Aggregation-deficient Mutants of Streptococcus gordonii Channon ...

3 downloads 0 Views 3MB Size Report
Aug 5, 1992 - (Difco)] containing I5 per cent (v/v) glycerol. A. naeslundii T14V, A. naeslundii ATCC 12104 and Actinomyces viscosus WVU267 were provided.
MICROBIAL ECOLOGY IN HEALTH AND DISEASE

VOL.

5: 277-289 (1992)

Aggregation-deficient Mutants of Streptococcus gordonii Channon Altered in Production of Cell-surface Polysaccharide and Proteins R. McNAB and H. F. JENKINSON* Depurtment of Oral Biology and Oral Pathology, University of Otago, PO Box 647, Dunedin, New Zealand Received I6 April 1992; revised 5 August 1992

Three spontaneous mutants of Streptococcus gordonii strain Channon were isolated that were deficient in cellkell aggregation. The mutants had reduced cell-surface hydrophobicity, were impaired in saliva-mediated cell aggregation. and had greatly reduced amounts of ruthenium red stained material a t their cell surfaces as visualised by electron microscopy. Alkali treatment of intact cells removed the ruthenium red-stained material together with a subset of polypeptides. One of these, a surface protein of molecular mass approximately 150 kDa, was present in the wild-type strain but not in the mutants. Parent and mutant strains secreted similar amounts of SSP-5 salivary agglutinin receptor, glucosyltransferase, and proteinase. The results implicate the ruthenium red-stained material and a 150-kDa polypeptide in hydrophobicity and aggregation properties of S. gordonii. KEY woRDs-Oral proteins.

streptococci; Streptococcus gordonii; Aggregation; Saliva; Ruthenium red layer; Cell-surface

INTRODUCTION

reduced cell-surface or that were defective in saliva-mediated aggregati~n’~ The cell surfaces of oral streptococci carry a variety or in coaggregation with A . naes~undii.20.22*” of structures and adhesins that enable the organisms Analysis of these mutants has identified polyto interact with host tissues and components, other peptides putatively involved in determining hydrobacterial cells, and to colonise the human oral phobicity, aggregation or coaggregation properties. cavity. Streptococcus gordonii is found in supra- The relationship of these polypeptides and surface gingival plaque and on oropharyngeal mucosa, properties with the presence of cell-surface strucespecially incaries-activeindivid~als.~~ The bacteria tures such as fibrils, which are commonly seen on do not produce immunoglobulin A, p r o t e a ~ e , ’ ~ S. g~rdonii,‘~ are unknown. Strains of S. gordonii a putative virulence component of other oral such as M5, FT2, MJ2, CN2814 and Blackburn streptococcal species such as Streptococcus sanguis, carry peritrichous fibrils of varying densities as Streptococcus oralis and Streptococcus mitis.26 revealed by electron microscopy of negatively However S. gordonii secretes a number of poly- stained cells. 14v4’ These bacterial cells aggregate in peptides that have been implicated in its ability to saliva,6 adhere to saliva-treated hydroxylapatite colonise the human mouth. These include adhesin beads4’ and they coaggregate with A . naeslundii, SPP-5* which interacts with human salivary glyco- but none of these properties could be directly related protein, cell-surface adhesins for coaggregation to fibril length or d e n ~ i t y . ’ ~ , ~ ~ with A ctinomyces naeslundii, 2 2 , 2 proteins that To try to correlate better electron microscopic interact with undefined components of saliva37 or observations with surface properties and composerum,’ and glucosyltransferase2 for glucan- sition we describe the isolation and characterization mediated accumulation of cells.39 of spontaneous mutants of S . gordonii strain To try to identify specific cell-surface components Channon. This organism was previously known as involved in colonisation-related functions, mutants S. sanguis antigenic type XI5 and its cell surface was of S. gordonii have been isolated that showed described as fimbriate.’ We have since reclassified *Author to whom correspondence should be addressed. this strain as belonging to the species S . g~rdonii.’~

’’

0891-060X/92/060277-13 $ 1 I .50 0 1992 by John Wiley & Sons, Ltd

278 The bacterial cells were hydrophobic, they aggregated during growth in complex medium, and ruthenium red staining of thin sections revealed a densely stained surface outer layer. Using an enrichment procedure selecting for slower-sedimenting cells, three independently isolated mutants are described that are defective in production of surface polysaccharide and that have altered cell-surface proteins. The results suggest that the properties of hydrophobicity and saliva-mediated aggregation are associated with ruthenium red-staining cellsurface material. METHODS Bacterial strains andgrowth conditions

S. gordonii Channon (NCTC 7869)36 and S. gordonii DL 1-Challis were laboratory stock strains and were stored frozen at - 80°C in BHY medium [37 g/l Brain Heart Infusion Broth (Difco Laboratories, Detroit, MI, USA), 5g/l Yeast Extract (Difco)] containing I 5 per cent (v/v) glycerol. A. naeslundii T14V, A . naeslundii ATCC 12104 and Actinomyces viscosus WVU267 were provided by Dr PS Handley (University of Manchester, UK) and were also kept as stocks at - 80°C. All bacteria were propagated on BHYN agar'' at 37°C in a candle jar, and grown without shaking in either BHY medium or TY-glucose medium24at 37°C in screw-capped tubes or bottles.

R. McNAB AND H. F. JENKINSON

cells that had not sedimented in culture. After 26 h incubation at 37°C individual colonies were picked from plates into microcentrifuge tubes containing 1 ml BHY medium, and these were then grown for 18 h at 37°C to screen for cultures in which cells were dispersed rather than flocculated. From approximately 200 isolates screened, 11 were selected that each sedimented to a lesser extent in culture than did the wild-type strain. Measurement of bacterial autoaggregation The optical density (OD) at 600 nm was measured of a sample removed from the surface of an undisturbed culture in BHY medium after 16h growth. The culture was then vortex-mixed and the OD,,, recorded. The degree of cell autoaggregation was then expressed as a percentage given by: [(mixed culture O D -settled culture OD)/mixed culture OD] x 100. Saliva-mediated aggregation assay

Unstimulated whole saliva was collected on ice from several human volunteers, pooled, and clarified by centrifugation (10 OOOg, 15 min, 4°C). Some saliva samples were heat treated at 100°C for 10 min and re-clarified by centrifugation. Streptococci were cultured in TY-glucose medium and harvested in the early stationary phase of growth by centrifugation (6000g, 10 min, 4°C). Cells were washed twice by suspension in 10mM Tris/HCl Isolation of mutants (pH 7.2) containing 5 mM CaC1, (TBC buffer) and Cells of S. gordonii Channon growing on the then suspended in TBC at an OD,,, of 5 (about surface of BHYN agar were harvested with a sterile 5 x lo9 cells/ml). Serial two-fold dilutions of saliva were made with TBC in round-bottom wells of glass pipette into 10ml PBSC (10mM K,HPO,/ KH,PO,, pH6.5, containing 0.15 M NaCl and microdilution plates (0.05 ml per well). Bacterial 1 mM CaCl,), and adjusted to a density of suspension (0.05 ml) was added to each well, plates approximately 5 x lo9 cells/ml. The suspension was were gyratory shaken at 25°C for 30min, and vortex-mixed and 1 ml portions were distributed aggregation titres were recorded as the reciprocals into microcentrifuge tubes and cells were allowed to of the highest dilution of saliva that caused complete settle for 2 h at room temperature. Portions bacterial agglutination. A second method used to assay saliva-mediated (0.05 ml) from the surface of each suspension were spread onto BHYN agar, the plates were incubated aggregation was based on rate sedimentation.6 for 24 h at 37"C, then the growth from the surface of Streptococcal cells were prepared as described each plate was harvested and suspended in 10ml above and were collected by centrifugation and BHY medium. These suspensions were used to suspended to an OD6,,=5 in PBSC. Cells were inoculate 10 ml portions of fresh BHY medium with diluted to a final OD,,, of approximately 1.0 in a 1-3 x lo7 cells. The cultures were incubated for plastic disposable cuvette containing PBSC, saliva 18 h at 37"C, after which time they contained heavy (0-1ml), and additional preparations to be tested for flocculates of cells. Portions (0.1 ml) from the inhibition of aggregation, to a final volume of 1 ml. surfaces of the undisturbed cultures were spread The contents were mixed, the cuvette was incubated onto BHYN agar in order to grow colonies from at 37°C and the OD,,, of the suspension was

279

AGGREGATION MUTANTS OF S. GORDONII

measured at intervals. All assays were run in duplicate or triplicate, and duplicate controls containing bacteria only (no saliva) were included. Coaggregation and hydrophobicity assays Coaggregation reactions of streptococci with actinomyces were performed in round-bottom wells of microtitre plates as described.23 Coaggregation titres of streptococci with actinomyces were expressed as the reciprocals of the highest dilution of streptococcal cells that caused coaggregation. Hydrophobicity was expressed as the percentage decrease in OD,,, of the aqueous phase after mixing cell suspension (OD,,, = 0.8, 3 ml) with hexadecane (0.2 ml). Iodination of cells Bacteria grown to early stationary phase in BHY medium were surface labelled with 1 2 , 1 using a lactoperoxidase-catalysed reaction as previously described. l9 Iodinated proteins were solubilised from cells with SDS extraction buffer (see below). Protein extraction To extract total cell proteins, cells were harvested from cultures (10 ml) in TY-glucose medium at late exponential phase of growth by centrifugation (6000g, 10 min, 4°C). Cells were washed once by suspension in distilled water, collected by centrifugation, and suspended in 0.5 ml TE buffer (10 mM Tris/HCl pH 7.5 containing 1 mM Na, EDTA). Glass beads (0.10-0.1 1 mm diam., 0.5 ml) were added and the suspension was vortex mixed for 1 min. Sodium dodecyl sulphate (SDS) extraction buffer (25 mM Tris/HCl pH 6.8, containing 2 per cent wjv SDS and 0.1 per cent v/v 2-mercaptoethanol) was added (0.5 ml), the suspension was heated at 80°C for 10 min, vortex mixed for 20 s, then centrifuged at 5000 g for 5 min. The supernatant (0.4 ml) containing solubilised proteins was mixed with 0.1 vol. loading dye (70 per cent vjv glycerol in water containing 0.05 per cent wjv bromophenol blue) in preparation for electrophoresis and was stored at - 80°C. To extract polypeptides from bacterial cells that had been surface labelled with '',I, bacteria were suspended in SDS extraction buffer (0.6 ml). heated at 80°C for IOmin, vortex mixed for 20s, then heated for a further 5 min at 80°C. The suspension was centrifuged (10 000 g, 10 min) and the super-

natant was mixed with loading dye and analysed immediately by electrophoresis. Culture fluid polypeptides were concentrated from growth medium by precipitation. Cell-free supernatant fluid from TY-glucose grown cultures (2.5 ml) was mixed with 10 ml acetone and stored at - 20°C for 2-3 h. The suspension was centrifuged (10 OOOg, 20 min, 4"C), the pellet was dried briefly under vacuum, and suspended in 0.15 ml SDS extraction buffer anddissolved by heatingat 70°C for 5 min. The extract was mixed with loading dye and proteins were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) as described below. Proteins and polysaccharide were extracted from bacterial cells with alkali as follows. Cells were harvested by centrifugation (6000g, 10 min) from 30ml early stationary phase culture in either TY-glucose or BHY medium. The bacteria were washed once only by suspension in distilled water, collected by centrifugation, and suspended in 0.6 ml ice-cold 0.1 N NaOH in water. Suspensions were incubated at 0 4 ° C on ice for 20 min with occasional vortex mixing, centrifuged (10 OOOg, 10 min, 4°C) and the supernatant (0.5 ml) was recentrifuged. The extract was neutralised with HCI and made to 10 mM with Tris and stored at - 80°C. It was found essential to wash the bacterial cells with distilled water before suspending them in alkali. Polypeptides were not effectively extracted from cells that were washed in 0.15 M NaCl or even in dilute buffer, e.g. 10 mM K,HP0,/KH2P0,, pH 7. In some experiments a portion of the alkali extract was deproteinised by incubating it with pronase (Sigma Chemical Co., St Louis, MO, USA; final concn. 10kg/ml) for 10min at 37°C. The extract was then mixed with an equal volume of ch1oroform:isoamyl alcohol (24: l), centrifuged to separate phases and the aqueous layer removed and extracted with an equal volume of chloroform. The pronase removed all polypeptides from the extract as assesed by SDS-PAGE of samples before and after protease treatment. SDS-PAGE Samples were subjected to SDS-PAGE using the system of Laemmli and Favre." Prior to electrophoresis samples were mixed with 0.5~01.SDS extraction buffer (unless they had already been prepared in this buffer) and with loading dye (see above), and heated at 70°C for 5min. Proteins were stained with Coomassie Blue R and gels were destained with methano1:acetic acid:water

280

R. McNAB AND H. F. JENKINSON

(20:10:70, by vol.). Molecular masses of proteins were calculated by reference of their mobilities to those of molecular mass marker proteins. The standards used were either from Sigma (1 80 to 27 kDa, prestained) or from Gibco BRL (218 to 15-3kDa, prestained; Gaithersburg, MD, USA). Gels containing iodinated proteins were soaked in destaining solution for 30 min, dried onto Whatman 3 MM paper and exposed to p-max film (Amersham International plc., Amersham, UK) at room temperature. Electroblotting and immunodetection Figured1. Flocculation ~ ~ lwere ~blotted~ from~ acrylamide ~ ~gels icells ~ ~of S. gordonii strain Channon or mutant grown in BHY medium. 1, Wild-type strain Channon; 2, onto nitrocellulose membrane in 25 mM Tris, mutant c2;3, c3;4, strain c8 192 mM glvcine, 20 Der cent (viv) methanol, UH 8.3, without ,;,or equilibration of the gel in this-buffer, at 18 V/cm for 2 h. Western blots were incubated tetroxide in 0.06 M cacodylate buffer containing with antiserum diluted 1:200 and antibody binding ruthenium red (3.3 mg/ml) for 4 h at room temperature. The cells were then washed three times with was detected with '2SI-labelled Protein A.19 cacodylate buffer, dehydrated in a graded series of ethanol and embedded in Procure 812 resin Detection of glucosyltransferase andproteinase (Probing and Structure. Kirwan., Old. Australia). . , Proteins from cell-free culture supernatants were Section; were cut on a 'Reichert-Jung Ultracut E precipitated with acetone as described above microtome and photographed using a Siemens 102 and separated by SDS-PAGE. Glucosyltransferase transmission electron microscope. Measurements activity was detected in situ.2 Proteinase activity of the thickness of the cell wall and ruthenium was detected after SDS-PAGE of samples in gels red-staining layers were made on 20 cells of each containing 0.2 per cent (w/v) gelatin.'* strain.

Determination of carbohydrate,protein and phosphorus

Carbohydrate was measured by the phenolsulphuric acid method' with D-glucose as standard. Protein was estimated with Folin reagent3' with bovine albumin as standard, and phosphorus was determined as inorganic p h ~ s p h a t e . ~ Electron microscopy

Cells were grown in BHY medium for 18 h, harvested by centrifugation (6000g, 5 min) and washed twice by suspension in 0.06 M cacodylate buffer pH 7.23, and centrifugation. To the cell pellet was added 1.5 ml glutaraldehyde (1.2 per cent v/v) in cacodylate buffer containing ruthenium red (3.3 mg/ml),16 the cells were suspended with a glass rod, and the suspensions were left at room temperature for 16 h. The cells were collected by centrifugation (10 000 g, 1 min), washed three times by suspension in cacodylate buffer and centrifugation, then suspended in 1.3 per cent (w/v) osmium

RESULTS Isolation and growth characteristics of non-aggregating mutants

Cells of S. gordonii Channon flocculated during growth in BHY medium and formed a loose sediment at the bottom of the culture vessel (Figure 1). Following three cycles of growth and enrichment for slower-sedimenting variant cells of S. gordonii Channon, as described in Methods, 11 strains were purified none of which flocculated to the same extent as the wild-type parent. After repeated subculture six of these clones were discarded because their flocculation phenotypes became indistinguishable from the wild-type. Of the remaining five that were stable on subculture, three were chosen for further study because their phenotypic properties suggested they contained different mutations. Stationary-phase cultures of mutant strains C2, C3 and C8 in BHY medium contained only small sediments of autoaggregated cells (Figure 1) and

28 I

AGGREGATION MUTANTS OF S. GORDON11

Table I . Aggregation properties of, and composition of alkali-extractablematerial from, cells of S. gurdunii Channon and derived mutants Alkali extract concentration of:$ Saliva-mediated aggregation titre i n t

Cell-cell aggregation

Hydrophobicity

Strain

(YO)*

(YO)*

Whole saliva

Heated saliva

Carbohydrate (mg/ml)§

Protein (mg/ml)§

Phosphorus (Pgiml)

Channon c2 c3 C8

85 6.2 28 k 5.4 46 4.8 27 & 3.5

78 & 7.0 55 & 4.2 37 k 6.3 46k7.1

25615 I2 4 418 4

32 0 0/2 012

0.29 (9.6%) 0.18 (6.4%) 0.20 (6.9%) 0.23 (8.5%)

0.36 (10.7%) 0.47 (144%) 0.46 (14.5%) 0.41 (12.5%)

6.10 5.00 5.50 5.45

+

*

* SE. ?Titre is expressed as the reciprocal of the highest dilution of saliva causing aggregation of 5 x lo8 streptococcal cells; duplicate titres were identical and ranges over experiments are shown. :Average of four determinations. SE within 7 per cent. SFigures in parentheses are percentages of total cell carbohydrate or protein.

were more uniformally turbid than cultures of the wild-type strain Channon. Autoaggregation of cells, leading to flocculation, was dependent to some extent upon the culture medium. The differences in aggregation between wild-type and mutants was not so apparent in TY-glucose medium. Mutants C2 and C3 had similar growth rates in BHY medium to the wild-type strain (doubling times approx. 45min), while strain C8 grew somewhat more slowly (doubling time approx. 55 min). All strains had similar colony morphologies that were indistinguishable on BHYN agar or on Mitis-Salivarius agar (Difco Laboratories, Detroit, MI, USA). When examined by phase-contrast light microscopy, all strains were found to produce chains of cocci from 2 to 40 cells in length, with an average chain length of 10. Aggregation properties and hydrophobicity Degrees of cell autoaggregation in wild-type and mutant strains were estimated from optical densities of BHY-grown stationary cultures taken before and after mixing. More than 85 per cent of the cells of strain Channon aggregated during growth, compared with only 3&50 per cent of cells of the mutant strains (Table 1). Since these differences in autoaggregation were likely to result from cellsurface alterations, other phenotypic properties associated with cell-surface composition were investigated. The abilities of wild-type and mutant strains to be aggregated by whole saliva were measured using a microtitre well plate dilution assay. Cells of S . gordonii Channon aggregated in saliva that had

been diluted 256-fold, whereas cells of all the mutant strains failed to aggregate in saliva that was diluted more than four-fold (Table 1). Controls containing cells only with no saliva were run alongside and under the conditions of the assay the cells alone did not autoaggregate. When saliva that had been heated (lOOOC, 10 min), to destroy immunoglobulin A and other heat-sensitive proteins, was used in the assay, the aggregation titre for wild-type strain Channon was reduced, and aggregation of the mutant strain cells was abolished (Table 1). Cell-surface hydrophobicity of wild-type and mutant strains was measured by the hexadecane partitioning assay.' All three mutants showed reduced hydrophobicity compared with the wildtype strain, with strain C3 being the least hydrophobic (Table I). Wild-type and mutant strains coaggregated to similar extents with A . naeslundii T14V and with A . viscosus WVU267, but poorly with A. naeslundii ATCC 12104 (results not shown). Cell-surface polypeptides Aggregation reactions and hydrophobicity of oral streptococci have been attributed to the activities of various cell-surface proteins. Therefore surface-exposed polypeptides present in wild-type and mutant strains were compared. Cells were surface labelled with 251 using lactoperoxidase catalysis and polypeptides were extracted by heating intact cells with SDS extraction buffer at 70°C. After SDS-PAGE separation of polypeptides, iodinated bands were detected by autoradiography. About 15 iodinated polypeptides were present in extracts from strain Channon and profiles of C2 and

282

Figure 2. Autoradiograph of polypeptides extracted from '''1labelled cells of S. gordonii and separated by SDS-PAGE through 10 per cent (w/v) acrylamide. I , Strain Channon; 2, strain C2; 3, strain C3; 4, strain C8. Similar amounts of protein (about 10 pg) were loaded onto each lane. Positions of molecular mass markers are indicated

C3 surface proteins were virtually identical to the Channon profile (Figure 2). The pattern of surfaceexposed polypeptides from strain C8 contained an additional faintly labelled band of about 130 kDa. Extracts from strain Channon contained a more strongly labelled band at 150 kDa (Figure 2, lane 1). Cell envelope fractions of the various strains were prepared and the polypeptides were subjected to SDS-PAGE and stained with Coomassie Blue. These extracts contained in excess of 80 polypeptide bands and no clear differences were apparent between profiles (not shown). After experimenting with different extraction methods for cells it was found that incubation of whole cells with 0.1 N NaOH at W " C solubilised a subset of polypeptides, in particular several of molecular mass>lOOkDa (Figure 3). More protein was extracted from the cells of the mutants with NaOH than from cells of strain Channon (Table 1). On SDS-PAGE of NaOH extracts a polypeptide of 150 kDa was present in the wild-type strain and was absent in the mutant strain extracts (Figure 3). Strain C8 was the only mutant strain to show additional differences in the 140- to 170-kDa range (Figure 3, lane 4). These protein differences detected in alkali extracts of the mutants correlated well with the differences observed in the surface-labelling

R. McNAB AND H. F. JENKINSON

Figure 3. SDS-PAGE (7.5 per cent w/v acrylamide) patterns of proteins, stained with Coomassie Blue, extracted with 0.1N NaOH from cells grown in BHY medium. I, Strain Channon; 2, strain C2; 3, strain C3; 4, strain C8. Approximately 40 pg protein was applied to each lane. Positions of molecular mass marker proteins are shown, and the arrow indicates a 150-kDa polypeptide band

patterns of polypeptides solubilised with SDS (in Figure 2).

Production of salivary adhesin Since saliva-mediated aggregation in S. gordonii has been shown to involve the interaction of a 205-kDa surface polypeptide (SSP-5) with salivary agglutinin,* it was of significance to determine if this adhesin was produced by the wild-type and mutant strains. To detect production of this polypeptide in S. gordonii we used rabbit polyclonal antiserum raised to SpaP polypeptide (Antigen 1/11) from S . mutans. This antiserum has been shown previously to react monospecifically with SSP-5 polypeptide in cell extracts of S. gordonii strain M5.6 The SSP-5 antigen in strain Channon was most clearly detected on Western blots of culture-medium proteins separated by SDS-PAGE. Wild-type and mutant strains were found to produce similar amounts of extracellular SSP-5 antigen visible as a band of approximate molecular mass 200 kDa reacting with anti-SpaP antibodies (Figure 4). The polypeptide was poorly extracted from intact cells with SDS

283

AGGREGATION MUTANTS OF S. GORDONII

1

2

3

4

5

6

7

8

kDa 180 1168458 48.5 36.5 -

Figure 4.Western blots of S . gordoniiculture fluid polypeptides (lanes 1 4 ) or of polypeptides extracted from disrupted cells (lanes 5-8) reacted with antiserum to SpaP protein of S . mutans. Lanes 1 and 5 , strain Channon; 2 and 6 , strain C2; 3 and 7, strain C3; 4 and 8, strain C8. Positions of molecular mass marker proteins are shown

extraction buffer and an antibody-reactive band was only faintly visible in extracts of wild-type and mutant strains (not shown). When cells were broken with glass beads and extracted with SDS, more proteins were solubilised but the SSP-5 antigen was fragmented, and anti-SpaP antibodies reacted with several bands primarily with one at 100 kDa (Figure 4).The SSP-5 antigen band was not present on Western blots of proteins extracted from cells with NaOH (not shown). There were no differences in the SDS-PAGE profiles ofculture medium polypeptides between the various strains (results not shown). Strain Channon is a low-activity producer of extracellular glucosyltransferase compared with S . gordonii strain Challis,' but wild-type Channon and mutants produced similar amounts of glucosyltransferase (Figure 5). All strains produced similar bands of extracellular proteinase activity as detected by gelatin SDS-PAGE (Figure 5).

Morphological ejects of alkali treatment When examined by transmission electron microscopy, thin sections of cells of strain Channon grown in BHY medium and stained with ruthenium

red and osmium tetroxide revealed a ruthenium red (RR)-stained material outside the cell wall layer (Figure 6). In the wild-type strain the weakly staining cell-wall layer was estimated to be 18+_ 3.2 nm diameter. The RR-stained material consisted of an electron dense layer (thickness 13.2& 2.1 nm) close to the cell wall and outside this, loosely intermeshed bundles of material projecting 120-240 nm from the cell wall (Figure 6a). All the mutant strains had significantly reduced amounts of RR-stained material on their cell surfaces (Figure 6 b d ) . Strain C2 had the least and strain C8 the most. In all mutant strains the electron-dense portion of the RR-stained material closest to the cell wall was similar in thickness to that in the wild-type strain. However in C2 the loosely meshed material projected only 30-40 nm from the cell wall (Figure 6b). In strain C3 the loosely meshed material projected 30-120 nm from the wall (Figure 6c), and in strain C8 it was 50-140 nm from the cell wall (Figure 6d). All cells in the populations had RR-stained material on their surfaces. Alkali treatment removed most of the RRstained material from the surface of the cells of the wild-type strain and left diffusely staining material 10-25nm outside an expanded cell wall layer

284

R. MrNAB AND H. F. JENKINSON

3 174 kDa-

4

5

1

2

3

4

5

1

B

A

Figure 5. Culture fluid proteins of S. gordonii separated by SDS-PAGE and gels incubated to detect glucosyltransferase (A) or proteinase (B) activities. Lane 1, strain Channon; 2, strain C2; 3. strain C3; 4, strain C8; 5, strain Challis. The molecular masses (kDa) indicate the major bands of enzymic activity in each case

(Figure 7a). The effect of alkali on the RR-stained layer was not so dramatic with the mutant strains, nevertheless in each instance the treatment reduced considerably the amount of material on the surfaces of cells (Figure 7 b d ) . After NaOH treatment of cells, the cytoplasmic membrane was still visible underneath the cell wall layer but it was convoluted and in places it appeared interrupted (Figure 7). Properties of alka Ii-ex tractable muter i d

Since ruthenium red has been shown to stain polysaccharides it was anticipated that the NaOH extracts of strain Channon would contain carbohydrate as well as protein. Measurements of amounts of protein and carbohydrate present in extracts revealed that alkali treatment of whole cells solubilised 64-9.6 per cent of total cell carbohydrate and 10.9-1443 per cent of total cell protein (Table 1). More carbohydrate (up to 35 per cent more) was extracted from strain Channon cells with alkali than from cells of the mutants (Table 1). Conversely, less total protein was extractable from strain Channon cells with alkali (Table 1) compared with the mutants. Extracts from all strains contained a small amount of phosphorus (Table 1). Alkali-extracted material from S . gordonii was tested for its ability to inhibit saliva-mediated aggregation of cells. In addition, alkali extracts were deproteinised and the remaining carbohydrate component was tested for aggregation-inhibitory activity. Alkali-extracted material from S . gordonii Channon was found to be inhibitory to salivainduced aggregation of strain Channon cells (Figure 8). Alkali-extracts prepared from the mutant strains and containing similar amounts of

protein were not inhibitory to saliva-mediated aggregation of strain Channon cells (not shown). The deproteinised alkali extract from strain Channon did not inhibit saliva-mediated cell aggregation of strain Channon (Figure 8). DISCUSSION Aggregation and coaggregation reactions of oral streptococci have been associated with activities of cell-surface proteins. The isolation of mutants defective in cell-surface proteins assists in defining the cell surface components involved in adherence and aggregation. Previously, chemical mutagens have been used to generate streptococcal mutants. Although experimental conditions such as duration of exposure to mutagen, and concentration of mutagen, were designed so as to reduce the likelihood of inducing multiple mutations, it was not possible to be certain that mutant phenotype was the result of a single mutation. The isolation of spontaneous mutants makes it more likely that they carry point mutations, or alternatively mutations in chromosomal hotspots involving for example insertion elements. The mutants described in this paper arose spontaneously and were enriched-for by inability of their cells to self-aggregate. It seems highly likely that the phenotypic properties of these mutants resulted in each case from a single mutation. A mutation affecting cell-surface morphology might for example reside in a gene encoding a component of secretory pathway, an enzyme involved in polysaccharide synthesis, or in a gene encoding an important structural protein of the cell wall. In support of the latter possibility, isogenic mutants of S. gordonii DLI -Challis deficient in

285

AGGREGATION MUTANTS OF S. GORDON11

Figure 6 . Transmission electron micrographs of thin sections of S. gordoniicells stained with ruthenium red and osmium tetroxide. a, Strain Channon; b, strain C2; c, strain C3; d strain C8. Sections were not post stained. Magnification x 250 000

a 290-kDa cell-surface polypeptide had muchreduced amounts of RR-stained surface material.34 Lindah13' utilised self-aggregation of cells to select for and isolate mutants of S. pyogenes deficient in fibrinogen and Immunoglobulin A binding. The mutants lacked the surface fibrillar material characteristic of the presence of M protein. Utilising ruthenium red to stain polysaccharidecontaining material as well as proteins in the cell outer layers, we have shown that a similar enrichment procedure in S. gordonii produces mutants deficient in production of a specific surface poly-

peptide (150 kDa) as well as in surface polysaccharide. Ruthenium red staining has proved valuable for the staining of S. salivarius cells grown under a variety of condition^.'^^^^ The stain reveals a layer in S. ~alivarius,'~S . p a r a ~ a n g u i s , ' and ~ in S. gordonii (this article) that is outside the cell wall layer and that is similar to the thickness of the wall itself. Ruthenium red staining reveals components outside the wall not seen by conventional fixation with glutaraldehyde and staining with osmium tetroxide.' It probably stains the carbohydrate

286

R. McNAB AND H. F. JENKINSON

Figure 7. Transmission electron micrographs of thin sections of S. gordonii or mutant cells treated with 0.1 N NaOH and stained with ruthenium red and osmium tetroxide. a, Strain Channon; b, strain C2; c, strain C3; d, strain C8. Sections were not post stained. Magnification x 250 000

components of surface fibrils that are present on was isolated from the cell wall of S. sanguis by S. salivarius, S. s a n g u i ~ ' ~and S. g o r d ~ n i i . ' ~ Emdur , ~ ~ et al.' and more recently several repeating However, fimbriae that were detected on cells of unit polysaccharides have been isolated from S.parasanguis FW213 by negative-staining were not S. oralis and their structures determined.'*3933 stained with ruthenium red, l 7 so the RR-stained Differences in surface polysaccharide production material on the surface of S. gordonii Channon by mutants were not related to glucosyltransferase might not correspond to the fimbriae previously activities which were similar in wild-type and described.' The composition of the RR layer in mutants. The mutants described in this paper were deS. gordonii Channon is currently under investigation. The presence of phosphorus and carbo- ficient in production of a polypeptide of molecular hydrate suggests it might contain lipoteichoic acid mass approximately 150 kDa, but in one mutant or other l i p ~ g l y c a n A . ~ ~phosphopolysaccharide there were several other protein differences. All

287

AGGREGATION MUTANTS OF S. GORDON11

100

80

-

h ._ I

’ Y

I

60-

3 .-

I

I

30

60

I

I

I

90 120 150 Time (min)

2

4

0

Figure 8. Aggregation of S. gordonii strain Channon cells in 10 per cent (v/v) heated saliva (0),in buffer only, no saliva ( O ) ,in heated saliva in the presence of 0.1 ml alkali-extract containing 40 pg protein and 30 pg carbohydrate ( A ) , in heated saliva in the presence of 0.1 ml deproteinised alkali extract containing 30 pg carbohydrate (A).Aggregation of mutant strain C2 cells in 10 per cent (v/v) heated saliva included for comparison (0)

mutants showed reduced RR-stained layers. We utilised alkali to extract this RR-staining material together with a subset of polypeptides. The alkali extraction method should prove useful for removal of surface carbohydrate and associated polypeptides from other streptococci. Alkali extraction has recently been shown to be ineffective in solubilising integral membrane lipoproteins in S. gordonii,21which suggests it does not disrupt the lipid bilayer. However the electron micrographs in this paper show some membrane convolution and interruption as a result of NaOH treatment. It is likely that at least some of the polypeptides that were alkali extracted were associated with the surface polysaccharide and RR-stained layer. The loss of RR-stained surface material in the mutants was concomitant with a reduction in cell surface hydrophobicity. A similar observation was made with mutants of S. parasanguis FW213 that

had reduced amounts of RR-stained material and all but one were reduced in hydrophobicity.’ These results contrast with S. salivarius HB and mutants where at higher growth rates in continuous cultures cells had thicker RR layer and decreased hydrophobicity.16 It would seem therefore that RRstained layer thickness and hydrophobicity are not necessarily correlated. While Harty et a f .l 7 could not show any surface protein differences in the S. parasanguis mutants, we have shown using alkali or SDS extraction of intact cells, alterations in surface protein composition in S. gordonii mutants. In ‘ S . milleri’ it has been shown recently that two major adherence factors, one proteinaceous and the other non-proteinaceous, contribute to cell attachment and aggregation. l o Spontaneous cellaggregating ability in ‘s.milleri’ was associated with higher hydrophobicity. It seems likely then that in S. gordonii and in ‘ S .milleri’ both polysaccharide and protein are involved in determining cell-surface properties. The alteration in surface protein and carbohydrate in mutants was associated with a reduction in saliva-mediated aggregation. This process in S. gordonii has been shown compellingly to involve surface protein SSP-56*7which is related to SpaP in S. mutans, SpaA in S. sobrinus, and to Antigen 1/11 shared by a number of other oral s t r e p t o c ~ c c i . ~ ~ Despite the evidence that SSP-5 polypeptide alone can initiate saliva-mediated cell a g g r e g a t i ~ n the ,~ results in this paper show that in S. gordonii Channon, production of SSP-5 antigen is not sufficient for saliva-mediated agglutination of cells to occur. Other cell-surface components, perhaps the 150-kDa polypeptide, are implicated as being involved in salivary agglutination of cells. In this respect it is already clear that saliva-mediated aggregation in S . gordonii Challis23 and in S . gordonii G9B3’ involves protein components not related to SSP-5 polypeptide. However, in S. gordonii Channon cell-surface polysaccharide was clearly not involved in saliva-mediated cell aggregation. The carbohydrate component of the alkali extract of S. gordonii Channon did not inhibit saliva-mediated aggregation of the wild-type cells, whereas the extract containing protein did inhibit. The alkali extracts of strains C2, C3 and C8 were all deficient in protein components necessary for saliva-induced aggregation since none of these extracts was significantly inhibitory to salivainduced aggregation of strain Channon cells. Because in strains C2 and C3 the only detectable protein difference was in the production of 150-kDa

288

polypeptide, these results suggest this polypeptide is involved in saliva-mediated cell aggregation. It is possible that the 150-kDa polypeptide is also involved in cell autoaggregation. Lamont et u Z . * ~ have suggested that SSP-5 adhesin may be involved in cell-cell aggregation and/or in interactions of streptococcal cells with other bacteria. The isolation of the mutants described here reveals that spontaneous mutation can result in the modification of cell-surface structure such that the cells no longer interact in the normal way with salivary components. Understanding the biochemical and genetic mechanisms operating these changes are important in the study of oral microbial adaptation to environment. In this case the spontaneous generation of genetic variants that are not recognised by salivary components, and whose cells remain non-aggregated, could select for bacteria resisting clearance. This might in turn assist streptococcal colonisation of oral surfaces.

R. McNAB AND H. F. JENKINSON

7.

8.

9.

10.

11.

ACKNOWLEDGEMENTS This work was supported by the Health Research Council of New Zealand and by The New Zealand Dental Research Foundation Board. We thank K. W. Knox (Institute of Dental Research, Chalmers St, Sydney, Australia) for kindly providing antiserum to SpaP polypeptide. The technical assistance of R. A. Easingwood and G. Gray-Young in electron microscopy, and of R. A. Baker is gratefully acknowledged.

12.

13. 14.

REFERENCES 1 . Abeygunawardana C, Bush CA, Cisar JO. (1990). Complete structure of the polysaccharide from Streptococcussanguis 522. Biochemistry 29,234-248. 2. Buchan RJ, Jenkinson HF. (1990). Glucosyltransferase production by Streptococcus sanguis Challis and comparison with other oral streptococci. Oral Microbiology and immunology 5,63-7 1 . 3. Cassels FJ, London J. (1989). Isolation of a coaggregation-inhibiting cell wall polysaccharide from Streptococcus sanguis H 1. Journal of Bacteriology 171,40194025. 4. Chen Jr PS, Toribara TY, Warner H. (1956). Microdetermination of phosphorus. Analytical Chemistry 28, 1756-1 758. 5. Cole RM, Calandra GB, Huff E, Nugent KM. (1976). Attributes of potential utility in differentiating among ‘group H’ streptococci or Streptococcus sanguis. Journal of Dental Research Special Issue A 55, A 142-AI 53. 6. Demuth DR, Davis CA, Corner AM, Lamont RJ, Leboy PS, Malamud D. (1988). Cloning and expression of a Streptococcus sanguis surface antigen that

15.

16.

17.

18.

interacts with a human salivary agglutinin. Infection and Immunity 56,2484-2490. Demuth DR, Berthold P, Leboy PS, Golub EE, Davis CA, Malamud D. (1989). Saliva-mediated aggregation of Enterococcus faecalis transformed with a Streptococcus sanguis gene encoding the SSP-5 surface antigen. Infection and immunity 57, 1470-1475. Demuth DR, Golub EE, Malamud D. (1990). Streptococcal-host interactions. Structural and functional analysis of a Streptococcus sanguis receptor for a human salivary glycoprotein. Journal of Biological Chemistry 265,7120-1126. Dubois M., Gilles KA, Hamilton JK, Rebers PA, Smith F. (1956). Colorimetric method for determination of sugars and related substances. Analytical Chemistry 28,350-354. Eifuku-Koreeda H, Yakushiji T, Kitada K, Inoue M. (1991). Adherence of oral ‘Streptococcus milleri’ cells to surfaces in broth cultures. Infection and Immunity 59,41034109. Emdur LI, Saralkar C, McHugh JG, Chiu TH. (1974). Glycerolphosphate-containing cell wall polysaccharides from Streptococcus sanguis. Journal of Bacteriology 120,724-734. Frandsen EVG, Pedrazzoli V, Kilian M. (1991). Ecology of viridans streptococci in the oral cavity and pharynx. Oral Microbiology and Immunology 6, 1 29- 133. Handley PS. (1990). Structure, composition and functions of surface structures on oral bacteria. Biofouling 2,239-264. Handley PS, Carter PL, Wyatt JE, Hesketh LM. (1985). Surface structures (peritrichous fibrils and tufts of fibrils) found on Streptococcus sanguis strains may be related to their ability to coaggregate with other oral genera. Infection and Immunity 47, 217-227. Handley PS, Hargreaves J, Harty DWS. (1988). Rutherium red staining reveals surface fibrils and a layer external to the cell wall in Streptococcus salivarius HB and adhesion deficient mutants. Journal of General Microbiology 134,3 165-3 172. Harty DWS, Handley PS. (1989). Expression of the surface properties of the fibrillar Streptococcus salivarius HB and its adhesion deficient mutants grown in continuous culture under glucose limitation. Journal of General Microbiology 135, 26 I 1-262 I , Harty DWS, Willcox MDP, Wyatt JE, Oyston PCF, Handley PS. (1990). The surface ultrastructure and adhesive properties of a fimbriate Streptococcus sanguis strain and six non-fimbriate mutants. Biofouling 2,75-86. Heussen C, Dowdle EB. ( I 980). Electrophoretic analysis of plasminogen activators in polyacrylamide gels containing sodium dodecyl sulfate and

289

AGGREGATION MUTANTS OF S. GORDONII

19.

20.

21.

22.

23.

24.

25.

26.

27.

28. 29.

copolymerized substrates. Analytical Biochemistry 102, 196202. Jenkinson HF. (1986). Cell-surface proteins of Streptococcus sanguis associated with cell hydrophobicity and coaggregation properties. Journal of General Microbiology 132, 1575-1 589. Jenkinson HF. (1987). Novobiocin-resistant mutants of Streptococcus sanguis with reduced cell hydrophobicity and defective in coaggregation. Journal of General Microbiology 133, 1909-191 8. Jenkinson H F. (1992). Adherence, coaggregation and hydrophobicity of Streptococcus gordonii associated with expression ofcell surface lipoproteins. Infection and Immunity 60, 122551228, Jenkinson HF, Carter DA. (1988). Cell surface mutants of Streptococcus sanguis with altered adherence properties. Oral Microbiology and Immunology 3,53-57. Jenkinson HF, Easingwood RA. (1990). Insertional inactivation of the gene encoding a 76-kilodalton cell surface polypeptide in Streptococcus gordonii Challis has a pleiotropic effect on cell surface composition and properties. Infection and Immunity 58,3689-3697. Jenkinson HF, Lala HC, Shepherd MG. (1990). Coaggregation of Streptococcus sanguis and other streptococci with Candida albicans. Infection and Immunity 58,1429-1436. Kilian M, Mikkelsen L, Henrichsen J. (1989). Taxonomic study of viridans streptococci: description of Streptococcusgordonii sp. nov. and emended descriptions of Streptococcus sanguis (White and Niven 1946), Streptococcus oralis (Bridge and Sneath 1982), and Streptococcus mitis (Andrewes and Horder 1906). International Journal of Systematic Bacteriology 39,471484. Kilian M, Reinholdt J, Nyvad B, Frandsen EVG, Mikkelsen L. (1989). IgAl proteases of oral streptococci: ecological aspects. Immunological Investigations 18, 161-170. Kolenbrander PE, Andersen RN. (1990). Characterization of Streptococcus gordonii (S. sanguis) PK488 adhesin-mediated coaggregation with Actinomyces naeslundii PK606. Infection and Immunity 58,3064-3072. Laemmli UK, Favre M. (1973). Maturation of the head of bacteriophage T4. I. DNA packaging events. Journal of Molecular Biology 80,575-599. Lamont RJ, Demuth DR, Davis CA, Malamud D, Rosan B. ( 1 991). Salivary-agglutinin-mediated

30.

31.

32.

33.

34.

35.

36. 37.

38. 39.

40.

adherence of Streptococcus mutans to early plaque bacteria. Infection and Immunity 59,34463450. Lindahl G . (1989). Cell surface proteins of group A streptococcus type M4: the IgA receptor and a receptor related to M proteins are coded for by closely linked genes. Molecular and General Genetics 216,372-379. Lowry OH, Rosehrough NJ, Farr AL, Randall RJ. (1951). Protein measurement with the Folin phenol reagent. Journal of Biological Chemistry 193, 265-275. Ma JK-C, Kelly CG, Munro G , Whiley RA, Lehner T. (1991). Conservation of the gene encoding streptococcal antigen 1/11 in oral streptococci. Infection and Immunity 59,26862694. McIntire FC, Crosby LK, Vatter AE, Cisar JO, McNeil MR, Bush CA, Tjoa SS, Fennessey PV. (1988). A cell wall polysaccharide from Streptococcus sanguis 34 that inhibits coaggregation of this organism with Actinomyces viscosus T14V. Journal of Bacteriology 170,2229-2235. McNab R, Jenkinson HF. (1992). Gene disruption identifies a 290 kilodalton cell-surface polypeptide conferring hydrophobicity and coaggregation properties in Streptococcus gordonii. Molecular Microbiology 6,in press. Nyvad B, Kilian M. (1990). Comparison of the initial streptococcal microflora on dental enamel in caries-active and in caries-inactive individuals. Caries Research 24,267-272. Pakula R, Walczak W. (1963). On the nature of competence of transformable streptococci. Journal of General Microbiology 31,125-1 33. Rosan B, Baker CT, Nelson GM, Berman R, Lamont RJ, Demuth DR. (1989). Cloning and expression of an adhesion antigen of Streptococcus sanguis G9B in Escherichia coli. Journal of General Microbiology 135,531-538. Sutcliffe IC, Shaw N. (1991). Atypical lipoteichoic acids of gram-positive bacteria. Journal of Bacteriology 173,7065-7069. Vickerman MM, Clewell DB, Jones GW. (1991). Sucrose-promoted accumulation of growing glucosyltransferase variants of Streptococcus gordonii on hydroxyapatite surfaces. Infection and Immunity 59, 3523-3530. Wyatt JE, Hesketh LM, Handley PS. (1987). Lack of correlation between fibrils, hydrophobicity and adhesion for strains of Streptococcus sanguis biotypes I and 11. Microbios 50,7-15.