In silico analysis as a strategy to identify candidate epitopes with human IgG reactivity to study Porphyromonas gingivalis virulence factors

Chronic periodontitis is a multifactorial and polymicrobial disease, which may negatively influence systemic diseases (Hajishengallis 2015). Its pathogenesis is related to host immune inflammatory factors and to a synergistic and dysbiotic oral microbiome (Hajishengallis et al. 2012; Hajishengallis 2014; Hajishengallis and Lamont 2014). In light of the diversity of the human oral microbiome (Proctor et al. 2018), immunoinformatics brings tools that provide faster analysis of virulence factors, considering the polymicrobial character of chronic periodontitis, and contribute to understanding the interaction between the oral microbiome and the host.

Porphyromonas gingivalis, a keystone pathogen related to periodontal dysbiosis (Hajishengallis 2014), produces various virulence factors that can act as immunogenic molecules. The host response to infection may be better understood by progressively investigating the immunogenicity of P. gingivalis virulence factors (Trindade et al. 2008, 2012a, b, 2013; Olczak et al. 2010; Bengtsson et al. 2015; Carvalho-Filho et al. 2016; Tada et al. 2016; Dashper et al. 2017).

Gingipains (cysteine proteases) are the main proteases of P. gingivalis (Guo et al. 2010), one of their main functions being heme acquisition (Smalley and Olczak 2017). These proteases are also immunogenic proteins, contributing to the pathogen’s ability to induce chronic periodontitis, since they can elicit a humoral immune response in humans, inducing higher serum levels of specific IgG in individuals with chronic periodontitis (O’Brien-Simpson et al. 2000; Inagaki et al. 2003; Nguyen et al. 2004).

Lys-gingipain (Kgp) is considered a major virulence factor of P. gingivalis (de Diego et al. 2014) and is also involved in the bacteria-host interaction through the production of cytokines, such as IL-17A and INF-γ (Bittner-Eddy et al. 2013). Kgp (1723 aa/187 kDa) is cleaved into four chains: the catalytic subunit and three adhesion domains (UniProt B2RLK2).

Although not as well studied as Kgp, neuraminidase (sialidase) of P. gingivalis is secreted by the microorganism for neuraminic acid (sialic acid) acquisition from sialoglycoconjugates of the host. Sialic acids are incorporated into the microorganism structures, thus mimicking the host cell and confounding the immune response, mainly under stressful microenvironmental conditions (Li et al. 2012; Xu et al. 2017). Sialidase deficiency in P. gingivalis increases sensitivity to hydrogen peroxide, decreases resistance to the action of complement and reduces virulence after subcutaneous injection in mice, likely by influencing capsule biosynthesis (Moncla et al. 1990; Aruni et al. 2011; Li et al. 2012; Xu et al. 2017).

Moreover, the sialidase activity may be involved in the production, maturation and secretion of gingipains and other virulence factors of P. gingivalis, probably due to their sialylation (glycosylation) (Aruni et al. 2011; Xu et al. 2017). Importantly, inhibitors of these two virulence factors studied herein could be promising for use in the treatment of chronic periodontitis and associated systemic diseases (Cueno et al. 2014; Olsen and Potempa 2014; Inaba et al. 2016; Xu et al. 2017).

The present study aimed to use in silico analysis as a strategy to identify potential immunogenic peptides from those P. gingivalis relevant proteins. Selected epitopes were evaluated concerning their immunoreactivity based on the IgG-mediated host response in order to contribute to immunogenicity studies of this keystone pathogen.

It is known that in silico models are used for understanding biological systems as well as to select, to complement, and to inspire the required laboratory experiments (Kollmann and Sourjik 2007; Setty 2014; Brodland 2015). In this context, immunoinformatics brings advances in immunology and can contribute to understanding the immune response (Lefranc 2014; Qiu et al. 2018). In the present study, the in silico analysis enabled the prediction and selection of immunoreactive peptides of P. gingivalis before being synthesized. Two virulence factors of P. gingivalis were analyzed: Kgp, which is widely studied, and neuraminidase, which is still being evaluated in a few studies.

The same HLA alleles were used to predict immunogenic peptides of virulence factors. The HLA alleles tested with neuraminidase peptides were also tested with Kgp peptides and the same pools of sera were used for the immunoreactivity test of the synthetic peptides. However, all tested Kgp peptides presented immunoreactivity to IgG, whereas neuraminidase peptides presented low immunoreactivity to IgG.

Besides that, Kgp12 distinguished between CP and WP pools. The coefficient obtained from the fraction CP pool/WP pool indicated that the absorbance of the CP pool was twice as high as that of the WP pool. Kgp12 was tested in another study to detect specific serum IgG of 71 subjects, and it distinguished the ones with gingivitis from those with chronic periodontitis (Cardoso, unpublished data, 2017).

All tested Kgp peptides presented immunoreactivity to IgG, whereas neuraminidase peptides presented low immunoreactivity to IgG and no neuraminidase peptide presented a coefficient between CP and WP sera that could differentiate subjects with CP; thus the low IgG reactivity of those neuraminidase peptides could be suggested. However, further studies need to be conducted to better define this characteristic of neuraminidase peptides tested herein.

Porphyromonas gingivalis extract and gingipains have the capability to induce immunogenicity since they are recognized by IgG from serum of individuals with chronic periodontitis (O’Brien-Simpson et al. 2000; Inagaki et al. 2003; Nguyen et al. 2004; Franca et al. 2007; Trindade et al. 2008, 2012a). The IgG-mediated response in humans to neuraminidase (sialidase) of P. gingivalis remains, to our knowledge, unknown. However, in the present study, based on the in silico analysis of tested peptides, under the conditions studied, the IgG reactivity of Kgp could be reaffirmed and the low reactivity of neuraminidase could be speculated.

Besides being asaccharolytic, P. gingivalis probably does not use N-acetylneuraminic acid (Neu5Ac), the most studied sialic acid, as a nutrient. This deduction was made because in the culture supplemented with Neu5Ac, it did not interfere with P. gingivalis planktonic growth, and the inactivation of the neuraminidase gene did not influence its growth (Li et al. 2012). Additionally, not only P. gingivalis but also other periodontopathogens, Tannerella forsythia and Treponema denticola, perform glycosylation of their proteins to evade the immune response, persist in the host and cause periodontal destruction (Stafford et al. 2012; Kurniyati et al. 2013; Settem et al. 2013). This is one reason why the low immunogenicity of its neuraminidase should be a favorable characteristic to P. gingivalis, since this virulence factor may be used to evade immune system inflammatory responses.

We also need to consider that, despite being primarily an extracellular pathogen, P. gingivalis has the ability to internalize non-phagocytic cells of the host (Nakagawa et al. 2006; Olsen and Progulske-Fox 2015) and to survive in a macrophage (Gmiterek et al. 2016; Yang et al. 2018) in order to evade the immune system. As in the case of other pathogens (Shtyrya et al. 2009; Banerjee et al. 2010; Freire-de-Lima et al. 2015), sialidase may be used by P. gingivalis as one of the cellular invasion mechanisms.

The ATCC 33277 strain of P. gingivalis (NCBI Taxonomy ID: 431947) is used for pathophysiological characterization of the microorganism, and it is considered a less virulent strain (Naito et al. 2008). Its genome was published by Naito et al. (2008), and the sequences of its proteins and its peptide epitopes have been deposited in online public databases. Epitopes from other antigens of this strain had already been tested and deposited at the IEDB. For example, Kgp peptide DKYFLAIGNCC (Epitope detailed search: Epitope ID 190728) induced IL-17A e IFN-γ production in Mus musculus C57BL/6 (Bittner-Eddy et al. 2013). On the other hand, in the IEDB, epitopes related to neuraminidase (sialidase) of P. gingivalis were not identified until our deposition. In the present study, eight neuraminidase peptides of P. gingivalis were originally deposited in the IEDB, as well as nine novel Kgp peptides, and they are available for public visualization and for use in other assays.

A limitation of the study is that binding to MHC is necessary but not sufficient for epitope recognition by T cells. However, the use of in silico analysis for prediction of immunogenic peptides allowed their selection for chemical synthesis and immunogenicity tests.

Porphyromonas gingivalis manipulates the host immune response using the diversity of its virulence factors (Hajishengallis and Lamont 2014; Dashper et al. 2017). The peptides tested herein could be useful for application in immunogenic studies of virulence factors, since it is more economical to use synthetic peptides than to use recombinant proteins. In addition, specific peptides decrease the risk of cross reactivity in the opposite way of the total extract of the microorganism.

The strategy used herein reduced the total analysis cost and the research expended time in the search for promising results. It could be applicable to a global understanding of P. gingivalis pathogenicity (as well as other periodontal pathogens) in an efficient and rapid manner using chimeric peptides as well. Thus, in silico tools provide comprehensive facilities for designing in vitro and in vivo immunology experiments. This can be a useful strategy in the study of the etiology and pathogenesis of human diseases, especially periodontitis.

In conclusion, here we report the identification of P. gingivalis epitopes. In silico analysis represented a viable strategy to obtain candidate epitopes with human IgG reactivity to study P. gingivalis virulence factors, because the IgG reactivity of Kgp could be reaffirmed and the low reactivity of neuraminidase could be suggested. Kgp12, Kgp17 and Kgp18 peptides were selected for subsequent assays. Besides that, the present study has provided immunogenic peptide sequences from P. gingivalis virulence factors, which may be tested by several different assays in order to contribute to the understanding of the host response to infection and to differentiate subjects with chronic periodontitis.