Electron Microscopy Studies of Cell-Wall-Anchored Cellulose (Avicel)-Binding Protein (AbpS) from Streptomyces reticuli

Abstract

Streptomyces reticuli produces a 35-kDa cellulose (Avicel)-binding protein (AbpS) which interacts strongly with crystalline cellulose but not with soluble types of cellulose. Antibodies that were highly specific for the NH₂-terminal part of AbpS were isolated by using truncated AbpS proteins that differed in the length of the NH₂ terminus. Using these antibodies for immunolabelling and investigations in which fluorescence, transmission electron, or immunofield scanning electron microscopy was used showed that the NH₂ terminus of AbpS protrudes from the murein layer of S. reticuli. Additionally, inspection of ultrathin sections of the cell wall, as well as biochemical experiments performed with isolated murein, revealed that AbpS is tightly and very likely covalently linked to the polyglucane layer. As AbpS has also been found to be associated with protoplasts, we predicted that a COOH-terminal stretch consisting of 17 hydrophobic amino acids anchors the protein to the membrane. Different amounts of AbpS homologues of several Streptomyces strains were synthesized.

Streptomycetes are gram-positive mycelium-forming bacteria which are very abundant in soil (1). They are optimally adapted to their natural environment, as they produce spores and are resistant to heat, dryness, and cold (12). Production of antibiotics and fungicides inhibits the growth of competing organisms. Many streptomycetes are able to degrade biopolymers, the most abundant carbon sources in soil (3). Starch, xylan, chitin, and cellulose are efficiently hydrolyzed due to the action of extracellular enzymes (18). A number of these enzymes have been well-characterized biochemically, and their genes have been identified. In contrast, there have been only a few studies on the regulation of these genes and the related signal transduction cascade, although monitoring the changes in external conditions and the ensuing intracellular response are the most important prerequisites for synthesis of specific catabolic enzymes. Bacteria have evolved several signalling systems for recognizing a variety of soluble substances, including the two-component systems (17) and the Fec system to regulate iron transport (4).

Contact between bacteria and surfaces is an additional signal which stimulates intracellular responses. Thus, Pseudomonas aeruginosa cells have been shown to produce an extracellular alginate matrix that protects the bacterium from antibiotics and the human defense system (6). Transcription of the algC gene (which encodes enzymes essential for the synthesis of the alginate matrix) is activated only by contact between the cells and Teflon or glass (6). Vibrio parahaemolyticus synthesizes accessory lateral flagella only when it is cultivated on agar surfaces. In liquid media, transcription of the flagellum-encoding gene laf is repressed (2).

Only when the mycelia of Streptomyces reticuli have been in contact with crystalline cellulose (Avicel) does the strain produce a cellulase (Avicelase or Cell) that is able to efficiently hydrolyze the biopolymer (22, 23, 30, 31).

Recently, we identified a 35-kDa Avicel-binding protein (AbpS) which interacts strongly with crystalline forms of cellulose (32). Other biopolymers are weakly recognized (chitin and Valonia cellulose) or not recognized at all (xylan, starch, and agar). The corresponding gene (abpS) was identified and sequenced. By analyzing the secondary structure of the deduced AbpS sequence, we found a large centrally located α-helical structure exhibiting low levels of homology with the tropomyosin protein family and the streptococcal M-proteins. In addition, it was predicted that a 17-amino-acid hydrophobic stretch which represents a putative transmembrane segment is present at the C-terminal end. In vivo labelling with fluorescein isothiocyanate (FITC), FITC-labelled secondary antibodies, and proteinase K treatment revealed that the protein is anchored to the cell wall and protrudes from the surface of the hyphae. Physiological studies showed that AbpS is synthesized during the late logarithmic phase, independent of the carbon source (32).

Furthermore, we examined the distribution of AbpS on the hyphae by performing immunomicroscopic investigations, demonstrated that AbpS interacts with peptidoglycan by biochemical studies, and investigated the occurrence of AbpS homologues in streptomycetes. The results are presented in this paper.

MATERIALS AND METHODS

Bacterial strains, plasmids, and cultivation.

Wild-type strain S. reticuli Tü45 described by Wachinger et al. (29) was obtained from H. Zähner, Tübingen, Germany. Streptomycetes were cultivated in pH-stable medium (MM3) supplemented with a carbon source (1%, wt/vol), as described previously (30). pUS1, a pUC18 derivative containing a 3.2-kb genomic SalI DNA fragment from S. reticuli on which the complete abpS gene is located, was described previously (32). The DNA sequence of abpS is available from the EMBL data bank under accession no. Z97071.

Escherichia coli plasmid pET21a (Novagen, Madison, Wis.) was used as a cloning vector for truncated abpS genes. Expression was performed in E. coli BL21(pLysS), which was obtained from Novagen, as the host.

Transformation of strains.

E. coli strains were transformed with plasmid DNA by the CaCl₂ method (21).

PCR.

Each 30-μl (final volume) PCR mixture contained primers, 10 ng of pUS1, 0.2 nM dATP, 0.2 nM dCTP, 0.2 nM dGTP, 0.2 nM dTTP, 10 mM KCl, 10 mM (NH₄)₂SO₄, 20 mM Tris-HCl (pH 8.8), 2 mM MgSO₄, and 0.1% Triton X-100. In order to reduce misreading, the Vent_R DNA polymerase, which has a 3′→5′ proof reading exonuclease activity, was used instead of the Taq polymerase. The following primers were used to introduce an NdeI site at the 3′ end of the abpS gene: P₀ (CAGGAACCATATGAGCGACAC), P₁ (GCTCGTCCATATGCGTGACAGCGCTCTCGCCCG), P₂ (CCAGGCCCATATGACGGACGCCGAG), and P₃ (GGCCAGCATATGCGCAACGACGCC).

Primer P₄ (GGGGTCGACCCGGGACTGCTGCGCCGGG) was utilized to replace the stop codon of abpS with an SalI site at the 5′ end of abpS.

The following cycling conditions were found to be optimal: 95°C for 90 s, 55°C for 60 s, and 72°C for 60 s. Thirty cycles were performed. The PCR products were purified by using a Qia quick spin PCR purification kit (Qiagen, Düsseldorf, Germany) as recommended by the supplier and were digested with NdeI and SalI.

Expression of the truncated abpS genes.

The digested PCR products were ligated in frame into the polycloning site of pET21a (linearized with NdeI and XhoI) and were transformed in E. coli BL21(pLysS). The transformants were grown at 37°C in SOC medium (20 g of Bacto Tryptone liter⁻¹, 5 g of yeast extract liter⁻¹, 0.5 g of NaCl liter⁻¹, 0.18 g of KCl liter⁻¹; 20 ml of 1 M glucose per liter was added after autoclaving, and the medium was supplemented with chloramphenicol [34 μg/ml] and ampicillin [100 μg/ml]). IPTG (isopropyl-β-d-thiogalactopyranoside) (final concentration, 1 mM) was added when the absorbance at 600 nm reached 0.6. After an additional 3 h of cultivation, the E. coli cells were harvested, washed, resuspended in sonification buffer (0.1 M NaH₂PO₄, 0.01 M Tris-HCl [pH 8.0], 8 M urea), and disrupted with a Branson model B12 Sonifier for 3 min by using 20-s intervals. After the cell debris was removed, Ni²⁺-nitrilotriacetic acid (NTA) (Qiagen) was added to bind the His₆ fusion protein. Unspecifically bound proteins were removed by consecutive washes with buffer (0.1 M NaH₂PO₄, 0.01 M Tris-HCl [pH 6.3], 8 M urea) containing 25 mM imidazole. The fusion protein was subsequently released by adding 0.5 M imidazole.

SDS-PAGE and Western blotting.

Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) was performed with 10% polyacrylamide gels in the presence of 1% SDS (13). For immunodetection, proteins were transferred onto nylon membranes after SDS-PAGE. The filters were incubated in phosphate-buffered saline (PBS) (80 g of NaCl liter⁻¹, 2 g of KCl liter⁻¹, 2 g of KH₂PO₄ liter⁻¹, 11.5 g of Na₂HPO₄ liter⁻¹; pH 7) containing a 1:100,000 dilution of the primary antisera or a 1:10,000 dilution of the purified immunoglobulin Gs (IgGs). After three washes, the blot was incubated with alkaline phosphatase conjugated with the AffinyPure F(ab′)₂ fragment of goat anti-rabbit IgG (Dianova, Hamburg, Germany). Color was developed as described by West et al. (33).

Isolation of the murein layer.

S. reticuli mycelia were harvested by centrifugation, washed in PBS, and then, with continuous stirring, mixed with the same volume of 8% SDS at 100°C. The mixture was boiled for 10 min and then stirred overnight at room temperature. After incubation for an additional 10 min at 100°C, murein was collected by centrifugation at 20,000 × g for 30 min and washed three times with PBS containing 10% 2-propanol and three times with PBS in order to remove the SDS.

Murein was treated with buffer (25 mM Tris-HCl [pH 8], 20 mM EDTA) containing 5 mg of lysozyme (Boehringer) per ml and incubated at 37°C for 16 h. After centrifugation at 15,000 × g for 15 min, an aliquot of the supernatant was analyzed by using an SDS-PAGE gel.

Another portion of the murein was incubated in buffer (Tris-HCl [pH 8], 50 mM EDTA) containing 5 mg of proteinase K per ml for various times at 37°C and washed 10 times with PBS containing 5 mM Pefabloc (a serine protease inhibitor; Boehringer) prior to digestion with lysozyme.

Purification of IgGs specific for the NH₂-terminal part of AbpS.

One of the truncated AbpS proteins, which lacked the smallest part of the NH₂ terminus of AbpS, was immobilized on a piece of nylon membrane. After blocking with PBS containing 1% bovine serum albumin, the blot was washed in 8 M urea, equilibrated with PBS, and incubated with the primary antiserum. The serum was obtained by immunizing a rabbit with AbpS isolated from S. reticuli (32). After incubation for 1 h at room temperature, the membrane was removed and washed with 8 M urea to release the bound antibodies. After equilibration in PBS, the membrane was again used to bind the IgGs present in the antiserum. The procedure was repeated until the retaining antiserum no longer contained IgGs specific for the immobilized protein, as shown by a Western blot analysis. The IgGs specific for the NH₂-terminal part of AbpS were found in the retaining antiserum.

General DNA techniques.

Restriction enzyme digestions, ligation, and analyses of DNA with polymerases were carried out by using the standard procedures (21). DNA sequencing was performed with a T7 sequencing kit and Cy5-labelled standard primers (Pharmacia).

In vivo immunolabelling of AbpS and fluorescence microscopy.

S. reticuli mycelia were incubated with PBS containing 1% bovine serum albumin for 1 h and then with PBS containing either a 1:1,000 dilution of the primary antiserum or a 1:100 dilution of the antibodies specific for the NH₂-terminal part of AbpS. To remove unbound antibodies, the mycelia were washed three times with PBS and then treated with PBS containing an FITC-labelled F(ab′)₂ fragment of goat anti-rabbit IgG. After three washes, the FITC-labelled mycelia were analyzed for fluorescence under UV light with an Axiovert microscope (Zeiss). For visualization, a charge-coupled device camera (SenSys; Photometrics) and the IPLab software were used.

Immunolabelling of the AbpS protein and analysis by TEM.

S. reticuli cells were incubated with the specific antibody (see above) and then with protein A-gold complexes (diameter, 10 nm; 1:50 dilution of the stock solution; British Biocell, Cardiff, Great Britain) for 1 h, washed three times with TE buffer (20 mM Tris-HCl, 1 mM EDTA; pH 7.0), absorbed onto a thin carbon film, washed with TE buffer, and air dried. Samples were examined with a Zeiss model TEM910 transmission electron microscope (TEM) at an acceleration voltage of 80 kV at calibrated magnifications.

Immunofield emission scanning electron microscopy of uncoated samples.

S. reticuli cells labelled with the specific antibody and 10-nm-diameter protein A-gold particles were adsorbed onto poly-l-lysine-coated coverslips, fixed with 3% glutaraldehyde in PBS for 15 min, washed three times with PBS, and then fixed with 2% osmium tetroxide at room temperature overnight. Samples were dehydrated with a graded acetone series and critical point dried with CO₂. Uncoated samples were examined with a Zeiss model FESEM DSM982 Gemini microscope at an acceleration voltage of 1.5 kV by using a 4-mm working distance and a built-in Everhart-Thornley SE detector.

Preembedding labelling of S. reticuli.

S. reticuli cells were labelled with the specific antibody and 10-nm-diameter protein A-gold complexes, fixed with 3% glutaraldehyde for 30 min at room temperature, and then washed and fixed with 1% osmium tetroxide for 1 h at room temperature. The cells were dehydrated with a graded acetone series and were embedded by using the protocol of Spurr (27). Ultrathin sections were counterstained with 4% uranyl acetate and lead citrate prior to examination with a Zeiss model TEM910 microscope at an acceleration voltage of 80 kV.

Postembedding labelling of S. reticuli.

Mycelia cultivated with glucose or Avicel as the carbon source were fixed in a fixation solution containing 0.2% glutaraldehyde and 0.5% formaldehyde for 1 h on ice. After several washes with PBS containing 10 mM glycine, the hyphae were dehydrated with a graded ethanol series on ice and embedded in LRWhite resin by using the following embedding schedule: 1 part of 100% ethanol and 1 part of LRWhite resin for 8 h, 1 part of ethanol and 2 parts of LRWhite resin overnight, and pure LRWhite resin for 1 day with several changes. Samples were then transferred into gelatin capsules which were filled with LRWhite resin and polymerized at 60°C for 2 days. Ultrathin sections were cut with a glass knife, collected on Formvar-coated nickel grids (300 mesh), and incubated with a 1:2 dilution of the specific antibody at 4°C for 14 h. After the grids were washed with PBS, they were incubated with protein A-gold complexes (diameter, 15 nm; 1:75 dilution of the stock solution) for 1 h at room temperature, washed with PBS containing 0.01% Tween 20, washed with distilled water, and air dried. Ultrathin sections were counterstained with 4% aqueous uranyl acetate for 10 min. Samples were examined with a Zeiss model TEM910 microscope.

RESULTS AND DISCUSSION

Production of truncated AbpS proteins.

PCR were used to amplify complete abpS genes or truncated abpS genes that were shortened at the 5′ end. To do this, we used a combination of four oligonucleotides (which introduced different NdeI sites at the 5′ end of the AbpS-encoding gene); one primer generated an SalI site at the 3′ end of the coding sequence. The four resulting DNA fragments were ligated in frame into pET21a that was cut with NdeI and XhoI. The plasmids which were constructed and isolated from E. coli transformants were analyzed with restriction enzymes. Sequencing of the religated sites revealed that the reading frames were preserved. Consequently, the corresponding plasmids encoded complete AbpS or truncated forms of AbpS that differed in the length of the NH₂ terminus. In addition, each of the proteins was tagged with a valine, a glutamine, and six histidines at the COOH terminus (Fig. 1). Logarithmically grown E. coli BL21(pLysS) transformants harboring the constructs were treated with IPTG (1 mM). Under these conditions, more than 90% of each of the fusion proteins was produced in the form of insoluble inclusion bodies. Therefore, the cells were disrupted in the presence of 8 M urea, and the proteins were subsequently isolated on the basis of their affinities to Ni²⁺-NTA (Fig. 1, left gel).

FIG. 1.

Truncated AbpS proteins. The predicted structure of AbpS is shown. α-Helical structures are indicated by shaded boxes, and β-sheets are indicated by black boxes. The amino acid sequence of the wild-type protein is compared to the amino acid sequences of the truncated AbpS forms. The four truncated AbpS proteins were separated by SDS-PAGE, transferred to a membrane, and stained with Coomassie (left gel). Alternatively, the proteins were immunologically tested with diluted (1:100,000) primary antiserum (middle gel) or with enriched antibodies (diluted 1:10,000) specific for the NH₂-terminal part of AbpS (right gel).

Isolation of specific antibodies.

The purified proteins were treated with the primary antiserum containing IgGs raised against AbpS (32). In contrast to the three larger AbpS forms (35.7, 32.3, and 29.1 kDa), the 23.5-kDa protein lacking the largest portion of the NH₂-terminal part was found to interact with a small number of immunoglobulins (Fig. 1, middle gel). This indicates that the central and carboxy-terminal parts of native AbpS comprise a low number of epitopes.

In order to isolate antibodies specific for the NH₂ terminus of AbpS, the truncated 32.3-kDa protein was immobilized on nitrocellulose membranes and used as a target for the anti-AbpS antibodies present in the polyclonal antiserum. When this method was used, all of the antibodies which had their epitopes in the central or C-terminal part of AbpS could be extracted from the antiserum. Western blot analysis revealed that the purified IgGs recognized the full-length AbpS but none of the truncated proteins, proving that the specificity of the enriched IgGs was high (Fig. 1, right gel). However, the levels of proteins present in the remaining antiserum were about 10% of the initial levels. This was due to extraction of the antibodies specific for the central and C-terminal parts of AbpS during the purification procedure.

Microscope studies.

Although electron microscopy is a very useful tool for studying ultrastructures, fixation and drying of the objects are required. This may lead to changes in the native structure, and in addition, during inspection the sample may be damaged by radiation (20). In contrast, samples can be analyzed without considerable destruction by fluorescence microscopy, but the resolution is rather low. Therefore, a combination of the two methods was used to determine the location of the S. reticuli AbpS.

S. reticuli mycelia grown in the presence of glucose were incubated with the primary antiserum (Fig. 2A) or with IgGs that specifically recognized the NH₂ terminus of AbpS (Fig. 2B) and then with the FITC-labelled Affinipure F(ab′)₂ fragment of goat anti-rabbit IgG, and then they were analyzed under UV light with a light microscope. Independent of the antibody type, fluorescence labels were detected at about equal intensities on the surfaces of some hyphae but not on the surfaces of hyphae which had been treated only with the FITC-labelled secondary antibody (32). As the murein layer is a barrier for proteins, these results clearly proved that the NH₂ terminus of AbpS protrudes from the S. reticuli murein layer. The COOH-terminal portion and the largest portion of the central part of AbpS seem to be covered by their environment.

FIG. 2.

In vivo FITC labelling of AbpS. Washed S. reticuli mycelia were incubated with primary antiserum diluted 1:1,000 in PBS (A) or with specific antibodies diluted 1:100 (B) for 4 h at room temperature. The unbound antibodies were removed, and secondary FITC-labelled anti-rabbit IgG2ab was used to determine the presence of AbpS–anti-AbpS antibody complexes on the surfaces of S. reticuli hyphae under UV light (left panels) and visible light (right panels) with an Axiovert microscope (Zeiss).

To obtain higher resolution, immuno-gold-labelled hyphae were analyzed with the TEM or the field scanning electron microscope. In contrast to the control experiments (in which labelling occurred without primary antibodies), ternary complexes consisting of AbpS, antibodies, and protein A-gold complexes were identified on the surfaces of the hyphae (Fig. 3A and B). Labelling increased 10 to 30% when 10-nm-diameter gold particles were used instead of 15-nm-diameter particles. AbpS was found to be regularly distributed on both tips and bases of the hyphae, and about 50 labelled AbpS molecules were detected per μm² of surface. In addition, thin sections were prepared from S. reticuli hyphae (cultivated with glucose or crystalline cellulose [Avicel]), treated with the specific anti-AbpS antibodies, and then incubated with protein A-gold particles (Fig. 3C through E). Examination with the TEM revealed that most labelling occurred on the surfaces of the hyphae; labelling was rarely observed in the lumen. Independent of the preculture conditions (glucose or Avicel), the perimeters of the hyphae (obtained after vertical cutting of the longitudinal axis of the hyphae) contained 7 to 10 labels. The distribution of the proteins was not influenced by the nutrient. This fact is consistent with the results obtained in physiological experiments, which showed that AbpS is synthesized during the late logarithmic phase, independent of the carbon source used (32).

FIG. 3.

Immunoelectron microscopic localization of AbpS in S. reticuli as determined by using preembedding labelling (A, B, and E) and postembedding labelling (C and D). (A and B) Gold particle labels indicate the localization of AbpS on the surface of a hypha. (A) TEM micrograph showing the contrast between a hyphae and gold particles (diameter, 10 nm). (B) Scanning electron micrograph of a critical-point-dried and uncoated S. reticuli hypha. Gold particles (diameter, 10 nm) bound to specific antibodies are visible as white dots. (C and D) Localization of AbpS in ultrathin sections after postembedding labelling. Most of the gold particles (diameter, 15 nm) occur at the cell periphery; in panel D a cell wall fragment is also densely labelled. (E) Ultrathin section of a preembedded hypha with labelled AbpS from S. reticuli cultivated with Avicel (A) as the sole carbon source. (A, B, and E) Bars = 1 μm. (C and D) Bars = 0.5 μm.

As shown in Fig. 3D, the labelled AbpS protein was also detected in thin sections of cell wall fragments of S. reticuli, which indicated that AbpS is tightly linked to the murein layer.

Interaction of AbpS with the cell wall.

To characterize the linkage of AbpS to murein, the peptidoglycan layer was extracted from S. reticuli hyphae by boiling a preparation in 4% SDS and then centrifuging it. Washing murein with H₂O, 8 M urea, 100 mM EDTA, or buffers adjusted to pH 2 or 9 did not result in removal of AbpS; neither did boiling in buffer containing 1% mercaptoethanol and 0.1% SDS. In contrast, treatment of murein with lysozyme or sonication resulted in release. Other murein-lytic enzymes (lysostaphine, mutanolysin) and proteinase K had no effect (Fig. 4A and B). Therefore, we propose that AbpS is covalently linked to murein.

FIG. 4.

Characterization of the AbpS-murein interaction. (A) Isolated S. reticuli murein was incubated in identical volumes of H₂O (lane 1), 0.1% SDS–1% mercaptoethanol at 100°C, (lane 2), 25 mM Tris-HCl (pH 7) containing 1% Tween 80 (lane 3), 8 M urea (lane 4), proteinase K solution (5 mg/ml) (lane 5), lysozyme solution (5 mg/ml) (lane 6), lysostaphine solution (5 mg/ml) (lane 7), 100 mM EDTA buffer adjusted to pH 2 (lane 10), and buffer adjusted to pH 11 (lane 11) for 2 h or was suspended in 25 mM Tris-HCl (pH 7) and sonicated for 5 min at 4°C with a Branson model B12 Sonifier (lane 9). After residual murein was removed by centrifugation, 50 μl of each sample was loaded onto an SDS-PAGE gel, and after blotting AbpS was detected immunologically. (B) A sample containing AbpS which was released from murein with lysozyme (panel A, lane 6) was applied to an SDS-PAGE gel and stained with Coomassie blue (lane 1). Purified AbpS was used as a control (lane 2). (C) Murein was incubated for 1 h (lane 1), 2 h (lane 2), and 8 h (lane 3) with proteinase K or for 8 h in the same buffer without the protease (lane C) at 37°C. Then the proteinase K was removed by washing the murein with 10 volumes of buffer containing 10 mM Pefabloc (Boehringer) and was inactivated by incubation for 10 min at 100°C. AbpS and the processed forms were released from the murein by adding lysozyme (5 mg/ml); after SDS-PAGE and blotting, they were detected immunologically with anti-AbpS antibodies. (D) S. reticuli mycelia grown in minimal media without glycine (lanes a, b, and c) or with 1% glycine (lanes 1, 2, and 3) were treated for 2 h at 37°C with 5 ml of buffer containing 5 mg of lysozyme per ml (lanes 1 and a). Subsequently, the protoplasts generated were recovered by centrifugation and were destroyed by adding 5 ml of 50 mM EDTA. The membranes and the associated proteins were separated from the soluble proteins (lanes 2 and b) by ultracentrifugation and solubilized in 0.5 ml of 1% SDS (lanes 3 and c). A 50-μl portion of each fraction was analyzed for the presence of AbpS by using anti-AbpS antibodies.

Covalently anchored surface proteins have been found in other bacteria. Immunoglobulin-binding protein A of Staphylococcus sp. was found to be covalently linked to the pentaglycine of the cell wall, and the NH₂-terminal part of the protein was found to protrude from the murein. It is thought that this protein plays a role in protecting the bacterium from the defense system of a host (24, 25). Internalin (InlA) of Listeria monocytogenes (a gram-positive, facultatively intracellular parasite) is covalently anchored by its COOH-terminal region, whereas the NH₂ terminus is exposed to the extracellular environment. InlA is a virulence factor, as it assists in adherence to surfaces and in invasion of the eukaryotic host cells, and it protects the bacterium from the defense system of the host (7, 14). Similarly, the immunoglobulin-binding P- and M-proteins of several streptococci have been shown to be cell wall associated (8, 19, 26, 28). In all of the examples given above, anchoring requires a COOH-terminal sorting signal (LPTXG) followed by a hydrophobic domain and a tail consisting of charged amino acids (11, 24). After secretion, an extracellular sortase modifies the surface proteins (16). First, the side chain consisting of threonine is covalently linked to the pentaglycine, and then the glycine and the hydrophobic domain are removed. The LPTXG sorting signal has not been found in the deduced amino acid sequence of AbpS, and no proteolytic processing of AbpS has been observed. These results, together with the fact that lysostaphine (which specifically cut the pentaglycine within the cell wall) was not able to release AbpS from the S. reticuli murein, indicate that the anchoring mechanism of AbpS is different.

However, when murein was pretreated for up to 8 h with proteinase K, subsequent lysozyme activity induced the release of a 21-kDa truncated form of AbpS (Fig. 4C). Additionally, previous proteinase K experiments revealed that soluble AbpS can be completely degraded by this protease and that only a 3-kDa peptide can be removed from AbpS in intact mycelia (32). Consequently, we speculated that proteinase K removes amino acids from both the NH₂ terminus and the COOH terminus of AbpS when it is in the polyglucane layer, and the remaining 21-kDa fragment appears to be protected from protease digestion by the surrounding murein structure.

In contrast, the largest portion of AbpS was found to be associated with protoplasts generated from S. reticuli hyphae by lysozyme treatment. The protein was released only by subjecting the protoplasts to osmotic shock, sonication, or detergent treatment, whereas small quantities were isolated in association with the membranes (Fig. 4D). The predicted COOH-terminal hydrophobic domain, which has all of the characteristics of a transmembrane segment, might anchor AbpS to the membrane of S. reticuli. This mechanism of anchoring is well-known for proteins of eukaryotic cells that lack the polyglucane layer (for a review see reference 5). The T-cell-associated IgGs, the class I and II MHC proteins, the CD4 and CD8 receptors of T-lymphocytes, or peripheral myelin proteins L1 and 0, which participate in cell-cell adhesion, were identified as membrane-anchored proteins, whereas their corresponding NH₂ terminus is extracellularly exposed (10).

Occurrence of AbpS homologues in different Streptomyces strains.

Total proteins of 10 different Streptomyces strains were tested for the presence of proteins that cross-reacted with anti-AbpS antibodies (Table 1). S. olivaceoviridis, S. coelicolor A(3)2, S. lividans 1326, S. albus, and S. flavogriseus produced approximately 75% as much AbpS as S. reticuli produced. In contrast, S. badius, S. vinaceus, and S. griseus synthesized relatively small quantities of the protein (Table 1). In addition, in the chromosomal DNA isolated from each of the strains investigated, an abpS homologue was identified by hybridization with an internal DNA fragment of abpS (data not shown).

TABLE 1.

Cellulolytic activities of and production of AbpS by streptomycetes^a

Organism	Growth in the presence of:			Production of AbpS (%)
	Carboxymethyl cellulose	Hydroxethyl cellulose	Avicel
S. reticuli	++	++	++	100
S. flavogriseus	++	++	++	75
S. lividans	++	++	−	75
S. coelicolor A3(2)	++	++	−	75
S. olivaceoviridis	++	++	−	75
S. albus	+	+	−	75
S. vinaceus	++	++	−	10
S. griseus	+	+	−	10
S. badius	+	+	−	10

^aOrganisms were cultivated in the presence of carboxymethyl cellulose, hydroxyethyl cellulose, or crystalline cellulose (Avicel) for 5 days at 30°C. Growth was determined microscopically (−, no growth; +, good growth; ++, very good growth). Production of AbpS was calculated from the results of a Western blot analysis; 30 μg of total proteins isolated from each organism was loaded onto an SDS-PAGE gel, blotted, and then treated with anti-AbpS antibodies. After scanning, the quantities were determined densitometrically by using the Cybertech program CAM 2.0. 

Additionally, the prototrophic strains were tested for growth in minimal medium supplemented with soluble carboxymethyl cellulose or hydroxyethyl cellulose, as well as crystalline cellulose (Avicel). Only S. reticuli and S. flavogriseus utilized Avicel, but each of the strains hydrolyzed both types of soluble cellulose (15, 22). These cellulolytic activities were found to be independent of the amounts of the AbpS homologues produced (Table 1). In the future it will be interesting to compare the specificities of binding of the different AbpS homologues to various types of native cellulose. Naturally occurring celluloses (which are the most abundant biopolymers in soil, which is the dominant habitat of streptomycetes) differ in crystallinity, chain length, and associated substances (9). In the future it will be interesting to test whether different binding specificities of AbpS homologues are correlated with the colonization properties of individual strains which are dominant in different ecological niches.