 3, triethylamine-eluted fraction. Arrows indicate the positions of ORF3p and CipA, as in Fig. 5. Numbers at right are in kilodaltons.

Localization of ORF3p by electron microscopy. When thin sections of C. thermnocellum cells were labeled with antiORF3p antibodies, labeling was almost exclusively associated with the fuzzy outer layer of the cell surface (Fig. 4A). This layer corresponds to the protuberance-forming, ferritin-binding layer described by Bayer and coworkers (2, 3) (Fig. 4B). ORF3p was also visualized in whole cells stained with ammonium molybdate (Fig. 4C). The protein was associated with a thick layer of amorphous material surrounding the cells. This layer appeared to shed into the medium fragments, which were also labeled with anti-ORF3p antibodies. This result suggests that ORF3p could be associated, at least in part, with a particulate structure. The pattern of labeling obtained with anti-ORF3p antibodies varied from cell to cell. In about half of the bacteria, labeling was strongest at both ends of the cell (Fig. 4D and E). Gel permeation and cellulose adsorption chromatography of ORF3p and of the cellulosome. To check whether ORF3p was associated with a particulate structure, the supernatant of sonicated cells was analyzed by gel permeation chromatography. Western blotting indicated that ORF3p was present exclusively in the fractions eluting in the voided volume of a Sepharose CL-4B column (exclusion volume, 2 x 107 Da for globular proteins) (Fig. 5A and B). In contrast, when fractions were tested for the presence of CipA (a marker for the cellulosome) by probing with "2I-labeled CelD, a significant proportion of CipA eluted between the void volume and the elution volume of thyroglobulin (669 kDa) (Fig. SB). This material, which did not contain ORF3p, probably corresponds to individual cell-bound cellulosomes, released in the soluble fraction by sonication (3, 16, 18). The set of bands recognized by "MI-labeled CelD and migrating between CipA and ORF3p may correspond, in part, to proteolysis products of CipA (22). No band was recognized on a similar blot probed with "Ilabeled 63-kDa CelD, which is devoid of the docking sequence (data not shown). The particulate fractions eluting with the void volume were pooled and assayed for binding to crystalline cellulose. CipA was only found in the cellulose-bound fraction, whereas only two-thirds of ORF3p was found in this fraction (Fig. 6). The

DISCUSSION The antibody directed against ORF3p synthesized in E. coli from pCT1700 specifically recognized a 59-kDa polypeptide of C. thermocellum which was able to bind to 1251I-labeled endoglucanase CelD bearing the docking sequence. This result confirmed that the 59-kDa band corresponded to ORF3p. The sizes of ORF3p estimated by SDS-PAGE were 57 kDa for the form synthesized in E. coli, which appeared to be a monomeric protein, and 59 kDa for the polypeptide detected in C. thermocellum. Both Mrs are higher than the Mw calculated from the expected polypeptide sequence (48,562 for the LacZ' fusion protein synthesized in E. coli and 45,318 for the mature protein expected in C. thermocellum after cleavage of the putative signal peptide). ORF3p contains a G/P/T/S-rich region which extends over 57 residues and which may confer unusual migration properties on the protein. The slower migration of ORF3p synthesized in C. thermocellum may be due to glycosylation. Indeed, several bacterial cell surface proteins are known to be glycosylated (14). Electron microscopy and biochemical studies showed that ORF3p was embedded in the outer layer of the cell surface. For this reason, we propose to term the ORF3 gene olpA (for outer layer protein) and its product OlpA. The localization of OlpA is in agreement with the hypothesis that the COORterminal triplicated domain of the polypeptide is specific for proteins associated with the cell surface of bacteria, as predicted in the model proposed by Fujino et al. (7). Screening of sequence data bases with the TFASTA program (21) indicates that similar motifs are present in the surface layer proteins of Bacillus brevis (GenBank M15364), Acetogenium kivui (GenBank M31069), and Thermus thermophilus (GenBank X57333), as well as in a xylanase of Thermoanaerobacter sp. strain B6A-R1 (GenBank M97882) and in an endoglucanase of alkalophilic Bacillus sp. strain KSM-635 (GenBank M27420). It would be of interest to check whether the latter two enzymes are also located on the cell surface of the respective bacteria. The outer layer appeared to consist of a thick matrix of soft, cationized ferritin-reactive material shedding into the medium particulate fragments with which OlpA was associated. This result probably explains why OlpA was found in the culture supernatant, in the supernatant of sonicated cells, and in the envelope fraction. Indeed, OlpA present in the supernatant of sonicated cells was exclusively associated with material with a molecular mass higher than 20 MDa. The majority of OlpAcontaining material could be adsorbed to cellulose. These properties coincide with those described for the "polycellulosomes" (13) and, in fact, CipA was found to be associated with the high-molecular-mass, cellulose-binding fractions containing OlpA. However, this does not mean that OlpA is directly associated with polycellulosomes. In fact, if polycellulosomes and OlpA are embedded in the same outer layer matrix, they may be present in the same fragments without directly interacting with each other. In addition, adsorption of OlpA to cellulose could be due to the association of OlpA with catalytic enzymes containing cellulose-binding domains. No OlpA was detected in fractions of 2 to 4 MDa containing individual cellulosomes, and a significant fraction of OlpA failed to bind to cellulose. Hence, if there is a direct interaction between OlpA and polycellulosomes, this interaction is probably weak. As shown in the accompanying paper (21a), the receptor domain of OlpA has about the same affinity for the duplicated segment of CelD as the seventh receptor domain carried by

2834

SALAMITOU ET AL.

CipA. Together with the location of OlpA in the outer cell layer, these results suggest that the role of OlpA might be to anchor individual cellulases or hemicellulases on the cell surface. Such noncellulosomal, cell-bound enzymes could contribute to cell-bound (hemi)cellulase activity. Alternatively, catalytic components of cellulosomes might be transiently bound to OlpA before being incorporated into the cellulosomes. ACKNOWLEDGMENTS S.S. was the recipient of a fellowship from Agence de l'Environnement et de la Maitrise de l'Energie under contract with the Lyonnaise des Eaux Company. M.L. was the recipient of a fellowship from the Ministere de l'Enseignement Superieur et de la Recherche. This work was supported by contracts CPL C462 and AIR1-CT-0321 from the Commission of the European Communities and by research funds from University of Paris 7. T.F. was on leave from Nagoya Seiraku Co. Ltd. REFERENCES 1. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl. 1990. Current protocols in molecular biology. Greene Publishing and Wiley Interscience, New York. 2. Bayer, E. A., and R. Lamed. 1986. Ultrastructure of the cell surface cellulosome of Clostridium thermocellum and its interaction with cellulose. J. Bacteriol. 167:828-836. 3. Bayer, E. A., E. Setter, and R. Lamed. 1985. Organization and distribution of the cellulosome in Clostridium thermocellum. J. Bacteriol. 163:552-559. 4. Beguin, P., J. Joliff, M. Juy, A. G. Amit, J. Millet, R. J. Poljak, and J.-P. Aubert. 1988. Crystalline endoglucanase D of Clostridium thermocellum overproduced in Escherichia coli. Methods Enzymol. 160:355-362. 5. Carlemalm, E., R. M. Garavito, and W. Villigier. 1982. Resin development for electron microscopy and an analysis of embedding at low temperature. J. Microsc. 126:123-143. 6. Fujino, T., P. Beguin, and J.-P. Aubert. 1992. Cloning of a Clostridium thermocellum DNA fragment encoding polypeptides that bind the catalytic components of the cellulosome. FEMS Microbiol. Lett. 94:165-170. 7. Fujino, T., P. Beguin, and J.-P. Aubert. 1993. Organization of a Clostridium thermocellum gene cluster encoding the cellulosomal scaffolding protein CipA and a protein possibly involved in the attachment of the cellulosome to the cell surface. J. Bacteriol. 175:1891-1899. 8. Gerngross, U. T., M. P. M. Romaniec, N. S. Huskisson, and A. L. Demain. 1993. Sequencing of a Clostridium thermocellum gene (cipA) encoding the cellulosomal SL-protein reveals an unusual degree of internal homology. Mol. Microbiol. 8:325-334. 9. Gibson, T. J. 1984. Studies on the Epstein-Barr virus genome. Ph.D. thesis. University of Cambridge, Cambridge, United Kingdom. 10. Grepinet, O., M.-C. Chebrou, and P. Beguin. 1988. Purification of

J. BAC-FERIOL.

Clostridium thermocellum xylanase Z expressed in Escherichia coli and identification of the corresponding product in the culture medium of C. thermocellum. J. Bacteriol. 170:4576-4581. 11. Hobot, J. A. 1989. Lowicryls and low temperature embedding for colloidal gold methods, p. 75-115. In M. A. Hayat (ed.), Colloidal gold: principles, methods, and applications. Academic Press, Inc., San Diego, Calif. 12. Hobot, J. A., M. A. Bjomsti, and E. Kellenberger. 1987. Use of on-section immunolabeling and cryosubstitution for studies of bacterial DNA distribution. J. Bacteriol. 169:2055-2062. 13. Hon-nami, K., M. P. Coughlan, H. Hon-nami, and L. G. Ljungdahl. 1986. Separation and characterization of the complexes constituting the cellulolytic enzyme system of Clostridium thermocellumr. Arch. Microbiol. 145:13-19. 14. Koval, S. F. 1988. Paracrystalline protein surface arrays on bacteria. Can. J. Microbiol. 34:1181-1189. 15. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680-685. 16. Lamed, R., and E. A. Bayer. 1988. The cellulosome of Clostridium thermocellum. Adv. Appl. Microbiol. 33:1-46. 17. Lamed, R., E. Setter, and E. A. Bayer. 1983. Characterization of a cellulose-binding, cellulase-containing complex in Clostridium thermocellum. J. Bacteriol. 156:828-836. 18. Lamed, R., E. Setter, R. Kenig, and E. A. Bayer. 1983. The cellulosome: a discrete cell surface organelle of Clostridium thermocellum which exhibits separate antigenic, cellulose-binding and various cellulolytic activities. Biotechnol. Bioeng. Symp. 13:163181. 19. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 20. Nolte, A., and F. Mayer. 1989. Localization and immunological characterization of the cellulolytic enzyme system in Clostridium thermocellum. FEMS Microbiol. Lett. 61:65-72. 21. Pearson, W. R., and D. J. Lipman. 1988. Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85:2444-2448. 21a.Salamitou, S., 0. Raynaud, M. Lemaire, M. Coughlan, P. Beguin, and J.-P. Aubert. 1994. Recognition specificity of the duplicated segments present in Clostridium thermocellum endoglucanase CelD and in the cellulosome-integrating protein CipA. J. Bacteriol. 176:2822-2827. 22. Salamitou, S., K. Tokatlidis, P. Beguin, and J. P. Aubert. 1992. Involvement of separate domains of the cellulosomal protein S1 of Clostridium thermocellum in binding to cellulose and in anchoring of catalytic subunits to the cellulosome. FEBS Lett. 304:89-92. 23. Tailliez, P., H. Girard, J. Millet, and P. Beguin. 1989. Enhanced cellulose fermentation by an asporogenous and ethanol-tolerant mutant of Clostridium thermocellum. Appl. Environ. Microbiol. 55:207-211. 24. Tokatlidis, K., S. Salamitou, P. Beguin, P. Dhurjati, and J.-P. Aubert. 1991. Interaction of the duplicated segment carried by Clostridium thermocellum cellulases with cellulosome components. FEBS Lett. 291:185-188.