Dynamics in Oxygen-Induced Changes in S-Layer Protein Synthesis ...

JOURNAL OF BACTERIOLOGY, Apr. 1996, p. 2108–2117 0021-9193/96/$04.0010 Copyright q 1996, American Society for Microbiology

Vol. 178, No. 7

Dynamics in Oxygen-Induced Changes in S-Layer Protein Synthesis from Bacillus stearothermophilus PV72 and the S-Layer-Deficient Variant T5 in Continuous Culture and Studies of the Cell Wall Composition ´ RA,1* BEATRIX KUEN,2 HARALD F. MAYER,1 FEDERICO MANDL,1 MARGIT SA KURT CHRISTIAN SCHUSTER,1 AND UWE B. SLEYTR1 Zentrum fu ¨r Ultrastrukturforschung und Ludwig Boltzmann-Institut fu ¨r Molekulare Nanotechnologie, Universita ¨t fu ¨r Bodenkultur, 1180 Vienna,1 and Institut fu ¨r Mikrobiologie und Genetik, Biozentrum der Universita ¨t Wien, 1030 Vienna,2 Austria Received 3 October 1995/Accepted 23 January 1996

Stable synthesis of the hexagonally ordered (p6) S-layer protein from the wild-type strain of Bacillus stearothermophilus PV72 could be achieved in continuous culture on complex medium only under oxygen-limited conditions when glucose was used as the sole carbon source. Depending on the adaptation of the wild-type strain to low oxygen supply, the dynamics in oxygen-induced changes in S-layer protein synthesis was different when the rate of aeration was increased to a level that allowed dissimilation of amino acids. If oxygen supply was increased at the beginning of continuous culture, synthesis of the p6 S-layer protein from the wild-type strain (encoded by the sbsA gene) was immediately stopped and replaced by that of a new type of S-layer protein (encoded by the sbsB gene) which assembled into an oblique (p2) lattice. In cells adapted to prolonged low oxygen supply, first, low-level p2 S-layer protein synthesis and second, synchronous synthesis of comparable amounts of both types of S-layer proteins could be induced by stepwise increasing the rate of aeration. The time course of changes in S-layer protein synthesis was followed up by immunogold labelling of whole cells. Synthesis of the p2 S-layer protein could also be induced in the p6-deficient variant T5. Hybridization data obtained by applying the radiolabelled N-terminal and C-terminal sbsA fragments and the N-terminal sbsB fragment to the genomic DNA of all the three organisms indicated that changes in S-layer protein synthesis were accompanied by chromosomal rearrangement. Chemical analysis of peptidoglycan-containing sacculi and extraction and recrystallization experiments revealed that at least for the wild-type strain, a cell wall polymer consisting of N-acetylglucosamine and glucose is responsible for binding of the p6 S-layer protein to the rigid cell wall layer.

preserved during growth in continuous culture on complex medium only under oxygen-limited conditions when glucose was used as the sole carbon source (37). When oxygen supply was increased up a level that allowed metabolism of amino acids as an additional carbon source, the structurally different S-layer proteins from the wild-type strains were replaced by a new common type of S-layer protein that assembled into an oblique (p2) lattice type. Strain variants producing the p2 Slayer protein with a molecular weight of 97,000 as the sole S-layer protein could be isolated from continuous culture of the different wild-type strains (37). The S-layer protein from B. stearothermophilus PV72 has a molecular weight of 130,000 and assembles into a hexagonally ordered lattice with a center-to-center spacing of the morphological units of 22.5 nm (38, 39, 43). The complete gene encoding the 130,000-molecular-weight S-layer protein from the wild-type strain (sbsA) and approximately the whole gene encoding the 97,000-molecular-weight S-layer protein from the p2 variant (sbsB) have been cloned and sequenced (22, 23). In this study, B. stearothermophilus PV72 and the p6-deficient variant T5 were chosen as a model system for further experiments elucidating oxygen-induced changes in S-layer protein synthesis in the course of continuous culture. The p6-deficient variant T5 was originally isolated after cultivation

Crystalline bacterial cell surface layers (S-layers) have been identified as outermost cell envelope component of numerous bacteria and represent an almost universal feature of archaea. The oblique (p1, p2), square (p4), and hexagonal (p3, p6) lattices are formed of assemblies of identical protein or glycoprotein subunits with molecular weights ranging from 40,000 to 200,000 (for reviews and compilations, see references 6, 7, 28, 41, and 42). S-layers from Bacillus stearothermophilus strains reveal a great diversity with regard to lattice type, lattice constants, molecular weights of the subunits, and the occurrence of glycosylated S-layer proteins (27, 43). On the other hand, great similarity was observed in the sizes of the pores and in the physicochemical properties of the S-layer surface (33–36). In addition, oxygen-induced changes in S-layer protein synthesis and oxygen-dependent variant formation seem to be a further common feature of B. stearothermophilus strains (33, 37). The S-layers characteristic of three different wild-type strains showing either oblique, square, or hexagonal lattice symmetry were

* Corresponding author. Mailing address: Zentrum fu ¨r Ultrastrukturforschung, Universita¨t fu ¨r Bodenkultur, Gregor-Mendelstr. 33, 1180 Vienna, Austria. Phone: 0043-1-47 654/2208. Fax: 0043-1-34 61 76. 2108

VOL. 178, 1996

S-LAYER PROTEIN SYNTHESIS AND CELL WALL COMPOSITION

of the wild-type strain for at least 10 passages at 678C instead of 578C (11, 40). MATERIALS AND METHODS Bacterial strain, growth in continuous culture, and isolation of the p2 vari¨ sterreichants. B. stearothermophilus PV72 was kindly provided by F. Hollaus (O isches Zuckerforschungs-Institut, Tulln, Austria). The organism was grown on 50 ml of SVIII medium (4) containing 5 g of yeast extract, 10 g of peptone, and 5 g of Lab Lemco per liter in 300-ml shaking flasks at 578C to mid-logarithmic growth. Two hundred milliliters of this suspension was used as the inoculum for 5 liters of SVIII medium sterilized in a Braun type Biostat E bioreactor (Braun, Melsungen, Germany). Before inoculation, 20 ml of a sterile glucose solution (6 g of glucose in total) was added. Cultivation was performed at 578C and at a stirring speed of 300 rpm. In continuous culture, the dilution rate was kept at 0.1 h21. For adaptation of the organism to low but constant oxygen supply, the rate of aeration was adjusted to 0.5 liter of air per min during the batch phase and stage 1 of continuous culture. For changing from oxygen-limited to non-oxygen-limited growth conditions in cultures adapted to constant oxygen supply for 160 h (corresponding to 16 volume exchanges), the rate of aeration was first increased from 0.5 to 5.0 liters of air per min in stage 2 of continuous culture and was finally kept constant at a partial O2 pressure (pO2) of 10% during stage 3 (corresponding to about 6.5 liters of air per min). In cultures not adapted to constant oxygen supply, the rate of aeration was increased from 0.5 to 1.5 liters of air per min 10 h after initiation of the continuous culture (corresponding to one volume exchange). The pH value of the culture was kept at 7.2 by addition of either 1 N NaOH or 2 N H2SO4. The cell density was measured at 600 nm in a Beckman model 25 spectrophotometer. Glucose and ammonium were determined with a test kit for glucose (Sigma 315-100) or test sticks for ammonium (Merck 10024-0001). For controlling the homogeneity of the culture, 10-ml samples were taken from the bioreactor at different times. Aliquots were plated on SVIII agar, and the grown biomass was used for sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), freeze-etching, and ultrathin sectioning, which was carried out as described previously (27). The relative amounts from both types of S-layer proteins were estimated from SDS-gels by densitometric evaluation (Elscript 400AT/SM; Hirschmann). Single-cell colonies grown on SVIII agar plates were subjected to SDS-PAGE for final identification. Growth of the p6-deficient variant T5 was performed as described for the wild-type strain. For inducing p2 S-layer protein synthesis, the rate of aeration was increased from 0.5 to 1.5 liters of air per min 10 h after initiation of the continuous culture. For obtaining biomass from the p2 variants derived either from the wild-type strain or from the p6-deficient variant T5, single-cell colonies isolated from continuous culture were transferred to SVIII agar slants and incubated at 578C for 18 h. The inoculum for continuous culture was prepared as described for the wild-type strain. The aeration was adjusted to 5.0 liters of air per min; the dilution rate was 0.1 h21. The biomass was harvested by continuous centrifugation at 14,000 3 g at 48C, washed once with 50 mM Tris-HCl buffer (pH 7.2), stored at 2188C, and used for chemical characterization of the cell wall fragments. To obtain biomass from the wild-type strain and the p6-deficient variant T5, both organisms were grown in continuous culture at a dilution rate of 0.1 h21 and at a rate of aeration of 0.5 liters of air per min. Chemical characterization of the p2 S-layer protein and the peptidoglycancontaining sacculi. Cell wall preparation and isolation and purification of the S-layer protein by recrystallization were performed as described previously (37). The molecular weight of the S-layer protein was estimated from SDS-gels by using a high-molecular-weight protein kit (Sigma, Munich, Germany) as the standard. Protein bands were visualized either by Coomassie blue R250 staining or by silver staining as described previously (33). Peptide mapping of the p2 S-layer protein was done with endoproteinase Glu-C (Staphylococcus aureus V8 protease; Sigma P-8400) under conditions given in a previous report (37). For N-terminal sequencing, 100 to 200 pmol of the p2 S-layer protein was dissolved in 20 to 60 ml of a solution of SDS and trifluoroacetic acid (1 and 5% in distilled water). N-terminal sequencing was performed on a reverse-phase high-pressure liquid chromatography (HPLC) system (Applied Biosystems) as described previously (37). Peptidoglycan-containing sacculi were obtained after extraction of the S-layer protein from cell wall fragments with guanidinium hydrochloride (GHCl; 5 M GHCl in 50 mM Tris-HCl buffer [pH 7.2]) under conditions described previously (37). After the peptidoglycan-containing sacculi were washed at least four times with 50 mM Tris-HCl buffer, 0.5 g (wet weight) of pellet obtained by centrifugation at 40,000 3 g for 20 min at 108C was resuspended in 8 ml of SDS solution (1% in distilled water) and incubated for 30 min at 1008C as described previously (25, 26). Subsequently, the peptidoglycan-containing sacculi were sedimented at 40,000 3 g for 15 min at 108C, washed six times with distilled water, frozen at 2188C, and lyophilized. Amino acid analysis of such purified peptidoglycancontaining sacculi was done as described previously (39). The extent of peptidoglycan cross-linking was determined by the method of Fordham and Gilvarg (15). Determination of phosphate as an assay for teichoic acids was performed by

2109

the method of Ames (2). The occurrence of neutral sugars in peptidoglycancontaining sacculi was investigated by HPLC analysis after hydrolysis with trifluoroacetic acid (33). For recrystallization experiments, GHCl- and SDS-treated, lyophilized peptidoglycan-containing sacculi were mixed with GHCl-extracted S-layer protein and dialyzed for 24 h against distilled water at 208C. Recrystallization of the S-layer protein was examined by ultrathin sectioning and negative staining. To remove putative secondary cell wall polymers, GHCl-extracted and SDS-treated peptidoglycan-containing sacculi from the wild-type strain were incubated with 0.5 M HCl at 608C for 30 min or with formamide at 608C, 1008C, and 1508C, each for 1 h as described by Masuda and Kawata (25). After being washed for at least six times with distilled water, the pellets were subjected to chemical analysis and used for recrystallization of the GHCl-extracted S-layer protein. Production of antisera against the p6- and p2 S-layer proteins and immunoblotting. For producing antisera against the p6 and p2 S-layer proteins purified by recrystallization, 1 mg of each was suspended in 100 ml of 0.1 N NaOH. After addition of 1.8 ml of sodium chloride solution (0.9% in distilled water), the pH was adjusted to 7.0 with 0.1 N HCl. Antisera were produced in two rabbits by subcutaneous injection of samples, each containing 20 mg of S-layer protein. Injections were performed three times at 3-week intervals. The antisera were collected 4 weeks after the final injection by bleeding of the ear vein (12). Inactivation of complement was achieved by heating at 568C for 15 min. The immunospecifity of antisera was checked by the immunoblot procedure (17). Incubation of SDS extracts from S-layer carrying cell wall fragments with the anti-p6 antiserum and the anti-p2 antiserum was carried out at dilutions of 1:3,000 and 1:5,000 in 3% bovine serum albumin in Tris-buffered saline at 378C for 2 h. Detection of proteins that bound the respective antibodies was accomplished by treatment of immunoblots with alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin G at a dilution of 1:2,000 for 1 h (17). Immunogold labelling of whole cells with anti-p2 antiserum. Immunogold labelling was carried out with cells harvested at different stages from continuous culture of the wild-type strain and the p6-deficient variant T5. Cell pellets from a 2-ml cell suspension (optical density [OD] adjusted to 1.5) were incubated with 500 ml of anti-p2 antiserum diluted 1:10 in phosphate-buffered saline (PBS) for 2 h at 208C. After centrifugation of the suspensions at 14,000 3 g for 10 min at 48C, the pellets were washed five times with PBS and incubated with 50 ml of protein A-colloidal gold conjugate (5 nm; Sigma) for 1 h at 208C. Unbound gold conjugate was removed by at least three washes with PBS and distilled water at 208C. Cells were immediately applied to electron microscopy grids which had been rendered hydrophilic by glow discharge. Electron microscopic examination was done with Philips CM12 and CM100 electron microscopes (Philips, Eindhoven, The Netherlands) at 80 kV, using a 30-mm objective aperture. Preparation of nucleic acids and hybridization analysis. Preparation of chromosomal DNA (3) and standard molecular techniques (32) were performed as described previously. All enzymes were used as recommended by the manufacturers. PCR was performed in a Bio-Med Thermocycler 60. Wobble oligonucleotides for PCR and hybridization analysis (for sbsB-specific assays) were derived from the N-terminal and four internal amino acid sequences given in a previous report (37) with the codon usage of the cloned sbsA gene (22, 23). Restriction enzyme-digested chromosomal DNA was electrophoresed, transferred to nylon microfiltration membranes (32), and hybridized at 658C under high-stringency conditions with radiolabelled PCR fragments which had been isolated from agarose gels with Gene Clean (Bio 101, La Jolla, Calif.). Radiolabelling was done by primer extension with random hexamer oligonucleotides (14).

RESULTS Specifity of antisera and immunoblotting. The S-layer proteins used for raising antisera for immunoblotting were isolated from biomass from the wild-type strain of B. stearothermophilus PV72 producing the p6 S-layer protein with a molecular weight of 130,000 and the p2 variant synthesizing the p2 S-layer protein with a molecular weight of 97,000. Figure 1 shows the S-layer proteins before (lanes a and c) and after purification (lanes b and d) by recrystallization. The purified preparations in which no contaminating protein bands could be visualized on SDS-gels by silver staining were used for producing the antisera. The immunospecificity of the anti-p6 antiserum for the p6 S-layer protein and the anti-p2 antiserum for the p2 S-layer protein was finally demonstrated by immunoblotting (Fig. 2, lanes a and c). Both S-layer proteins revealed a strong immunological reaction with the respective antiserum at a dilution of 1:3,000. Only certain contaminating proteins in the cell wall fragments showed a cross-reaction, but no cross-reaction between the p2 S-layer protein and the anti-p6 antiserum (lane b)

2110

´ RA ET AL. SA

J. BACTERIOL.

FIG. 1. SDS-PAGE patterns of purified, silver-stained S-layer proteins from B. stearothermophilus PV72. Lanes: a and b, S-layer protein from the wild-type strain (Mr, 130,000) before (lane a) and after (lane b) purification by recrystallization; (c and d) S-layer protein from the p2 variant (Mr, 97,000) before (lane c) and after (lane d) purification by recrystallization. Samples in lanes b and d were used for preparing polyclonal rabbit antisera.

and the p6 S-layer protein and the anti-p2 antiserum (lane d) was observed. Further confirmation that there was no unspecific binding of the anti-p2 antiserum to the p6 S-layer lattice came from immunogold labelling of whole cells and electron microscopic examination (data not shown). Consequently, the anti-p2 antiserum could be used for visualizing the newly synthesized p2 S-layer subunits on the cell surface of the wild-type strain or the p6-deficient variant T5. Dynamics in oxygen-induced changes in S-layer protein synthesis in B. stearothermophilus PV72 cells adapted to constant oxygen supply in continuous culture. In all experiments, continuous cultures used for examining the dynamics in oxygeninduced changes in S-layer protein synthesis were started 5 h after inoculation at an OD of 1.4. At this stage, the rate of aeration was 0.5 liter of air per min and the dilution rate was kept at 0.1 h21. For adaptation of the wild-type strain of B. stearothermophilus PV72 to low oxygen supply, the organism was grown in continuous culture for 160 h at a rate of aeration of 0.5 liter of air per min. Although during this initial stage (stage 1) both oxygen and glucose were completely metabolized, the production of organic acids (acetic acid and lactic acid) indicated that the organism was grown under oxygen-limited conditions. Control samples taken during the batch phase and at the end of stage 1 of the continuous culture confirmed the homogeneity (see Fig. 3, point A). On SVIII agar plates, uniform 2- to 3-mm colonies typical of the wild-type strain developed. Further, on SDS-gels, the SDS extracts of the biomass revealed the characteristic p6 S-layer protein band with an apparent molecular

FIG. 2. Immunoblots demonstrating the immunospecificity of polyclonal rabbit antisera raised against the S-layer proteins. Lanes: a and b, anti-p6 antiserum applied to the p6 S-layer protein (lane a) and to the p2 S-layer protein (lane b); c and d, anti-p2 antiserum applied to the p2 S-layer protein (lane c) and to the p6 S-layer protein (lane d). In lanes b and d, the faint bands represent contaminations from either the p6 (lane b) or the p2 (lane d) S-layer protein. Dilution of the antiserum was 1:3,000.

FIG. 3. Protocol of growth from B. stearothermophilus PV72 in continuous culture. – – –, OD at 600 nm; – · – · –, acid (expressed in terms of added dosing time in seconds); ——, percent pO2; - - -, percent p2 S-layer protein from total S-layer protein (as derived from SDS-PAGE patterns of whole-cell extracts; see Fig. 4g).

weight of 130,000. In freeze-etched preparations, only the hexagonally ordered S-layer lattice typical of the wild-type strain could be detected (Fig. 4a). Immunogold labelling of whole cells with anti-p2 antiserum gave negative results, demonstrating that only the original p6 S-layer-protein was synthesized (Fig. 4b). Stage 2 of continuous culture was initiated by increasing the rate of aeration to 5.0 liters of air per min, which led to an increase in the OD from 1.2 to 3.2 within the following 5 h (Fig. 3). During this period, the organism started to metabolize amino acids as an additional carbon source, which led to complete utilization of oxygen (pO2 5 0%) and liberation of ammonium being responsible for the constant pH value of the culture. When the OD of the culture was 3.2, acid consumption per unit of time remained constant until the end of stage 2 of continuous culture (Fig. 3, point C), which indicated stabilization of the metabolic pathways. Although during stage 2 of continuous culture oxygen was completely metabolized, immunogold labelling of whole cells with anti-p2 antiserum revealed that p2 S-layer protein synthesis had started approximately 30 h after the rate of aeration was increased (Fig. 3, point B; Fig. 4c). At this time, the density of gold particles on whole cells was very low (Fig. 4c), but labelling of all cells in the culture indicated a rather synchronized initiation of p2 S-layer protein synthesis. However, no changes were observed in SDS-PAGE patterns of whole-cell SDS extracts (Fig. 4g, lanes A to C) or in freeze-etched preparations. This low-level synthesis of p2 S-layer protein, which as a result of the detection limit of protein bands on SDS-gels could represent no more than 3% of the total S-layer protein, remained constant during the following 20 h until the end of stage 2 (Fig. 3, point C). Stage 3 of continuous culture was initiated by adjusting the rate of aeration (6.5 liters of air per min) in a way that a constant pO2 of 10% was maintained in the culture (Fig. 3, points C to E). In response to a surplus of dissolved oxygen in the culture medium, significant changes in SDS-PAGE patterns from whole-cell SDS extracts were observed. A protein band with an apparent molecular weight of 97,000 which after 15 hours represented about 50% of the total S-layer protein content appeared on SDS-gels (Fig. 3, point D; Fig. 4g, lane D). In freeze-etched preparations, the surface from whole cells revealed a granular structure (Fig. 4d). Because all cells in the

VOL. 178, 1996


2111

FIG. 4. Electron micrographs and SDS-PAGE patterns demonstrating oxygen-induced changes in S-layer protein synthesis during growth fromB. stearothermophilus PV72 in continuous culture. Cells were completely covered with the hexagonally ordered S-layer lattice (a) and could not be labelled with anti-p2 antiserum (b) during stage 1 of continuous culture (see also Fig. 3, point A). Immunogold labelling (c) confirmed that the beginning of p2 S-layer protein synthesis was 30 h after an increase in the rate of aeration to 5.0 liters of air per min (see Fig. 3, point B) and stayed at this low level until the rate of aeration was further increased (see Fig. 3, point C). In panel d (corresponding to point D in Fig. 3), the hexagonally ordered S-layer lattice was replaced by an amorphous layer which could uniformly be labelled with anti-p2 antiserum (e). Only on a few cells, separate patches with oblique and hexagonal lattice symmetry were arranged in a monomolecular layer (f). Bars, 100 nm. On SDS gels (g), the p2 S-layer protein (Mr, 97,000) became visible in whole-cell extracts. Lanes A to E correspond to points A to E in Fig. 3. Molecular weights are indicated in thousands.

2112

´ RA ET AL. SA

J. BACTERIOL.

TABLE 1. Decrease in p6 S-layer protein content of whole cells from B. stearothermophilus PV72 in continuous culture at a dilution rate of 0.1 h21 after increasing the rate of aeration from 0.5 to 1.5 liters of air per mina % p6 S-layer protein Time (h)

0 5 7 10 17 24 40

OD (600 nm) Calculated

SDS-PAGE

100 61 50 36 18 9 2

100 60 ND ND 20 ,5 ND

1.6 1.8 ND ND 1.9 1.9 1.9

a This point of time corresponds to time zero. Values given were either calculated according to equation 1 or estimated from SDS-PAGE patterns by densitometric evaluation of whole-cell extracts from samples taken at different time from continuous culture. ND, not determined.

culture could uniformly be labelled with anti-p2 antiserum (Fig. 4e) and both S-layer proteins were present on SDS-gels in comparable amounts (Fig. 4g, lane D), synchronous synthesis of the p6 and the p2 S-layer proteins within single cells must have occurred. The random distribution of the gold particles on the cell surface indicated that there was no preference in insertion of the new type of S-layer protein into the existing hexagonally ordered lattice (Fig. 4e). Samples taken from continuous culture up to 35 h later (Fig. 3, point E) led to identical results (Fig. 4e and g, lane E). Only a few cells revealed patches with either hexagonal or oblique lattice symmetry in a monomolecular layer (Fig. 4f). According to the equation x 5 x0 z e2Dt,

(1)

describing dilution in continuous culture, the amount of p6 S-layer protein should have decreased to 50% within 7 h if production was immediately stopped at the time when p2 S-layer protein synthesis was induced (this calculation does not consider the slight increase in cell density during stage 3 of continuous culture). The time actually required was approximately twice as much (15 h), which was ascribed to reducing p6 S-layer protein synthesis to a level of about 50% and starting p2 S-layer protein synthesis to a comparable extent. Dynamics in changes in S-layer protein synthesis in B. stearothermophilus PV72 cells not adapted to constant oxygen supply in continuous culture. The batch and continuous culture was started under conditions described above. After one volume exchange in continuous culture when the cell density had reached an OD of 1.6, the rate of aeration was increased from 0.5 to 1.5 liters of air per min. Since at this stage amino acids were metabolized as an additional carbon source, the pO2 rapidly decreased to zero. The liberation of ammonium and production of organic acids were responsible for the constant pH value of the culture. A slight increase in cell density, from 1.6 to 1.9, was observed during the following 5 h (Table 1). According to equation 1, the theoretical decrease in the p6 S-layer protein content from whole cells was calculated by assuming that its synthesis was immediately stopped upon increasing the rate of aeration. The calculated values for the p6 S-layer protein content (Table 1) fitted well those determined by SDS-PAGE for biomass samples collected 5, 17, 24, and 40 h after the rate of aeration was increased (Fig. 5). These data clearly showed that p6 S-layer protein synthesis was immediately stopped upon an increase in the rate of aeration in cul-

tures not adapted to low oxygen supply and was completely replaced by p2 S-layer protein synthesis. Induction of p2 S-layer protein synthesis in continuous culture of the p6-deficient variant T5 not adapted to constant oxygen supply. Cultivation conditions for the p6-deficient variant T5 were identical to those for the wild-type strain. The rate of aeration was increased from 0.5 to 1.5 liters of air per min 10 h after initiation of the continuous culture. Subsequently, the cell density, the pO2, and the pH value of the culture changed in the way described for the wild-type strain. As confirmed by immunogold labelling, p2 S-layer subunits could be localized on the cell surface for the first time 3 h after the rate of aeration was increased, which was considered to be the beginning of p2 S-layer protein synthesis. However, at this time, the corresponding p2 S-layer protein band was not detectable on SDS-gels, nor was a crystalline structure present on the surface of freeze-etched whole cells (Fig. 6a). Ten hours after the rate of aeration was increased, randomly distributed crystalline patches characteristic of the p2 S-layer protein lattice could be localized on the cell surface (Fig. 6b). A complete coverage of the cell surface with the p2 S-layer protein lattice was observed 24 h after the rate of aeration was increased (Fig. 6c). The appearance of the p2 S-layer protein with an apparent molecular weight of 97,000 was confirmed by SDS-PAGE (Fig. 7). As described for the wild-type strain, the time course of covering the cell surface with p2 S-layer protein fitted well the theoretical values calculated according to equation 1 (Table 1). Stability of the p2 variants. Both the p2 variant which developed in the course of continuous culture of the wild-type strain and the p6-deficient variant T5 showed stable growth without changes in S-layer protein synthesis on SVIII medium in shaking flasks, in batch culture, in continuous culture, and on agar plates. Under oxygen-limited growth conditions, glucose was used as the preferred carbon source. Even after several passages under oxygen-limited growth conditions, the p6 S-layer protein was not synthesized anymore. Chemical characterization of the p2 S-layer proteins. The p2 S-layer proteins synthesized by the wild-type strain and the p6-deficient variant T5 were compared in the present study. Both N-terminal sequencing (Ala-Ser-Phe-Thr-Asp-Val-AlaPro-Gln-Tyr-Lys-Asp-Ala-Ile) and peptide mapping with endoproteinase Glu-C (Fig. 8) confirmed that the S-layer proteins were identical. Chemical characterization of peptidoglycan-containing sacculi. Data from chemical characterization of the peptidoglycan-containing sacculi from the wild-type strain, the p6-deficient variant T5, and the p2 variants derived from both organisms are summarized in Table 2. The molar ratios of

FIG. 5. SDS-PAGE patterns of SDS extracts from whole cells of B. stearothermophilus PV72 demonstrating oxygen-induced changes in S-layer protein synthesis. Cells were not adapted to low oxygen supply and were harvested before (0 h) and 5, 17, and 24 h after an increase in the rate of aeration from 0.5 to 1.5 liters of air per min.

VOL. 178, 1996


2113

FIG. 7. SDS-PAGE patterns of whole-cell extracts from the p6-deficient variant T5. Samples were taken before (lane a) and 24 h after (lane b) an increase in the rate of aeration. In lane b, the p2 S-layer protein (Mr, 97,000) is visible.

acid was 1.5:1. For peptidoglycan-containing sacculi of the p2 variants, a molar ratio of even 2.5:1 was determined. This excess N-acetylglucosamine corresponded to 6.1% of the dry weight from peptidoglycan-containing sacculi from the wildtype strain or the p6-deficient variant T5 and represented 8.8% of the dry weight from peptidoglycan-containing sacculi from the p2 variants. No difference could be observed in the extent of peptidoglycan cross-linking, which was in the range of 50%. The glucose content of peptidoglycan-containing sacculi from the wild-type strain and the p6-deficient variant T5 was 5.1%, whereas only 1.8% glucose was detected in peptidoglycan-containing sacculi of both p2 variants. The excess of Nacetylglucosamine and glucose, which together represented about 10% of the peptidoglycan-containing sacculus dry weight, was attributed to the presence of a secondary cell wall polymer. In such a polymer, the molar ratio of N-acetylglucosamine to glucose would be 1:1 for the wild-type strain and the p6-deficient variant T5 and change to 4:1 in the p2 variants. Recrystallization of the isolated p6 and p2 S-layer protein onto peptidoglycan-containing sacculi. Recrystallization of the isolated p6 and p2 S-layer protein onto fragments of peptidoglycan-containing sacculi was performed for studying the binding specificity of the S-layer proteins to the respective supporting layer. Generally, the peptidoglycan-containing sacculi exhibited no differences as templates for S-layer protein recrystallization when treated with SDS after the GHCl extraction procedure. Ultrathin sectioning and negative staining revealed that in vitro recrystallization of isolated p6 S-layer protein onto peptidoglycan-containing sacculi from the wild-type strain, the p6deficient variant T5, and the p2 variants was possible. Both faces of the peptidoglycan-containing sacculi were completely covered with the hexagonally ordered S-layer lattice typical of the wild-type strain (not shown). In contrast, isolated p2-S-

FIG. 6. Electron micrographs demonstrating changes on the cell surface of the p6-deficient variant T5. During stage 1 of continuous culture, the cell surface revealed an amorphous structure (a). Ten hours after an increase in the rate of aeration, crystalline patches (arrows) were visible (b); the patches completely covered the cell surface (c) 24 h after an increase in the rate of aeration. Bars, 100 nm.

N-acetylmuramic acid, alanine, glutamic acid, and diaminopimelic acid were identical for all peptidoglycans and corresponded to the directly cross-linked meso-diaminopimelic acid type (31). The only exception was found for N-acetylglucosamine. In case of the wild-type strain and the p6-deficient variant T5, the molar ratio of N-acetylglucosamine to muramic

FIG. 8. SDS-PAGE patterns of the p2 S-layer protein synthesized either by the wild-type strain (lane a) or by the p6-deficient variant T5 (lane b) after proteolytic cleavage with endoproteinase Glu-C. Molecular weights are given in thousands.

2114

´ RA ET AL. SA

J. BACTERIOL.

TABLE 2. Chemical analysis of peptidoglycan-containing sacculi from B. stearothermophilus PV72 (wild-type strain), the p6-deficient variant T5, and p2 variants derived either from the wild-type strain or from the p6-deficient variant T5a Molar ratio of constituents Strain or variant

Wild-type strain and p6-deficient variant T5 p2 variants

MurNAc to Glu

Ala to Glu

DAP to Glu

GlcNAc to MurNAc

Extent of cross-linking (%)

0.99 0.96

1.51 1.54

1.01 1.10

1.52 2.45

52 57

a The surplus of N-acetylglucosamine in comparison with muramic acid was attributed to the presence of secondary cell wall polymers. MurNAc, N-acetylmuramic acid; Glu, glutamic acid; Ala, alanine; DAP, meso-diaminopimelic acid; GlcNAc, N-acetylglucosamine.

layer subunits were capable of binding and recrystallizing into an oblique lattice onto peptidoglycan-containing sacculi from which they were originally extracted but could not bind to those from the wild-type strain. Since only the molar ratio of excess N-acetylglucosamine to glucose had changed upon the switch, these results strongly indicated that both types of sugars could be involved in binding of the S-layer protein to the rigid cell wall layer. This assumption was finally supported by extraction and recrystallization experiments performed with peptidoglycan-containing sacculi from the wild-type strain and the p6 S-layer protein. By using 0.5 M HCl at 608C and formamide at 60 and 1008C as extracting agents, up to 60% of glucose and excess N-acetylglucosamine could be removed from the peptidoglycan-containing sacculi, which had no effect on the recrystallization of the p6 S-layer protein. When more than 90% of N-acetylglucosamine and glucose was extracted with formamide at 1508C, binding and recrystallization of the p6 S-layer protein ceased. The integrity of the peptidoglycan after the various extraction procedures was examined by chemical analysis and electron microscopy. Since neither the molar ratio of the peptidoglycan constituents nor the morphology of the sacculi was changed upon the different extraction procedures (Table 3), a cell wall polymer consisting of glucose and N-acetylglucosamine in a molar ratio of 1:1 seems to be responsible for binding of the p6 S-layer protein to the peptidoglycan-containing sacculi from the wild-type strain. Hybridization analysis. To obtain more insight into the genomic organization of the genes encoding the p6-S-layer protein (sbsA [22, 23]) and the p2 S-layer protein (sbsB) from B. stearothermophilus PV72, Southern hybridization analysis was performed. For this purpose, the total genomic DNA from the wild-type strain, the p6-deficient variant T5, and the p2 variant (derived from the wild-type strain) was isolated and subjected to restriction enzyme digestion with the enzymes BamHI and HindIII. For hybridization analysis, two sbsA-

specific PCR fragments were constructed by using sbsA-specific primers (22, 23). Probe A corresponded to the N-terminal region (260 bp) of sbsA, and probe B corresponded to the C-terminal region (1,700 bp). The sbsB-specific probe C was constructed by PCR with wobble oligonucleotides derived from the previously determined N-terminal region and one of four internal amino acid sequences (37). The resulting PCR fragment (1,100 bp) was gel purified. To confirm the specifity of this fragment, oligonucleotides derived from the three remaining internal amino acid sequences were used to hybridize to the PCR fragment. Two of these oligonucleotides which hybridized with the 1,100-bp PCR fragment, corresponding to the sbsB gene N-terminal sequences, were amplified and further used in hybridization studies. Hybridizations were performed separately with radiolabelled probes A, B, and C under stringent conditions. As shown in Fig. 9A, the wild-type strain (lanes 3 and 4) and the p6-deficient variant T5 (lanes 5 and 6) contained one copy of the N-terminal region of sbsA before induction of sbsB expression. The hybridization patterns of the wild-type strain and the p6-deficient variant T5 were different, indicating deletion of 800 bp in the 59 upstream region of sbsA in the p6-deficient variant T5. Probe B, corresponding to the C-terminal region of sbsA, hybridized strongly to two bands in the BamHI-restricted DNA of the wild-type strain (Fig. 9B, lanes 3 and 4) and strongly to one band of the p6-deficient variant T5 (Fig. 9B, lanes 5 and 6). With probe C, correspond-

TABLE 3. Results from recrystallization experiments of the p6 S-layer protein from B. stearothermophilus PV72 onto peptidoglycancontaining sacculi treated with SDS (native) and after extraction with HCl or formamide at different temperaturesa Condition

Native 0.5 M HCl, 608C Formamide 608C 1008C 1508C

Ala/DAP ratio

Glc (mg/mg [dry wt])

1.48 1.48

51 16

60

50

1

1.47 1.45 1.48

36 36 ,5

30 30 .90

50 60 .90

1 1 2

% Glc extracted

% GlcNAc extracted

Recrystallization

a Ala, alanine; DAP, meso-diaminopimelic acid; Glc, glucose; GlcNAc, Nacetylglucosamine. Integrity checked by negative staining was positive in all cases.

FIG. 9. Hybridization of the radiolabelled N-terminal sbsA fragment (A), the C-terminal sbsA fragment (B), and the N-terminal sbsB fragment (C) to the total genomic DNA of the p2 variant (lanes 1 and 2), the wild-type strain (lanes 3 and 4), and the p6-deficient variant T5 (lanes 5 and 6). Chromosomal DNA was digested with BamHI (lanes 1, 3, and 5) and HindIII (lanes 2, 4, and 6). The positions of BstEII-digested lambda phage DNA size markers (in kilobases) are indicated at the left.

VOL. 178, 1996


ing to the N-terminal region of sbsB (Fig. 9C), a hybridization signal was obtained with the p2 variant (Fig. 9C, lanes 1 and 2) but not with the wild-type strain (Fig. 9C, lanes 3 and 4) and the p6-deficient variant T5 (Fig. 9C, lanes 5 and 6). On the other hand, the sbsA gene could not be detected in the p2 variant (Fig. 9A and B, lanes 1 and 2). Identity of the wild-type strain and the variants. Sequence analysis of the 16S rRNA gene in the variable region of about 450 bases performed with the wild-type strain and the variants confirmed that they are 100% identical. DISCUSSION Continuous culture under well-defined conditions clearly demonstrated that stable synthesis of the p6 S-layer protein from the wild-type strain of B. stearothermophilus PV72 was possible only under oxygen-limited growth when glucose was used as the sole carbon source. When the rate of aeration was increased, low-level p2 S-layer protein synthesis (maximum of 3% of the total S-layer protein) was observed in cultures that had been adapted to low oxygen supply. By further increasing the rate of aeration to a pO2 of 10%, prolonged synchronous synthesis of comparable amounts of both types of S-layer proteins within single cells could be induced. The time course of covering the cell surface with p2 S-layer protein confirmed that down regulation of p6 S-layer protein synthesis to half and induction of p2 S-layer protein synthesis to a comparable extent must have occurred. When samples from such cultures were plated on SVIII agar, single-cell colonies showing either the p6 or the p2 S-layer protein but not mixed colonies with both types of S-layer proteins developed. This result indicated that synchronous S-layer protein synthesis is not a stable process in B. stearothermophilus PV72 cells. It seems that the phase of transition from the wild-type strain to the p2 variant could artificially be extended under steady-state conditions. A similar observation that low-level expression of alternative S-layer genes is possible was reported for the first time for the pathogenic Campylobacter fetus strains (8, 16, 30). The major S-layer protein from three clinical isolates showed a molecular weight of either 149,000, 127,000, or 98,000, but the presence of low amounts of the other S-layer proteins on SDS-gels indicated that individual strains can express up to three S-layer proteins. Although changes in S-layer protein synthesis were interpreted as antigenic variation, which is a mechanism used among pathogens to survive the bactericidal activity of the immune system (9, 50), the determinants of which of the Slayer proteins was predominantly expressed are still unknown. The molecular mechanism for changes in S-layer protein expression could be elucidated for C. fetus 23D (47, 48). This strain produced a predominant S-layer protein with a molecular weight of 98,000 but produced a 127,000-molecular-weight S-layer protein after subculturing on blood agar. The hybridization data indicated that changes in S-layer protein synthesis were accompanied by chromosomal rearrangement and sitespecific reciprocal recombination. The results from Southern hybridization obtained in this study for B. stearothermophilus PV72 suggest that chromosomal rearrangement is responsible for S-layer protein variation in this organism as well. The 59 region of the sbsB gene was found only in the p2 variant which developed under conditions of increased oxygen supply. The inability to detect the sbsA gene in the p2 variant and, conversely, the absence of sbsB in the wild-type strain and the p6-deficient variant T5 before induction of sbsB expression strongly indicate that rearrangement of partial coding sequences had occurred. Hybridization (Fig. 9) as well as PCR analysis (not shown) of the sbsA gene

2115

in the p6-deficient variant T5 revealed a full-length copy of the structural sbsA gene with differences in the 59 upstream promoter region. In comparison with the 59 upstream region of sbsA in the wild-type strain, an 800-bp deletion had occurred. This finding suggests that the inability of the p6-deficient variant T5 to synthesize the p6 S-layer protein must be related to the 59 end of sbsA or adjacent upstream regions. A similar observation was made for the S-layer proteins of C. fetus (47) and Aeromonas salmonicida (5). Considering that the p6-deficient variant T5 is missing the 59 upstream region of sbsA which corresponds to the regulatory region but that expression of sbsB after increasing oxygen supply was possible, more than one promoter must exist for S-layer gene expression in B. stearothermophilus PV72. These findings are in contrast to the situation of S-layer protein variation in C. fetus. For this organism, it was suggested that variation involves rearrangement of silent gene cassettes into a single expression locus (10). Stable synchronous synthesis of comparable amounts of two different S-layer proteins that were arranged in a monomolecular layer was described for the pathogenic Clostridium difficile (19, 45). Although the S-layer proteins from this organism revealed different molecular weights and showed no crossreaction with polyclonal antisera, they were capable of assembling into a square S-layer lattice. In contrast, a common in vivo or in vitro self-assembly could never be observed for the two S-layer proteins produced by B. stearothermophilus PV72. Organisms with two superimposed S-layers such as Aquaspirillum serpens (20, 12) and Bacillus brevis 47 (46, 49) are capable of synchronous synthesis of structurally different S-layer proteins that assemble sequentially. The genes encoding the outer and middle wall protein from B. brevis 47 are preceded by a common promoter, but the transcript revealed two ribosomebinding sites (1). When the rate of aeration was increased after the batch phase or after the initiation of continuous culture of B. stearothermophilus PV72, p6 S-layer protein synthesis was immediately stopped and was completely replaced by that of p2 S-layer protein. This observation confirmed that cells which were not adapted to low oxygen supply were much more sensitive to higher oxygen concentrations. Considering the solubility of oxygen at elevated temperatures, it can be assumed that thermophilic aerobic bacteria are adapted to very low concentrations only in their natural habitats (44). Increasing the rate of aeration in bacterial cultures not only led to an increase in oxygen tension in the medium but also was found to be responsible for increased levels of intracellular oxygen radicals such as superoxide anion, hydrogen peroxide, and the highly reactive hydroxyl radical which are formed as by-products of the normal aerobic metabolism (51). Studies with Escherichia coli and Salmonella typhimurium showed that cells adapted to sublethal doses of hydrogen peroxide were less sensitive to toxic doses because of induction of a 10-fold increase in catalase in E. coli (18) or production of increased levels of catalase and superoxide dismutase in S. typhimurium (13). A similar observation that pretreatment of Bacillus subtilis with sublethal concentrations of hydrogen peroxide protected the cells against a later challenge by lethal concentrations was reported by Murphy et al. (29). The S-layer protein from the wild-type strain and the p2 variants showed different N-terminal regions and led to quite different cleavage products upon peptide mapping, indicating that they are encoded by different genes (37). A significant change upon switching to the p2 variants was also observed in the molar ratio of excess N-acetylglucosamine to glucose in peptidoglycan-containing sacculi, which are most probably constituents of secondary cell wall polymers. However, the

2116

´ RA ET AL. SA

J. BACTERIOL.

peptidoglycan composition and the extent of peptidoglycan cross-linking were unaffected by the switch. Extraction and recrystallization experiments revealed that at least for the wildtype strain, N-acetylglucosamine and glucose are involved in binding of the S-layer protein to the peptidoglycan-containing sacculi. More recently, a polymer with a molecular weight in the range of 20,000 composed of both types of sugars could be isolated from the peptidoglycan-containing sacculi (30a). A similar observation that secondary cell wall polymers are responsible for binding of the S-layer protein to the rigid cell wall layer was reported for Lactobacillus buchneri (25, 26). Many of the S-layer proteins sequenced to date showed a high homology at the N-terminal region, leading to the assumption that S-layer homologous domains could be responsible for binding the subunits to the peptidoglycan (24). Such a relatively unspecific binding mechanism would imply that at least within one species, the S-layer proteins would recognize the peptidoglycan-containing sacculi from different strains, which to date has not been experimentally demonstrated. ACKNOWLEDGMENTS This work was supported by the Austrian Science Foundation (projects S72/02 and S72/08) and by the Federal Ministry of Science, Research and the Arts, Republic of Austria. We thank Christoph Hotzy for excellent technical assistance and Sonja Zayni for amino acid and sugar analysis. REFERENCES 1. Adachi, T., H. Yamagata, N. Tsukagoshi, and S. Udaka. 1989. Multiple and tandemly arranged promoters of the cell wall protein gene operon in Bacillus brevis 47. J. Bacteriol. 171:1010–1016. 2. Ames, B. N. 1965. Assay for inorganic phosphate, total phosphate and phosphate. Methods Enzymol. 8:115–118. 3. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl (ed.). 1987. Current protocols in molecular biology. John Wiley & Sons, New York. 4. Bartelmus, W., and F. Perschak. 1957. Schnellmethode zur Keimzahlbestimmung in der Zuckerindustrie. Z. Zuckerind. 7:276–281. 5. Belland, R. J., and T. J. Trust. 1987. Cloning of the gene for the surface array protein of Aeromonas salmonicida and evidence linking loss of expression with genetic deletion. J. Bacteriol. 169:4086–4091. 6. Beveridge, T. J. 1994. Bacterial surface-layers. Curr. Opin. Struct. Biol. 4: 204–212. 7. Beveridge, T. J., and S. F. Koval (ed.). 1993. Advances in paracrystalline surface layers. Plenum Press, New York. 8. Blaser, M. J. 1993. Biology of Campylobacter fetus S-layer proteins, p. 173– 180. In T. J. Beveridge and S. F. Koval (ed.), Advances in paracrystalline surface layers. Plenum Press, New York. 9. Dubreuil, J. D., M. Kostrzynska, J. W. Austin, and T. J. Trust. 1990. Antigenic differences among Campylobacter fetus S-layer proteins. J. Bacteriol. 172:5035–5043. 10. Dworkin, J., M. K. R. Tummuru, and M. J. Blaser. 1995. A lipopolysaccharide binding domain of Campylobacter fetus S-layer protein resides within the conserved N-terminus of a family of silent and divergent genes. J. Bacteriol. 177:1734–1741. 11. Eder, J. 1983. Versuche zur Aufkla¨rung der Funktion parakristalliner Proteinmembranen bei Bacillus stearothermophilus. Ph.D. thesis. University of Agriculture, Vienna. 12. Egelseer, E., I. Schocher, M. Sa ´ra, and U. B. Sleytr. 1995. The S-layer from Bacillus stearothermophilus DSM 2358 functions as an adhesion site for a high-molecular-weight amylase. J. Bacteriol. 177:1444–1451. 13. Farr, S. B., and T. Kogoma. 1991. Oxidative stress responses in Escherichia coli and Salmonella typhimurium. Microbiol. Rev. 55:561–585. 14. Feinberg, A., and B. Vogelstein. 1983. A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132:6–13. 15. Fordham, W. D., and C. Gilvarg. 1974. Kinetics of cross-linking of peptidoglycan in Bacillus megaterium. J. Biol. Chem. 249:2478–2482. 16. Fujimoto, S., A. Takade, K. Amako, and M. J. Blaser. 1991. Correlation between molecular size of the surface array protein and morphology and antigenicity of the Campylobacter fetus S layer. Infect. Immun. 59:2017– 2022. 17. Harlow, E., and D. Lane. 1988. Antibodies: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

18. Imlay, J. A., and S. Linn. 1987. Mutagenesis and stress responses induced in Escherichia coli by hydrogen peroxide. J. Bacteriol. 169:2967–2976. 19. Kawata, T., A. Takeoka, K. Takumi, and K. Masuda. 1984. Demonstration and preliminary characterization of a regular array in the cell wall of Clostridium difficile. FEMS Microbiol. Lett. 24:323–328. 20. Kist, M., and R. G. E. Murray. 1984. Components of the regular surface array of Aquaspirillum serpens MW5 and their assembly in vitro. J. Bacteriol. 157:599–606. 21. Koval, S. F., and R. G. E. Murray. 1981. Cell wall proteins of Aquaspirillum serpens. J. Bacteriol. 146:1083–1090. 22. Kuen, B., M. Sa ´ra, and W. Lubitz. Heterologous expression and self-assembly of the S-layer protein SbsA of Bacillus stearothermophilus in Escherichia coli. Mol. Microbiol., in press. 23. Kuen, B., U. B. Sleytr, and W. Lubitz. 1994. Sequence analysis of the sbsA gene encoding the 130 kDa surface layer protein of Bacillus stearothermophilus PV72. Gene 145:115–120. 24. Lupas, A., H. Engelhardt, J. Peters, U. Santarius, S. Volker, and W. Baumeister. 1994. Domain structure of the Acetogenium kivui surface layer revealed by electron crystallography and sequence analysis. J. Bacteriol. 176: 1224–1233. 25. Masuda, K., and T. Kawata. 1981. Characterization of a regular array in the wall of Lactobacillus buchneri and its reattachment to the other wall components. J. Gen. Microbiol. 124:81–90. 26. Masuda, K., and T. Kawata. 1985. Reassembly of a regularly arranged protein in the cell wall of Lactobacillus buchneri and its reattachment to cell walls: chemical modification studies. Microbiol. Immunol. 29:927–938. 27. Messner, P., F. Hollaus, and U. B. Sleytr. 1984. Paracrystalline cell wall surface layers of different Bacillus stearothermophilus strains. Int. J. Syst. Bacteriol. 34:202–210. 28. Messner, P., and U. B. Sleytr. 1992. Crystalline bacterial cell-surface layers. Adv. Microb. Physiol. 33:213–275. 29. Murphy, P., B. Dowds, D. McConnell, and K. M. Devine. 1987. Oxidative stress and growth temperature in Bacillus subtilis. J. Bacteriol. 169:5766– 5770. 30. Pei, Z., and M. J. Blaser. 1990. Pathogenesis of Campylobacter infections. Role of surface array proteins in virulence in a mouse model. J. Clin. Invest. 85:1036–1043. 30a.Ries, W. Unpublished observation. 31. Rogers, H. J., H. R. Perkins, and J. B. Ward. 1980. Microbial cell walls and membranes. Chapman and Hall, London. 32. Sambrook, J., E. F. Fristch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 33. Sa ´ra, M., D. Pum, S. Ku ¨pcu ¨, P. Messner, and U. B. Sleytr. 1994. Isolation of two physiologically induced variant strains of Bacillus stearothermophilus NRS 2004/3a and characterization of their S-layer lattices. J. Bacteriol. 176: 848–860. 34. Sa ´ra, M., and U. B. Sleytr. 1987. Production and characteristics of ultrafiltration membranes with uniform pores from two-dimensional arrays of proteins. J. Membr. Sci. 33:27–49. 35. Sa ´ra, M., and U. B. Sleytr. 1987. Charge distribution on the S layer of Bacillus stearothermophilus NRS 1536/3c and importance of charged groups for morphogenesis and function. J. Bacteriol. 169:2804–2809. 36. Sa ´ra, M., and U. B. Sleytr. 1987. Molecular sieving through S layer of Bacillus stearothermophilus strains. J. Bacteriol. 169:4092–4098. 37. Sa ´ra, M., and U. B. Sleytr. 1994. Comparative studies of S-layer proteins from Bacillus stearothermophilus strains expressed during growth in continuous culture under oxygen-limited and non-oxygen-limited conditions. J. Bacteriol. 176:7182–7189. 38. Schuster, K. C. 1994. Kultivierung von Bacillus stearothermophilus PV72 mit definierten Zellwandeigenschaften fu ¨r biotechnologische Anwendungen von S-Schichten. Ph.D. thesis. University of Agriculture, Vienna. 39. Schuster, K. C., H. F. Mayer, R. Kieweg, W. Hampel, and M. Sa ´ra. 1995. A synthetic medium for continuous culture of the S-layer carrying Bacillus stearothermophilus PV72 and studies on the influence of growth conditions on cell wall properties. Biotechnol. Bioeng. 48:66–77. 40. Sleytr, U. B., P. Messner, D. Pum, and J. Eder. 1982. Struktur und Morphogenese periodischer Proteinmembranen bei Bakterien. Mikroskopie 39:215– 232. 41. Sleytr, U. B., P. Messner, D. Pum, and M. Sa ´ra. 1993. Crystalline bacterial cell surface layers. Mol. Microbiol. 10:911–916. 42. Sleytr, U. B., P. Messner, D. Pum, and M. Sa ´ra (ed.). 1988. Crystalline bacterial cell surface layers. Springer, Berlin. 43. Sleytr, U. B., M. Sa ´ra, Z. Ku ¨ pcu ¨ , and P. Messner. 1986. Structural and chemical characterization of S-layers of selected strains of Bacillus stearothermophilus and Desulfotomaculum nigrificans. Arch. Microbiol. 146: 19–24. 44. Sonnleitner, B. 1983. Biotechnology of thermophilic bacteria—growth, products and application. Adv. Biochem. Eng. Biotechnol. 28:69–138. 45. Takeoka, A., K. Takumi, T. Koga, and T. Kawata. 1991. Purification and characterization of S layer proteins from Clostridium difficile GAI 0714. J. Gen. Microbiol. 137:261–267.

VOL. 178, 1996


46. Tsuboi, A., R. Uchihi, R. Tabata, Y. Takahashi, H. Hashiba, T. Sasaki, H. Yamagata, N. Tsukagoshi, and S. Udaka. 1986. Characterization of the genes coding for two major cell wall proteins from protein-producing Bacillus brevis 47: complete nucleotide sequence of the outer wall protein gene. J. Bacteriol. 168:365–373. 47. Tummuru, M. K. R., and M. J. Blaser. 1992. Characterization of the Campylobacter fetus sapA promoter: evidence that the sapA promoter is deleted in spontaneous mutant strains. J. Bacteriol. 174:5916–5922. 48. Tummuru, M. K. R., and M. J. Blaser. 1993. Rearrangement of sapA homologs with conserved and variable regions in Campylobacter fetus. Proc.

2117

Natl. Acad. Sci. USA 90:7265–7269. 49. Udaka, S., N. Tsukagoshi, and H. Yamagata. 1989. Bacillus brevis, a host bacterium for efficient extracellular production of useful proteins. Biotechnol. Gen. Eng. Rev. 7:113–146. 50. Wang, E., M. M. Garcia, M. S. Blake, Z. Pei, and M. J. Blaser. 1993. Shift in S-layer protein expression responsible for antigenic variation in Campylobacter fetus. J. Bacteriol. 175:4979–4984. 51. Watson, K. 1990. Microbial stress proteins. Adv. Microb. Physiol. 31:183– 223.