Community Development between Porphyromonas gingivalis and Candida albicans Mediated by InlJ and Als3

ABSTRACT The pleiomorphic yeast Candida albicans is a significant pathogen in immunocompromised individuals. In the oral cavity, C. albicans is an inhabitant of polymicrobial communities, and interspecies interactions promote hyphal formation and biofilm formation. C. albicans colonizes the subgingival area, and the frequency of colonization increases in periodontal disease. In this study, we investigated the interactions between C. albicans and the periodontal pathogen Porphyromonas gingivalis. C. albicans and P. gingivalis were found to coadhere in both the planktonic and sessile phases. Loss of the internalin-family protein InlJ abrogated adhesion of P. gingivalis to C. albicans, and recombinant InlJ protein competitively inhibited interspecies binding. A mutant of C. albicans deficient in expression of major hyphal protein Als3 showed diminished binding to P. gingivalis, and InlJ interacted with Als3 heterologously expressed in Saccharomyces cerevisiae. Transcriptional profiling by RNA sequencing (RNA-Seq) established that 57 genes were uniquely upregulated in an InlJ-dependent manner in P. gingivalis-C. albicans communities, with overrepresentation of those corresponding to 31 gene ontology terms, including those associated with growth and division. Of potential relevance to the disease process, C. albicans induced upregulation of components of the type IX secretion apparatus. Collectively, these findings indicate that InlJ-Als3-dependent binding facilitates interdomain community development between C. albicans and P. gingivalis and that P. gingivalis has the potential for increased virulence within such communities.

of C. albicans. Screening mutants of P. gingivalis lacking minor fimbrial adhesin Mfa1 or hemagglutinin HagB found no effect on coadhesion with C. albicans (not shown), implicating InlJ as the predominant P. gingivalis adhesin in this interaction.
Role of InlJ in heterotypic biofilm formation. As cells in biofilms can display phenotypes that are distinct from those seen with their free-living counterparts, we investigated early biofilm formation using C. albicans and P. gingivalis. C. albicans biofilm formation was initiated on saliva-coated glass coverslips, and, as shown in Fig. 1C, C. albicans cells attached to the surface and formed hyphal filaments to which P. gingivalis cells clearly bound. Adherence of P. gingivalis ΔinlJ to hyphal filaments was diminished, consistent with the planktonic condition. The complemented strain of P. gingivalis, the cΔinlJ mutant, adhered to hyphal filaments under biofilm conditions at the same level as the wild type. These findings support the idea that InlJ is a mediator of P. gingivalis binding to C. albicans under both planktonic and sessile conditions. Inhibition of P. gingivalis interaction with C. albicans by InlJ. To provide further insight into the role of InlJ in P. gingivalis-C. albicans coadhesion, recombinant protein was expressed as a His-tagged fusion and tested for inhibition of P. gingivalis binding to C. albicans. Figure 2A shows that recombinant InlJ (rInlJ) inhibited P. gingivalis binding in the suspension assay in a dose-dependent manner and that up to 65% inhibition was seen in the presence of 20 g rInlJ. Control proteins bovine serum  albumin (BSA) and rLtp1, an irrelevant phosphatase protein from P. gingivalis (42), did not display inhibitory activity. The same inhibitory effect was observed when dualspecies biofilms were developed in the presence of rInlJ. P. gingivalis adherence to hyphae was reduced in a concentration-dependent manner in the presence of rInlJ protein (Fig. 2B). As relatively large amounts of soluble protein were required for inhibition, the presentation of InlJ on the bacterial surface may be necessary for the maintenance of optimal active structure. These findings support the model that InlJ mediates the attachment of P. gingivalis to C. albicans.
Als3 is necessary for C. albicans interactions with P. gingivalis. Previous studies demonstrated that the hypha-specific adhesin Als3 is important for C. albicans biofilm formation and for adhesion to host tissue and to the oral early plaque colonizer S. gordonii (43,44). Further, Als3 is associated with hyphae (45), to which P. gingivalis preferentially binds. The potential involvement of Als3 in P. gingivalis binding was examined using an als3Δ als3Δ-URA3 mutant (designated als3Δ). In the suspension assay, binding of C. albicans als3Δ to P. gingivalis decreased 58% (Fig. 3A). The als3Δ mutant formed sparse biofilms with few hyphae (Fig. 3B), consistent with the role of this protein in biofilm formation (46). Nonetheless, binding of P. gingivalis to the als3Δ mutant hyphae that were present was less than that seen with wild-type C. albicans. These results do not exclude the possible involvement of other Als proteins or, indeed, of other candidal adhesins, which may play a complementary or overlapping role in binding with Als3.
InlJ interacts with Als3. Our findings suggested that C. albicans Als3 can act as a component of an adhesin-receptor system with P. gingivalis. To investigate whether Als3 interacts directly with InlJ, we utilized Saccharomyces cerevisiae cells expressing C. albicans adhesins. In an enzyme-linked immunosorbent assay (ELISA), rInlJ bound to S. cerevisiae cells expressing Als3 (derived from either the large or small ALS3 allele) P. gingivalis-C. albicans Communities ® but not to control cells or cells expressing S. cerevisiae wall protein Cwp1 (Fig. 4A). Moreover, binding of P. gingivalis whole cells to S. cerevisiae expressing Als3 occurred only in the presence of InlJ (Fig. 4B). In control experiments, all strains of S. cerevisiae attached to the ELISA plates to the same degree (Fig. S2). These findings support a model whereby InlJ binds to hyphal Als3 to effectuate P. gingivalis-C. albicans association.
Transcriptional profiling of InlJ-dependent P. gingivalis-C. albicans interactions. RNA sequencing (RNA-Seq) was utilized to examine the transcriptional responses of P. gingivalis in InlJ-dependent communities with C. albicans under planktonic coculture conditions. Comparing the P. gingivalis wild type to P. gingivalis ΔinlJ coincubated with or without Candida, 256 P. gingivalis genes in the parental strain were downregulated in the coincubation, with 96 of these being unique (i.e., not regulated without Candida) (Fig. 5A). Among those 96 genes, 58 had a Log2 fold change level of Ͼ1 (see Table S1 in the supplemental material). Overall, there were markedly fewer differences in expression under the coculture condition than under the monoculture condition, and there were over twice as many genes downregulated in the parental strain as in the InlJ-deficient cells (Table S1). Similarly, 125 genes were upregulated in the P. gingivalis wild type in the presence of C. albicans, with 57 of these being unique to the coincubation (Fig. 5B). Among those 57 unique genes, 21 had a Log2 fold change level of Ͼ1 (Table S1).
The 57 unique coculture InlJ-dependent upregulated genes were assigned to 31 Gene Ontology (GO) terms with an overenrichment P value of Ͻ0.05 (Fig. 6A). Among those 31 GO terms, biological process terms comprised 10, cellular component terms comprised 2, and molecular function terms comprised 19. Among the biological process terms, the most significantly overrepresented terms comprised cell wall organization terms, cell cycle terms, and cell division terms. The most significantly enriched cellular component terms were ribosome terms and cell wall terms, and terms corresponding to metallopeptidase activity, structural constituent of the ribosome, and uracil DNA N-glycosylase activity were those most significantly enriched among the molecular function terms. Collectively, these results suggest that InlJ-dependent association with C. albicans increases growth and division of P. gingivalis. Consistent with this, 3 genes involved in peptidoglycan biosynthesis, murE, murC, and murG, were upregulated with InlJ present, although only one of those genes, murG, had a Log2 fold change level of Ͼ1. An illustration depicting the results of a STRING network analysis ( Fig. 7A) depicts genes corresponding to nodes, namely, ribosomal protein genes, peptidoglycan biosynthesis genes, and genes of the type IX secretion system (T9SS), according to k-means clustering, and also shows the potential for interactions among the products of these differentially regulated genes. There have been 18 components of the T9SS recognized in P. gingivalis to date, and the machinery is responsible for the translocation of over 30 proteins from the periplasm across the outer membrane (47)(48)(49). Targets of the T9SS include a number of virulence-associated proteins, including the gingipain proteases (47)(48)(49). Genes encoding 9 components of the T9SS machinery were upregulated in an InlJ-dependent manner; those genes included porPKLMN, representing an operon whose members are cotranscribed (Fig. 7B). The porPKLMN operon is controlled by the PorXY two-component system, which operates through SigP, an extracytoplasmic function (ECF) sigma factor. Genes encoding PorY and SigP were also upregulated by C. albicans. Although only two of the T9SS cargo proteins, PGN_1437 and the thiol protease PGN_0900, were upregulated (Table S1), C. albicans could potentially increase the pathogenicity of P. gingivalis indirectly through induction of type IX-dependent secretion of virulence factors, without affecting gene expression.
The 96 unique coculture InlJ-dependent downregulated genes were assigned to GO terms (Fig. 6B). Among those 96 genes, 25 showed overrepresentation (P value of Ͻ0.05). Of the 25 genes, 9 were part of the biological process GO category, with the term ЉpathogenesisЉ being the most highly represented followed by Љcell redox ho-meostasisЉ and then Љmetal ion transport.Љ Note that the data corresponding to the term ЉpathogenesisЉ were not P. gingivalis specific but were assembled from a variety of organisms. Overpopulation among the members of this category is derived from the genes for the fimbrial adhesins (fimA and mfa1) and the hagA gene encoding a hemagglutinin adhesin. Expression of these adhesion-associated genes may be up-FIG 5 Venn diagram representing differentially expressed genes that were either (A) downregulated or (B) upregulated in the P. gingivalis 33277 WT strain compared to the ΔinlJ mutant, with and without coincubation with C. albicans. Yellow shading represents unique genes in coculture, red shading represents unique genes in monoculture (Mono), and orange shading represents genes that appeared under both conditions. P. gingivalis-C. albicans Communities ® regulated in the mutant to compensate for the loss of InlJ. Data corresponding to the term Љtetrahydrofolate interconversionЉ also showed significant regulation, and this pathway has been found to have a significant impact on the pathogenicity of P. gingivalis-S. gordonii dual-species communities (50). Within the cellular component category, 3 terms were significantly enriched from the upregulated genes, namely, the terms Љcell outer membrane,Љ Љcell,Љ and Љpilus.Љ The majority belonged to the molecular function category, with 13 terms being enriched. The term Љprotein disulfide oxi-doreductaseЉ was shown to be the most significantly overrepresented term, followed by Љdrug transmembrane transport activity.Љ

DISCUSSION
C. albicans is an opportunistic fungal pathogen which colonizes the gut, genital tract, and oral cavity of healthy individuals. Disruption of immune surveillance or broad-spectrum antibiotic therapy can allow overgrowth and realization of pathogenic potential by the organism. C. albicans can cause superficial infections, such as oropharyngeal candidiasis (thrush) and vulvovaginal candidiasis, and also life-threatening systemic infections (17,51,52). In addition, C. albicans biofilm infections are common on prosthetic devices such as urinary or intravascular catheters, artificial joints, and voice boxes (53). C. albicans is a pleiomorphic fungus and can transition among three distinct morphological states: yeast cells, pseudohyphae, and filamentous hyphae. Different morphological states are associated with colonization and growth, and the hyphal form enables biofilm formation (54). Several virulence properties contribute to the pathogenic potential of the organism. Expressed cell wall adhesins, including the members of the agglutinin-like sequence family (Als) and hyphal wall protein (Hwp1), are crucial for C. albicans attachment to host tissue and for multispecies biofilm formation (44,55,56). C. albicans secretes a number of hydrolytic enzymes, such as lipases, esterases, and secreted aspartyl proteinases (Saps), that affect a variety of processes, including biofilm formation with streptococci, tissue invasion, and immune evasion (57)(58)(59). A newly reported candidalysin, a cytolytic peptide toxin secreted by C. albicans hyphae, causes damage to oral epithelial cells by intercalation, permeabilization, and calcium influx; triggers a proinflammatory signaling pathway response; and activates epithelial immunity (60).
The oral carriage rate of C. albicans in healthy subjects ranges from 25% to 60% (43). Successful colonizers of the oral microbiota generally interact synergistically with one another (61). P. gingivalis, for example, can attach to S. gordonii and accumulate into heterotypic communities, a process facilitated by several adhesin-receptor interactions and phosphotyrosine-dependent signaling within P. gingivalis (12,62). Similarly, C. albicans is usually present in heterotypic communities (16) and interacts synergistically with a variety of other organisms, including the oral streptococci (17,19,63). C. albicans can be isolated from periodontal pockets along with P. gingivalis (64,65). In this study, we found that P. gingivalis and C. albicans can coadhere, both in suspension and in sessile communities, and that interspecies binding in both contexts is mediated by the InlJ internalin-family protein on the surface of P. gingivalis interacting with the candidal Als3 hyphal protein.
Internalins belong to a multigene family characterized by variable numbers of leucine-rich repeats (LRRs). The internalins InlA and InlB in Listeria are major virulence factors and mediate attachment and bacterial uptake by nonprofessional phagocytic epithelial cells (66). InlJ comprises a distinct class of internalins, and the LRR consensus sequence contains 21 leucine residues, compared to the standard 22 residues found in other internalins. In addition, a hydrophobic residue in one of the LRRs is replaced by a cysteine in InlJ, and InlJ thus possesses a total of 14 cysteine residues (67). In Listeria, ЉporЉ (genes corresponding to secretion systems), "rp" (genes corresponding to ribosomal proteins), and "mur" (corresponding to peptidoglycan biosynthesis). The genes indicated are those upregulated in P. gingivalis with InlJ in the context of a community shared with C. albicans. Edges between nodes represent protein interactions between nodes. The greater the number of edges, the larger the evidence base for identification of a functional link. These edges are drawn from curated databases (light blue), from experimental data (purple), and from predicted interactions (green, red, and blue). Other edges are also drawn and are from text mining, coexpression, and protein homology data. (B) Differential expression of T9SS genes in the P. gingivalis 33277 (WT) strain relative to the ΔinlJ mutant in response to C. albicans expressed as Log2 fold change. See Materials and Methods for statistical thresholds. (68,69), and oral infection with an inlJ mutant results in reduced Listeria levels in the intestine, mesenteric lymph nodes, liver, and spleen (70). In P. gingivalis, the InlJ homologue is upregulated following contact with gingival epithelial cells (71), and an inlJ mutant is deficient in homotypic biofilm formation by P. gingivalis (41). In the current study, we found that InlJ was required for maximal attachment of P. gingivalis to the hyphae of C. albicans. In addition, soluble recombinant InlJ protein was able to compete with whole P. gingivalis cells for attachment to Candida. These results establish a new role for InlJ of P. gingivalis as an adhesin mediating attachment to hyphae of C. albicans. In addition, an interdomain binding function extends the repertoire of internalin-family activities.

InlJ is a sortase-LPXTG anchored adhesin upregulated during infection in vivo (68). Listerial InlJ can bind to MUC2 (the major component of intestinal mucus) and to a variety of human cells in vitro
In order to identify the C. albicans receptor for P. gingivalis, we first tested the involvement of the Als3 hyphal protein. C. albicans defective for Als3 showed a significantly reduced ability to bind P. gingivalis. The ability of recombinant InlJ and of P. gingivalis expressing InlJ to bind to S. cerevisiae strains expressing Als3 corroborated the role of Als3 in mediating attachment through interactions with InlJ. Als3 is able to bind extracellular matrix (ECM) proteins and epithelial and endothelial cells (44,(72)(73)(74), induce endocytosis through adherence to E-or N-cadherins (75), and mediate trafficking to the brain (76). Als3 also mediates attachment to S. gordonii through binding to the SspA/B streptococcal surface proteins, and this interaction stimulates the development of a mixed bacterium-fungus community with a potentially increased risk for candidiasis (43,77). Interestingly, P. gingivalis also bound to the SspA/B proteins of S. gordonii (78)(79)(80), raising the possibility that P. gingivalis and C. albicans could compete for binding to a streptococcal substratum.
To further delineate the role of the P. gingivalis InlJ in the interaction with C. albicans, a global transcriptional approach was undertaken. Here, 381 P. gingivalis genes were shown to be InlJ regulated, among which 153 (79 with a Log2 fold change Ͼ1) were unique to coincubation with C. albicans. While this level of community-dependent regulation is similar to data reported for S. gordonii in association with C. albicans (81), the differentially expressed genes were functionally different, indicating organismspecific responses of bacteria to Candida. Overall, GO analysis indicated that InlJ may play a role in increased growth and cellular division during coculture. Indeed, these analyses showed that peptidoglycan biosynthesis potential was increased during coculture, a characteristic demonstrated within complex polymicrobial biofilms from periodontitis patients (82). Alternatively, the close association between the organisms mediated by InlJ may facilitate generation of an anaerobic environment by C. albicans which enhances the growth of P. gingivalis, as has been demonstrated with Candida and Bacteroides species (83). Perhaps the most notable cluster of coassociated subnetworks of genes that were uniquely upregulated in coculture were those from the T9SS, which is widely distributed in the Fibrobacteres-Chlorobi-Bacteroidetes superphylum and secretes cargo proteins that are often cell associated and possess a conserved C-terminal domain (48). Many of the substrate proteins are considered major virulence factors in P. gingivalis, including the gingipains and other proteases; peptidylarginine deiminase (PAD), which catalyzes the conversion of peptidylarginine to peptidyl citrulline; and InlJ itself. Upregulation of T9SS components in P. gingivalis-C. albicans communities is thus consistent with elevated community pathogenicity, or nososymbiocity (11). Similarly, communities of P. gingivalis with the accessory pathogen S. gordonii, which are synergistically pathogenic (13), show an increase in expression of genes encoding T9SS components (84). Periodontal diseases are polymicrobial infections, and it is the heterotypic community that is considered the fundamental unit of pathogenicity (61). As an inhabitant of these complex multispecies biofilms, P. gingivalis, which is a keystone pathogen, may thus have evolved mechanisms to sense the community environment and respond through upregulation of the secretion system which can modulate virulence potential.
Complementation of strain ⌬inlJ. For complementation of the ⌬inlJ mutant, the DNA sequence containing the promoter and the coding region of inlJ was amplified from P. gingivalis 33277 chromosomal DNA using primers F1 (AATAGGATCCGTCCCGACTTTCCGATATATAAG) (containing a BamHI restriction site) and R2 (AATAGTCGACTTACGGCATCGCGGTTTTG) (containing a SalI restriction site). The shuttle vector pT-COW plasmid was digested with the appropriate restriction enzymes to allow cloning of the amplified PCR product into the tetC region. The resulting plasmid, pT-COW:inlJ, was transformed into E. coli TOP10 and selected with ampicillin. Purified pT-COW:inlJ was introduced into the ΔinlJ strain by conjugation as described previously (85). The presence of the pT-COW:inlJ plasmid and of the ermF gene on the chromosome of the transconjugants was confirmed by PCR and sequencing. The resulting strain was designated cΔinlJ. As determined by quantitative reverse transcription-PCR (qRT-PCR), the expression levels of the inlJ gene were similar (P Ͼ 0.05) in strains cΔinlJ and 33277.
Expression of recombinant InlJ protein (rInlJ). InlJ protein was expressed as a His-tagged fusion protein using an Expressway cell-free E. coli expression system (Invitrogen, Carlsbad, CA). Briefly, the entire coding region of inlJ (PGN_1611) was amplified from a P. gingivalis 33277 genomic template using primers F1 (ATGAAAAGAAAACCGCTATTCTCAG) and R1 (TTACGGCATCGCGGTTTTGATCG), cloned into pEXP5-NT/TOPO, and transformed into E. coli TOP10 cells. Following confirmation by sequencing, soluble His-tagged protein was obtained using MagneHis particles (GE Healthcare, Pittsburgh, PA). The purity of the resulting protein was verified by SDS-PAGE electrophoresis.
C. albicans-P. gingivalis interactions in the planktonic phase. Binding interactions between C. albicans and P. gingivalis in suspension were measured essentially as described previously (43). C. albicans cells were grown for 16 h in YPD medium, harvested by centrifugation (4,000 ϫ g, 10 min), washed twice with YPT medium, and suspended at an optical density at 600 nm (OD 600 ) of 1.0 (~1 ϫ 10 7 cells/ml). Aliquots (0.2 ml) of cell suspension were then incubated in YPT-Glu (1.8 ml) at 37°C for 2 h with shaking to induce formation of hyphae. P. gingivalis was cultured for 16 h in TSBHM medium, harvested by centrifugation (4,000 ϫ g, 10 min), washed with phosphate-buffered saline (PBS), suspended in 1.5 mM fluorescein isothiocyanate (FITC) solution, and incubated at 20°C for 30 min. After two washes with PBS to remove excess FITC, cells were suspended in YPT-Glu supplemented with hemin (5 g/ml) and menadione (1 g/ml) (YPT-GluHM) at an OD 600 of 0.5. FITC-labeled bacteria were then added to the yeast cell suspension and incubated for 1 h at 37°C with shaking. Samples (50 l) of the suspension were applied to microscope slides and visualized by light and fluorescence microscopy. Images were analyzed using Zeiss Zen imaging software. Attachment of P. gingivalis to C. albicans was categorized as representing either "binding" (extensive attachment of bacteria to hyphae with bacterial clumping and bacterial cells aligned along hyphae in distinct patches) or "no binding" (sparse or no interactions between bacteria and hyphae) (see Fig. S3 in the supplemental material). The numbers of hyphae within these categories were expressed as the percentages of the total number of hyphae counted from 4 independent experiments. One hundred hyphal cells were counted for each assay. For inhibition assays, C. albicans was incubated with rInlJ or control protein at 37°C for 2 h, prior to addition of P. gingivalis. Dual-species biofilm formation. Biofilm formation by C. albicans and P. gingivalis was assayed as described previously (43). Sterile glass coverslips were incubated with filter-sterilized 10% saliva for 16 h at room temperature and washed twice with PBS. C. albicans cells were grown for 16 h in YPD medium, harvested by centrifugation (4,000 ϫ g, 10 min), washed twice with YPT medium, and suspended to an OD 600 of 1.0. Cells (1 ϫ 10 6 ) were added to wells of 12-well plates containing saliva-coated coverslips and YPT-Glu medium (0.9 ml) and were incubated at 37°C for 2 h with gentle shaking to induce formation of hyphae. The YPT-Glu medium was replaced with YPT-GluHM, and FITC-labeled P. gingivalis cells (5 ϫ 10 6 ) were added. Dual-species cultures were incubated for a further 1 h at 37°C with gentle shaking (50 rpm). Unbound bacteria in suspension were removed, and 1 ml of YPT medium containing hexidium iodide (Sigma-Aldrich, St. Louis, MO) was added for 5 min to fluorescently stain C. albicans. Coverslips were washed twice with PBS, mounted with Prolong Gold (Invitrogen), and imaged with a Leica SP8 confocal microscope. Images were analyzed using Volocity 6.3 software (PerkinElmer, Waltham, MA). For inhibition assays, C. albicans was incubated with rInlJ or control protein at 37°C for 2 h, prior to addition of P. gingivalis.
Whole-cell enzyme-linked immunosorbent assay (ELISA). S. cerevisiae cells were grown for 16 h in CSM-Glu medium, harvested by centrifugation (4,000 ϫ g, 10 min), washed twice in PBS, and suspended at an OD 600 of 1.0. Microtiter plates were coated with 100 l of cell suspension at room temperature for 1 h. The coated plate was washed twice with 0.1% Tween-PBS followed by blocking performed for 1 h with 100 l of 10% skim milk-PBS and was further washed as described above. S. cerevisiae cells were then reacted for 1 h with either rInlJ (5 g) or P. gingivalis cells (1 ϫ 10 7 ). After a washing step, bound rInlJ protein was detected using a 1:2,000 dilution of His-tagged monoclonal antibody (MAb) (Cell Signaling, Inc., Danvers, MA) and P. gingivalis was detected with P. gingivalis whole-cell antibodies (1:10,000). After 1 h of incubation, reactions were developed with goat anti-rabbit horseradish peroxidase (HRP)-conjugated IgG (Cell Signaling) and TMB substrate (Invitrogen). The reaction was stopped with 100 l of 1 N HCl, and the OD value was determined at 450 nm.
RNA sequencing (RNA-Seq). C. albicans was induced to form hyphae in YPT-GluHM for 2 h, as described above. Equal numbers of P. gingivalis and C. albicans cells (5 ϫ 10 8 ) were then incubated together in the planktonic phase in YPT-GluHM for 1 h. The two species were treated identically in monocultures in separate experiments. Cells were harvested by centrifugation and suspended in ice-cold RLT buffer (Qiagen, Manchester, United Kingdom) containing 2-mercaptoethanol. Acid-washed Biospec glass beads (0.6 ml) were added, and cells were disrupted by alternating shaking (30 s) using a FastPrep-25 bead beater (MP Biomedicals, Santa Ana, CA) and incubating for 1 min on ice (repeated 3 times). RNA was extracted and purified using an RNeasy minikit (Qiagen) with on-column DNase digestion (Qiagen). rRNA was depleted with a RiboZero Magnetic Gold kit (Epicentre, Illumina Inc., Madison, WI), and lllumina sequencing libraries were prepared using ScriptSeq v2 (Epicentre) with 10 cycles of PCR amplification. Paired-end sequencing of 100 bp was undertaken using a HiSeq 2500 system (Illumina) in high-output mode with Truseq v3 reagents.
FASTQ data were filtered using the fastq-mcf command from the EA-UTILS suite to remove adapter sequences and low-quality bases (86). Filtered data were aligned against the reference using Bowtie v.2.2.6. The resulting aligned reads were processed with SAMtools (v0. 1.19), and gene features were counted using SAM files and the function htseq-count from Python package HTSeq v0.9.1 (https://pypi .python.org/pypi/HTSeq). The DESeq2 package was then used to apply a negative binomial model to account for dispersion between samples, before assessing differential expression between variables. P values were calculated using DESeq2, and Benjamini-Hochberg adjusted P values of Ͻ0.05 were considered significant. Following identification of differentially expressed P. gingivalis genes, unique genes (with an adjusted P value of Ͻ0.05) that were differentially expressed only during the coincubation with C. albicans were discerned. P. gingivalis genes were annotated with their associated Gene Ontology identifiers from UniProt (http://www.uniprot.org/). GO enrichment/overrepresentation analysis was performed by the use of the R GOSeq package, which implements a Wallenius hypergeometric distribution to account for bias based on gene length. Protein interaction networks were drawn from significantly differentially expressed genes using STRING (https://string-db.org/cgi/input.pl).