Porphyromonas gingivalis HmuY and Streptococcus gordonii GAPDH—Novel Heme Acquisition Strategy in the Oral Microbiome

The oral cavity of healthy individuals is inhabited by commensals, with species of Streptococcus being the most abundant and prevalent in sites not affected by periodontal diseases. The development of chronic periodontitis is linked with the environmental shift in the oral microbiome, leading to the domination of periodontopathogens. Structure-function studies showed that Streptococcus gordonii employs a “moonlighting” protein glyceraldehyde-3-phosphate dehydrogenase (SgGAPDH) to bind heme, thus forming a heme reservoir for exchange with other proteins. Secreted or surface-associated SgGAPDH coordinates Fe(III)heme using His43. Hemophore-like heme-binding proteins of Porphyromonas gingivalis (HmuY), Prevotella intermedia (PinO) and Tannerella forsythia (Tfo) sequester heme complexed to SgGAPDH. Co-culturing of P. gingivalis with S. gordonii results in increased hmuY gene expression, indicating that HmuY might be required for efficient inter-bacterial interactions. In contrast to the ΔhmuY mutant strain, the wild type strain acquires heme and forms deeper biofilm structures on blood agar plates pre-grown with S. gordonii. Therefore, our novel paradigm of heme acquisition used by P. gingivalis appears to extend to co-infections with other oral bacteria and offers a mechanism for the ability of periodontopathogens to obtain sufficient heme in the host environment. Importantly, P. gingivalis is advantaged in terms of acquiring heme, which is vital for its growth survival and virulence.


Introduction
The oral cavity of healthy individuals is inhabited by commensal microbiota, with species of Streptococcus being the most abundant and prevalent in sites not affected by periodontal diseases [1,2]. Streptococcus gordonii, a member of the viridans group of streptococci, is one of the initial Gram-positive colonizers of oral mucous surfaces [3]. Despite being a part of the normal microflora, the viridans streptococci (e.g., S. gordonii FSS2) are primary etiologic agents of infective endocarditis [4,5]. The development of periodontal diseases is linked with an environmental shift in the oral microbiome, leading to the domination of Gram-negative pathogenic bacteria over early Gram-positive commensal colonizers [6]. In the initial stages of periodontitis, the early Gram-negative anaerobic colonizers (e.g., Prevotella intermedia) are responsible for the initiation and progression of periodontitis. Advanced

S. gordonii SgGAPDH Binds Heme
A comparison of amino acid sequences of known GAPDH proteins has shown a wide range of identity (46-92%), but crystallographic analysis has revealed almost identical three-dimensional protein structures. Our theoretical analysis of SgGAPDH corroborated those findings (Figures 1 and 2 and Figure S1). It has been also demonstrated that eukaryotic GAPDH proteins bind heme in a non-canonical manner, typical for heme-binding proteins involved in heme sensing and transport [29,35,36]. Figure 1. A comparison of amino acid sequences of selected GAPDH proteins. Amino acid residues involved in heme binding in several GAPDH proteins are marked in red. Conserved heme regulatory motif is marked in blue. His53/51 -conserved histidine residue engaged in heme coordination in human and rabbit GAPDH, respectively. His43, His109, His111, His137, His179, and His192 histidine residues replaced singly by an alanine residue in this study in SgGAPDH using site-directed mutagenesis are marked in red. Amino acid sequences were identified using BLAST search and compared using the Clustal Omega and the BoxShade servers.

Figure 2.
Localization of conserved amino acid residues in the theoretical three-dimensional model of the SgGAPDH protein structure. Potential heme-coordinating histidine residues are shown in black and indicated with arrows, the potential heme-binding regulatory motif is shown in pink, and the NAD molecule is shown in blue. Two orientations of the modeled protein structure are shown.
We showed that the Soret maximum determined for SgGAPDH-Fe(III)heme complex was present at 412 nm, and Q band maxima at 536 and 557 nm (Figure 3a). In addition, a charge transfer band at 622 nm and a smaller peak in the Soret region at 351 nm were also visible. The difference spectrum analysis demonstrated four well-resolved peaks at 416, 536, 559, and 653 nm (Figure 3b). After reduction with sodium dithionite, the Soret peak maximum of SgGAPDH red-shifted to 422 nm (Figure 3a). In addition, weak and less resolved Q bands and an additional band at 382 nm could be observed. The difference spectrum analysis showed clear maxima at 425, 529, and 559 nm (Figure 3b). These results support the ability of the SgGAPDH to bind heme, preferentially under oxidizing conditions. The binding stoichiometry of heme to SgGAPDH was 1:1 ( Figure S2), which is similar to that for S. pneumoniae GAPDH [30]. Also, the heme dissociation constant value for SgGAPDH (K d = 1.7 × 10 −6 M ± 2.8 × 10 −7 M) was similar to those reported for well-characterized eukaryotic GAPDH proteins (i.e., from 10 −6 M to 10 −8 M) [19,29].
Amino acid comparison of selected GAPDH proteins showed that SgGAPDH possesses conserved histidine residues, except that corresponding to conserved His53 or His51 in human or rabbit GAPDH, respectively ( Figure 1, Figure S3). In contrast to eukaryotic GAPDH, SgGAPDH contains the region (KTVVFNTNH) (Figure 1), which is similar to heme-binding motifs exposed on the surface of several non-canonical heme-binding regulatory proteins [30,37]. To study the effects of specific amino acids on heme iron coordination in SgGAPDH, amino acid residues, which could bind heme, were first theoretically identified. Such an approach was based on amino acid sequence comparisons (Figure 1), available crystallographic data of GAPDH proteins, and theoretical modeling of SgGAPDH (Figure 2, Figures S1 and S3). Selected conserved His residues were then systematically replaced by an Ala residue, and the ability of the purified protein variants to bind heme was analyzed. The stability and integrity of the purified wild type SgGAPDH and site-directed mutagenesis protein variants were confirmed by SDS-PAGE ( Figure 4) and CD spectroscopy ( Figure S8). His179 and His192 are located in the NAD-binding pocket (Figure 2), suggesting that they might not be engaged in heme iron coordination. Indeed, as shown in Figure 4, the replacement of these amino acids by Ala did not influence the heme-binding ability. A similar effect was observed in the case of His137 ( Figure 4). It is worth noting that the protein variant His109Ala was highly unstable during overexpression and purification, suggesting significant changes in the tertiary protein structure. However, the heme-binding ability was still observed by visual inspection of red-colored protein samples. In addition, His43 and His111 are predicted to be located on the surface of the theoretical SgGAPDH protein structure and may participate in heme binding ( Figure 2). As shown in Figure 4, only the SgGAPDH His43Ala protein variant was significantly affected in heme-binding ability, suggesting engagement of this amino acid in Fe(III)heme coordination. In eukaryotic GAPDH proteins (human and rabbit GAPDH), one heme molecule is bound to the GAPDH tetramer, whereas in prokaryotic GAPDH proteins (GAPDH from S. pneumoniae) heme is usually bound to the monomer [29,30,36]. The three-dimensional protein structure solved by crystallographic analysis also suggests tetramer formation in the case of Streptococcus agalactiae GAPDH [25,28]. Therefore, we theoretically modeled the SgGAPDH protein structure to examine the possibility of heme coordination through two His residues located in separated monomers. As shown in Figure 3, only His43 might be able to form such binding because His111 is exposed in the SgGAPDH regions that are not involved in potential tetramer formation. However, our experimental analysis demonstrated that SgGAPDH binds one heme molecule to a monomer rather than to a tetramer ( Figure S2). Based on our data, one may conclude that monomeric SgGAPDH coordinates heme iron using His43 only.

S. gordonii SgGAPDH Could be a Player in Heme Acquisition Strategy Used by Periodontopathogens
To obtain a desired breakthrough in the fields of understanding and treatment of chronic periodontitis, as well as the roles played by periodontopathogens in systemic diseases, it is necessary to characterize their proteins that are crucial to growth survival and virulence. We have shown that P. gingivalis displays a novel paradigm in its heme acquisition from hemoglobin [34], whereby oxyhemoglobin must be firstly oxidized to methemoglobin. In the case of P. gingivalis, this process employs the arginine-specific gingipain protease A (HRgpA) [38], and in the case of P. intermedia the protease interpain A (InpA) [39]. The bacteria are then able to fully proteolyze the more susceptible methemoglobin substrate to release free heme, mostly through P. gingivalis lysine-specific gingipain K (Kgp) activity and via heme sequestration mediated by the HmuY protein [40][41][42]. The latter protein is the first representative of a novel family of hemophore-like proteins which have been recently characterized and which include Tfo produced by T. forsythia, as well as PinO and PinA produced by P. intermedia [34,40,42,43]. These proteins are similar to classical hemophores used by Serratia marcescens [44] or Yersinia pestis [45], which bind heme with high affinity and deliver it to the outer-membrane receptors for subsequent transport into the cell. However, in contrast to secreted hemophores, HmuY and its homologs are associated with the outer membrane and can be shed from the bacterial cell surface by limited proteolytic processing [40,42,43].
Our novel paradigm of heme acquisition, which is displayed by the black-pigmented anaerobes, appears to extend also to co-infections with other bacteria and offers a synergistic mechanism for the ability of P. gingivalis to obtain sufficient heme in the host environment. We showed that the presence of Pseudomonas aeruginosa pyocyanin facilitates the extraction of heme from hemoglobin by the P. gingivalis HmuY by oxidizing oxyhemoglobin to methemoglobin [46]. Moreover, we demonstrated that HmuY homologs produced by T. forsythia (Tfo) and P. intermedia (PinO and PinA) also bind heme, but importantly the heme-bound proteins may provide the P. gingivalis HmuY with heme, resulting in the increased virulence of the latter species [42,43].
It has been demonstrated that S. gordonii produces hydrogen peroxide, which converts oxyhemoglobin into methemoglobin, both under microaerobic and anaerobic conditions [32,47]. Therefore, the ability of S. gordonii to mediate hemoglobin oxidation may support heme acquisition during co-aggregation with P. gingivalis [32]. S. gordonii may increase heme availability by promoting hemolysis and formation of methemoglobin, which can be utilized as a heme source via proteolytic degradation by the action of P. gingivalis Kgp or via extraction by the HmuY [34]. Indeed, we have previously shown that HmuY is able to sequester heme bound to methemoglobin, the latter formed by hydrogen peroxide produced by S. gordonii [32]. Importantly, the HmuY-heme complex formed was stable at the highest concentrations of hydrogen peroxide produced by this bacterium [32]. In this study, we showed that SgGAPDH is not able to sequester heme bound to methemoglobin ( Figure S4) and may, at least in part, exchange the heme bound to serum albumin ( Figure S5). It is possible that in the early phase of disease in the presence of abundant gingival crevicular fluid which contains albumin as its major serum protein [40], any free heme may remain bound to albumin, and any heme pre-bound to the SgGAPDH could be transferred to albumin, thus forming a dynamic heme reservoir, from which the heme might be easily accessed by periodontopathogens as a result of bacterial protease activity or hemophore-like activity of HmuY and its homologs [38,39,42,43].
To assess any possible syntrophy between P. gingivalis HmuY or its homologs and S. gordonii SgGAPDH, we first examined the interactions between HmuY, PinO, PinA, Tfo apo-proteins, and SgGAPDH-heme complex. As shown in Figure 5, HmuY, PinO and Tfo sequestered Fe(III)heme which had been complexed to SgGAPDH. In contrast, P. intermedia PinA was not able to sequester heme bound to SgGAPDH. In the case of Fe(II)heme, all the hemophore-like proteins examined were able to sequester heme bound to SgGAPDH ( Figure 6). Under air (oxidizing) conditions apo-SgGAPDH was able to capture heme bound only to PinA ( Figure S6). It is worth noting, however, that if P. intermedia grows more abundantly in the early phase of disease it could degrade both the albumin-heme complex and the SgGAPDH-heme complex ( Figure 6). Under reducing conditions, apo-SgGAPDH was not able to sequester heme bound to any of the hemophore-like proteins examined ( Figure S7). Therefore, we hypothesize that heme bound to secreted or surface-associated SgGAPDH might represent a heme reservoir for P. gingivalis or other periodontopathogens.  Our results confirmed that SgGAPDH is exposed on the cell surface of S. gordonii as well as is being secreted into the culture medium ( Figure 7). Therefore, the heme bound to SgGAPDH might be accessed by P. intermedia in the early phases of colonization when S. gordonii and P. intermedia dominate over late colonizers, but that later, as more anaerobic and reducing conditions develop in the periodontal pocket, it could be used by HmuY and Tfo, aiding colonization of both P. gingivalis and T. forsythia. It is worth noting that viridans streptococci secrete higher levels of GAPDH when the pH increases [48], which could be advantageous for periodontopathogens, especially for the late colonizers residing in deep periodontal pockets where the environmental pH is increased above neutrality [49]. Experiments demonstrating the growth of selected bacteria in the presence of the purified proteins showed that although SgGAPDH was not proteolytically digested during growth in S. gordonii and T. forsythia cultures, proteases produced by P. gingivalis and P. intermedia caused its degradation ( Figure S8), additionally aiding heme release. Our data suggest that periodontitis-associated bacteria, especially the better adjusted to both anaerobic and aerobic conditions P. gingivalis, could also benefit from commensal members of the oral microbiome. This hypothesis is supported, at least in part, by the observation of co-localized P. gingivalis and S. gordonii in various layers of supragingival plaque [50]. Identification of SgGAPDH forms produced by S. gordonii. SgGAPDH was visualized using (a) ELISA or (b,c) Western blotting. The protein was detected on the S. gordonii cell surface (a,b) or in the culture medium (a,c) using rabbit polyclonal anti-SgGAPDH antibodies and HRP-conjugated anti-rabbit IgG antibodies. The protein associated with the bacterial cell surface was examined in the native form or after treatment with proteinase K and/or SDS. Soluble secreted protein was detected in~20× concentrated culture medium. (b,c) SgGAPDH was examined in two independent samples (#1 and #2). (c) To examine putative cross-reactivity of anti-SgGAPDH antibodies, 20× concentrated culture medium alone was also examined (sample #3). SgGAPDH protein is indicated by asterisks ( * ).

P. gingivalis HmuY May Participate in Co-Aggregation and Biofilm Formation with S. gordonii
Streptococci occupy a broad range of oral habitats where they co-localize and co-aggregate with other bacteria [50][51][52]. There is growing evidence demonstrating that interaction between P. gingivalis and oral streptococci is crucial for biofilm formation and colonization of P. gingivalis [50,[53][54][55][56][57][58]. Interaction between S. gordonii and P. gingivalis employs the major (FimA) and minor fimbrial (Mfa1) proteins of P. gingivalis and outer-membrane adhesins (SspA and SspB) of S. gordonii [58][59][60][61][62][63][64][65][66]. In addition, in vitro studies showed that S. gordonii might promote the penetration and invasion of dentinal tubules by P. gingivalis by providing an adhesive platform on intra-tubular dentine collagen [67]. It has been suggested, however, that those interactions may not be sufficient for the formation of biofilms composed of S. gordonii and P. gingivalis. In this study, we corroborated the co-aggregation ability occurring between S. gordonii and selected periodontopathogens (Figure 8). Interestingly, we showed that such inter-bacterial interaction might employ the HmuY protein. As shown in Figure 9, the ∆hmuY mutant TO4 strain exhibited a lower co-aggregation tendency as compared with the wild type A7436 strain. Recently, we demonstrated that HmuY homologs produced by P. intermedia and T. forsythia bind heme preferentially under anaerobic conditions, suggesting their employment in heme acquisition in deep periodontal pockets [42,43]. In contrast to Tfo, PinO, and PinA, HmuY is very efficient in heme binding under both reducing and air (oxidizing) conditions [41,42]. Although P. gingivalis is an anaerobe, the bacterium can also grow in microbiome structures exposed to aerobic conditions [50]. Therefore, co-localization of P. gingivalis and S. gordonii could be beneficial for P. gingivalis growing in the oral microbiome under a range of higher oxygen levels in which it may not otherwise survive alone.  . Interaction of S. gordonii with P. gingivalis strains. The tendency of S. gordonii to co-aggregate with P. gingivalis wild type (A7436) and mutant strains lacking hmuY (TO4), hmuR (WS1) and both hmuY and hmuR genes was monitored by measuring optical density at 600 nm at the indicated time points (left panel) or after 6 h (right panel). A7436, wild type P. gingivalis strain; TO4, ∆hmuY mutant strain; TO2, ∆hmuY∆hmuR mutant strain; WS1, ∆hmuR mutant strain. All mutant strains were constructed in the A7436 wild type strain. ** p < 0.01; *** p < 0.001.
Visual inspection demonstrated that, in contrast to the ∆hmuY mutant strain, the wild type P. gingivalis strain grew faster and formed bigger biofilm structures on blood agar plates pre-grown with S. gordonii ( Figure 10). Interestingly, we also observed a different ability to form pigment. The ∆hmuY mutant strain became black pigmented faster, whereas the wild type strain, although forming pigment, did so but with significantly lesser ability, even after prolonged growth ( Figure 10). We suspect that such a feature could result from higher expression and/or better exposure of hemagglutinin domains present in proteins, which are also known to participate in pigment formation [34]. Co-culture of P. gingivalis with S. gordonii resulted in increased hmuY gene expression (fold change of 3.2 ± 0.45 as determined by RT-qPCR), indicating that HmuY might be required for efficient inter-bacterial interactions in heme uptake by P. gingivalis. Other studies also demonstrated increased expression of the hmuY gene during synergistic interactions between bacterial members in the dental plaque of periodontitis patients [68]. In contrast, decreased expression of the hmuY gene was found by others testing P. gingivalis and S. gordonii co-cultures, but in those studies the less virulent P. gingivalis strain ATCC 33277 and different culturing conditions were used [69]. A decreased co-aggregation ability was also observed in this study in the case of ∆hmuR mutant WS1 strain and the double ∆hmuY∆hmuR mutant TO2 strains (Figure 9). Although the HmuR protein was suggested to be important for efficient community formation between P. gingivalis and Fusobacterium nucleatum [64], no significant change in the hmuR gene expression was observed in our study (fold change of −1.22 ± 0.51 as determined by RT-qPCR) during co-culture of P. gingivalis and S. gordonii. Interestingly, when the ∆hmuR (WS1) and ∆hmuY∆hmuR (TO2) mutant strains were subjected to growth on blood agar plates pre-grown with S. gordonii, slower biofilm formation but significantly faster pigmentation was observed as compared to the wild type (A7436) or ∆hmuY mutant (TO4) strain ( Figure 10). One explanation for this could be the lack of the outer-membrane HmuR transporter, which causes accumulation of the pigment on the cell surface instead of heme uptake. Our data may suggest that the hmu operon, with a leading role played by the hmuY gene, may be important for efficient host colonization by P. gingivalis.
To examine biofilm formation, S. gordonii was grown on blood agar plates for 24 h under an increased concentration of CO 2 . Then, 50 µL of P. gingivalis cultures adjusted in PBS to OD 600 = 1.0 was applied (starting from 5 × 10 7 cells/50 µL/spot and further diluted) on blood agar plates alone or on blood agar plates pre-grown with S. gordonii. Bacterial growth was continued under anaerobic conditions for 12 days. The formation of biofilm structure and pigmentation efficiency were monitored by visual inspection.

Co-Aggregation Assay
Co-aggregation assay was performed as described previously [72]. Briefly, overnight liquid cultures of P. gingivalis, P. intermedia, T. forsythia and S. gordonii were centrifuged (4000× g, 20 min, 4 • C) and washed with 20 mM sodium phosphate buffer, pH 7.4, containing 140 mM NaCl (PBS). Subsequently, optical density at 600 nm (OD 600 ) of the cultures was adjusted to 1.0 using PBS. The cultures of the respective bacteria were mixed at a 1:1 ratio and incubated anaerobically for 6 h. The co-aggregation rate was monitored by measurement of the decrease in OD 600 . The sum of the auto-aggregation of the tested mono-species cultures was used as a control.

Overexpression, Purification and Site-Directed Mutagenesis
The DNA sequence encoding SgGAPDH protein (NCBI accession no. SQF26726.1) was cloned into the NdeI and EcoRV restriction sites of pTXB1 plasmid (New England Biolabs) or into the XcmI and BamHI restriction sites of pTriEx-4 plasmid (Merck, Darmstadt, Germany) using primers listed in Table S1 and NEBuilder ® HiFi DNA Assembly (New England Biolabs). Resulting pTXB1_SgGAPDH or pTriEx_SgGAPDH expression plasmids were used to overexpress SgGAPDH protein. Recombinant protein possessing C-terminal intein with chitin-binding domain was purified using affinity chromatography with chitin resin (New England Biolabs) according to the manufacturer's instructions. SgGAPDH protein was released from the resin by on-column cleavage in the presence of 50 mM β-mercaptoethanol at 4 • C for 24 h, followed by affinity chromatography using Cibacron Blue 3G-A (Sigma). The protein was bound to the chromatographic resin in 50 mM Tris/HCl buffer, pH 7.6, containing 30 mM NaCl and eluted using 50 mM Tris/HCl buffer, pH 7.6, containing 1 M NaCl. Recombinant protein possessing N-terminal His-tag with the Factor Xa recognition site was purified using TALON resin (Clontech, Mountain View, CA, USA) according to the manufacturer's instructions.
Selected amino acids of SgGAPDH with a potential ability to coordinate heme iron were substituted by an alanine, resulting in overexpressing of single point mutation protein variants. Point mutations were introduced into expression plasmids using a QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies) and primers listed in Supplementary Table S1. The resulting plasmids were sequenced and used to transform E. coli cells.
The P. gingivalis A7436 HmuY protein (NCBI accession no. CAM 31898) lacking the signal peptide (MKKIIFSALCALPLIVSLTSC) and additional (GKKK) N-terminal amino acid residues present in the full-length protein sequence was overexpressed using E. coli ER2566 cells (New England Biolabs), and purified from a soluble fraction of the E. coli cell lysate as described previously [73].
The DNA sequence encoding T. forsythia ATCC 43,037 Tfo protein (NCBI accession no. CEH11291), lacking the predicted signal peptide (MKMRNVMTLALVALSLAFVGC) was amplified using primers listed in Table S1. Then, using NEBuilder ® HiFi DNA Assembly (New England Biolabs), the gene was cloned into the XmnI and BamHI restriction sites of pMAL-c5x plasmid, modified as reported previously [74]. Overexpression and purification of Tfo protein were performed as described for PinA and PinO proteins.
The concentration of SgGAPDH protein was determined spectrophotometrically using the empirical molar absorption coefficient (ε 280 = 23.97 mM −1 cm −1 ), calculated in this study as reported previously [40].

Susceptibility of SgGAPDH to Proteolytic Digestion
To examine the susceptibility of the SgGAPDH to proteases produced by the examined bacterial species, bacterial cells were grown under rich, high-iron/heme conditions (Hm), ensuring proper cell viability and efficient proteolytic activity [71,73], in the presence of added purified SgGAPDH at final 5 µM concentration [42,43]. All cultures (10 ml) were started at OD 600 = 0.2 (time 0), grown and collected at 24 h. The number of bacterial cells at the starting point was~2 × 10 8 per ml of the culture medium and increased during cultivation. As controls, culture media with the addition of the purified protein were analyzed. At the indicated time points, aliquots of samples were examined by SDS-PAGE and Coomassie Brilliant Blue G-250 staining [73].

Analysis of Heme Binding
Heme (hemin chloride; ICN Biomedicals) solutions were prepared as reported previously [38,42]. The formation of the heme-protein complexes was examined in PBS or 100 mM Tris-HCl buffer, pH 7.5, containing 140 mM NaCl (TBS). UV-visible spectra were recorded in the range 250-700 nm with a double beam Jasco V-650 spectrophotometer using cuvettes with 10 mm path length. Titration curves were analyzed using the equation for a one-site binding model, and dissociation constant (K d ) values determined as reported earlier [42] using OriginPro 8 software (OriginLab, Northampton, MA, USA).
To determine heme binding ability of site-directed mutagenesis variants of SgGAPDH, 5 µM protein was mixed with 4 µM Fe(III)heme, and spectra were recorded and compared with spectra obtained for the wild type protein.
To confirm the heme to protein binding ratio, an additional approach was applied. For this purpose, SgGAPDH was saturated with Fe(III)-heme and the excess heme was removed by desalting on a Zeba Spin Desalting Column 7 K MWCO (Thermo Scientific). 4 µM SgGAPDH-Fe(III)heme complex was mixed with 25 µM apo-HmuY in PBS, incubated for 30 minutes, and the UV-visible spectrum was recorded. In parallel, 25 µM apo-HmuY in the presence of 4 µM apo-SgGAPDH was titrated with Fe(III)heme. The UV-visible spectrum was examined after increasing the heme concentration in the sample by 1 µM. The absorbance at 411 nm that corresponds to the HmuY-Fe(III)heme complex Soret maximum was used to create a standard curve of the amount of heme bound to the HmuY protein.
Excess HmuY protein in the sample ensured that all heme would be sequestered from SgGAPDH. The amount of heme bound to the SgGAPDH monomer was determined using the standard curve.
To analyze the redox properties of the heme with iron in the Fe(II) state, 10 mM sodium dithionite was used as the reductant with or without mineral oil overlay [38,40].

Detection of SgGAPDH Protein
To detect SgGAPDH present on the S. gordonii cell surface or secreted into the culture medium, enzyme-linked immunosorbent assay (ELISA) was first used. Bacteria were grown in 5% TSB under an increased concentration of CO 2 for 24 h, centrifuged (4000× g, 12 min), washed with PBS, and suspended in 50 mM Tris/HCl buffer, pH 8.0, to an OD 600 = 1. Bacteria were examined in the native form or after treatment for 1 h at 37 • C with proteinase K (20 µg/mL) and/or 0.5% SDS. Subsequently, bacteria were washed 3 times with PBS and suspended in PBS to the original volume. The culture medium was filtrated through 0.22 µm filters, ultracentrifuged (100,000× g, 2 h), and concentrated~20× using 10 kDa cut-off filter units (Millipore, Billerica, MA, USA). Then, 96-well polystyrene plates (Sarstedt, Waltham, MA, USA) were coated with either 100 µL of bacterial suspension or 100 µL of concentrated culture medium. The protein was probed with rabbit polyclonal anti-SgGAPDH antibodies, raised in this study against purified SgGAPDH protein (ProteoGenix, Schiltigheim, France), and complexes formed detected using HRP-conjugated anti-rabbit IgG antibodies (Sigma) and a substrate solution as described previously [75].
To confirm the presence of SgGAPDH, Western blotting was employed. Bacterial suspension in the native form and/or treated as described above was subjected to lysozyme digestion (2 µg/mL, 10 min, 37 • C), and sonication (20 sec) in SDS-PAGE denaturation buffer. Proteins present in bacterial cell lysates (20 µL of the OD 600 = 0.02 samples) or in 40 µL of concentrated culture medium were separated by SDS-PAGE, transferred onto a nitrocellulose membrane (Millipore), probed with rabbit anti-SgGAPDH antibodies, and complexes formed detected using HRP-conjugated anti-rabbit IgG antibodies and chemiluminescence staining (Perkin Elmer, Waltham, MA, USA) [73].

Heme Sequestration Experiments
Co-incubation of HmuY, Tfo, PinO, or PinA with SgGAPDH, each protein in apo-or holo-form, was carried out under air (oxidizing) or reducing (10 mM sodium dithionite) conditions in PBS at 37 • C and monitored by UV-visible spectroscopy using each protein at 10 µM concentration [40,42,43]. Co-incubation of methemoglobin or human serum albumin with SgGAPDH was carried out using each protein at 5 µM concentration. Protein:heme complexes were formed at a 1:1.2 ratio, and the excess heme was removed by desalting with a Zeba Spin Desalting Column 7 K MWCO (Thermo Scientific).

Reverse Transcriptase-Quantitative Polymerase Chain Reaction (RT-qPCR)
Total RNA was purified from 1 ml of P. gingivalis culture or 1 ml of P. gingivalis and S. gordonii co-cultures using the Total RNA Mini Kit (A&A Biotechnology, Gdynia, Poland) and the Clean-Up RNA concentrator kit (A&A Biotechnology). Reverse transcription was carried out using 1 µg of RNA and SensiFAST cDNA synthesis kit (Bioline, London, UK). qPCR was performed using SensiFAST SYBR No-ROX kit (Bioline) and the LightCycler 96 system (Roche, Basel, Switzerland). The amplification reaction was carried out as follows: an initial denaturation at 95 • C for 2 min, 40 cycles of denaturation at 95 • C for 5 s, primer annealing at 60 • C for 10 s, and extension at 72 • C for 20 s. The melting curves were analyzed to monitor the quality of PCR products. Relative quantification of the hmuY or hmuR gene was determined in comparison to the 16SrRNA gene of P. gingivalis A7436 (PGA7_00000960) as a reference, using the ∆∆C t method and LightCycler 96 software (Roche). All samples and controls were run in triplicate in three independent experiments for the target and reference genes. All primers are listed in Table S1.

Circular Dichroism (CD) Analysis
Far-UV CD spectroscopy was carried out using a Jasco J-1500 spectropolarimeter (Jasco, Tokyo, Japan). Protein samples at 2.5 µM concentration in 10 mM sodium phosphate buffer, pH 7.5, containing 20 mM NaCl, were examined in a 2-mm path-length cuvette. The spectra were recorded over a wavelength range of 200-260 nm by signal averaging of three spectra. Scanning speed of 100 nm/min, bandwidth 2 nm, and response time 2 s were used. All spectra were baseline corrected for the buffer.

Bioinformatics and Statistical Analyses
The sequences were aligned by the Clustal Omega tool from the EMBL-EBI server (the European Bioinformatics Institute). The statistical analysis was performed using Student's t or one-way ANOVA test. Data were expressed as mean ± standard deviation (mean ± SD). For statistical analysis, the GraphPad Prism 8.0 software (GraphPad Sooftware Inc., San Diego, CA, USA) was used. The theoretical three-dimensional model of SgGAPDH protein structure was constructed using Phyre2 software [76] based on GAPDH protein structure from Streptococcus agalactiae (PDB: 4QX6). Subsequently, the SgGAPDH model was submitted to ModRefiner server [77]. Final protein structure of SgGAPDH protein was visualized by Swiss-PdbViewer [78].

Conclusions
We are aware that bacteria grown in liquid cultures or solid media under laboratory conditions may display different properties compared with those grown in polymicrobial consortia within dental plaque biofilms. However, the model presented here has allowed for important observations that have demonstrated interactions between P. gingivalis and S. gordonii. Our observations clearly showed that P. gingivalis is advantaged during biofilm co-culture in terms of acquiring heme, which could be vital for its growth survival and virulence.