Inherently and conditionally essential protein catabolism genes of Porphyromonas gingivalis

Abstract

Porphyromonas gingivalis proteases are key to the virulence of this periodontitis agent and emerging Alzheimer’s disease, cancer and arthritis pathogen. Transposon sequencing has been employed to define the core essential genome of this bacterium and genes conditionally essential in multiple environments - abscess formation; epithelial colonization; and cigarette smoke toxin exposure; as well as to elucidate genes required for iron acquisition and a functional type 9 secretion system. Protein catabolism genes identified include a combination of established virulence factors and a larger set of seemingly more mundane proteolytic genes. The functions and relevance of genes that share essentiality in multiple disease-relevant conditions are examined. These common stress-related genes may represent particularly attractive therapeutic targets for the control of P. gingivalis infections.

P. gingivalis proteases and virulence

Porphyromonas gingivalis is a causative agent of periodontitis that has also been associated with multiple systemic conditions, including pre-term birth [1, 2], arthritis [3, 4], vascular problems [5, 6] and, more recently, head and gastrointestinal tract cancers [7, 8] as well as Alzheimer’s disease [5, 9]. P. gingivalis is an asaccharolytic anaerobe that is reliant on proteolysis for metabolic function and subsequent virulence in the oral cavity and, presumably, beyond. In addition to provision of carbon [10, 11], the proteases and peptidases of this bacterium play critical roles in iron acquisition [12, 13], adhesion, colonization and invasion [14, 15], immune subversion [16–18] and tissue destruction [15, 17]. This cacophony of virulence traits, as well as other non-proteolytic mediators of pathogenesis, has led to the recognition of P. gingivalis as a key pathogen. It has been further posited that, even at low incidence, P. gingivalis is able to induce community dysbiosis, at least in mice [19, 20].

Therefore, proteinases have oft been proposed as therapeutic targets. However, while certain protein catabolism gene products of P. gingivalis are well characterized, others remain, essentially, unstudied. Transposon mutagenesis has facilitated insight into which proteolytic enzymes of P. gingivalis are essential for survival per se and which are essential in several disease-relevant environments. Of particular focus, herein, will be stress-related genes that make critical contributions to fitness in multiple niches. As delineated below, the common conditionally essential stress-related proteolytic gene set encodes a combination of established virulence factors and a larger set of seemingly more mundane and / or uncharacterized proteolytic genes that may yet prove to represent particularly attractive therapeutic targets for the control of P. gingivalis infections.

Transposon sequencing and P. gingivalis

The emergence of transposon sequencing (TnSeq) technology allows researchers to rapidly “carry out genome-wide studies in a wide range of bacterial species, under a multitude of conditions, with unprecedented depth” [21]. Initial attempts at library building used technologies that resulted in mutagenesis at non-random sites or hotspots [22–24]. Several P. gingivalis-specific TnSeq libraries, however, have now been generated by independent groups that exhibit both random insertion into the chromosome and high genome saturation [12, 14, 25–28]. Such successful libraries have been built using the Himar 1 Mariner mini-transposon system originally designed for Bacteroides thetaiotaomicron and utilizing the pSAM_Bt vector [29] or the plasmid pMI07-based system, originally designed for Bacteroides spp, in general [30]. To date, the highly penetrative libraries, containing between 54,000 and 80,000 mutant clones each, are specific to P. gingivalis ATCC 33277 [12, 25–27], the most comprehensively studied P. gingivalis strain. While we have had partial success with P. gingivalis HG66 (unpublished data), penetration into the P. gingivalis W83 genome has proven problematic, perhaps due to significant bacterial capsulation. However, Klein et al have successfully generated a smaller 12,000-transposon insertion P. gingivalis W83 library which has been employed to screen for pigmentation-related genes [12]. Therefore, unless otherwise stated, the literature discussed herein specifically refers to P. gingivalis ATCC 33277.

Absolutely essential proteases of the P. gingivalis

A comparison of the TnSeq libraries for P. gingivalis ATCC 33277 independently generated by Klein et al [26] and Hutcherson et al [25] identified 281 common genes, out of a total of 2155 (13%) in the chromosome, as essential for planktonic growth in complex media. Perhaps surprisingly, and presumably due to a lack of selective pressure, only 5 proteolytic genes, as presented in Table 1, were found to be part of this core essential genome. PGN_0202 (PG_2157; PGTDC60_1267) encodes a putative leucine aminopeptidase precursor, related to the M28 group of metalloproteinases, that likely functions to release N-terminal amino acids, preferably leucine, from polypeptides. PGN_0250 (PG_0137; PGTDC60_0414), a pepD-like peptidase gene, whose product likely targets N-terminal neutral or hydrophobic residues from Xaa-His peptide or peptide fragments [31]. PGN_0993 (PG_0956; PGTDC60_0878) encodes a M23/M27 metalloendopeptidase, also likely to have a broad specificity. The reasons for non-conditional essentiality of PGN_0202, PGN_0250 and PGN_0993 are not immediately obvious, yet they may nevertheless represent attractive, previously unheralded, therapeutic targets. On the other hand, LepB (PG_1598; PGTDC60_0703) is a single copy signal peptidase I that cleaves N-terminal signal sequences from secreted and periplasmic proteins. Lsp A (signal peptidase II, PG_0515; PG_2001; PGTDC60_0274), also encoded by a single copy gene, is responsible for the removal of signal peptides from prolipoproteins. The fundamental roles of LepB and Lsp A in protein transport likely explains core essentiality.

Table 1.

Essential proteases of the P. gingivalis ATCC 33277 core genome.

Gene	Annotation */**		EC/ WP Number***
PGN_0202		Leucine aminopeptidase precursor M28 group metalloproteinase	BAG32721 WP_012457324.1
PGN_0250		pepD-like aminoacyl-histidine dipeptidase	3.4.13.- WP_012457365.1
PGN_0515		IspA; signal peptidase II	3.4.23.36 WP_004584054.1
PGN_0993		M23/M27 family peptidase	BAG33512 WP_012457940.1
PGN_1946		lepB; signal peptidase I	3.4.21.89 WP_012458631.1

^*Cross library comparison of genes essential for planktonic growth in complex media led to the identification of 251 common genes in total [25, 26], of which 5 were related to protein catabolism, as determined by investigation of the GenBank (www.ncbi.nlm.nih.gov/genbank/) and KEGG (www.kegg.jp) databases, enhanced by MEROPS (www.ebi.ac.uk/merops).

^**Structural models were generated using the SWISS-MODEL homology-modelling server (https://swissmodel.expasy.org/).

^***Enzyme Commission and NCBI Reference Sequence accession numbers, respectively, are provided. If an Enzyme Commission number is unavailable, the GenBank accession BAG code is presented.

Conditionally essential P. gingivalis protease genes

In addition to the identification of a core genome [25, 26], TnSeq has been used to determine P. gingivalis genes that determine fitness in a number of specific environments, as presented in Figure 1. These are epithelial colonization [14], murine abscess formation [14] and survival of cigarette smoke toxins [28], as well as to pinpoint genes required for black pigmentation [12], requisite for gingipain transport and hemin acquisition, and, relatedly, for a functional type IX secretion system (T9SS) [27]. Herein, we specifically delineate and summarize those established or putative protein catabolism genes that endow P. gingivalis with fitness in such diverse niches. The specific protease-related genes essential for each condition are presented in Supplemental Table 1a–e.

Figure 1: Exploitation of transposon sequencing libraries to identify P. gingivalis genes essential for fitness in multiple environments.

Highly saturated P. gingivalis TnSeq libraries have been used to identify genes that are conditionally essential for (A) murine abscess formation [14]; (B) epithelial colonization [14]; (C) survival of cigarette smoke toxins [28]; (D) black pigmentation [12]; and (E) maintenance of a functional T9SS [27]. High throughput sequencing allows for the identification and quantification of multiple transposon insertion sites simultaneously among a large mixture of mutants. Varying library generation and analytical methods have been employed, while the technology is rapidly evolving, e.g., [21, 79–81]. Generally, however, subtraction of the sequenced output library (non-essential mutants) from the larger input library (essential and non-essential mutants) allows elucidation of those individual strains that are required for survival under a specific condition of interest and, so, unable to propagate, are absent or under-represented in the output library. Herein, we focus particularly on those P. gingivalis protease-encoding genes that are essential in several disease-related environments, assigning high potential therapeutic value to communal environmental fitness determinant(s).

P. gingivalis proteases essential under four disease-relevant conditions (n = 2)

There are two P. gingivalis proteolysis genes that are essential in 4/5 conditions in which TnSeq libraries have been examined. They are an established virulence factor, heat shock-related protein A, and an essentially uninterrogated papain-like cysteine endopeptidase, bleomycin hydrolase (PGN_1777; PG_1788; PGTDC60_0106). While bleomycin hydrolase is conserved across kingdoms, its function in bacteria is largely unknown. P. gingivalis, however, has a large number of immune defense evasion mechanisms, many protease-based [32–34]. Mammalian bleomycin hydrolase has been shown to contribute to the regulation of chemokine release by in human innate cells. More specifically, the bleomycin hydrolase signal inversely correlates with expression of CXCL8 and GROα in several inflammatory skin conditions in humans, a phenomenon reflected in keratinocytes in which bleomycin hydrolase has been knocked down [35]. While the mechanisms underlying the bleomycin hydrolase-CXCL8 axis remain to be elucidated, should the equivalent P. gingivalis enzyme function similarly, it is possible that the bacterial bleomycin hydrolase could reduce local chemokine activity. This could be in addition to the direct hydrolysis of CXCL8 by gingipains [36, 37]. The advantage, presumably, would be reduced neutrophil, or other innate cell, chemotaxis towards P. gingivalis. Bleomycin is a Streptomyces verticillus-derived antibiotic used as component of treatment for specific cancers, including several squamous cell cancers. P. gingivalis has been associated with multiple oral and gastrointestinal epithelial cancers, but the use of bleomycin in in such patients is not routine. However, while speculative, a role for PGN_1777 orthologues in other bacterial-associated cancers where bleomycin is indeed a chemotherapeutic drug of choice, may be worthy of investigation. What is clear, is that further research is required to better understand the reliance of P. gingivalis on bleomycin hydrolase for survival in multiple, highly variable environments.

Heat shock-related protein A (PGN_0637; PG_0593; PGTDC60_1718) is intimately involved in the transmembrane transport of proteins [38], helps protect against oxidative stress [39, 40] and has been shown to be important for P. gingivalis virulence in mice [40], as presented in Figure 2. More specifically, HtrA, a predominantly outer membrane protease with preference for unfolded proteins, is requisite for the degradation of T9SS cargo proteins that are not protected by the presence of an IgG-domain [38]. Consequently, in the absence of HtrA, incorrectly processed proteins, including gingipains, accumulate near the bacterial cell surface [38, 41]. Such a fundamental role in the quality control of transported proteins likely explains common conditional essentiality for PGN_0637. The ability to withstand oxidative stress is particularly relevant in murine abscess, epithelial colonization and tobacco exposure environments. It is interesting, then, that a P. gingivalis W83 htrA mutant strain exhibit increased sensitivity to hydrogen peroxide, as is the case in several other bacterial species, while gingipain activity was compromised upon heat shocking [39, 40]. LuxS, critical for the formation of autoinducer II of P. gingivalis, is a negative regulator of htrA gene activation in strain W83, suggesting a potential role for HtrA in adaptation to higher density conditions [42]. Indeed, htrA encodes a known protein which appears to be an important virulence factor. Yuan et al have shown that a heat shock-related protein A P. gingivalis W83 mutant is outcompeted by the replete strain following subcutaneous injection in mice while the mutant exhibits reduced lethality compared to the parent bacterium [40]. Therefore, the biological rationale for essentiality in multiple conditions is more obvious for htrA than it is for bleomycin hydrolase.

Figure 2: Virulence-related attributes of the commonly conditionally essential stress-related P. gingivalis ATTCC 332277 gene, htrA (PGN_0637).

Heat shock-related protease A is an essential fitness gene under four different experimental conditions (abscess formation; epithelial colonization; tobacco smoke survival; and black pigmentation), as determined using TnSeq technology [12, 14, 28]. Interestingly, HtrA has been shown to be important to several virulence-related functions of P. gingivalis, including (A) quorum sensing [42]; (B) a functional T9SS and gingipain transport [38, 41]; (C) murine abscess formation [40]; (D) oxidative stress protection [39, 40]; (E) murine lethality [40]; and (F) temperature stability [39].

P. gingivalis proteases essential under two or more disease-relevant conditions (n = 24)

In addition to PGN_0637 and PGN_1777, a number of genes are conditionally essential in multiple conditions tested, again as noted in Supplemental Table 1. Those genes essential in three conditions (n = 16) are PGN_0271 (pepO); PGN_0295 (C-terminal domain of Arg- and Lys-gingipain proteinase), recently shown to be required for appropriate gingipain transmembrane transport and maturation [43]; PGN_0303 (an M16 family zinc protease); PGN_0561 (PrtT protease); PGN_0607 (dipeptidyl peptidase 11); PGN_0754 and PGN_0771 (NLP/P60 family proteins); PGN_0780 (PrtQ protease); PGN_0788 (peptidyl-dipeptidase); PGN_0900 (thiol protease / hemagglutinin / periodontain); PGN_1335 (hypothetical peptidase); PGN_1349 (a dipeptidyl aminopeptidase / acylpeptidyl oligopeptidase [44]); PGN_1466, the classic P. gingivalis virulence-related protease, rgbB; PGN_1479 (S46 family dipeptidyl-peptidase; PGN_1694 (alanyl dipeptidyl peptidase); and PGN_2064 (Oma1, M48-family peptidase).

Those of particular interest include PGN_0271 (PG_0159; PGTDC60_0435), an established virulence-related gene encoding an endopeptidase that is common among 40 global P. gingivalis isolates [45]. Interestingly, pepO activity is upregulated during, and is requisite for, invasion of epithelial cells, as has been previously shown [46]. A mechanistic role for PepO in epithelial invasion is also implied by aberrant cytoskeleton rearrangements seen on the interaction of a pepO mutant with gingival epithelial cells [46]. Mutation of the P. gingivalis 381 orthologue similarly resulted in reduced invasion of HeLa cells [47]. PepO, which exhibits high homology to the cardiovascular disease risk factor, human endothelin-converting enzyme 1 [46–48], and can convert big endothelin-1 to functional endothelin-1 [48], a potent vasoactive peptide. PepO is also packaged into outer membrane vesicles [49] and, therefore, likely to be distributed in the periodontal tissues themselves and may circulate beyond.

PGN_0561 (PG_1427; PGTDC60_0751), prtT, encodes a secreted cysteine protease and streptopain homologue, that also possesses hemagglutinin activity, with at least two distinct interstrain catalytic domain allele variants [50, 51]. PGN_0900 encodes an extacellular cysteine protease, known as periodontain, a potent inactivator of α1-antrypsin inhibitor which exhibits high homology to PrtT [51]. Such functionalities would be expected to increase the local burden of endogenous serine proteases, particulalrly neutrophil elastase, leading to augmented host-mediated collateral tissue damage and nutrient availability to asaccharolytic bacteria. Clearly, the ability to acquire iron and peptide-based carbon sources is critical to bacterial survival. In keeping, PrtT has been shown to be an important virulence factor in vivo. Essentially, a non-human primate P. gingivalis 3079 strain lacking trypsin-like activity, unlike the parent strain, was non-lethal and unable to induce lesions following subcutaneous inoculation [52]. Further, inactivation of rgpA results in upregulation of prtA gene activity [53], suggesting proteinase T may also act as a compensatory gingipain. Despite the potential of periodontain to represent an important P. gingivalis virulence factor, PGN_0900-related pathogenesis has not yet been studied in animal models, to the best of our knowledge.

P. gingivalis produces several dipeptidyl peptidases which are considered fundamental for asaccharolytic carbon metabolism [10, 11]. Orthologues of PGN_0607 (DPP11), an Asp/Glu-specific periplasmic exopeptidase, are common among bacteria but not so in other kingdoms, increasing the attractiveness of this peptidolytic enzyme as a therapeutic target [10]. Further, aspartate and glutamate peptides are the amino acids primarily utilized by P. gingivalis [54], while disruption of PGN_0607 reduces bacterial growth even in rich medium [11], shedding light on the essentiality of this particular dipeptidase.

Finally, while gingipain-encoding rgbB (PGN_1466) did not form part of the core P. gingivalis genome, its conditional essentiality emerged upon the imposition of several environmental stresses. As the best characterized P. gingivalis proteases, there are several gingipain reviews already available, not limited to [10, 34, 55–58]. The gingipains account for ~85% of the proteolytic potential of P. gingivalis [59]. The 2 Arg-X-specific gingipains (RgpA and RgpB) contain nearly identical catalytic domains, however RgpB lacks the large C-terminal hemagglutinin domain of RgpA and the Lys-X-specific gingipain, Kgp. The absence of this adhesion-associated domain in RgpB suggests essential functions are adhesion-independent and likely specifically related to proteolytic activity. The many pathogenicity associated roles of RgpB are well characterized and includes complement evasion by degrading several complement proteins [60], degradation of host defensins [61], proteolytic destruction of numerous host receptors and effectors that dysregulates host immune responses [62].

Those genes essential in two conditions (n = 8) are PGN_0022 (porU, cysteine proteinase), PGN_1149 (prolyl tripeptidase A), PGN_1409 (a putative M20 peptidase), PGN_1416 (pepK, lysyl endopeptidase precursor), PGN_1469 (dipeptidyl peptidase IV), PGN_1728 (the classic P. gingivalis virulence-related protease, kgp), PGN_2035 (putative M16B subfamily peptidase) and PGN_2065 (Lys- and Rgp-gingipain domain protein).

PGN_0022 encodes PorU, a cysteine proteinase and integral component of the T9SS of P. gingivalis. Considering the critical role of the T9SS in Bacteroides physiology, not least the delivery of functional membrane-associated gingipains, it is perhaps not surprising that the PorU protease has been identified as conditionally essential.

PGN_1416 (pepK) encodes a cysteine proteinase, likely LPS-anchored, that appears to function as an Rgp-activated Lys-specific serine endopeptidase [63] whose translocation is reliant upon an intact T9SS [64]. The reason for the essentiality of PepK, predominantly found in the Bacteroidetes, remains elusive. However, we have shown that the specific proteolytic activity of PGN_1416, an outer membrane-associated surface protein [63], is clearly required for epithelial colonization and abscess formation [14].

Dipeptidyl peptidase IV (DppIV) activity has been shown to play an important role in P. gingivalis virulence [65–68]. The 99% identical (www.kegg.jp) dppIV serine protease encoding W83 orthologue of PGN_1469 targets a broad spectrum of X-Pro and X-Ala dipeptides at peptide N-termini and has been shown to contribute to subcutaneous murine lesion formation and lethality [65, 66, 69]. Interestingly, P. gingivalis-derived DppIV has been hypothesized to cleave RANTES into a less potent innate cell chemokine that may actually antagonize full length RANTES, and suggested to processes the CXC12 (SDF1) into a less potent lymphokine [65]. In keeping, neutrophil infiltration into P. gingivalis-induced murine lesions is reduced with dppIV mutant bacteria comparted to wild type cells [68]. Further, DppIV may contribute to periodontal connective tissue destruction through an ability to augment endogenous MMP-1,−2, −8 and −9 and subsequent degradation of resultant gelatins [67]. DppIV also contributes to the adhesion of P. gingivalis to fibronectin [66, 68], which may assist in bacterial colonization, while concomitantly inhibiting fibroblast binding to fibronectin [68], an important event in the healing process. Thus, the identification of PGN_1469 as a common conditionally essential gene is in keeping with the extant literature.

Similar to the prior discussion of gingipains and the rgpB gene, detailed physiological roles of the gingipains are extraordinarily well documented. PGN_1728-encodes the Lys-X-specific gingipain, Kgp. Kgp-mediated proteolysis of provides essential nutrients to asaccharolytic bacteria, but also functions to dysregulate immune recognition and response systems through degradation of ECM components, immunoglobulins, and cytokines [70]. Specific inhibition of Kgp activity in a mouse model significantly reduced virulence, suggesting Kgp may contribute more to P. gingivalis-associated pathogenicity than any other protease [71].

P. gingivalis proteases essential to a specific disease-relevant condition (n = 6)

It is important to consider genes that are conditionally essential to a unique condition because it is likely more straightforward to dissect their pathogenic importance, using gain and loss of function approaches, than for those genes essential in multiple conditions. P. gingivalis genes identified as critical to fitness in a single experimental environment were PGN_0335 (PG_0232; PGTDC60_0510), encoding a M14 family zinc carboxypeptidase domain protein essential for CSE survival [28]; PGN_0508 (PG_1605; PGTDC60_0696), an aminopeptdase C-encoding gene essential for epithelial colonization [14]; PGN_0952 (PG_1060; PGTDC60_1149), PGN_1914 (orthologues also annotated as PG_1060; PGTDC60_1149) and PGN_1970 (PG_2024; PGTDC60_0300) encoding, two serine peptidases and the arginine-specific cysteine proteinase A gene, respectively, and essential for black pigmentation [27]; and PGN_1434 (PG_0537; PGTDC60_1655), which encodes a PepD-like aminoacyl-histidine dipeptidase, required for murine abscess formation [14].

Other than rgpA, the mechanisms underlying conditional essentiality are not immediatelyobvious. However, in silico homology searches, primarily using www.kegg.jp, do present some clues. PGN_0335, for example, is unexamined in P. gingivalis, but orthologues of this M14-family peptidase represent T9SS C-terminal target domain-containing proteins in several Aquimarina species. PGN_0508 exhibits high homology to bleomycin hydrolase [EC:3.4.22.40], whose significance has been discussed above in the context of PGN_1777, an essential gene in multiple environments. PGN_0952 and PGN_1914 each encode a ctpA-related carboxyl-terminal processing protease. While not investigated in P. gingivalis specifically, CtpA proteins, which are not common in bacteria, are known to function in transmembrane protein transport and maturation in some Gram-negative species [72, 73]. Clearly, this may well be related to their yet be determined essential role in the sequestration of hemin at the cell surface, leading to black pigmentation, and the acquisition of iron.

The emergence of rgpA as a gene required for hemin acquisition means that all three gingipain genes have been identified as conditionally essential. RgpA and Kgp have both been shown to bind heme, porphyrins and metalloporphryins [74]. Additionally, RgpA is known to degrade transferrin, hemoglobin and hemopexin [75], demonstrating its essential role in heme acquisition. Importantly, RgpA is known to cleave pro-Kgp, leading to Kgp activation, and is involved in proteolytic processing of many cell-associated proteins. Thus RgpA plays direct roles in iron and heme recruitment but additional indirect functions in protein maturation.

Concluding Remarks and Future Perspectives

To improve our understanding of the biological relevance of TnSeq-identified essential protease genes, a number of future initiatives are envisaged, as noted in the Outstanding Questions section. Most obviously, there is a need to develop libraries in a variety of P. gingivalis strains as, currently, the data are ATCC 33277-centric. Secondly, there is a need to generate prioritized individual mutants for phenotypic characterization. Such an approach has been employed to confirm the reduced fitness of genes deemed requisite by TnSeq for epithelial colonization [14], survival of cigarette smoke-induced stress [28], murine abscess formation [14] and black pigmentation [12]. It should be noted that the deletion of single putatively essential gene can [12, 14, 27, 28], but does not necessarily, result in a definitive, growth-compromised phenotype per se. However, reduced fitness, under the relevant conditions, is often noted upon competition of the mutant of interest with the parental bacterial strain [14, 28]. When trying to gain a deeper understanding of essential genes in situ, there are emergent issues to consider. With complex datasets, it may be more efficient to test a combination of mutants in the condition of interest. Perhaps more importantly, it should be remembered that periodontal biofilms are highly intricate structures and contain spatially-oriented and temporally shifting subsets of interacting bacteria among the multitude of species comprising the overall plaque microbiome [76, 77]. Therefore, it will be important to ascertain if non-P. gingivalis bacteria may contribute compensatory biomolecules. Finally, the design of new transposon libraries is likely to be abetted by the recently launched Transposon Registry, which aims to catalogue and assign a searchable transposon number to all prokaryotic transposable elements [78]. It must be envisaged that theTransposon Registry can be utilized alongside the growing Database of Essential Genes (www.essentialgene.org).

Meantime, there are multiple obvious strengths of transposon sequencing technology, in the context of the identifying absolute and essential protease-related genes of the model periodontal pathogen P. gingivalis. They include (a) high genome saturation of the 33277 library; (b) the generation of similar libraries by several independent groups; (c) the deposition of the relevant information in the expanding Database of Essential Genes; (d) library exploitation in multiple disease-relevant conditions; and, in particular, (e) the identification of a stress-related protein catabolism gene set common to multiple environments that both confirms the significance of established virulence factors while identifying previously unheralded pathogenesis genes. Perhaps the best examples are the two genes common to 80% of the conditions tested to date, PGN_637,encoding HtrA, and PGN_1777, encoding a bleomycin hydrolase. These common stress-related genes may represent particularly attractive therapeutic targets for the control of P. gingivalis infections – an established oral pathogen with growing systemic disease relevance.