Coassociation between Group B Streptococcus and Candida albicans Promotes Interactions with Vaginal Epithelium

ABSTRACT

Group B Streptococcus (GBS) is a leading cause of neonatal sepsis, pneumonia, and meningitis worldwide. In the majority of cases, GBS is transmitted vertically from mother to neonate, making maternal vaginal colonization a key risk factor for neonatal disease. The fungus Candida albicans is an opportunistic pathogen of the female genitourinary tract and the causative agent of vaginal thrush. Carriage of C. albicans has been shown to be an independent risk factor for vaginal colonization by GBS. However, the nature of interactions between these two microbes is poorly understood. This study provides evidence of a reciprocal, synergistic interplay between GBS and C. albicans that may serve to promote their cocolonization of the vaginal mucosa. GBS strains NEM316 (serotype III) and 515 (serotype Ia) are shown to physically interact with C. albicans, with the bacteria exhibiting tropism for candidal hyphal filaments. This interaction enhances association levels of both microbes with the vaginal epithelial cell line VK2/E6E7. The ability of GBS to coassociate with C. albicans is dependent upon expression of the hypha-specific adhesin Als3. In turn, expression of GBS antigen I/II family adhesins (Bsp polypeptides) facilitates this coassociation and confers upon surrogate Lactococcus lactis the capacity to exhibit enhanced interactions with C. albicans on vaginal epithelium. As genitourinary tract colonization is an essential first step in the pathogenesis of GBS and C. albicans, the coassociation mechanism reported here may have important implications for the risk of disease involving both of these pathogens.

INTRODUCTION

Streptococcus agalactiae (group B Streptococcus [GBS]) is a leading cause of invasive disease (sepsis, pneumonia, and meningitis) in neonates and is responsible for life-threatening infections in elderly and immunocompromised individuals (1–3). GBS is an opportunistic pathogen of the female genitourinary (GU) tract, with a carriage rate in Western countries of approximately 30% (2). The primary route of transmission to neonates is from the mother during or preceding birth, with estimated transmission rates of up to 50% (2). Among neonates that are colonized, about 1% develop severe GBS disease, resulting in significant infant morbidity or mortality (2, 4).

A variety of proteins that may promote colonization of host mucosae have been identified on the surface of GBS. These include pili (5), alpha C protein (6), BibA (7), serine-rich repeat proteins (Srr1/2) (5, 8), FbsA (9), Lmb (10), and the recently characterized antigen I/II (AgI/II) family proteins, designated BspA to -D (11, 12). Many of these surface proteins have been shown to target receptors expressed directly on the cervical or vaginal epithelia, while others bind extracellular matrix (ECM) proteins, such as collagen, fibrinogen, fibronectin, or laminin (5, 8–10). An additional colonization strategy for GBS, but one that remains poorly understood, is via interactions with other members of the vaginal microbiota. It is widely accepted that a “healthy” vaginal microbiota is dominated (ca. 70%) by the genus Lactobacillus, but Gram-positive bacteria (e.g., streptococci and staphylococci), Gram-negative bacteria (e.g., Escherichia coli), and yeasts (e.g., Candida albicans) are also frequently isolated (13). Of particular relevance to GBS colonization is a growing body of evidence indicating an association with the fungus C. albicans. In both developed and developing countries, vaginal carriage of C. albicans has been shown to be an independent risk factor for vaginal colonization by GBS (14–18).

C. albicans accounts for the fourth highest rate of systemic nosocomial infection in the United States (19), and as an opportunistic pathogen of the oropharynx and the female GU tract, it is the predominant cause of both oral and vaginal thrush. Key risk factors for C. albicans infection are immunosuppression, use of oral contraceptives, hormone therapy, antibiotics, diabetes, and pregnancy (20). A number of colonization determinants have been implicated in promoting candidal adhesion to and invasion of mucosae. These include proteins that are expressed on the surfaces of both morphological forms (blastospore and hypha) of C. albicans, such as Als1, Eap1, Eno1, Pra1, and Tdh1 (21–25). Other major candidal adhesins, including Hwp1, Als3, and Ssa1, are expressed exclusively on the filamentous hyphae (22, 26, 27). Similar to the case with GBS, epithelial receptor molecules (e.g., CEACAMs and cadherins) and ECM proteins (e.g., fibronectin and laminin) have been identified as targets of these C. albicans adhesins (25, 27–29).

Synergistic polymicrobial interactions have already been described for C. albicans and a number of Gram-positive bacteria. For example, the oral bacterium Streptococcus gordonii produces nutrient by-products that are stimulatory to C. albicans, enhancing the length of hyphal filaments (30). In turn, S. gordonii benefits from the reduced oxygen environment generated by C. albicans metabolism (31). Physical coadhesion of these two microbes also serves to promote retention of C. albicans within the oral cavity; the molecular basis of this was identified as recognition of the C. albicans adhesin Als3 by the S. gordonii AgI/II family protein SspB (32). Similar interactions have been reported for Streptococcus mutans and C. albicans, for which the S. mutans glucosyltransferase GtfB has been shown to bind mannans on the candidal cell surface, promoting robust cross-kingdom biofilm formation within the oral cavity of rats (33). In addition to niche colonization, interkingdom interactions may modulate disease progression. Streptococcus oralis and C. albicans synergize within the oropharynx to promote breakdown of epithelial tight junctions, resulting in enhanced systemic dissemination of C. albicans (34, 35). Likewise, Staphylococcus aureus has a high affinity for binding C. albicans hyphae and can “piggyback” on these filamentous forms as they infiltrate host cells to gain access to deeper tissues (36). Again, staphylococcal recognition of the C. albicans hyphal protein Als3 is critical for this coadhesion (37).

We recently characterized an AgI/II family polypeptide of GBS designated BspA. Alongside binding to the salivary pellicle and vaginal epithelium, BspA was shown to promote coaggregation of GBS strain NEM316 with C. albicans under planktonic conditions (12). The present study therefore aimed to build on these initial observations, determine in more detail the interkingdom interactions between GBS and C. albicans, and investigate the potential of these interactions to modulate the colonization or pathogenic capabilities of these two microbes within the GU tract.

RESULTS

Planktonic interactions of GBS and C. albicans.

The first step in exploring the interactions of GBS and C. albicans was to confirm their capacity to coaggregate under planktonic conditions. The following two strains of GBS that represent two of the most common capsular serotypes associated with neonatal disease were tested: GBS strain 515 (capsular serotype Ia) and strain NEM316 (capsular serotype III) (Table 1). C. albicans was fluorescently labeled with calcofluor white, while GBS strains were labeled with fluorescein isothiocyanate (FITC). Suspensions were then incubated together for 1 h before visualization by fluorescence microscopy. Both GBS strains were able to coaggregate with C. albicans, indicating that these interactions are not restricted to a single capsular serotype (Fig. 1). Furthermore, as reported previously (12), GBS strain NEM316 exhibited a tropism for C. albicans hyphae rather than blastospores. This binding pattern was also apparent with GBS strain 515, although higher levels of association were seen overall with strain NEM316 (Fig. 1). Taken together, these data confirmed that GBS can undergo planktonic interactions with C. albicans but implied that levels of coaggregation may be strain dependent.

TABLE 1.

Microbial strains and plasmids used in this study

Strain or plasmid	Unique ID	Relevant genotype	Reference or source
Strains
C. albicans
SC5314	UB1843	Wild type	Neil Gow, University of Aberdeen
	UB1941	Δals3	39
	UB1940	Δals3 pUL.als3	39
S. cerevisiae
BY4742	UB2156	pBC542-ALS3sm	40
S. agalactiae
NEM316	UB1931	Wild type	Shaynoor Dramsi, Institut Pasteur
515	UB2410	Wild type	Victor Nizet, University of California, San Diego
515	UB2873	ΔbspC	This study
L. lactis
NZ9800	UB2635	pMSP	12
	UB2658	pMSP.bspA	12
	UB2659	pMSP.bspC	This study
Plasmids
pMSP7517		E. coli-Enterococcus shuttle vector containing Enterococcus faecalis prgB under the control of the nisA promoter; erythromycin resistance	54
pMSP.bspC		pMSP7517-derived plasmid containing bspC from GBS 515 in place of prgB
pR326		Chloramphenicol resistance	58
pHY304		Temp-sensitive E. coli-Streptococcus shuttle vector	53

FIG 1.

Fluorescence micrographs of planktonic interactions between C. albicans and GBS. C. albicans SC5314 was grown in YNBPTG for 2 h at 37°C and 220 rpm before addition of GBS strain NEM316 (A) or GBS strain 515 (B) and incubation for a further 1 h. GBS was labeled with FITC (green), and C. albicans was labeled with calcofluor white (blue). Bars, 20 μm.

GBS-C. albicans interactions with VECs.

Since GBS and C. albicans are able to coaggregate, we hypothesized that such interactions can influence the capacity of these microbes to associate with vaginal epithelium. To this end, an in vitro assay was developed using the vaginal epithelial cell (VEC) line VK2/E6E7. In the first instance, epithelial cell monolayers were exposed either to GBS alone for 1 h or to C. albicans for 1 h to initiate hypha formation followed by GBS for a further 1 h. Associated GBS organisms were then enumerated by viable counts (CFU) from epithelial cell lysates. While GBS strain NEM316 showed a higher level of association (1.51 × 10⁵ CFU/monolayer) than that of strain 515 (7.57 × 10⁴ CFU/monolayer) (Fig. 2), both strains exhibited a strong affinity for the VEC monolayers. However, significantly larger numbers of bacteria were recovered for both strains in the presence of C. albicans. Numbers of GBS cells recovered from the epithelium were 1.9-fold higher for strain NEM316 and 2.1-fold higher for strain 515 than those from their respective monospecies samples (Fig. 2). These augmented effects were verified by confocal microscopy, although a slightly longer incubation period (5 h) was needed to obtain bacterial cell numbers that were of sufficient abundance to be clearly visible (Fig. 3). For monospecies samples, both GBS strains were evenly distributed across the VECs, but numbers of GBS cells per field of view were higher for strain NEM316 than for strain 515 (Fig. 3, columns 1 and 2). In the presence of C. albicans, an increase in the number of GBS cells associated with the VECs was apparent for both strains (Fig. 3, columns 3 and 4) compared to those for the monospecies equivalents. This was verified by quantification of GBS biovolume (Fig. 4A). In the presence of C. albicans, GBS biovolume levels were 4.6-fold and 2.8-fold higher for strains NEM316 and 515, respectively, than for their monospecies equivalents (Fig. 4A). Many GBS cells could be seen interacting with C. albicans hyphae, which formed extensive mats that overlaid the epithelial monolayers (Fig. 3, white arrows). However, there was also a visible increase in the number of GBS cells interacting with the epithelium in areas that were not seemingly colonized by C. albicans (Fig. 3, red arrows). This pattern was seen for both GBS strains. Augmentation of the GBS association with VECs after 5 h by C. albicans was further supported by enumeration of GBS cells from recovered epithelial lysates, and the effects were demonstrated more strikingly with GBS strain 515 (Fig. 4B).

FIG 2.

Effects of C. albicans SC5314 on association of GBS with VECs. VEC monolayers were incubated with GBS suspensions (MOI = 2.5) for 1 h (open bars) or with C. albicans (MOI = 2.5) for 1 h followed by GBS for a further 1 h (black bars). Monolayers were then lysed, and associated GBS cells were enumerated by serial dilution onto THY agar supplemented with 50 μg/ml nystatin. *, P < 0.05 compared to monospecies controls, as determined by unpaired Student's t test (n = 4).

FIG 3.

Representative confocal micrographs of C. albicans-GBS association with VECs. VEC monolayers were incubated with GBS alone for 5 h (columns 1 and 2) or with C. albicans for 1 h followed by GBS for a further 5 h (columns 3 and 4). Cells were then fixed, stained, and mounted onto glass slides. GBS was labeled using an Alexa Fluor 488-conjugated antibody (green), C. albicans was labeled with calcofluor white (blue), and VECs were labeled with phalloidin-TRITC (red). GBS strains NEM316 (top panels) and 515 (bottom panels) were tested. Columns 2 and 4 are duplicates of columns 1 and 3, respectively, in which the red filter (i.e., the VECs) has been removed (VK2/E6E7 off). Bars, 100 μm. White arrows indicate areas where GBS binds C. albicans hyphae, while red arrows indicate areas where GBS is found in the absence of C. albicans.

FIG 4.

Effects of C. albicans on association of GBS with VECs following 5 h of incubation. (A) Quantification of GBS cells from the confocal micrographs illustrated in Fig. 3. Images were processed using Volocity software, and Imaris software was used to calculate GBS biovolumes (in cubic micrometers). (B) Quantification of GBS cells by viable counts. VEC monolayers were incubated with GBS suspensions for 5 h (open bars) or with C. albicans for 1 h followed by GBS for a further 5 h (black bars). Monolayers were then lysed, and associated GBS cells were enumerated by serial dilution onto THY agar supplemented with 50 μg/ml nystatin. **, P < 0.01 compared to monospecies controls, as determined by unpaired Student's t test (n = 4).

To investigate the potential for a reciprocal relationship between GBS and C. albicans, the effects of GBS on the C. albicans association with vaginal epithelium were then explored using the same in vitro assay. For both strains tested, the presence of GBS resulted in a 4-fold elevation in the levels of C. albicans recovered from the VEC monolayers compared to those of C. albicans alone (Fig. 5). These data imply that a synergistic relationship exists between C. albicans and GBS and that each microbe can enhance association of the other with the vaginal epithelium.

FIG 5.

Effects of GBS on association of C. albicans SC5314 (Ca WT) with VECs. VEC monolayers were incubated with C. albicans cells for 1 h to allow production of hyphae. GBS suspensions were then added for a further 1 h before monolayers were lysed. Associated C. albicans cells were enumerated by serial dilution onto SAB agar supplemented with 5 μg/ml erythromycin. *, P < 0.05 compared to the monospecies control, as determined by unpaired Student's t test with Bonferroni correction (n = 3).

Role of diffusible signals in GBS-C. albicans interactions.

One potential mechanism for the enhanced recovery of both GBS and C. albicans cocultured with vaginal epithelium might be that each microbe releases some form of diffusible, chemical signal that either stimulates growth of the other or promotes its capacity to associate with VECs. To explore the first possibility, growth levels of GBS and C. albicans in single- and dual-species suspensions were compared. These studies were performed under conditions similar to those of the in vitro cell culture assay, using keratinocyte serum-free medium (K-SFM) and incubation periods of 1 to 2 h. No significant differences in numbers of CFU were seen for either species (Fig. 6A and B), regardless of whether they were grown under mono- or dual-species conditions. This implies that the presence of C. albicans does not affect the overall growth rate of GBS, and vice versa.

FIG 6.

Growth of C. albicans or GBS in mono- or dual-species suspension culture. K-SFM broth cultures were inoculated with C. albicans SC5314 (Ca WT) at 37°C and 220 rpm for 1 h before addition of GBS and incubation for a further 1 h (black bars). Alternatively, broth cultures were inoculated with C. albicans or GBS alone and incubated for 2 h or 1 h, respectively (open bars). The numbers of C. albicans CFU per milliliter were then determined by viable counts on SAB agar supplemented with 5 μg/ml erythromycin (A), while numbers of GBS CFU per milliliter were determined by viable counts on THY agar supplemented with 50 μg/ml nystatin (B). NS, P > 0.05 compared to the monospecies control, as determined by unpaired Student's t test (n = 3).

To determine if diffusible signals were modulating microbial interactions with the vaginal epithelium, GBS was incubated with VEC monolayers in K-SFM or in spent medium harvested from C. albicans grown in K-SFM for 1 h in the presence or absence of VECs. After 1 h of incubation with VEC monolayers, associated GBS cells were enumerated by viable counts from epithelial cell lysates. No significant differences in numbers (CFU per monolayer) of GBS cells recovered were observed across the different conditions (Fig. 7A).

FIG 7.

Role of contact-independent mechanisms or fixation in modulating interactions of C. albicans or GBS with VECs. (A) C. albicans SC5314 (Ca WT) was grown on VEC monolayers, or planktonically in K-SFM medium, for 1 h before spent media were collected and filter sterilized. GBS cells were incubated in these spent media on VECs for 1 h before the monolayers were lysed and associated GBS cells enumerated by serial dilution onto THY agar. NS, P > 0.05 compared to the blank K-SFM control, as determined by unpaired Student's t test (n = 3). (B) C. albicans was grown on VEC monolayers for 1 h before GBS suspensions or K-SFM alone was placed into transwell baskets suspended above. After a further 1 h of incubation, C. albicans was enumerated by serial dilution onto SAB agar. (C and D) VEC monolayers were fixed with 2% paraformaldehyde and then incubated with GBS suspensions (MOI = 2.5) for 1 h (open bars) or with C. albicans (MOI = 2.5) for 1 h followed by GBS for a further 1 h (black bars). Monolayers were then lysed, and numbers of GBS CFU per milliliter were determined by viable counts on THY agar supplemented with 50 μg/ml nystatin (C), while numbers of C. albicans CFU per milliliter were determined by viable counts on SAB agar supplemented with 5 μg/ml erythromycin (D). NS, P > 0.05; **, P < 0.01 compared to monospecies controls, as determined by unpaired Student's t test with Bonferroni correction (n = 4 [A and B] or 3 [C and D]).

For the reciprocal study, C. albicans was incubated on VEC monolayers for 1 h, and then suspensions of GBS or K-SFM alone were placed in transwell inserts above the VECs. Viable counts of C. albicans were determined after a further 1 h of incubation. Again, no significant differences in C. albicans association levels with VECs were seen in the presence or absence of GBS (Fig. 7B).

One final possibility explored was that GBS or C. albicans modulated the permissiveness of VECs to association with the other microbe via an active but contact-dependent mechanism. This was investigated by repeating the association assays with paraformaldehyde-fixed VECs. Fixation reduced the numbers of GBS cells recovered from the cell lysates overall. Nonetheless, the presence of C. albicans again resulted in elevated association levels of GBS (Fig. 7C), and the reciprocal effect was seen for levels of C. albicans recovered in the presence of GBS (Fig. 7D). Taken together, these data imply that neither intermicrobial diffusible signals nor active modulation of the VEC receptor profile is required for enhanced coassociation of GBS or C. albicans with the vaginal epithelium.

Role of Bsp protein in GBS-C. albicans interactions.

Oral streptococci have been shown to promote the colonization and retention of C. albicans within the oral cavity, and this is mediated in large part by coadhesion between the microbes (31). Having demonstrated similar coadhesion between GBS and C. albicans, the next step was to determine the molecular basis for this physical interaction and its contribution to the synergistic effects seen with vaginal epithelium.

We recently showed that the AgI/II family protein BspA of GBS strain NEM316 promotes coaggregation with C. albicans under planktonic conditions (12). We therefore wanted to build on this observation and determine if the Bsp adhesin family is important in GBS-augmenting interactions of C. albicans with VECs. In the first instance, a ΔbspC knockout mutant was generated in GBS strain 515, which carries only a single copy of the bspC gene (a homologue of bspA). This strain displayed only a modest (ca. 15%) reduction in association with VECs compared to that of parent strain 515 (Fig. 8A). However, it was reported previously that streptococci can compensate for loss of AgI/II family proteins by upregulation of alternative adhesins (38). To further explore the role of Bsp adhesins, inhibition studies were therefore performed using specific antisera. Anti-Bsp sera reduced the association of wild-type GBS strains NEM316 and 515 with VECs by 46% and 63%, respectively, compared to that with preimmune control serum (Fig. 8B). Together these data support previous evidence that Bsp adhesins have the capacity to promote GBS interactions with vaginal epithelium (12), but they indicate that there are additional determinants utilized by GBS for this purpose.

FIG 8.

Effects of C. albicans or Bsp antisera on the association of GBS wild-type and isogenic mutant strains with VECs. (A) VEC monolayers were incubated with GBS wild-type (WT) strain 515 or mutant ΔbspC cell suspensions (MOI = 2.5) for 1 h (open bars) or with C. albicans SC5314 (MOI = 2.5) for 1 h followed by addition of strain 515 suspensions for a further 1 h (black bars). Monolayers were then lysed and associated GBS cells enumerated by serial dilution onto THY agar supplemented with 50 μg/ml nystatin. **, significance relative to monospecies controls; Ω, significance relative to wild-type monospecies control; §, significance relative to the wild type in the presence of C. albicans. (B) GBS cell suspensions were preincubated with preimmune (open bars) or anti-Bsp (black bars) sera before incubation with VEC monolayers (MOI = 2.5) for 1 h and enumeration from cell lysates by viable counts. **, significance relative to preimmune controls. Significance indicates that the P value was <0.01 as determined by unpaired Student's t test with Bonferroni correction (n = 3).

In the presence of C. albicans, a more significant difference was seen between the parent strain 515 and the ΔbspC knockout strain. C. albicans significantly promoted recovery of both GBS strains from the epithelium compared to that with their respective monospecies samples. However, numbers of bacteria recovered were approximately 30% lower for mutant strain 515 ΔbspC than for parent strain 515 (Fig. 8A). These data imply that BspC plays a role in mediating GBS coassociation with C. albicans. However, additional adhesins must be involved and may compensate for the lack of BspC in strain 515 ΔbspC.

Given this apparent adhesin redundancy, surrogate Lactococcus lactis strains expressing BspA or BspC were then employed in coassociation assays. This allowed the functional properties associated with the individual AgI/II family proteins to be explored in greater detail. For monospecies L. lactis samples, once again, only a modest increase in numbers of bacteria recovered from VECs was seen for L. lactis strains expressing BspA or BspC compared to the numbers with the empty vector control strain (Fig. 9). However, for dual-species samples, recoveries of L. lactis strains expressing BspA and BspC were promoted 1.8-fold and 3-fold, respectively, by C. albicans (Fig. 9), while vector-only L. lactis control recovery was increased only slightly (<0.5-fold). Overall this implies that GBS AgI/II family proteins have the capacity to promote GBS association with vaginal epithelium directly, but they likely play a greater role by promoting association via C. albicans.

FIG 9.

Effects of C. albicans on association of L. lactis Bsp surrogate expression strains with VECs. VEC monolayers were incubated with suspensions of the L. lactis pMSP vector control, pMSP.bspA, or pMSP.bspC (MOI = 2.5) for 1 h (open bars) or with C. albicans SC5314 (MOI = 2.5) for 1 h followed by addition of L. lactis suspensions for a further 1 h (black bars). Monolayers were then lysed, and associated L. lactis cells were enumerated by serial dilution onto GM17 agar supplemented with 50 μg/ml nystatin. **, significance relative to monospecies controls; Ω, significance relative to the pMSP empty vector control; §, significance relative to the pMSP empty vector control in the presence of C. albicans. Significance indicates that the P value was <0.01 as determined by unpaired Student's t test with Bonferroni correction (n = 4).

Role of Als3 protein in GBS-C. albicans interactions.

A possible receptor for the Bsp proteins of GBS was the candidal glycoprotein Als3, since this adhesin is hypha specific (22) and has been shown to bind the AgI/II family protein SspB of S. gordonii to mediate interkingdom interactions (32). A C. albicans strain with both alleles of the als3 gene deleted (39) and a corresponding complemented strain (Δals3+als3) were used to determine if Als3 is involved in interactions between GBS and C. albicans. This was first investigated under planktonic conditions, and levels of coaggregation were determined semiquantitatively according to numbers of GBS cells associated with individual hyphae. Both GBS strains exhibited strong interactions with C. albicans wild-type strain SC5314 and the C. albicans Δals3+als3 strain, with the majority of hyphae recorded as binding 6 to 20 bacteria or >20 bacteria (Fig. 10). In contrast, the majority of C. albicans Δals3 hyphae were either devoid of bacterial cells or bound only 1 to 5 GBS cells (Fig. 10). Thus, the expression of Als3 on candidal hyphae is required to mediate strong physical interactions with GBS under planktonic conditions.

FIG 10.

Role of Als3 in planktonic interactions between C. albicans and GBS. (A) Fluorescence micrographs of planktonic interactions between the C. albicans Δals3 strain (left panels) or the Δals3+als3 complemented strain (right panels) and GBS strain NEM316 (top panels) or 515 (bottom panels). C. albicans was grown in YNBPTG for 2 h at 37°C and 220 rpm before addition of GBS and incubation for a further 1 h. GBS was labeled with FITC (green), and C. albicans was labeled with calcofluor white (blue). Bars, 20 μm. Note that interactions of GBS strains with wild-type C. albicans SC5314 are shown in Fig. 1. (B) Semiquantitation of C. albicans hyphae with 0, 1 to 5, 6 to 20, or >20 interacting GBS cells, based on approximately 40 randomly selected images for each experimental group. *, P < 0.05; NS, P > 0.05 (as determined by linear regression analysis of data sets) (n = 4).

The various C. albicans strains were then used to determine if Als3-mediated interactions were required to modulate GBS association with VECs. Interestingly, numbers of C. albicans Δals3 cells associated with VECs were not significantly different from those recovered for wild-type SC5314 or the C. albicans Δals3+als3 strain. This contrasts with observations made by others (39) in studies of oral epithelium and implies that Als3 may exhibit tissue-specific tropism. Unlike the phenomenon observed with wild-type C. albicans, there was no enhanced association of GBS with C. albicans Δals3 in the presence of VECs, and numbers of GBS cells recovered were comparable to those from monospecies samples (Fig. 11). In contrast, complementation of the Δals3 mutation in the C. albicans Δals3+als3 strain restored the capacity of C. albicans to significantly promote GBS association with vaginal epithelium relative to that of GBS monospecies samples (Fig. 11).

FIG 11.

Role of Als3 in synergistic effects of C. albicans on association of GBS with VECs. VEC monolayers were incubated with GBS suspensions for 1 h (open bars) or with C. albicans SC5314 (wild type [WT]) (black bars), Δals3 (gray bars), or Δals3+als3 (striped bars) for 1 h followed by GBS for a further 1 h. Monolayers were lysed, and then associated GBS cells were enumerated by serial dilution onto THY agar supplemented with 50 μg/ml nystatin. **, P < 0.01; NS, P > 0.05 (as determined by unpaired Student's t test with Bonferroni correction) (n = 4).

A similar scenario was seen for reciprocal studies to determine the role of Als3 in GBS modulation of C. albicans interactions with VECs. Numbers of C. albicans Δals3 cells recovered from epithelial monolayers were comparable for monospecies samples and dual-species samples incorporating either of the two GBS strains (Fig. 12). In contrast, both GBS strains enhanced the recovery of C. albicans Δals3+als3 2.5-fold (Fig. 12) relative to that of the monospecies control. These effects were similar to those seen previously with C. albicans wild-type strain SC5314 (Fig. 5). Thus, Als3 expression by C. albicans is required for both GBS and C. albicans to modulate coassociation with vaginal epithelium.

FIG 12.

Role of Als3 in synergistic effects of GBS on association of C. albicans with VECs. VEC monolayers were incubated with C. albicans SC5314 (wild type [WT]) (open bars), Δals3 (black bars), or Δals3+als3 (striped bars) for 1 h followed by GBS for a further 1 h. Monolayers were lysed, and associated C. albicans cells were enumerated by serial dilution onto SAB agar supplemented with 5 μg/ml erythromycin. **, P < 0.01; NS, P > 0.05 (as determined by unpaired Student's t test with Bonferroni correction) (n = 4).

Finally, studies were performed to investigate if Bsp polypeptides of GBS can bind directly to Als3 of C. albicans. Again, to avoid potential issues with adhesin redundancy, surrogate expression strains were utilized. A strain of Saccharomyces cerevisiae that expresses the small allele of C. albicans Als3 on its cell surface was previously generated (40). This S. cerevisiae (Als3⁺) strain was fluorescently labeled with FITC, while L. lactis strains expressing BspA, BspC, or an empty vector control were labeled with tetramethyl rhodamine isocyanate (TRITC). Suspensions were then incubated together for 1 h before visualization by fluorescence microscopy. No interactions were seen between S. cerevisiae (Als3⁺) and the L. lactis control (Fig. 13A). In contrast, coaggregation could clearly be seen with S. cerevisiae (Als3⁺) and L. lactis strains expressing either BspA (Fig. 13B) or BspC (Fig. 13C). Thus, GBS polypeptides BspA and BspC are direct binding partners for Als3 of C. albicans.

FIG 13.

Fluorescence micrographs of planktonic interactions between S. cerevisiae Als3 and L. lactis Bsp surrogate expression strains. S. cerevisiae (Als3⁺) was grown in YNBPTG for 3 h at 30°C and 220 rpm before addition of L. lactis(pMSP) (control) (A), L. lactis(pMSP.bspA) (B), or L. lactis(pMSP.bspC) (C) and incubation for a further 1 h. L. lactis was labeled with TRITC (red), and S. cerevisiae was labeled with FITC (green). Bars, 20 μm.

DISCUSSION

Intermicrobial interactions occur at most sites of colonization within the human body, and according to the National Institutes of Health, biofilms underpin approximately 80% of infections (41). In some instances, these interactions have antagonistic outcomes, such as those between C. albicans and Pseudomonas aeruginosa. Other partnerships are seemingly synergistic in nature, such as the interactions between C. albicans and S. gordonii, S. oralis, S. mutans, and S. aureus (32, 35, 37, 42). Several studies have reported the cooccurrence of GBS and C. albicans within the GU tract (14–18), and we recently provided evidence for coaggregation of these two microbes (12). The aim of this study was therefore to further define the interkingdom interactions of these two microbes and their capacity to modulate GU tract colonization, an essential step in the pathogenesis of both microorganisms.

Using the VEC line VK2/E6E7 as a model system, this study provides evidence that a reciprocal, synergistic relationship exists between GBS and C. albicans and may serve to promote their cocolonization of the vaginal mucosa. Specifically, when the organisms were incubated together, numbers of both microbes associated with the epithelial monolayers were found to be significantly larger than the numbers recovered from equivalent monospecies samples. Confocal microscopy revealed extensive hyphal “mats” of candidal cells overlaying the epithelial monolayers to which GBS cells were attached. This implies that direct physical contact (i.e., coadhesion) between GBS and C. albicans is a key mechanism that contributes to their synergistic interplay. Thus, GBS may bind directly to epithelium or to adherent C. albicans cells, and vice versa.

To identify the mechanistic basis of coadhesion between GBS and C. albicans, our studies focused on the hypha-specific adhesins of C. albicans, and specifically the adhesin Als3, since a distinct tropism for candidal hyphae was observed for both GBS strains tested. Use of Als3 knockout and complemented strains of C. albicans confirmed that recognition of this glycoprotein by GBS is required for effective coaggregation of these two microbes under planktonic conditions and for coassociation with vaginal epithelium. This correlates well with the interactions of C. albicans and streptococci within the oral cavity reported to date (32, 34) and thus may imply that Als3 recognition represents a common mechanism for C. albicans engagement by members of the Streptococcus genus. The addition of GBS to the list of microbes that utilize Als3 as a receptor, alongside other streptococci, S. aureus, and Rothia dentocariosa (32, 37, 43), also adds support to the notion that Als3 plays a major role in the capacity of C. albicans to mediate a diverse range of polymicrobial interactions.

In addressing the GBS side of this synergistic partnership, this study provides evidence for the role of GBS AgI/II family (Bsp) adhesins in this process. Previous work implicated BspA in facilitating coaggregation of GBS strain NEM316 with C. albicans under planktonic conditions (12). These data are supported here and were developed to include the adhesin BspC, implying that these capabilities may represent functions that are shared across the Bsp adhesin family. Moreover, loss of BspC impaired GBS coassociation with C. albicans, while expression of BspC by L. lactis enabled C. albicans to promote association of this surrogate host with VEC monolayers. This extends our current understanding of the properties of the adhesin family and implies that Bsp adhesins are determinants of GBS that facilitate coassociation with C. albicans on vaginal epithelium. Moreover, coaggregation of surrogate hosts expressing Als3 and Bsp adhesins adds support to the hypothesis that direct binding between Bsp polypeptides of GBS and Als3 of C. albicans is a mechanism that underpins, at least in part, the synergy in epithelial cell interactions between these two microbes. Interestingly, while deletion of bspC did not ablate the coassociation of GBS and C. albicans, deletion of both als3 alleles effectively prevented the interaction. This indicates a role for additional GBS determinants in mediating the interkingdom relationship and implies that these determinants may also target the candidal receptor Als3. This supports the evidence that Als3 has the capacity to bind multiple, diverse ligands (44).

Based on primary sequence, the AgI/II family polypeptides of GBS can be divided into four homologues: BspA and -B, which share 90% sequence identity, and BspC and -D, which share 99% sequence identity (12). The highest level of variation between BspA/B and BspC/D is seen within the N-terminal alanine-rich and proline-rich domains. In contrast, the V domain shares 96 to 100% sequence identity across all four Bsp homologues (12). The V domain has been identified as the functional region of a number of AgI/II family polypeptides (45–47), including BspA, in which it was shown to promote binding of GBS NEM316 to the scavenger receptor agglutinin glycoprotein-340 (12). If the V domain is also responsible for GBS coassociation with C. albicans, then the high level of sequence similarity may explain why both BspA and BspC display comparable functional properties. Delineating the precise domains within Bsp that are required for engagement with candidal Als3 will be the focus of future studies.

It is clear that direct physical contact between C. albicans and GBS plays a significant role in their coassociation with VECs. We also considered the possibility that intermicrobial signals played a role in the processes described here. However, no evidence was found for diffusible molecules released by either C. albicans or GBS having the capacity to significantly modulate microbial interactions with vaginal epithelium. Nonetheless, provision of additional intermicrobial binding sites may not be the only mechanism involved in the synergy with VECs. For example, in dual-species images, there were patches of epithelium that were heavily colonized by GBS while seemingly devoid of C. albicans (Fig. 3). Fixation of VECs did not inhibit coassociation of GBS and C. albicans, implying that these effects are not dependent upon modulation of epithelial cell biology (e.g., receptor availability). Nonetheless, it remains possible that GBS engagement with C. albicans alters the GBS receptor profile such that the bacteria are subsequently more permissive to interactions with VECs. The large impact of als3 gene deletion on the GBS-C. albicans-VEC coassociation raises the prospect that Als3 may mediate such effects. Future studies will explore these possible explanations.

To conclude, this study identifies a synergistic interplay between GBS and C. albicans that enhances the capacity of both microorganisms to associate with vaginal epithelial cells. Molecular determinants critical to this coassociation mechanism were identified as Bsp adhesins of GBS and Als3 of C. albicans. GU tract colonization is an essential first step in the pathogenesis of some diseases, such as vaginal thrush, and is a significant risk factor for neonatal GBS disease due to vertical transmission. Coassociation of GBS and C. albicans may therefore have important implications for disease risk for both of these opportunistic pathogens. This coassociation also raises the intriguing possibility of utilizing a convergent immunity approach to develop novel intervention strategies, as has been explored for C. albicans and S. aureus (48). There is currently no vaccine against GBS disease. Furthermore, while use of intrapartum antibiotic prophylaxis (IAP) has been successful in decreasing the incidence of early-onset neonatal GBS disease in some countries, the logistics of IAP make it an unrealistic control strategy for rural and developing countries, and IAP has had no impact on the rate of late-onset GBS infection (49, 50). The data presented here imply that better control of vaginal colonization by C. albicans may restrict or reduce GBS colonization, which in turn would reduce the risk of GBS transmission. Hence, vaccines against C. albicans, such as the promising rAls3 vaccine that has completed phase 1 clinical trials (51), may concomitantly help to reduce the burden of neonatal GBS disease.

MATERIALS AND METHODS

Microbial strains and culture conditions.

The microbial strains used in this study are listed in Table 1. GBS strains were cultured in Todd-Hewitt broth with 0.5% yeast extract (THY) or on THY agar plates at 37°C and 5% CO₂. L. lactis was cultured in GM17 broth (M17 broth supplemented with 0.5% glucose) or on GM17 agar plates at 30°C in a candle jar. Escherichia coli was cultured aerobically in Luria-Bertani (LB) broth or on LB agar plates at 37°C. Media were supplemented with 5 μg/ml erythromycin or with 50 μg/ml (E. coli) or 5 μg/ml (GBS) chloramphenicol, as appropriate. Heterologous protein expression in L. lactis was induced from nisin-inducible plasmids by the addition of 10 ng/ml nisin. Cells from GBS and L. lactis broth cultures were harvested by centrifugation at 5,000 × g for 7 min.

C. albicans was cultured in YPD medium (1% yeast extract, 2% mycological peptone, 2% glucose) at 37°C with shaking (220 rpm) or maintained on Sabouraud dextrose (SAB) agar plates incubated aerobically at 37°C. C. albicans cells were harvested from broth cultures by centrifugation at 5,000 × g for 5 min. S. cerevisiae was cultured in complete supplement medium (CSM) without uracil (ForMedium), supplemented with 0.67% yeast nitrogen base (YNB; Difco) and 2% glucose, at 30°C with shaking.

Generation of GBS knockout and L. lactis surrogate expression strains.

A ΔbspC mutant was generated in GBS strain 515 by in-frame allelic replacement with a chloramphenicol resistance gene cassette by homologous recombination, using a previously described method (52). Briefly, a knockout construct was generated by amplifying flanking regions directly upstream and downstream of the bspC gene from GBS strain 515 genomic DNA by using primer pairs bspC.F1/bspC.R1 and bspC.F2/bspC.R2, respectively (Table 2). A cat cassette was amplified from the chloramphenicol resistance plasmid pR326 by use of primers cat.F and cat.R (Table 2). Upstream and downstream bspC and cat amplicons were then combined by stitch PCR, using primers bspC.F1 and bspC.R2. The resultant amplicon was cloned into vector pHY304 (53) via XbaI and BamHI sites and propagated in E. coli Stellar cells (Clontech) prior to isolation and electroporation into GBS 515.

TABLE 2.

Primers used in this study

Primer name	Sequencea
bspC.F1	GCTCTAGAGCAATTAGCAGATGCACAG
bspC.R1	TAAAATCAAAGGAGAAAATATGAACTTTA
bspC.F2	GCTTTTATAATCAATATTCAGAAGCACTTG
bspC.R2	CGGGATCCGAGCCAAATTACCCCTCC
cat.F	AGAAAATATGAACTTTAATAAAATTGATTTAG
cat.R	TGAATATTGATTATAAAAGCCAGTCATTAGG
pMSP.bspC.F	CATGCCATGGAGGAGGAAATATGTATAAAAATCAAAAC
pMSP.bspC.R	CCGCTCGAGGCAGGTCCAGCTTCAAATC

^aRestriction endonuclease sites are underlined.

An L. lactis strain expressing BspA had been generated previously (12), and a similar methodology was employed here to generate an L. lactis strain expressing BspC. In brief, the bspC gene was amplified from GBS strain 515 genomic DNA by use of primers pMSP.bspC.F and pMSP.bspC.R (Table 2). The resultant amplicon was then cloned into the nisin-inducible expression vector pMSP7517 (54) via NcoI and XhoI sites, generating plasmid pMSP.bspC. This construct was transformed directly into electrocompetent L. lactis NZ9800 as described previously (12). Transformants were confirmed by plasmid isolation and PCR, while expression of BspC in L. lactis was verified by dot immunoblotting.

Tissue culture.

Experiments were conducted using VK2/E6E7 cells (ATCC CRL-2616), an immortalized human VEC line with a protein profile similar to that of the natural tissue (55, 56). VECs were cultured in K-SFM (Gibco) supplemented with 0.4 mM CaCl₂, 0.05 mg/ml bovine pituitary extract, and 0.1 ng/ml epidermal growth factor. Upon reaching 70 to 80% confluence, cells were disassociated by use of TrypLE Express trypsin replacement reagent (Gibco) before being harvested and resuspended in K-SFM. Appropriate volumes of cells were seeded in fresh flasks or in assay plates, as required.

Visualization of dual-species planktonic interactions.

Cells from 16-h cultures of C. albicans were harvested, washed in YNBPT (1× YNB, 20 mM Na₂HPO₄, 0.02% tryptone, adjusted to pH 7), and suspended to an optical density at 600 nm (OD₆₀₀) of 1.0 (equivalent to 10⁶ cells/ml) in YNBPT. This suspension was diluted 1:10 in YNBPTG (YNBPT supplemented with 0.4% glucose) and incubated at 37°C and 220 rpm for 2 h (2-ml final volume). These growth conditions have previously been shown to induce candidal hypha formation (57).

GBS cells were harvested from 16-h cultures, washed in YNBPT, suspended in 1.5 mM fluorescein isothiocyanate (FITC) dissolved in carbonate buffer (100 mM NaCl, 50 mM Na₂CO₃), and incubated for 30 min with gentle agitation. GBS cells were harvested and washed three times in carbonate buffer, and the pellet was suspended and adjusted to an OD₆₀₀ of 0.5 (equivalent to 5 × 10⁷ cells/ml) in YNBPTG. GBS suspension (1 ml) was added to that of C. albicans and incubated at 37°C for a further 1 h with shaking. Calcofluor white (0.00001% in distilled water [dH₂O]) was added before visualization of 10-μl samples by fluorescence microscopy.

For quantification assays, approximately 40 images of randomly selected hyphae were taken for each experimental group. Hyphal interactions were counted and placed into one of the following four groups, similar to a method reported previously (32): 0 interacting bacteria, 1 to 5 bacteria, 6 to 20 bacteria, and >20 bacteria per hypha.

In a variation of this assay, S. cerevisiae cells were harvested from a 16-h overnight broth culture in CSM broth, washed once in YNBPT (5 ml), and stained with 1.5 mM FITC for 30 min with gentle agitation. S. cerevisiae cells were harvested and washed three times in carbonate buffer. The pellet was suspended and adjusted to an OD₆₀₀ of 1.0 (equivalent to 10⁶ cells/ml) in YNBPTG before 1:5 dilution in YNBPTG (final volume, 2 ml). This suspension was incubated at 30°C and 220 rpm for 3 h. L. lactis cells were harvested from a 16-h overnight broth culture and washed once in YNBPT before suspension in 2 ml TRITC (0.1 mg/ml in carbonate buffer) and incubation for 30 min with gentle agitation. L. lactis cells were harvested, washed three times in carbonate buffer, and adjusted to an OD₆₀₀ of 0.5 in YNBPTG (equivalent to 5 × 10⁷ cells/ml). The adjusted L. lactis suspension (1 ml) was added to S. cerevisiae and incubated for a further 1 h at 30°C and 220 rpm before visualization of 10-μl samples by fluorescence microscopy.

Microbial growth in dual-species broth cultures.

Cells from an overnight (16 h) C. albicans suspension culture were harvested and washed once in phosphate-buffered saline (PBS). The pellet was suspended and adjusted to an OD₆₀₀ of 1.0 in K-SFM before being diluted 1:10 in K-SFM (2-ml final volume) and incubated at 37°C and 220 rpm for 2 h. Cells from overnight GBS broth cultures were harvested, washed once in PBS, and suspended in K-SFM to an OD₆₀₀ of 0.5. GBS suspension (1 ml) was added to the C. albicans suspension, and the mixture was incubated at 37°C for a further 1 h. Planktonic suspensions were vortex mixed for 15 s before being serially 10-fold diluted in THY broth. Numbers of microorganisms were detected by viable counts (CFU) on either THY agar plates (GBS) supplemented with 50 μg/ml nystatin to inhibit C. albicans growth or SAB agar plates (C. albicans) supplemented with 5 μg/ml erythromycin to inhibit GBS growth.

Epithelial association assay.

Epithelial association assays were conducted as described previously (5), with a few modifications. VECs were seeded into a 24-well plate at 2 × 10⁵ cells/well and incubated at 37°C and 5% CO₂ until confluent (48 to 72 h). C. albicans cells were diluted in K-SFM to obtain approximately 5 × 10⁵ cells/ml, while GBS or L. lactis cells were diluted in K-SFM to obtain approximately 5 × 10⁵ cells/ml.

Wells containing VEC monolayers were washed once with PBS, and approximately 5 × 10⁵ bacteria or C. albicans cells (1 ml; multiplicity of infection [MOI] = 2.5) were then added to each well. Bacterial suspensions were incubated at 37°C and 5% CO₂ for 1 h, while C. albicans suspensions were incubated for 2 h. For dual-species assays, C. albicans suspensions were incubated for 1 h before the medium was replaced by GBS or L. lactis and incubated for a further 1 h. For all assays, wells were then washed three times with PBS before incubation for 15 min with TrypLE, followed by two ice-cold water incubations, lasting 20 min each, to lyse the VECs. Lysates were serially diluted onto THY (GBS), GM17 (L. lactis), or SAB (C. albicans) agar plates and viable counts determined as described above. It was confirmed both visually and by monitoring levels of lactate dehydrogenase (LDH) released into the culture supernatants that epithelial monolayers remained intact and viable over the periods of the mono- or dual-species association assays.

In a variation of this assay, VEC monolayers were fixed in 2% paraformaldehyde overnight prior to incubation with cell suspensions of C. albicans and/or GBS. Alternatively, GBS suspensions were prepared as described above and preincubated at room temperature with 10 μg/ml rabbit preimmune or anti-Bsp sera (Eurogentec) for 30 min prior to incubation at 37°C for 1 h with VEC monolayers.

Spent medium studies.

VECs were seeded in a 24-well plate and grown to confluence. C. albicans was prepared as described above and then incubated with the VECs or grown planktonically in K-SFM medium for 1 h. The C. albicans medium was then collected and sterilized by filtration through a 0.2-μm filter. GBS suspensions, prepared as described above, were adjusted to an OD₆₀₀ of 1.0 in K-SFM before being diluted 1:200 in either fresh K-SFM, K-SFM from C. albicans planktonic growth, or K-SFM from C. albicans growth on VK2/E6E7 monolayers. Aliquots (1 ml) were added to VEC monolayers and incubated for 1 h. VECs were disassociated and lysed as described above, and numbers of GBS CFU were determined by serial dilution and viable counts on THY agar plates.

Transwell studies.

VECs were seeded in a 24-well plate and grown to confluence. C. albicans cells were prepared as described above and then incubated with VEC monolayers for 1 h before the medium was replaced with 1 ml K-SFM. Transwell inserts with high-density, 0.4-μm pores (Sarstedt) were placed into wells. GBS suspensions in K-SFM (OD₆₀₀ = 1.0) were diluted 1:100 in K-SFM. Aliquots (0.5 ml) were added to the transwell inserts, and the plates were incubated for a further 1 h. The inserts were removed, remaining VECs were disassociated and lysed as described above, and numbers of C. albicans CFU were determined by serial dilution and viable counts on SAB agar plates.

Confocal microscopy.

For visualization by confocal microscopy, VEC monolayers were grown on 19-mm glass coverslips in a 12-well plate until confluent. The epithelial association assay was then carried out as described above, except that the time was extended by 4 h. Calcofluor white (1 μl) was added to stain the chitin in the C. albicans cell wall, and the coverslips were then fixed in 2% paraformaldehyde. Triton X-100 (0.3%) was used to permeabilize the epithelial cells before blocking in 2% bovine serum albumin (BSA). Bacteria were stained with a mouse anti-GBS antibody (1.B.501; Santa Cruz Biotechnology) followed by an Alexa Fluor 488-conjugated goat anti-mouse antibody (Fisher), both of which were used at a dilution of 1:200. The F-actin of the epithelial cells was stained with phalloidin-TRITC (Sigma). Coverslips were then mounted onto glass slides by use of Vectashield reagent (Vector Laboratories) and imaged on a Leica SP5-AOBS confocal laser scanning microscope (CLSM) attached to a Leica DM I6000 inverted epifluorescence microscope. Images were processed using Volocity software, and Imaris v7.5 software (Bitplane AG, Zurich, Switzerland) was used to calculate biovolumes (in cubic micrometers).

Statistical analyses.

All assays were performed in triplicate unless otherwise stated. Data were analyzed using unpaired Student's t tests with Bonferroni correction, as appropriate.