mutants of actinomyces do not have these fim- briae (6, 47). The fimbrial subunit gene, fimA, has been cloned from A. naeslundii (A. viscosus). T14V (48) and A.
Coaggregation:
specific
bacteria P. E. KOLENBRANDER,’
N. GANPSHKUMAR,
Laboratory of Microbial Ecology, National Maryland 20892, USA
P.
Nearly all human oral bacteria exhibit coaggregation, cell-to-cell recognition of genetically distinct cell types. Clumps or coaggregates composed of the two kinds of cells are formed immediately upon mixing two partner cell types. Members of all 18 genera tested exhibit lactose-reversible coaggregation. Many of these interactions appear to be mediated by a lectin on one cell type that interacts with a complementary carbohydrate receptor on the other cell type. A lactose-sensitive adhesin has been isolated from Prevotella loesdzeii PK1295, and it exhibits the adherence properties observed with whole cells. Other adhesins have been identified and the genes for some of them have been cloned and sequenced. One Streptococcus sanguis adhesin is a lipoprotein that appears to have a dual function of recognizing both a bacterial carbohydrate receptor and a receptor in human saliva. Carbohydrate receptors for some adhesins have been purified from five oral streptococci, and they specifically block the coaggregations with the sfreptococcal partners that express the complementary adhesins. Coaggregation offers an explanation for the temporally related accretion of dental plaque and bacterial recognition of mucosal surfaces. Early colonizers of the tooth surface coaggregate with each other and late colonizers of the tooth surface coaggregate with each other, but with few exceptions, early colonizers do not recognize late colonizers. Furthermore, bacteria that colonize mucosal surfaces coaggregate with each other, indicating the high degree of specificity of coaggregation in the oral bacterial population.-Kolenbrander, P. E., Ganeshkumar, N., Cassels, F. J., Hughes, C. V. Coaggregation: specific adherence among human oral plaque bacteria. FASEBJ. 7: 406-413; 1993. ABSTRACt’
Key Words: adherence
coaggregation
dental
plaque
.
adhesins/receptors
Coaggregation is prevalent among bacteria isolated from the human oral cavity and was first reported by Gibbons and Nygaard in 1970 (1). There have been only occasional reports of this kind of interbacterial interaction in other ecosystems (for reviews, see ref 2, 3). Here, notions regarding coaggregation will be placed in the perspective of the temporal appearance of different taxonomic groups of bacteria on a freshly cleaned tooth surface and the coaggregations among these groups. Considerable information has accumulated concerning coaggregation partnerships and the taxonomic identification of colonizing bacteria. In contrast, less is known
406
J. CASSEIS2
AND
C. V. HUGI1!.
of.D#{243}n*alReseanh, Nadonld
Institute ...
.
lflstitutes
of.Health.
Bethesda,
.
about the mediators of coaggregation. This review will focus on human oral bacterial coaggregation and will not discuss salivary aggregation and mediation of adherence to oral surfaces by salivary molecules (see ref 4 for a recent review). It is the future of studies of bacterial coaggregation to understand the surface recognition phenomena, and this review will reflect the current information available on coaggregation-mediating adhesins and their complementary carbohydrate receptors.
DEFINITION AND COAGGREGATION
PRINCIPLES
OF
Coaggregation is the result of cell-to-cell recognition between distinct cell types. Macroscopically, the phenomenon can usually be detected as clumping when the different cell types are mixed. Microscopically, the clumps of cells formed consist of a network of interacting cell types. The recognition may be intrageneric, intergeneric or multigeneric in nature (2). In all three kinds of coaggregations, the cells appear to interact independently of other cells in the population (5). Viable as well as dead cells coaggregate, indicating that interacdons are dependent on existing surface molecules and not on a response by viable cells. Each cell type has specific surface components that mediate coaggregation. They are stably expressed by oral bacteria: fresh isolates and stock culture collection strains both exhibit the same range of coaggregation properties. Adhesins, proteinaceous surface molecules, on one cell type recognize carbohydrate receptors on partner cell types in many of the coaggregations studied so far. Many coaggregations between members of all 18 genera tested are inhibitable by lactose or other -galactoside derivatives. But there also are many coaggregations that are not inhibited by sugars.
‘To whom correspondence should be addressed, at: Bldg. 30, Room 310, National Institutes of Health, 9000 Rockville Pike, Bethesda,
MD 20892, USA
2Current address: Department of Gastroenterology, Division of Medicine, Walter Reed Army Institute of Research, Washington, DC 20307, USA 3Current address: Department of Pediatric Dentistry, Indiana University, Indianapolis, IN 46202, USA. Abbreviations: Rha, L-rhamnose; Gal, D.galactose; GalNac, Nacetyl-n-galactosamine; Glc, n-glucose; Glyc, glycerol; P04-, phosphate; OAc, 0-acetyl; p. pyranose; f furanose. 0892-6638/93/0007-0406/$O1 .50. © FASEB
The most frequently isolated species of bacteria from dental plaque belong to these 18 genera. Thus, most dental plaque bacteria coaggregate. Coaggregating pairs exhibit properties of competition and cooperation found among populations in other ecosystems. If two or more cell types recognize a common coaggregation partner, these cells may compete for binding sites on the common partner. Cell types compete if they coaggregate with the common partner by the same mechanism, e.g., lactose-sensitive coaggregation (6, 7). In contrast, when strains coaggregate with the common partner by different mechanisms, the common partner can act as a coaggregation bridge among the three participating cell types (7). The structure of coaggregates is dependent on the shape of the cell types and the ratio of the coaggregating partners. When approximately equal numbers of coaggregating cells are mixed, large amorphous coaggregates of interacting cell types are formed. If one of the cell types is in a 10-fold or greater excess, morphological shapes of corn cobs and rosettes can be formed (2). For example, when a long rod such as Fusobacterium nucleatum is mixed with an excess of spherical cells like Streptococcus sp., corn cob-shaped coaggregates can be detected where the central fusobacterial cell is surrounded by streptococci (3, 8). Such corn cob-like structures were originally observed in 1972 in scanning electron micrographs of plaque (9). Corn cobs can also be formed between fusobacteria and a variety of other oral bacteria (3). When one cell is relatively small and the partner is rod shaped
and in 10-fold or greater excess, rosettes can be formed (2, 3). The central cell of a rosette is sequestered by the exterior cells, and it is not available for coaggregation with a third partner cell type (10). In effect, the exterior cells compete successfully with newly added third-partner cells. However, partners of the exterior cells can act as a bridge between rosettes and contribute to the formation of large multigeneric coaggregates (5, 10). These cell-tocell interactions presumably play a role in bacterial plaque accretion, as they are routinely observed at the periphery of dental plaque (11). Our understanding of the cell types that participate in coaggregations has guided the course of further investigations into identifring adherence-relevant mechanisms. The highly specific nature of coaggregation partnerships has made it easier to predict the properties of the coaggregation mediators. Adhesins and their genes are being isolated. Carbohydrate receptors are being purified and characterized. The objective is to mimic coaggregation with purified complementary molecules and determine the role of these molecules in mediating colonization of human oral bacteria. Also, inhibitors of these complementary interactions are being developed in an effort to interrupt bacterial colonization of oral surfaces.
BACTERIAL
PLAQUE
ACCRETION
The fact that nearly all oral plaque bacteria from the 18 genera so far examined exhibit coaggregation suggests that this property contributes to bacterial plaque accretion. Additional support comes from the correlation of coaggregation partners and the temporal appearance of
ORAL BACTERIAL ADHERENCE
different bacterial genera on a freshly cleaned tooth surface. About 60-80% of the bacteria that colonize freshly cleaned teeth in the first 4 h are viridans streptococci (12). The remainder are mostly actinomyces, veillonellae, and haemophili. Subgingival plaque samples from sites on healthy teeth contain primarily streptococci, actinomyces, veillonellae, and fusobacteria (13, 14). If plaque is left undisturbed, an increasingly complex population of bacteria develops with a shift from a primarily Gram-positive to a primarily Gram-negative flora (15). This shift is often accompanied by the advent of gingivitis or more severe periodontal diseases. Such temporal changes in some of the genera of accreting bacteria are presented diagrammatically in Fig. 1, where early colonizers are shown adjacent to the tooth surface and late colonizers are depicted distally. Cell types shown in contact with each other are known coaggregation
partners. Due to the constraints of a two-dimensional representation, there are many additional coaggregations that cannot be shown. In general, early colonizers coaggregate with other early colonizers and with fusobacteria, and late colonizers (for example, Actinobacillus, Treponema, and Wolinella) coaggregate almost exclusively with fusobacteria. Thus, coaggregating partners are temporally related with respect to colonization of the tooth surface.
The
colonizers of a freshly cleaned tooth surto the acquired pellicle (Fig. 1), a mostly proteinaceous coating that immediately covers the tooth. The acquired pellicle consists of host-derived proline-rich proteins, statherin, a-amylase, mucins, and bacterial components such as enzymes, cell wall fragments, and membrane molecules. Some host molecules are known to be receptors for bacterial recognition. For example, proline-rich proteins are recognized by actinomyces and streptococci, but not fusobacteria (16, 17, 18). Statherin is recognized by fusobacteria and actinomyces but not streptococci, which bind to a-amylase (16, 17, 18, 19). After the tooth surface is covered with the earliest colonizers, each newly accreted cell becomes a nascent surface for recognition by unattached cells. Accreted cells may also interact with cells that are juxtaposed in plaque. With respect to streptococci, it would be to their advantage to recognize other already adherent streptococci, and indeed, in vitro, they exhibit intrageneric coaggregalion (20) as do a few actinomyces. In stark contrast, none of the other nine genera tested showed intrageneric coaggregation. Presumably, streptococci use intrageneric coaggregation in combination with cell growth to establish large populations at these early times. Later colonizers do not need to possess this property, because the population has become more diverse and the available variety of cell surfaces is greater. Although fusobacteria are not among the earliest colonizers, they predominate in samples of dental plaque from both healthy and diseased sites (13, 14). They are shown in Fig. 1 as the major cell type available for recognition by early and late colonizers. Their ability to coaggregate with both early and late colonizers suggests that they could serve as adherent bridges between these two groups of bacteria. Generally, their coaggregation with Gram-negative partners is lactose inhibitable, whereas none of numerous sugars tested could inhibit their coagface
earliest
must
attach
407
L1MsCfIr’..
ACQUIRED
TOOTH
SURFACE
Figure 1. Diagrammatic representation of proposed temporal relationship of bacterial accretion and multigeneric coaggregation during the formation of dental plaque. The tooth surface is indicated to the left and is coated initially with a thin layer called the acquired pellicle, which is subsequently colonized by bacteria.
gregation with Gram-positive partners (21). Fusobacteria primarily ferment certain amino acids and the energy derived from this process can be used to transport glucose, galactose, or fructose (22). These features and a readily available source of sugars and amino acids, coupled with their unusual global coaggregation, may promote their numerical predominance independent of the state of health or disease in the oral cavity of the host.
Another ubiquitous group of bacteria in samples taken from healthy, gingivitis and other periodontitis sites is the veillonellae (13, 14). They possess the advantageous metabolic property of utilizing the lactic acid end product of sugar fermentation of streptococci, actinomyces and many other oral bacteria. These organisms coaggregate with other early colonizers like streptococci, actinomyces, Propionibacterium acnes, and Rothia dentocariosa (23), but they only coaggregate with fusobacteria among the late colonizers. Hughes et al. (23) isolated veillonellae from the tooth and tongue surfaces and compared their coaggregation patterns. Those from dental plaque coaggregated with streptococci and actinomyces also isolated
408
Vol. 7
March
1993
The
from the tooth, but those veillonellae from the tongue coaggregated mostly with streptococci from the tongue. Moreover, the 68 strains of veillonellae from the tongue were either V. atypica or V. dispar, whereas greater than 90% of the veillonellae from dental plaque were V. parvula (13, 14). These results strongly support the notion that coaggregation contributes to successful colonization of both the hard tooth surface and the mucosal tongue surface. Several other genera of bacteria are depicted in Fig. 1, and each most certainly contributes to plaque population dynamics. With the exception of Eubacterium, the later colonizers are Gram-negative. The ecological significance of coaggregation has been discussed in detail previously (3, 24). ADHESINS
An interesting Fig. 1 is the
bridging
partners.
of the
FASEB Journal
One
feature
of
the
coaggregations
between first
examples
one
shown type and of a coaggregation
cell
KOLENBRANDER
in its
Ef AL.
bridge (7) was Prevotella loescheii PK1295, which can serve as a bridge between Streptococcus oralis 34 and Actinomyces israelii PK14, two Gram-positive oral bacteria that are otherwise unable to coaggregate (25). Coaggregation with S. oralis 34 is inhibited by lactose, whereas coaggregation with A. israelii PK14 is not. A comparison of the properties of oral bacterial adhesins with other bacterial adhesins appears in a recent review (26); in the current review only the oral bacterial adhesins will be discussed. Adhesins
from
Gram-negative
bacteria
A coaggregation-mediating adhesin may mediate other kinds of adherence such as hemagglutination of erythrocytes (27). The first oral adhesin that was purified to electrophoretic homogeneity was isolated from P. loescheii PK1 295 (28). This adhesin is a lectinlike protein that is sensitive to 3-galactosides. It has a native molecular weight of 450,000, consists of six identical Mr 75,000 subunits, and is a basic protein with a p1 between 7.4 and 8.0. When in an aggregated form at pH 6.8, it mediates agglutination of S. oralis 34 and of a wide variety of sialidase-treated mammalian erythrocytes. When it is unaggregated at pH 4.6, it blocks coaggregation between P. loescheii PK1295 and S. oralis 34. Both this adhesin and a second prevotella adhesin, which mediates coaggregation with A. israelii PK14, are expressed simultaneously on P. loescheii PK1295 (29). Monoclonal antibodies to the lactose-sensitive adhesin block only lactose-sensitive coaggregations, and monoclonal antibodies to the second adhesin block only the coaggregation with A. israelii PK14 (30). The adhesins are minor fimbria-associated proteins and are presented on the cell surface at or near the tip of
thin fimbrial structures extending from the cells. These fimbriae are easily sheared from the prevotellae by mild sonication and the actinomyces-recognizing adhesin can be purified by adding cells of the coaggregation partner A. israelii PK14 as an affinity surface (J. London, unpublished results). By using monoclonal antibodies to the latter adhesin in immunoblot analyses, the adhesin size was determined to be 43 kDa (31). The theoretical maximum number of adhesin molecules on a cell surface is 300 and 500 molecules per cell for the 43- and 75-kDa adhesins, respectively (31). In contrast to the prevotella adhesins, the adhesin of a different Gram-negative bacterium, Porphyromonas gingivalis ATCC 33277, appears to be associated with the 40kDa fimbrillin monomer (32). Other Gram-negative bacteria that have no fimbriae appear to bear their adhesins as part of the outer membrane. Monoclonal antibodies that block coaggregation have been used to identif’ a 150-kDa adhesin of Capnocytophaga gingivalis DR2001 (33, 34) and a 155-kDa adhesin of Capnocytophaga ochracea ATCC 33596 (35). Anti-adhesin antibodies to other purified adhesins that recognize streptococci have been used to identif’ a 34-kDa protein from Haemophilus parainfluenzae HP-28 (36) and a 39.5-kDa protein from Fusobacterium nucleatum ATCC 10953 (8). The fusobacterium-recognizing receptors on the streptococci are polarly located, and when an excess of streptococcal cells is mixed with fusobacteria, the coaggregates look like corn
cobs.
Another fusobacterial corn-cob partner is Veillonella atypica PK191O, which also coaggregates with certain oral ORAL BACTERIAL
ADHERENCE
streptococci in a lactose-inhibitable way. The streptococci were used as an affinity matrix to bind veillonella adhesins released by mild sonication of veillonella cells (37). A 45-kDa veillonella protein was released from the streptococci only after treatment with 100 mM lactose. Coaggregation-defective mutants that are unable to coaggregate with the streptococci do not bear this 45-kDa adhesin. Both the mutant and the parent cells possess fimbriae (37). Thus, the 45-kDa lactose-sensitive adhesin may be like the 75-kDa lactose-sensitive adhesin of P. loescheii PK1 295, appearing as a minor accessory protein at or near the tips of fimbriae rather than as a fimbrial subunit (29). These adhesins are functionally similar because they both specifically recognize the same cluster of streptococcal partner strains involved in lactose-sensitive coaggregations (P. E. Kolenbrander and J. London, unpublished observations). And coaggregation-defective mutants of the streptococcal partner, which were selected by their failure to coaggregate with the prevotella, also fail to coaggregate with the veillonella (P. E. Kolenbrander and J. London, unpublished results). Similarly, coaggregation-defective mutants of the streptococcal
partner, which were selected by their failure to coaggregate with the veillonella (38), also fail to coaggregate with the prevotella (C. V. Hughes and P. E. Kolenbrander, unpublished observations). Adhesins
from
Gram-positive
bacteria
Studies of coaggregation-mediating adhesins among the Gram-positive oral bacteria have been limited to only a few streptococci and actinomyces. Streptococcus salivarius HB presents a 91 nm surface fibril that is absent on S. salivarius HB-V5, a mutant that is unable to coaggregate with Veillonella parvula Vi (39). The presumptive adhesin, antigen B, is 380 kDa under native and denatured conditions and is the largest adhesin reported from oral bacteria (40). Insertional inactivation of a streptococcal gene encoding a 76-kDa adhesin that mediates coaggregation with actinomyces was used to study adherence-related functions of S. gordonii DL1 (41). Gene-inactivated mutants were deficient in other adherence recognition functions such as binding to immunoglobulin A and cell aggregalion in serum and saliva, which suggests that the 76-kDa protein is important in the general functional conformation of the cell surface. This surface-exposed adhesin termed SarA is a lipoprotein that is tightly associated with the cytoplasmic membrane (42).
The idea that lipoproteins in Gram-positive bacteria may function as adherence-mediating proteins is a rapidly developing area of research. The sequence Leu-X-XCys that corresponds to the consensus cleavage site for bacterial prolipoproteins by signal peptidase II is also present in the FiniA and SsaB proteins from S. parasanguisFW2l3 (43) and S. sanguis 12 (44), respectively. Both proteins are associated with binding to salivary molecules and the SsaB protein is thought also to mediate coaggregation with Actinomyces naeslundii PK606 (44). Although these two proteins show a very limited identity (15%) to SarA (42), they are both recognized by antiserum to another streptococcal coaggregation adherence protein, ScaA, from S. gordonii PK488 (45; R. N. Andersen et al., unpublished results). ScaA is 38 kDa and 409
mediates coaggregation with A. naeslundii PK606 (45), and its gene, scaA, has been cloned in E. coli (R. N. Andersen et al., unpublished results). Oligonucleotides based on the scaA and ssaB sequences were used to probe the presence of related sequences in other oral streptococci. Streptococci that coaggregate with A. naeslundii PK606 all possess DNA fragments that hybridize to both scaA and ssaB probes (R. N. Andersen et aL, unpublished results). There is a growing awareness that these Grampositive lipoproteins may be homologous to components of binding protein-dependent transport systems of Gramnegative bacteria (42, 46). Many of the coaggregations between actinomyces and streptococci are lactose-inhibitable (24) and involve type two fimbriae on the actinomyces (47). Coaggregationdefective mutants of actinomyces do not have these fimbriae (6, 47). The fimbrial subunit gene, fimA, has been cloned from A. naeslundii (A. viscosus) T14V (48) and A. naeslundii WVU45 (49) and sequenced. These genes encode proteins of about 59 kDa, which are more than 70% homologous (50). Monoclonal antibodies to these proteins react with the fimbriae on the cell surface but do not block coaggregation between the actinomyces and streptococci, suggesting that the 59-kDa proteins are type 2 fimbrial subunits and not the adhesin molecules. In an attempt to identify the adhesin molecule (or molecules), antiserum against whole wild-type actinomyces cells and absorbed with coaggregation-defective mutant cells was tested for its ability to block coaggregation. The absorbed antiserum blocked coaggregation (C. M. Klier and P. E. Kolenbrander, unpublished results) and will be used in immunoblot analyses of cell-surface preparations from isogenic parent and mutant cells. This approach was first used successfully to identify putative coaggregationmediating adhesins in Capnocytophaga gingivalis DR2001 (33) and subsequently with Prevotella liiescheii PK1295 (25), Streptococcus gordonii PK488 (45), and Veillonell.a atypica PK191O (37). In each case the absorbed antiserum blocked coaggregation specifically with the parent and the partner used to select for the isogenic coaggregationdefective mutant, whereas certain other coaggregations between the parent and unrelated partners were unaffected by the absorbed antiserum. While spontaneously occurring coaggregation-defective mutants are useful in identifying putative adhesins, their shortcoming is that the gene (or genes) encoding the adhesin (or adhesins) are not easily identified. To address this deficiency, transposon mutagenesis has been used to identify coaggregation-relevant genes (D. L. Clemans and P. E. Kolenbrander, unpublished results). This method is based on inactivating target genes (encoding coaggregation mediators) by inserting the transposon Tn 916, which contains the tetM gene encoding tetracycline resistance. Bacteria expressing drug resistance and the coaggregation-defective phenotype can then be used to map the coaggregation-relevant gene. This approach in combination with examination of spontaneously occurring mutants should yield information concerning both the adhesin molecule and the genetic organization of coaggregation-relevant genes. The coupling of transposonderived mutants and naturally occurring mutants with site-directed mutagenesis and monoclonal antibodies that specifically block only certain coaggregations is likely to
410
Vol. 7
March
1993
result in defining the domains binding sites of adhesins.
specifically
involved
in the
RECEPTORS Carbohydrate receptors that mediate coaggregation have been isolated and extensively characterized from five streptococci (Fig. 2). All five of the carbohydrate receptors are linear cell wall polysaccharides, each having been removed from the cell wall by the action of mutanolysin, a mureinolytic enzyme. Several structural features are common to these receptor polysaccharides. All contain galactopyranose and -galactofuranose, all are hexasaccharides or heptasaccharides, and all contain phosphodiester linkages. Coaggregation-defective mutants of two of these strains have been isolated and neither possesses these receptors. The top four of the five polysaccharides (Fig. 2) are similar in structure and in biological activity and will be discussed as a group. Two common features of these four polysaccharides are 1) the oligosaccharide repeating units are linked by phosphodiester bonds to the 6-carbon of the nonreducing terminal sugar, and 2) all contain N-acetylgalactosamine. The top two streptococcal polysaccharides contain --Galfi 1-36)-f3-GalpNAc ( 1-*3)-aGalp(1-*) and the next two contain -J-GalJ(1--*6)-fr Galp( 1 -i.3)-ct-GalpNAc ( 1-i). They differ in the order of the two sugars near the reducing end (GalNAc-Gal vs. Gal-GalNAc), but the glycosidic linkages are the same. This disaccharide region of the four oligosaccharides is considered to be the active actinomyces-adhesin binding site. This highly conserved receptor region has been described by Abeygunawardana et al. (51). All four of these streptococci are partners of actinomyces and all exhibit lactose-inhibitable coaggregalions with some of the actinomyces. Yet the degree of coaggregation of the streptococcus-actinomyces coaggregation varies with each streptococcus and any common-partner actinomyces. For example, S. sanguis C104 coaggregates poorly with A. naeslundii T14V, whereas the other three show strong coaggregation. One noticeable difference in the carbohydrate structure of strain C104 is that it is a ribitol teichoic acid, whereas the other three polysaccharides contain phosphodiester linkages involving glycosyl phosphate residues. The presence of the ribitol and an additional galactofuranose residue apparently gives a high degree of internal motion by conferring greater flexibility to the polysaccharide as evidenced in two-dimensional NMR experiments (51). This greater flexibility may be responsible for the differential coaggregation patterns of the streptococci with A. naeslundii T14V. Other studies of receptor specificities by A. naeslundii ATCC 12104 and A. naeslundii (A. viscosus) LW toward GalNAcf-containing glycoconjugates revealed similar results of subtle differences in coaggregation patterns of the two actinomyces (52). But large differences such as the presence and absence of coaggregation between S. oralis MPB1 and actinomyces strains ATCC 12104 and LW, respectively, were also observed. Both actinomyces strains bind to receptors on other cells in a GalNAc13GalcsOethyl-sensitive fashion but the effect of neighboring sugar groups on this binding is different with the two
The FASEB Journal
KOLENBRANDER
ET AL.
Streptococcus [-i
oralis
34:
PO4-.6)-a-D-GalpNAc(1--3)--L-Rhap(1-.4)-5-D-Glcp(1--.6)--D-Galj1-.6)-b-D-GalpNAc(1
Streptococcus
sanguis
[-PO4--46)-b-Galf(1--’
Streptococcus
]n
Cl 04: 3)-D-Galp(1--’6)-D-Galf(1--
oralis
-.3)-a-D-Galp(1--
6)--GalpNAc(1--’3)-a-Galp(1--.
1)-ribitol(5--.],,
J 22:
],
[-i’
ci-L-Rhap(1-.2)1
Streptococcus
oralis
AT CC 10557:
[-PO4---.
3)-a-GalpNAc(1 -.15 6 I (OAc)o.33
2
I (OAc)o.33
Streptococcus [-.3)-a-L-Rhap(1--.
oralis
AT CC 55229:
2)-a-L-Rhap(1-.3)-a-D-Galp(1--.
3)--D-Galp(1
-.4)--D-Glcp(1
-.3)--D-Galf(1
-.
Glyc(1-’P04-.6)1
Figure 2. Structures
of carbohydrate
receptors
recognized
in coaggregations
strains. The notion of proper configuration of adjacent and nearby sugar residues of carbohydrate receptors as sites for bacterial adherence has been reviewed by Karlsson (53). In contrast, the fifth streptococcal polysaccharide listed in Fig. 2 acts as a receptor for a Gram-negative bacterium (54). Of the five streptococci listed in Fig. 2, only S. oralis ATCC 55229 (S. sanguis Hi) coaggregates with C. ochracea ATCC 33596 and the polysaccharide blocks this coaggregation (54). 5. oralis ATCC 55229 contains a quite different
surface
receptor
that
lacks
N-acetylated
sugars
(55). The repeating hexasaccharide is not linked by phosphodiester bonds, but rather the phosphate is found as a glycerol phosphate substitution on an internal galactose moiety (56). The coaggregation is sensitive to lactose and galactose, 16-fold more sensitive to L-rhamnose, but insensitive to GalNAc. The most likely explanation for these data is that the C. ochracea ATCC 33596 adhesin contains a combining site for the unique di-rhamnose moiety of the S. oralis ATCC 55229 hexasaccharide. This Rhaai-2Rhaa is quite distinct from the di-rhamnose moiety found on S. oralis J22, which has Rha on the hexasaccharide core and Rhaa branching from it. The receptors of S. oralis strains 34 and ATCC 10557 have only a single Rha, and Rha is absent from the strain C104 carbohydrate. In all of the five receptor polysaccharides the putative adhesin binding site is on the reducing end of a galactofuranose residue. Additional sugar inhibition studies including oligosaccharides containing the galactofuranose moiety are needed to ascertain any role the
ORAL BACTERIAL ADHERENCE
among
human
oral bacteria.
may play in adhesin binding of all of the partners utilizing these streptococcal receptors. The receptor polysaccharide from S. oralis ATCC 55229 has a more striking structural similarity to S. pn#{128}umoniae type 23F and type 11 capsular polysaccharides than to the upper four polysaccharides from Fig. 2 (56). The upper four polysaccharides also have a high degree of structural similarity with S. pneumoniae type polysaccharides, including types 20, 19A, 29, and 13 (51). S. oralis is genetically related to S. pneumoniae (57), and therefore these similarities are not entirely surprising. galactofuranose
CONCLUSIONS Coaggregation is highly specific in that partners of one cell type may not be partners of a closely related cell type. All coaggregations occur between genetically distinct bacteria, including those that occur between bacteria of the same genus. The latter intrageneric coaggregations seem to be limited to those among streptococci and a few actinomyces; both of these groups of bacteria are initial colonizers and may use intrageneric coaggregation as a critical function in adherence. Recent work in this field has focused on identifying the mediators of coaggregation. Research approaches using transposon insertional inactivation of adherence-relevant genes are being developed and are expected to clarify the organization and regulation of associated genes. The possibility that some of these genes have been exchanged be-
411
tween organisms that colonize a specific habitat and that environmental factors regulate the expression of adherence-relevant genes may now be studied. Many coaggregations are inhibitable by specific sugars and are mediated by a lectin on one partner and a complementary carbohydrate receptor on the other. It is striking that several genetically diverse bacteria that colonize the tooth surface at a similar stage in plaque development have evolved similar mechanisms for adherence. The specificity of the lectins for certain receptors and the presentation of receptors on cell surfaces are active areas ges in the exposed sufficient for allowing
of research. receptor only certain
Relatively subtle chansurface appear to be cell types to recognize
a receptor. Coaggregations for which no inhibitors are known are equally prevalent and amenable to study. Some of these adhesins may recognize receptors on bacterial partner cells and similar receptor domains of host molecules that coat the freshly cleaned human tooth surface.
The authors thank Roxanna Andersen, Daniel Clemans, Jacob Donkersloot, and Jack London for helpful comments during
of the review.
the preparation
REFERENCES R. J., and Nygaard, M. (1970) Interbacterial aggregation of plaque bacteria. Arch. Oral Biol. 15, 1397-1400 2. Kolenbrander, P. E. (1989) Surface recognition among oral 1. Gibbons, bacteria:
multigeneric
coaggregations
and their mediators.
Ceit. Rev. MicrobioL 17, 137-159 3. Kolenbrander, P. E. (1991) Coaggregation: adherence in the human oral microbial ecosystem. In Microbial cell-cell Interactions (Dworkin, M., ed) pp. 303-329, American Society for Microbiology, Washington, D.C. 4. Hasty, D. L., Ofek, I., Courtney, H. S., and Doyle, R. J. (1992) Multiple adhesins of streptococci. Infect. Immun. 60,
2147-2152 5. Kolenbrander, P. E., and Andersen, R. N. (1986) Multigeneric aggregations among oral bacteria: a network of independent cell-to-cell interactions. J. BacterioL 168, 851-859 6. Kolenbrander, P. E., and Andersen, R. N. (1985) Use of coaggregation-defective mutants to study the relationships of cell-to-cell interactions and oral microbial ecology. In Molecular Basis of Oral Microbial Adhesion (Mergenhagen, S. E., and Rosan, B., eds) pp. 164-171, American Society for
Microbiology,
Washington, D.C. P. E., Andersen, R. N., and Holdeman, L. V. (1985) Coaggregation of oral Bacteroid.es species with other bacteria: central role in coaggregation bridges and competi-
7. Kolenbrander,
13. Moore, L. V. H., Moore, W. E. C., Cato, E. P., Smibert, R. M., Burmeister,J. A., Best, A. M., and Ranney, R. R. (1987) Bacteriology of human gingivitis. j Dent. Res. 66, 989-995
14. Moore, W. E. C., Holdeman, L. V., Cato, E. P., Smibert, R. M., Burmeister, J. A., Palcanis, K. C., and Ranney, R. R. (1985) Comparative bacteriology Infect. Immun. 48, 507-519
10. Kolenbrander,
P. E., and Andersen,
generic
rosettes:
sequestered
human
periodontal
bacteria.
R. N. (1988)
Inter-
surface recognition among Appl. Environ. Microbiol. 54,
1046-1050
11. Listgarten, M. A., Mayo, H., and Amsterdam, M. (1973) Ultrastructure of the attachment device between coccal and filamentous microorganisms in “corn cob” formation of dental plaque. Arch. Oral Biol. 18, 651-656 12. Nyvad, B., and Kilian, M. (1990) Comparison of the initial streptococcal microflora on dental enamel in caries-active and in caries-inactive individuals. Caries. Res. 24, 267-272
412
Vol. 7
March
1993
periodontitis.
15. Theilade, E., Wright, W. H., Jensen, S. B., and L#{244}e, H. (1966) Experimental gingivitis in man. II. A longitudinal clinical and bacteriological investigation. J. Penodont. Res. 1, 1-13 16. Gibbons, R. J., Hay, D. I., Cisar, J. 0., and Clark, W. B. (1988) Adsorbed salivary proline-rich protein-i and statherin: receptors for type 1 fimbriae of Actinomyces vizcosus TI4V-J1 on apatitic surfaces. Infect. Immun. 56, 29902993
17. Gibbons, R.J., Hay, D. I., and Schlesinger,
D. H. (1991) Delineation of a segment of adsorbed salivary acidic prolinerich proteins which promotes adhesion of Streptococcus gordonii to apatitic surfaces. Infect. Immun. 59, 2948-2954 18. Xie, H., Gibbons, R. J., and Hay, D. I. (1991) Adhesive properties of strains of Fusobacterium nudeatum of the subspecies nucleatum, vincentii and polymorphum. Oral Microbiol. ImmunoL 6, 257-263 19. Scannapieco, F. A., Bergey, E. J., Reddy, M. S., and Levine, M. J. (1989) Characterization of salivary a-amylase binding to Streptococcus sanguis. Infect. Immun. 57, 2853-2863 20. Kolenbrander, P. E., Andersen, R. N., and Moore, L. V. H.
(1990) Intrageneric
coaggregation
among strains of human
oral bacteria: potential role in primary colonization of the tooth surface. Appi. Environ. MicrobioL 56, 3890-3894 21. Kolenbrander, P. E., and Andersen, R. N. (1989) Inhibition of coaggregation between Fusobacterium nucl.eatum and Porphyromonas (Bacteroides) gingivalis by lactose and related sugars. Infect. Immun. 57, 3204-3209 22. Robrish, S. A., Oliver, C., and Thompson, J. (1987) Amino acid-dependent transport of sugars by Fusobacterium nucleatum ATCC 10953. j Bacteriol. 169, 389 1-3897 23. Hughes, C. V., Kolenbrander, P. E., Andersen, R. N., and
Moore, L. V. H. (1988) Coaggregation properties of human oral Veillonella spp.: relationship to colonization site and oral ecology. AppL Environ. MicrobioL 54, 1957-1963 24. Kolenbrander, P. E., and London,J. (1992) Ecological significance of coaggregation among oral bacteria. Adv. Microb. EcoL 12, 183-217 25. Weiss, E. I., Kolenbrander, P. E., London, J., Hand, A. R., and Andersen, R. N. (1987) Fimbria-associated proteins of Bacteroides loescheii PK1295 mediate tions. J. BacterioL 169, 4215-4222
intergeneric
coaggrega-
26. London, J. (1991) Bacterial adhesins. Annu. Reports Med. Chem. 26, 239-247 27. Weiss, E. I., London,J., Kolenbrander, P. E., and Andersen,
28. 29.
tions. Infect. Immun.
48, 741-746 8. Kaufman, J., and DiRienzo, J. M. (1989) Isolation of a corncob (coaggregation) receptor polypeptide from Fusobactevium nudeatum. Infect. Immun. 57, 331-337 9. Jones, S. J. (1972) A special relationship between spherical and filamentous microorganisms in mature human dental plaque. Arch. Oral Biol. 17, 613-616
of juvenile
30.
31.
R. N. (1989) Fimbria-associated adhesin of Bacteroides loeschei that recognizes receptors on procaryotic and eucaryotic cells. Infect. Immun. 57, 2912-2913 London, J., and Allen, J. (1990) Purification and characterization of a Bacteroid.es loeschei adhesin that interacts with procaryotic and eucaryotic cells. j BacterioL 172, 2527-2534 London, J., Hand, A. It, Weiss, E. I., and Allen, J (1989) Bacteroides loeschei PK1 295 cells express two distinct adhesins simultaneously. Infect. Immun. 57, 3940-3944 Weiss, E. I., London, J., Kolenbrander, P. E., Andersen, R. N., Fischler, C., and Siraganian, R. P. (1988) Characterization of monoclonal antibodies to flmbna-associated adhesins of Bacteroides loescheii PK1295. Infect. Immun. 56, 2 19-224 Weiss, E. I., London, J., Kolenbrander, P. E., Hand, A. R., and Siraganian, R. (1988) Localization and enumeration of fimbria-associated adhesins of Bacteroides loescheii. j Bacteriol.
170, 1123-1128 32. Goulbourne, P. A., and Ellen, it P. (1991) Evidence that Pophyromonas (Bacteroides) gingivalis fimbriae function in adhesion to Actinomyces viscosus. j BacterioL 173, 5266-5274 33. Kagermeier, A. S., and London, J. (1986) Identification and preliminary characterization of a lectinlike protein from Capnocytc.phaga gingivalis (emended). Infect. Immun. 51, 490-494
The FASEBJournal
KOLENBRANDER ET AL.
34. Tempro,
P., Cassels, F., Siraganian, R., Hand, A. R., and (1989) Use of adhesin-specific monoclonal antibodies to identii’ and localize an adhesin on the surface of C.apnocytophaga gingivalis DR2001. Infect. Immun. 57, London,
J.
3418-3424 35. Weiss, E. I., Eli, I., Shenitzki, B., and Smorodinsky, N. (1990) Identification of the rhamnose-sensitive adhesin of Capnocytophaga ochracea ATCC 33596. Arch. Oral Bzol. 35, 127s-130s 36. Lai, C-H., Bloomquist, C., and Liljemark, W. F. (1990) Purification and characterization of an outer membrane protein adhesin from Haemophilus parainfiuenzae HP-28. Infect. Immun. 58, 3833-3839
37. Hughes,
C. V., Andersen, R. N., and Kolenbrander, P. E. Characterization of Veillonella atypica PKi9iO adhesin-mediated coaggregation with oral Streptococcus Spp. In(1992)
fect. Immun. 60, 1178-1 186 38. Hughes, C. V., Roseberry, C. A., and Kolenbrander, P. E. (1990) Isolation and characterization of coaggregationdefective mutants of Veillonella atypica. Arch. Oral BioL 35,
i23S-125S 39. Weerkamp,
A. H., Handley, P. S., Baars, A., and Slot, J. W. (1986) Negative staining and immunoelectron microscopy of adhesion-deficient mutants of Streptococcus salivarius reveal that the adhesive protein antigens are separate classes of cell surface fibril. J. BacterioL 165, 746-755 40. Weerkamp, A. H., van der Mei, H. C., and Liem, R. S. B. (1986) Structural properties of fibrillar proteins isolated from the cell surface and cytoplasm of Streptococcus salivanus (Kp) cells and nonadhesive mutants. J. BadenoL 165, 756762 41. Jenkinson, H. F., and Easingwood, it A. (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. Infect. Immun. 58, 3689-3697 42. Jenkinson, H. F. (1992) Adherence, coaggregation, and hydrophobicity of Streptococcus gordonii associated with expression of cell surface lipoproteins. Infrct. Immun. 60, 1225-1228 43. Fenno, J. C., LeBlanc, D. J., and Fives-Taylor, P. (1989) Nucleotide sequence analysis of a type 1 fimbrial gene of Streptococcus sanguis FW213. Infect. Immun. 57, 3527-3533 44. Ganeshkumar, N., Hannam, P. M., Kolenbrander, P. E., and McBride, B. C. (1991) Nucleotide sequence of a gene coding for a saliva-binding protein (SsaB) from Streptococcus sanguis 12 and possible role of the protein in coaggregation with actinomyces. Infect. Immun. 59, 1093-1099
ORAL BACTERIAL ADHERENCE
45. Kolenbrander, P. E., and Andersen, it N. (1990) Characterization of Streptococcus gordonii (S. sanguis) PK488 adhesinmediated coaggregation with Actinoinyces naeslundii PK606. lnfect. Immun. 58, 3064-3072 Russell, R. R. B., Aduse-Opoku, J., Sutcliffe, I. C., Tao, L., and Ferretti, J. J. (1992) A binding protein-dependent transport system in Streptococcus mutans responsible for multiple sugar metabolism. j BioL Chem. 267, 4631-4637 47. Cisar,J. 0., Curl, S. H., Kolenbrander, P. E., and Vatter, A. 46.
E. (1983) Specific absence
of type 2 fimbriae
on a coag-
gregation-defective mutant of Actinomyces viscosus TI4V. Infect. Immun. 40, 759-765 48. Donkersloot, J. A., Cisar, J. 0., Wax, M. E., Harr, R. J., and Chassy, B. M. (1985) Expression of Actinomyc.es viscosus antigens in Escherichia coli: cloning of a structural gene (fimA) for type 2 fimbriae. j BactetioL 162, 1075-1078 49. Yeung, M. K, and Cisar, J. 0. (1988) Cloning and nucleotide sequence of a gene for Actinomyces naeslundii WVU45 type 2 fimbriae. j BacterioL 170, 3803-3809 50. Yeung, M. K (1992) Conservation of an Actinomyces viscosus T14V type 1 fimbrial subunit homolog among divergent groups of Actinomyces spp. Infect. Immun. 60, 1047-1054 51. Abeygunawardana, C., Bush, C. A., and Cisar, J. 0. (1991)
Complete Streptococcus
structure
of the cell surface
oralis C104: a 600-MHz
polysaccharide
NMR study.
of
Biochemistry
30, 8568-8577 52.
53. 54. 55.
56.
57.
Stromberg, N., and Bor#{233}n, T. (1992) Actinomyces tissue specificity may depend on differences in receptor specificity for GalNAcb-containing glycoconjugates. Infect. Immun. 60, 3268-3277 Karlsson, K-A. (1989) Animal glycosphingolipids as membrane attachment sites for bacteria. Annu. Rev. Biochem. 58, 309-350 Cassels, F. J., and London, J. (1989) Isolation of a coaggregation-inhibiting cell wall polysaccharide from Streptococcus sanguis Hi.j BacterioL 171, 4019-4025 Cassels, F. J., Fales, H. M., London, J., Carlson, R. W., and van Halbeek, H. (1990) Structure of a streptococcal adhesin carbohydrate receptor. J. BioL Chem. 265, 1412714135 Gluska,J., Cassels, F.J., Carlson, it W., and van Halbeek, H. (1992) Complete structure of the adhesin receptor polysaccharide of Streptococcus oralis ATCC 55229 (S. sanguis Hi). Biochemistry In press Kilpper-Balz, R., Wenzig, P., and Schleifer, K H. (1985)
Molecular
relationships
and classification
of some viridans
streptococci as Streptococcus oralis and emended description of Streptococcus oralis (Bridge and Sneath 1982). Int. J. Syst. Bacteriol. 35, 482-488
413
