Comparative analysis of Porphyromonas gingivalis A7436 and ATCC 33277 strains reveals differences in the expression of heme acquisition systems

ABSTRACT Porphyromonas gingivalis strains exhibit different phenotypes in vitro, different virulence potential in animal models, and different associations with human diseases, with strains classified as virulent/more virulent (e.g., A7436 and W83) or as less virulent/avirulent (e.g., ATCC 33277). In this study, we comparatively analyzed the A7436 and ATCC 33277 strains to better understand their variability. Global gene expression analysis in response to heme and iron limitation revealed more pronounced differences in the A7436 than in the ATCC 33277 strain; however, in both strains, the largest changes were observed in genes encoding hypothetical proteins, genes whose products participate in energy metabolism, and in genes encoding proteins engaged in transport and binding proteins. Our results confirmed that variability between P. gingivalis strains is due to differences in the arrangement of their genomes. Analysis of gene expression of heme acquisition systems demonstrated that not only the availability of iron and heme in the external environment but also the ability to store iron intracellularly can influence the P. gingivalis phenotype. Therefore, we assume that differences in virulence potential may also be due to differences in the production of systems involved in iron and heme acquisition, mainly the Hmu system. In addition, our study showed that hemoglobin, in a concentration-dependent manner, differentially influences the virulence potential of P. gingivalis strains. We conclude that iron and heme homeostasis may add to the variability observed between P. gingivalis strains. IMPORTANCE Periodontitis belongs to a group of multifactorial diseases, characterized by inflammation and destruction of tooth-supporting tissues. P. gingivalis is one of the most important microbial factors involved in the initiation and progression of periodontitis. To survive in the host, the bacterium must acquire heme as a source of iron and protoporphyrin IX. P. gingivalis strains respond differently to changing iron and heme concentrations, which may be due to differences in the expression of systems involved in iron and heme acquisition. The ability to accumulate iron intracellularly, being different in more and less virulent P. gingivalis strains, may influence their phenotypes, production of virulence factors (including proteins engaged in heme acquisition), and virulence potential of this bacterium.

Many studies have shown that P. gingivalis strains exhibit different phenotypes and pathogenicity, which are caused mainly by variations in capsular polysaccharides, the major and minor fimbriae, adhesion domains of lysinespecific gingipain (Kgp) and hemagglutinin A (HagA) .Although low sequence divergence within genes encoding catalytic domains of gingipains has been found (21), P. gingivalis clinical isolates exhibited different rates and distribution of Kgp and argininespecific gingipain (RgpA and RgpB) activities (23).Sequencing of several P. gingivalis genomes revealed that the bacterium can exchange chromosomal DNA by specific gene rearrangements and allelic exchange through horizontal gene transfer using natural transformation and conjugation (20,21,(44)(45)(46).Importantly, the spontaneous generation of phenotypically different sub-strains, caused in part by differences in genes encoding hemagglutinin domains of Kgp and HagA, can also occur, resulting in significant attenuation in a mouse virulence model (47)(48)(49)(50).On the other hand, many genes are highly conserved in P. gingivalis strains and located in the core genome (22), including genes encoding heme acquisition Hmu and Iht systems (22,(51)(52)(53).The best characterized, and important for P. gingivalis virulence, is a heme uptake (Hmu) system composed of HmuY (a heme-binding hemophore-like protein), HmuR (a typical TonB-dependent receptor transporting heme through the outer membrane), and four proteins with unknown functions (HmuSTUV) (52,54,55).The Hmu system works in synergy with gingipains and their production is correlated (55)(56)(57).Apart from nutrient acquisition for this asaccharolytic pathogen and for avoiding the host immune response (58), gingipains participate in heme acquisition by RgpA-dependent oxidation of oxyhemoglobin to methemoglobin, allowing direct heme sequestration by HmuY (57) or heme release by proteolytic digestion of methe moglobin, the latter activity carried out mainly by Kgp (58,59).Another P. gingivalis system that is considered to be an alternative one involved in heme and/or iron uptake is the Iht system (53); however, it is hardly characterized.The Iht system comprises the TonB-dependent outer membrane receptor (IhtA), putative heme-binding lipoprotein (IhtB), periplasmic binding protein (IhtC), permease (IhtD), and cytoplasmic ATP binding protein (IhtE).Due to IhtB homology to CbiK chelatase, it is thought that IhtB could remove iron from heme and transfer it to the IhtA protein for subsequent iron transfer to the periplasm.Iron transport to the cytoplasm could be performed by IhtC-E proteins (55,60).
P. gingivalis does not synthesize heme de novo, and consequently it produces numerous heme acquisition systems, allowing this bacterium to efficiently take up heme, as a source of both iron and protoporphyrin IX (PPIX), to survive inside the host and initiate the pathogenic process.Interestingly, it has been shown that P. gingivalis strains respond differently to changing iron and heme concentrations in the external environment (61).Therefore, we believe that differences in the virulence potential of P. gingivalis strains selected in this study may also be due to differences in the expression of systems involved in iron and heme acquisition.To analyze differences between P. gingivalis strains in their phenotypes and response to iron and heme starvation, we comparatively examined the less virulent, non-capsulated, highly fimbriated ATCC 33277 (33277) strain and more virulent, capsulated, poorly fimbriated A7436 and W83 strains.

Differences in genome organization of P. gingivalis strains
Analysis of known P. gingivalis genomes revealed significant differences among laboratory strains and clinical isolates.According to the NCBI database (www.ncbi.nlm.nih.gov; as of June 26, 2023), over 90 P. gingivalis genomes have been fully or partially sequenced, and some examples of representative genomes are listed in Table 1; Fig. 1A.To perform a detailed analysis, we chose laboratory strains used in our previous studies (A7436 and 33277), and in addition, in some experiments, a reference W83 strain was employed.First, we constructed a phylogenetic tree based on nucleotide sequences of the hmu operon of selected strains (Fig. 1A).In contrast to the high homology of the hmuY gene, other genes of the hmu operon exhibited lower homology (Fig. 1B), thus confirming the high significance of HmuY protein.Furthermore, we analyzed the location of selected genes responsible for iron and heme uptake, gene regulation, and replication.Differences in their location in respective strains (Fig. 1C) might result from chromosome inversions and insertions (Fig. 1D).Moreover, our comparative analysis of A7436, 33277, and W83 genomes showed differences in the hmu operon location in relation to other genes (Fig. 1C and D), which might result mostly from chromosome rearrangements.

Response to iron and heme limitation varies in P. gingivalis A7436 and 33277 strains
First, global gene expression analysis was carried out in P. gingivalis A7436 and 33277 strains.Relative differences in gene expression in bacteria grown in iron-and hemedepleted conditions (DIP medium) in relation to iron-and heme-replete conditions (Hm medium) were examined using microarray analysis.In the A7436 strain, the expression of 117 genes was up-regulated and the expression of 128 genes was down-regulated, whereas in the 33277 strain changes in gene expression of the lower number of genes were observed (43 and 66 genes, respectively).The most pronounced differences were found in genes belonging to the group of hypothetical proteins, in genes whose products participate in energy metabolism, and in genes encoding proteins engaged in transport and binding (Fig. 2).Proteins ascribed to those groups are listed in Tables S1

Strain
GenBank sequence ID hmu operon locus ID (hmuY-hmuV) and S2.It is also worth noting here that although the microarray analysis was designed based on the A7436 genome, which could result in no detection of changes in the expression of genes specific to the 33277 strain only, genes important for P. gingivalis virulence are located in the core genome (22).Although the expression of a different number of genes was changed in A7436 and 33277 strains, the expression of the same 40 genes was up-regulated under iron and heme starvation in both examined strains (Fig. 2).Among them are genes encoding proteins responsible for iron and heme uptake and their homeostasis (e.g., Hmu system, FeoB, Dps), oxidative stress response (e.g., AhpC, AhpF, flavodoxin), transport (e.g., TonB-dependent receptors, ABC transporters, metal transporters, TolC-like protein), regulation and signal transduction (e.g., FeoA, PorX, LuxR, SigH) (Table S1).Interestingly, changes in the expression of many genes were strain specific but less pronounced (Fig. 2; Table S1).In the A7436 strain cultured under iron and heme starvation, increased expression of genes encoding some proteases, including rgpB (FC ~2.5×),PG1788 (FC ~2.2×),PG1855 (FC ~2.7×), and PG2029 (FC ~2.1×) genes, as well as genes encoding proteins involved in transport, including membraneassociated heme-binding protein HBP35 (FC ~2.5×), type IX secretion system (T9SS) protein PorS (FC ~3.3×), putative sulfur exporter (PG2004; FC ~2.6×), and CorA family protein (FC ~3.2×), was detected.Surprisingly, only in the A7436 strain was an expression of the gene encoding HmuT protein changed (FC ~57.1×), while the expression of the rest of the genes of the hmu operon increased in both strains.Genes whose expression increased only in the 33277 strain include mostly ones encoding hypothetical proteins and ribosomal proteins (e.g., rsmH, rprR, rpmI genes).
Greater discrepancies between A7436 and 33277 strains in gene expression were detected in genes whose expression decreased when bacteria were grown under iron and heme starvation (Table S2).Only 26 genes are common to tested strains whose expression decreased, being ~20% and ~60% of total genes whose expression decreased in A7436 and 33277 strains, respectively.They include mainly genes encoding proteins involved in energy metabolism (e.g., oxidoreductases, ferredoxin, electron transport complex proteins).The greatest decrease in gene expression, especially in the 33277 strain (FC ~−13.3× and ~ −61.2× in A7436 and 33277 strains, respectively), was observed for a gene encoding a protein from the rubrerythrin family (PG0195), which is employed in the response to oxidative stress (62,63).Decreased gene expression only for the A7436 strain included genes encoding proteins involved in energy metabolism (e.g., cytochrome c biogenesis proteins, IspH, TpiA), transport (e.g., HusA, HusB, PorT), and cell envelope and virulence (e.g., WecC, PG0109, PG0411-hemagglutinin, LptC, FimA, FimB).On the other hand, the expression of only a few genes changed specifically in the 33277 strain under iron and heme starvation, including genes encoding hemolysin secretion protein D (PG0093; FC ~−2.1×), ferredoxin (FC ~−3.0×),HemG (FC ~−2.8×), and GltA (FC ~−2.9×).In addition, the expression of two genes (PG2038 and PG2040) decreased (FC for both ~−2.2×) and increased (FC for both ~2.5×) in A7436 and 33277 strains, respectively, in response to iron and heme limitation (Tables S1 and S2).The two genes seem to be co-transcribed and encode N-acetylmuramoyl-L-alanine amidase and histidinol phosphate phosphatase.
To validate the results obtained by microarray analysis, we examined the expression of selected genes using RT-qPCR in bacteria cultured for 24 hours (early stationary growth phase) under iron and heme starvation in comparison to iron-and heme-replete conditions (DIP medium vs Hm medium).Moreover, to understand in more detail the expression pattern of genes involved in iron and heme uptake, we also analyzed gene expression in bacteria cultured for 10 hours (late exponential growth phase).As shown in Table 2, the results obtained with RT-qPCR are in general consistent with the results obtained with microarray analysis.As our previous studies have shown (52,64,65), expression of the hmuY gene in the A7436 strain increased under iron and heme starvation.Furthermore, our analysis showed that the expression of hmuY and hmuR genes increased during prolonged growth in both A7436 and 33277 strains.A similar effect was observed for the feoB genes (PG1043 and PG1294), whose products are engaged in ferrous iron/manganese transport (66)(67)(68)(69).In addition, the feoB gene (PG1043), encoded close to the feoA gene (PG1044), was shown to be engaged in protection from oxidative stress (67,68).Expression of ihtB, rgpA, rgpB, and kgp genes changed throughout the culture time, with a decrease after 10 hours and an increase after 24 hours in the DIP medium in comparison to the Hm medium.Although the changes were more evident in the A7436 strain, they were not statistically significant for both examined strains.

Main heme acquisition systems are expressed differently in A7436 and 33277 strains
Genetic differences and different responses to iron and heme starvation between A7436 and 33277 strains have led us to compare the production of selected proteins involved in iron and heme uptake by P. gingivalis.First, we determined the expression of hmuY and ihtB genes in bacteria grown in different conditions (Fig. 3A).As expected, the transcript encoding the HmuY protein significantly increased during incubation in both A7436 and 33277 strains.Although expression of the ihtB gene was higher in the 33277 strain in the DIP medium and after 24 hours in the Hm medium, the observed differences were not statistically significant.Then, we compared transcript levels of hmuY and ihtB genes in the 33277 strain in comparison to the A7436 strain (Fig. 3B).We observed ~100× higher hmuY transcript production in the 33277 strain, as compared to the A7436 strain, after 10 hours in bacteria grown in iron-and heme-replete conditions (Hm medium).After 24 hours, expression of the hmuY gene did not differ significantly between the strains in bacteria cultured in both conditions (Hm and DIP media).
Next, using Western blotting with anti-HmuY and anti-IhtB antibodies, we analyzed the production of HmuY and IhtB proteins (Fig. 3C; Fig. S1).We showed that HmuY protein production was ~6× higher in the 33277 strain in comparison to the A7436 strain when culturing bacteria in iron-/heme-replete conditions, whereas the production of HmuY protein in the DIP medium was the same in both tested strains.The amount of IhtB protein was ~2× higher in the A7436 strain when bacteria were grown in the DIP medium, which is consistent with ihtB transcript production after 24 hours (Table 2).On the other hand, we did not observe differences in IhtB protein production in the 33277 strain between DIP and Hm media.However, we found ~2-6× higher IhtB protein production in the 33277 strain in comparison to the A7436 strain (Fig. 3C; Fig. S1).
Using microarray analysis, we detected an increase in rgpB expression (Table S1) in the A7436 strain when bacteria had been starved of iron and heme.Moreover, we found that expression of all genes encoding gingipains increased after 24 hours in iron-and hemedepleted conditions (Table 2).No differences in both conditions and between the examined strains were found in the expression of the hagA, the gene encoding the main hemagglutinin A (HagA), of which hemagglutinin/adhesion domains are present in RgpA and RgpB (Table 2).Therefore, we decided to analyze rgpB gene expression at the protein level.Expression of the cell-associated RgpB (70,71), gingipain produced without hemagglutinin/adhesion domains, was similar in both examined strains (Fig. 3C and D; Fig. S1 and S2).However, some differences could be observed in the case of pro-RgpB protein associated with the outer membrane (Fig. 3C, D, and 4C), although neither rgpB gene length, amino acid sequence, nor theoretical RgpB molecular weight significantly differed in the examined strains (Fig. 4A and B).
To present in more detail differences in the production of HmuY, IhtB, and RgpB proteins in A7436 and 33277 strains, we tested protein levels in bacteria cultured on blood agar (ABA) plates, containing hemoglobin as the main heme source, and serum heme-binding proteins, mainly albumin.These conditions also allow P. gingivalis to deposit excess heme on the cell surface, providing the bacterium with the ability to maintain iron and heme homeostasis, similar to that found in the natural environment in advanced stages of the disease.As shown in Fig. 3D; Fig. S2, HmuY protein production was ~3× higher in the 33277 strain in comparison to the A7436 strain, whereas produc tion of both IhtB and RgpB did not differ between examined strains.Our data demon strate that production of the main components of heme acquisition systems may vary depending on the strain, therefore revealing bases of additional differences between P. gingivalis strains.

Intracellular iron content influences P. gingivalis phenotype
Phenotypic characterization of P. gingivalis growth in liquid culture media in iron-and heme-depleted or iron-and heme-replete conditions confirmed that the A7436 strain grew faster and its growth resulted in higher biomass production, as compared to the 33277 strain (Fig. 5A; Table S3).This result is not surprising because differences in the growth rate, including W83 and A7436 strains, have also been observed by others (72), suggesting that even strains belonging to the same group may phenotypically differ.When bacteria had been starved of iron and heme, subsequent growth in the presence of heme (DIP + Hm) or hemoglobin (DIP + Hb) was more efficient in the case of the A7436 strain as compared to the 33277 strain, the latter often being unable to recover and proliferate (Fig. 5B; Table S3).These results may indicate that the 33277 strain is less resistant to iron and heme starvation due to disturbed iron uptake and/or homeosta sis.To explain this phenomenon, we determined intracellular iron content.The 33277 strain contained ~25% less intracellular iron than the A7436 strain (Fig. 5C).When P. gingivalis was grown on ABA plates, no significant differences in the formed pigment were observed (data not shown).The only difference was ~50% higher hemolysis caused by the 33277 strain in comparison to the A7436 strain (Fig. 5D).To find the bases for this phenotypic difference, we analyzed gingipain activity.Neither Rgp nor Kgp activity differed significantly when bacteria were grown either in iron-and heme-replete or in iron-and heme-depleted conditions (Fig. 5E), confirming that gingipain activity does not depend on iron and heme availability but rather on growth phase and bacterial Bacterial growth was monitored over time by measuring the optical density at 600 nm (statistical analysis is presented in Table S3).density (73).The Rgp activity was similar in both tested strains; however, the 33277 strain exhibited two times higher Kgp activity than the A7436 strain in the two tested conditions (Fig. 5E).Since mainly Kgp is responsible for hemoglobin degradation, these results may explain the higher hemolytic activity of the 33277 strain.Production of some P. gingivalis proteins, including outer membrane virulence factors (such as HmuY), is dependent on iron and heme availability (55).We observed that the 33277 strain, compared to the A7436 strain, bound more heme when cultured in ironand heme-replete conditions (Fig. 5F).When both strains were cultured in iron-and heme-depleted conditions, although the general amount of bound heme was higher in 33277 than A7436 strain, the differences were not statistically significant (Fig. 5F).
In addition, we performed a comparative analysis of A7436 and 33277 strains in terms of their ability to form biofilm. Generally, biofilm formation by P. gingivalis is enhanced when bacteria are grown with reduced iron and heme content in the external environment.As expected (74), the more fimbriated 33277 strain formed larger biofilm structures, as compared to the A7436 strain (Fig. 5G).Moreover, both strains formed larger biofilm structures when bacteria had been starved of iron and heme; however, relative biofilm formation ability increased by ~2.25× for the A7436 strain and only ~1.45× for the 33277 strain (Fig. 5G).Although the results are not that profound, it is worth noting that we observed a higher increase in heme-binding ability caused by culturing the A7436 strain in iron-and heme-depleted conditions (~1.35× increase) as compared to iron-and heme-rich conditions, while for the 33277 strain, the increase was only ~1.2×.Moreover, these results can explain, at least in part, the differences in HmuY levels.Relative HmuY production in the DIP medium was ~35× and ~6× higher in A7436 and 33277 strains, respectively, in comparison to the Hm medium (Fig. 3C; Fig. S1).Altogether, this shows that iron and heme depletion has a greater impact on the phenotype presented by the A7436 strain than the 33277 strain.Therefore, our results suggest that not only the availability of iron and heme in the external environment but also the ability to store iron intracellularly can influence the P. gingivalis gene expression and phenotype.

Hemoglobin influences the interaction of A7436 and 33277 strains with keratinocytes differently
The availability of hemoglobin, as the main source of iron and PPIX, varies at different stages of periodontitis, being highly available in advanced stages of the disease (55,56).Therefore, in this study, we not only compared the interaction of A7436 and 33277 strains with human gingival keratinocytes but also examined the influence of hemoglobin concentration on the adhesion to and invasion of keratinocytes.Also for this experiment, we used the reference W83 strain.The general interaction ability of all strains with keratinocytes was as follows: 33277 > W83 > A7436 (Fig. 6A through C).Increasing concentration of hemoglobin increased the interaction ability of more virulent A7436 and W83 strains.In the case of A7436, hemoglobin increased its invasion ability, and in the case of the W83 strain both adhesion and invasion abilities (Fig. 6A and B).The addition of 2 µM and 5 µM hemoglobin resulted in an increase in the growth of A7436 and W83 strains in the presence of keratinocytes by ~25% and ~50%, respectively (Fig. 6D).Conversely, increasing the concentration of hemoglobin decreased the growth, adhesion, and invasion abilities of the 33277 strain by ~50% (Fig. 6C and D).This finding was quite surprising in light of higher hemolytic and Kgp activities presented by the 33277 strain (Fig. 5D).As a control, we supplemented P. gingivalis-keratinocyte co-cul tures with purified apo-SgGAPDH protein.However, we did not observe any significant changes in this case (Fig. 6D).Overall, our results showed that hemoglobin availability may influence the adhesion to and invasion of keratinocytes of some P. gingivalis strains in a concentration-dependent manner.

DISCUSSION
P. gingivalis is one of the most important microbial agents involved in the initiation and progression of periodontitis (1,(14)(15)(16)(17).The bacterium produces numerous virulence factors, enabling it to survive in different niches of the human body and infect host cells (55).However, it has been shown that P. gingivalis can also be present in healthy humans (e.g., 28,30,32,42), which may suggest that not all P. gingivalis strains are pathogenic or their contribution to this process is of minor importance (75).Indeed, it has been shown that during evolution P. gingivalis strains may have diverged into groups varying in their genotypes and phenotypes, resulting in different virulence potentials (e.g., 28-30, 32-35, 40-43, 76, 77).The diversity of P. gingivalis strains results from their genetic variability caused mainly by the rearrangement of the bacterial chromosome.This seems to be possible because P. gingivalis encodes many transposon sequences that are responsible for this process and their number differs between strains (29).Moreover, recent studies showed that clustered regularly interspaced short palindromic repeat (CRISPR) and CRISPR-associated genes (Cas) (CRISPR-Cas) can influence virulence due to their engagement in transcription regulation or biofilm formation (44,(78)(79)(80)(81).

FIG 6
Influence of hemoglobin on interaction with host cells of P. gingivalis A7436 (A), W83 (B), and ATCC 33277 (C) strains.The ability to interact with, adhere to, and invade host cells was analyzed using a P. gingivalis-gingival keratinocytes co-culture model.The number of viable bacteria was shown as the number of colony-forming units (CFU) per mL of the culture medium.Adhesion-live bacteria attached to keratinocytes; invasion-live bacteria that invaded keratinocytes; interaction-the total number of live bacteria that adhered to and invaded keratinocytes.(D) Summary of the correlation between hemoglobin concentration and P. gingivalis interaction with human keratinocytes.The average interaction of P. gingivalis with keratinocytes in samples without hemoglobin was set as 100% for each tested strain.As a protein control, purified apo-SgGAPDH protein instead of hemoglobin was used.Results are shown as mean ± SD (A through C) or as mean ± SE (D).Hb-hemoglobin; SgGAPDH-glyceraldehyde-3-phosphate dehydrogenase from Streptococcus gordonii.*P < 0.05, **P < 0.01, ****P < 0.0001.
While the genomes of the A7436 and W83 strains are similar, significant differences between them and the 33277 genome were reported (44,82,83,92,93).For example, in comparison to the 33277 strain, both A7436 and W83 strains can survive much longer intracellularly after entering human endothelial cells in vitro (28).However, in contrast to the W83 strain, the A7436 strain traffics through different intracellular pathways and is not able to induce autophagy (26).Also increased colonization of rat tissues caused by the A7436 strain, as compared to the 33277 or W83 strain, has been reported (94,95).Although W83 and A7436 genomes are quite similar, the strains exhibit some differences in the observed phenotypes and virulence potential (26).
Heme is a limiting factor for P. gingivalis growth and is essential for its survival.It is worth mentioning here that heme availability changes during the disease progress.Moreover, the source of heme depends on the niche occupied by the bacterium (55,56).In addition, different heme concentrations affect various P. gingivalis strains differently, resulting in different tolerance to heme (61).To date, several studies have documented the influence of iron and heme availability on gene expression in P. gingivalis (e.g., 96,97).Our additional analyses, including the A7436 strain, not previously studied in this aspect, showed that iron and heme may contribute differently to gene expression in P. gingivalis strains (Tables S1 and S2; Table 2).Some of the genes only partially overlapped (Fig. 2), which may indicate different mechanisms leading to the adaptation of different P. gingivalis strains due to the changing availability of iron and heme.
Heme uptake systems play a key role in the adaptation of P. gingivalis to the niche it occupies.Interestingly, the genetic variation in these systems is negligible between P. gingivalis strains, showing that they are crucial for the bacterium.Our phenotypic characterization of examined strains grown in iron-and heme-replete or in iron and heme-depleted conditions showed that expression of the Hmu system in both tested strains is highly dependent on iron and heme availability, whereas expression of other P. gingivalis putative heme uptake systems, such as Iht and Hus systems, is not iron-and heme-dependent (Fig. 3).Iron-dependent regulation in bacteria is typically mediated by a ferric uptake regulator (Fur).Although we have shown that a P. gingivalis Fur homolog (PgFur) does not function like the classic Fur protein, it regulates the production of HmuY protein (98,99).Moreover, we observed that PgFur has a different effect on the regulation of virulence factors in 33277 and A7436 strains (100).It seems that the 33277 strain exhibits a lower ability to accumulate iron intracellularly (Fig. 5C), which may directly affect its susceptibility to starvation from this element, and the production of proteins whose regulation is dependent on this ion.One example is the HmuY protein, the production of which is directly related to the availability of iron, being produced in higher amounts after iron and heme starvation (55).We found that in iron-and hemereplete conditions production of HmuY is higher in the 33277 strain as compared to the A7436 strain (Fig. 3).Such an effect might be caused by the lower amount of intracel lular iron in the 33277 strain, which stimulates the bacterial cell to produce proteins responsible for iron acquisition, including higher expression of the HmuY protein.This might result, at least in part, in the higher adsorption of heme on the surface of the 33277 strain grown in iron-and heme-rich conditions (Fig. 5F).We observed a similar effect for the IhtB protein, production of which is in general higher in the 33277 strain than in the A7436 strain (Fig. 3).However, in contrast to HmuY, IhtB production is not regulated by PgFur (98).An additional potential explanation of the differences observed between the examined strains in iron and heme homeostasis and their influence on the phenotype is the lack of the complete two-component regulatory system (HaeSR) in the 33277 strain, which is engaged in other strains in the regulation of expression of the hmu and iht operons, as well as in the regulation of expression of genes encoding gingipains (101).
The interaction of P. gingivalis with other bacteria and host cells is essential for an efficient virulence process, often dependent on the proteins that are components of the outer membrane, for example, fimbriae, hemagglutinins, and gingipains.In general, P. gingivalis tends to form higher biofilm structures under iron and heme starvation in comparison to conditions of iron and heme excess (74), a feature also confirmed in this study.Moreover, we also found that iron and heme starvation leads to a greater increase in biofilm formation by the A7436 strain in comparison to the 33277 strain.We think that differences in iron and heme homeostasis play an additional role in different P. gingivalis strains' phenotypes.Moreover, the ability of the A7436 strain to accumulate more iron intracellularly may provide it with a greater capacity to adapt to changing environmental conditions, including stress conditions, such as iron and heme deficiency.Consequently, it could lead to quantitative and qualitative differences in the composition of the outer membrane of particular strains, thus influencing the ability to interact with host cells differently.In the advanced stage of periodontitis, the main source of heme for P. gingivalis is hemoglobin, which is present in large amounts after lysis of erythro cytes.Therefore, its concentration may affect the expression of heme-dependent genes in P. gingivalis.The presence of many proteins, including gingipains, hemagglutinins, HmuY, and HmuR, affects the binding of heme and hemoglobin on the surface of the P. gingivalis cell.Our data showed that the interaction with host cells of the strains can vary depending on the amount of free hemoglobin in the external environment.We noted an increase in the adhesion to and invasion of keratinocytes of A7436 and W83 strains with an increasing hemoglobin concentration but the opposite effect for the 33277 strain (Fig. 6).Another explanation of this effect could be the lower resistance of the 33277 strain, as compared to the A7436 strain, to oxidative stress (102).The addition of hemoglobin under aerobic conditions used in our study may generate an increase in reactive oxygen species, which might affect the survival of bacteria.This may prove the better adaptation ability of the A7436 strain to the changing conditions of the external environment, as compared to the 33277 strain.
We found that the 33277 strain produced a higher amount of HmuY in iron-and heme-replete conditions, and generally produced higher amounts of IhtB, as well as Kgp.The influence of the latter protein, besides higher Kgp enzymatic activity, could also be responsible for the increased hemagglutinating activity.Therefore, a higher amount of heme-and hemoglobin-binding proteins on the surface of the 33277 strain may result in higher heme and hemoglobin binding, thus preventing adhesion to host cells and subsequent entry into keratinocytes.Moreover, besides proteinaceous components of the outer membrane, differences in LPS structures, mainly in lipid A, might be responsi ble for strains' variability (89,(101)(102)(103)(104).It was reported that differences in both fimbriae and LPS between P. gingivalis strains can result in different hydrophobicity of strains, which may cause various recognition of molecules produced by cohabitating bacteria, different co-aggregation abilities, and different capacities for biofilm formation, and possibly differences in heme and hemoglobin binding by outer membrane molecules (103)(104)(105)(106)(107)(108).Gingipains are modified by the APS (58,90,91,109,110).Therefore, we assumed that differences in molecular weight observed for pro-RgpB (Fig. 3 and 4C) could be due to the changes in LPS structures.Both A7436 and 33277 strains differ in glycosylation (100), which could, at least in part, explain the effect observed in the examined strains (Fig. 3C, D, and 4CD, ).
In conclusion, accumulating evidence demonstrates that P. gingivalis strains exhibit different phenotypes in vitro, different virulence potential in animal models, and different associations with human diseases.We also observed that the less virulent 33277 strain and more virulent A7436 strain exhibit different phenotypes in regard to iron and heme acquisition, which may be caused by their genetic divergence.We think that the ability to accumulate iron intracellularly, and resistance to iron and heme starvation being different in the examined P. gingivalis strains, may exert a significant impact on their observed phenotypes, influencing the production of virulence factors and the virulence potential of P. gingivalis strains.We believe that the results presented in this study shed new light on differences between P. gingivalis strains and their virulence potential to cause dysbiosis in the oral cavity.

Bacterial strains and growth conditions
Porphyromonas gingivalis wild-type A7436, ATCC 33277 (33277) and W83 strains were maintained on Schaedler blood agar (ABA) plates (Biomaxima, Lublin, Poland) in anaerobic conditions (80% N 2 , 10% H 2 , and 10% CO 2 ) (Whitley A35 anaerobic work station; Bingley, UK) at 37°C for 3-5 days.Then, bacteria were inoculated into liquid basal medium (BM) prepared from 3% trypticase soy broth (Becton Dickinson, Sparks, MD, USA) and 0.5% yeast extract (Biomaxima), supplemented with 3.6 mM L-cysteine hydrochloride (Carl Roth, Karlsruhe, Germany), 0.5 mg/L menadione (Sigma-Aldrich, St. Louis, MO, USA), and 7.7 µM hemin chloride (Fluka, Munich, Germany) (Hm medium).To examine the influence of iron and heme on gene expression, bacteria were grown in two types of liquid media.To imitate bleeding occurring in individuals with an advanced stage of periodontitis (characterized by high concentrations of iron and heme in the oral cavity), bacteria were grown in Hm medium.To imitate the conditions of a healthy oral cavity (characterized by trace amounts of iron and heme), 160 µM of the iron chelator 2,2-dipyridyl (Sigma-Aldrich) (DIP medium) was used instead of hemin chloride.To monitor P. gingivalis growth, bacteria were inoculated into Hm or DIP medium with a starting optical density at 600 nm (OD 600 ) equal to 0.2.P. gingivalis growth was monitored using a Stratus plate reader (Cerillo, Charlottesville, VA, USA) as described previously (56).

Biofilm formation
To compare the biofilmforming capacity of P. gingivalis 33277 and A7436 strains, a sterile 96-well plate (Corning, New York, US) was inoculated with 200 µL of bacterial cultures in Hm or DIP medium per well at starting OD 600 equal to 0.2.Bacteria were cultured at 37°C under anaerobic conditions for 24 hours.Then, wells were washed three times with 20 mM sodium phosphate buffer, pH 7.4, containing 140 mM NaCl (phosphatebuffered saline, PBS) to remove unattached bacteria.The biofilm was stained with 1% crystal violet (Carl Roth, Karlsruhe, Germany) for 10 minutes, then washed five times with PBS and de-stained with 99.8% ethanol.Absorbance at 570 nm (A 570 ) was measured using a GloMax Discover plate reader (Promega, Madison, WI, USA).The obtained values were normalized to the OD 600 of the liquid bacterial cultures equal to 1.

Microarray analysis of gene expression
P. gingivalis A7436 and 33277 strains were grown in three biological replicates in Hm and DIP media for 20 hours as described above.The microarray analysis was performed in IMGM laboratories (Martinsried, Germany) as described previously (99).Briefly, the online tool eArray (http://earray.chem.agilent.com/;Agilent Technologies, Santa Clara, CA, USA) was used to design an Agilent Custom Porphyromonas gingivalis A7436 Gene Expres sion Microarray (8 × 15K format).Probes were prepared based on P. gingivalis transcrip tome information derived from the NCBI reference sequence NZ_CP011995.1.Total RNA isolation, RNA quantity, and quality were determined as described previously (99).For internal labeling control, the total RNA was spiked with in vitro synthesized polyade nylated transcripts (One-Color RNA Spike-In Mix; Agilent Technologies).Subsequently, samples were reverse transcribed into cDNA and then converted into cyanine-3-labeled complementary RNA (cRNA) with Low Input Quick-Amp Labeling Kit One-Color (Agilent Technologies).For microarray hybridization, a Gene Expression Hybridization Kit (Agilent Technologies) was used.Labeled cRNA was hybridized for 17 hours at 65°C on Agilent Custom GE 8 × 15K Microarrays, washed according to the manufacturer's protocol, and dried with acetonitrile (Sigma-Aldrich).The fluorescence of samples was detected with Scan Control A.8.4.1 software (Agilent Technologies) on the Agilent DNA Microarray Scanner (Agilent Technologies) and extracted from the images using Feature Extraction 10.7.3.1 software (Agilent Technologies).For data analysis, Feature Extraction 10.7.3.1 (Agilent Technologies), GeneSpring GX 13.1.1(Agilent Technologies), and Excel 2010 (Microsoft, Redmond, WA, USA) were used.For statistical analysis, Welch's approximate t-test was used.Differences in gene expression are shown as fold change values (FC).The average was calculated from the normalized signal values and they were transformed from the log 2 to the linear scale.Increases and decreases in gene expression are shown as positive and negative numbers, respectively.The fold change in gene expression was considered significant for FC ≥ 2 or FC ≤ −2 and P-value ≤ 0.05.

Gene expression analysis examined using reverse transcriptase-quantitative polymerase chain reaction
P. gingivalis strains were maintained for 10 or 24 hours in Hm and DIP media.Total RNA was isolated with the Total RNA Mini Kit (A&A Biotechnology, Gdańsk, Poland), followed by the Clean-up RNA concentrator Kit (A&A Biotechnology).cDNA was obtained by reverse transcription using a LunaScript RT SuperMix Kit (New England Biolabs, Ipswich, MA, USA).For PCR, the SensiFAST SYBR no-ROX Kit (Bioline, London, UK) and LightCycler 96 (Roche, Basel, Switzerland) were used.The PCR program consisted of initial denatura tion at 95°C for 120 seconds, 40 cycles of denaturation at 95°C for 5 seconds, primer annealing at 60°C for 10 seconds, and extension at 72°C for 15 seconds.For PCR product quality, the melting curves were generated and analyzed.All samples were examined in triplicate.Fold change in gene expression was calculated with LightCycler 96 software (Roche).All primers used in this study are listed in Table S4, including the P. gingivalis 16S rRNA gene used as the reference gene.

IhtB protein purification and anti-IhtB antibody production
To generate anti-IhtB antibodies, the recombinant IhtB protein (GenBank accession number AKV64486), lacking the predicted signal peptide (MKKLILATLGLMAIAMLSCS), was purified.Briefly, the ihtB gene was PCR amplified using primers listed in Table S4.The PCR product was cloned into XcmI and BamHI restriction sites of a pTriEx-4 plasmid (Sigma-Aldrich) using NEBuilder HiFi DNA Assembly (New England Biolabs), enabling IhtB protein production with the 6×His tag fused to the N-terminus with the possibil ity to be removed with Xa factor.IhtB protein was overexpressed using Escherichia coli BL21-CodonPlus (DE3)-RIL strain (Agilent Technologies, Santa Clara, CA, USA) and purified from the soluble E. coli cell lysate using TALON Superflow resin according to the manufacturer's protocol (Sigma-Aldrich).Subsequently, the 6×His-tag was cut off using the Xa factor according to the manufacturer's protocol (New England Biolabs) and removed by washing and protein concentration using Amicon Ultra-4 Centrifugal Ultracel-10KDa filter units (Millipore).
Anti-IhtB antibodies were produced in rabbits using purified IhtB protein as an antigen, according to the company's protocol (ProteoGenix, Schiltigheim, France).
To analyze potential differences in glycosylation of macromolecules, a lectin assay was performed as reported previously (100).Briefly, whole cell lysates were separated by SDS-PAGE, transferred onto nitrocellulose membranes, and probed with JACALIN lectin (1:5,000) specific to galactose linked with β−1,3 linkage to N-acetylgalactosamine and mono-or di-sialylated form of this structure (Vector Laboratories, Newark, CA, USA).Lectin binding was detected using avidin D conjugated with horseradish peroxidase (1:10,000; Vector Laboratories).Equal loading of proteins was visualized on nitrocellulose membranes by staining with Ponceau S (data not shown).
Western blotting and lectin assay signals were visualized using chemiluminescence staining (Perkin Elmer, Waltham, MA, USA) and a ChemiDoc Imaging System (Bio-Rad Laboratories, Hercules, CA, USA).Densitometric analysis (relative quantification) was performed using Image Lab 6.0.1 software (Bio-Rad).

Determination of intracellular iron content
Intracellular iron concentration was determined using the ferrozine-based method as described previously (56).Briefly, aliquots of overnight bacterial cultures (BM + Hm; 20 mL) were centrifuged and washed with PBS, and bacterial suspensions (4 mL) at OD 600 equal to 2.0 were sonicated.Samples were then treated according to the procedure reported by Ceriotti and Ceriotti (111).The absorbance of the supernatants was measured at 562 nm.The iron content was determined from a standard curve prepared similarly to bacterial samples using FeCl 3 (Sigma Aldrich).

Heme binding by the whole P. gingivalis cells
The ability of heme binding to whole P. gingivalis cells was performed as described previously (56).Briefly, aliquots of overnight bacterial cultures (BM + Hm or BM + DIP; 20 mL) were centrifuged and washed with PBS, and bacterial suspensions (0.8 mL) at OD 600 equal to 1.25 were mixed with 0.2 mL of hemin chloride in PBS (50 µg/mL).After incubation for 1 h at 37°C, the samples were centrifuged and the absorbance of the supernatant was measured at 385 nm.Control samples contained all components except bacteria.The amount of heme bound to bacterial cells was determined from the difference in absorbance between the control and bacterial samples.

Determination of gingipain activity
Gingipain activity was measured as reported by others (112) with minor modifications.Briefly, to 150 µL of the reaction buffer (20 mM TRIS pH 7.5, supplemented with 150 mM NaCl, 0.05% Tween 20, 5 mM CaCl 2 , and 10 mM L-cysteine hydrochloride, neutralized with NaOH), 10 µL of P. gingivalis cultures was added and samples were incubated for 15 minutes at 37°C.The reaction was started by addition of 50 µL of reaction buffer, supplemented with 2 mM N-(p-tosyl)-Gly-Pro-Lys 4-nitroanilide acetate salt (Sigma-Aldrich) for determination of the Kgp activity or 2 mM Nα-benzoyl-DL-argi nine p-nitroanilide hydrochloride (BApNA; Sigma-Aldrich) for determination of the Rgp activity.The samples were incubated for 2 hours at 37°C and the reaction was monitored by measuring the increase of absorbance at 405 nm (A 405 ) over time using a GloMax Discover plate reader (Promega).Rgp or Kgp activities are shown as an increase in absorbance at 405 nm (A 405 ) per minute per 1 µL of the bacterial culture at an optical density at 600 nm (OD 600 ) equal to 1 (mOD/min/μL).

Hemolytic activity
Aliquots of 25 µL of overnight liquid bacterial cultures (BM + Hm) at OD 600 equal to 2 were applied on ABA plates in spots of ~1 cm diameter and cultured for 5 days.Bacteria were removed from plates and relative hemolytic activity was determined with the ChemiDoc Imaging System (Bio-Rad) using densitometric analysis (relative quantifi cation) with the Image Lab 6.0.1 software (Bio-Rad).
To determine the influence of hemoglobin on P. gingivalis interaction with keratino cytes, human hemoglobin (Sigma-Aldrich) was used.As a control, another heme-bind ing protein, namely Streptococcus gordonii glyceraldehyde-3-phosphate dehydrogenase (SgGAPDH), was used in its apo form.SgGAPDH was overexpressed and purified as described previously (114).To remove endotoxins, Detoxi-Gel Endotoxin Removing Columns were used according to the manufacturer's protocol (Thermo Scientific).
P. gingivalis was grown to the early stationary phase in Hm medium, centrifuged (4,000 × g, 20 minutes, 4°C), and washed twice with PBS.Keratinocytes were washed three times with PBS and a fresh DMEM medium (Sigma-Aldrich) or DMEM medium supplemented with a sterile solution of human hemoglobin or SgGAPDH at the final concentration of 2 µM or 5 µM.The cells were treated with P. gingivalis with the multiplicity of infection (MOI) of 100 and incubated for 4 hours at 37°C in an atmosphere of 5% CO 2 .Then, the wells were washed three times with PBS.To determine the bacteria present inside the cells and attached to them, a fresh DMEM medium was added.To kill the external bacteria and to determine only live bacteria inside the cells, DMEM medium supplemented with 300 µg/mL gentamicin (Sigma-Aldrich) and 200 µg/mL metronida zole (Sigma-Aldrich) was added.After 1 hour, the wells were washed three times with PBS, and lysed with sterile distilled water.The cell lysates were used to prepare serial dilutions, which were plated on ABA plates.The plates were incubated for 7 days at 37°C in anaerobic conditions to determine colony-forming units (CFU).

Statistical analyses
Except for microarray analysis, all statistical analyses were performed with GraphPad software (GraphPad Prism 8.0 Inc., San Diego, CA, USA) using the unpaired Student's t-test or one-way analysis of variance (ANOVA) test with post hoc Tukey's test.All experiments were conducted at least three times with at least two biological replicates and collected data are shown as mean ± standard deviation (mean ± SD) or as mean ± standard error (mean ± SE).

FIG 4
FIG 4 Comparison of RgpB produced by P. gingivalis A7436, 33277, and W83 strains.(A) Amino acid sequence alignment.(B) Gene length analysis.The whole DNA sequence of the rgpB gene was amplified using genomic DNA of A7436, 33277, or W83 strains and primers listed in Table S4.(C) Differences in RgpB molecular weight were examined in whole cell lysates by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting using anti-RgpB antibody.(D) Differences in the production of glycans were examined using reactivity with JACALIN (the lectin preferentially recognizes galactose linked with β−1,3 linkage to N-acetylgalactosamine and mono-or di-sialylated form of this structure).

FIG 5
FIG5 Phenotypic analysis of P. gingivalis A7436 and ATCC 33277 (33277) strains.(A) Growth of bacteria in liquid culture media containing iron and heme (iron-and heme-replete conditions, Hm medium) or without heme and supplemented with the iron chelator 2,2-dipyridyl (iron-and heme-depleted conditions, DIP medium).(B) Recovery of bacteria after prior heme and iron starvation.Bacteria were cultured for two passages in a DIP medium and then transferred to a fresh DIP medium supplemented with heme (DIP + Hm) or hemoglobin (DIP + Hb).Growth without an iron source was used as a control (DIP medium).
(C) The intracellular amount of iron was measured using a ferrozine-based assay.(D) Hemolytic activity was examined on anaerobic blood agar (ABA) plates after 5 days by visual inspection, semi-quantitatively evaluated by densitometric analysis, and shown as the relative hemolytic activity in the 33277 strain in relation to the A7436 strain (the latter set as 1.0).To better visualize and analyze hemolysis, bacterial colonies were removed.BM medium was used as a negative hemolysis control.(E) Gingipain activities of whole P. gingivalis cultures were measured using argininespecific (Rgp) and lysinespecific (Kgp) substrates.Enzymatic activity was determined as an increase in absorbance at 405 nm per 1 minute per 1 µL of bacterial culture of optical density at 600 nm equal to 1. (F) Heme-binding ability by P. gingivalis cells.(G) Biofilm formation on the abiotic surface was examined using crystal violet staining and absorbance measurement at 570 nm.Insets in the graphs show a relative increase in heme binding (F) or biofilm formation (G) in iron-and heme-depleted conditions (DIP medium) in comparison to iron and heme-replete conditions (Hm medium).*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

TABLE 1
Genome and hmu operon access IDs of selected P. gingivalis strains

TABLE 2
Gene expression analyzed using RT-qPCR.Relative transcript levels were determined in bacteria grown in iron-and heme-depleted conditions (DIP medium) in comparison to iron-and heme-replete conditions (Hm medium)