Periodontal diseases are developed because of colonization of the subgingival area by multiple bacterial species (Page & Kornman, 1997). Socransky et al. (1998) have determined that three bacterial species, Porphyromonas gingivalis, Tannerella forsythia, and Treponema denticola, are mostly responsible for the development and advancement of periodontitis, and proposed to call them the “red complex.” Among these species, P. gingivalis, a Gram-negative anaerobic bacterium forming characteristic black-pigmented colonies on blood agar, has been extensively studied for its pathogenicity (Gibbons & Macdonald, 1960; Macdonald & Gibbons, 1962; Macdonald, Gibbons & Socransky, 1960), and accumulated evidence indicates its critical role in periodontitis (Lamont & Jenkinson, 1998; Socransky & Haffajee, 2002). Furthermore, although the proportion of P. gingivalis in the periodontal biofilm is low, it could lead to dysbiosis at the periodontal site, which prompted Hajishengallis et al. (2011) to call P. gingivalis a keystone pathogen. Still, it is well known that P. gingivalis can also be detected in healthy people (Amano et al., 2004; Griffen et al., 1998; Haffajee et al., 1998; Teanpaisan et al., 1996; Ximenez-Fyvie, Haffajee & Socransky, 2000), suggesting that its presence may not necessarily cause periodontitis. This discrepancy is suggested to be attributed to heterogenic virulence of P. gingivalis (Griffen et al., 1999; Igboin, Griffen & Leys, 2009; Tribble, Kerr & Wang, 2013), which shows a high degree of genetic clonal diversity (Enersen, 2011).
Although P. gingivalis expresses a number of potential virulence factors (How, Song & Chan, 2016), fimbriae, filamentous proteinaceous appendages on the bacterial surface, are one of the most important because they play a pivotal role in P. gingivalis colonization through association with other bacteria and host tissues (Hospenthal, Costa & Waksman, 2017; Lamont & Jenkinson, 2000). P. gingivalis generally expresses two distinct types of fimbriae: FimA and Mfa1 (Yoshimura et al., 2009). FimA fimbriae are primarily composed of polymers of the FimA protein encoded by the fimA gene (Dickinson et al., 1988; Yoshimura et al., 1984), whereas Mfa1 fimbriae are mostly composed of the Mfa1 protein encoded by the mfa1 gene (Hamada et al., 1996). In addition, several minor accessory components are incorporated into the respective fimbriae (Hasegawa et al., 2013; Nishiyama et al., 2007).
Based on fimA sequence variability, the gene is classified into six genotypes (I, Ib, II, III, IV, and V) (Amano et al., 2004; Nakagawa et al., 2002b), and the encoded proteins exhibited distinct antigenicity, with the exception of subtypes I and Ib (Nagano et al., 2013; Nakagawa et al., 2002b). Several studies indicate that strains with fimA genotype II are the most prevalent in patients with periodontitis, whereas those with genotype I are predominantly detected in healthy individuals (Amano et al., 2004; Enersen, Nakano & Amano, 2013; Kuboniwa, Inaba & Amano, 2010; Missailidis et al., 2004; Miura et al., 2005), indicating that the genotype-II P. gingivalis may have higher pathogenicity compared with genotype-I bacteria. Furthermore, genotype-II strains showed higher adhesion and invasion ability in human epithelial cells (Nakagawa et al., 2002a) and in a mouse abscess model (Nakano et al., 2004). However, other reports indicated that fimA genotypes were not associated with adhesion to and invasion of host cells (Inaba et al., 2008; Umeda et al., 2006); moreover, there are studies showing that genotype-II strains had rather low rates of adhesion to and invasion of epithelial cells (Eick et al., 2002) and induced significantly less alveolar bone resorption in mice compared to genotype-I strains (Wang et al., 2009). Collectively, these data indicate that P. gingivalis pathogenicity cannot be defined based on the fimA genotype.
In contrast to fimA, there are few clinical data regarding mfa1 genotypes. Recently, we found that the mfa1 gene had at least two variants (Nagano et al., 2015), encoding proteins with molecular weights about 70 kDa (67 (Arai, Hamada & Umemoto, 2000; Hamada et al., 1996) or 75 kDa (Park et al., 2005)) and 53 kDa (Arai, Hamada & Umemoto, 2000; Nagano et al., 2015), hereafter called Mfa70 and Mfa53 (mfa70 and mfa53, respectively, for the genes). In this study, we investigated the distribution of P. gingivalis mfa1 as well as fimA genotypes in clinical specimens.
Materials and Methods
A total of 100 patients, who visited Aichi Gakuin University Dental Hospital at Nagoya, Japan, for periodontal treatment from September 2016 to March 2017, participated in this study. The study was approved by the institutional review board (Aichi Gakuin University, School of Dentistry, Ethics Committee, approval numbers 460 and 478), and written informed consent was obtained from all participants.
Clinical oral examination, and consolidation and maintenance treatments
Among the 100 participants, 81 could be examined for clinicopathological parameters of periodontitis at the first visit. Clinical oral examination was performed according to the guidelines published by The Japanese Society of Periodontology (2015). Probing pocket depth (PD) and bleeding on probing (BOP) analyzed in six sites per tooth (buccal-mesial, mid-buccal, buccal-distal, lingual-mesial, mid-lingual, and lingual-distal) for all remaining teeth. The PD and BOP values were utilized to calculate periodontal inflamed surface area (PISA) and periodontal epithelial surface area (PESA), which reflect the surface area of bleeding epithelium and total pocket epithelium (in mm2), respectively, using a free spreadsheet (downloaded from www.parsprototo.info) (Nesse et al., 2008, 2009). Consolidation and maintenance treatments mainly consisted of professional scaling and cleaning. Patients visited the hospital for consultation every 1–6 months.
Collection of subgingival plaques
Subgingival plaque samples were collected by a sterile hand scaler and transferred in either one ml of sterile reduced transfer fluid (RTF) consisting of 0.01% dithiothreitol in PBS, pH 7.4 or one ml of distilled water. The samples were immediately placed at 4 °C and analyzed within 4 h.
Isolation and identification of black-pigmented bacteria
The samples collected in RTF were thoroughly suspended, serially diluted, and aliquots spread on blood agar consisting of Brucella HK agar (Kyokuto Pharmaceutical Industrial Co., Ltd, Tokyo, Japan), 5% laked rabbit blood, and 100 μg/ml kanamycin, as anaerobic bacteria, including P. gingivalis, are typically kanamycin resistant (Jousimies-Somer et al., 2002). Plates were cultured at 37 °C under anaerobic conditions for a week, and the emerged black-pigmented colonies were streaked on fresh plates to ensure isolation of single clones, which were then subjected to species identification. For this, genomic DNA was purified using Wizard Genomic DNA Purification Kit (Promega Corporation, Madison, WI, USA) and analyzed for 16S rRNA-encoding genes by polymerase chain reaction (PCR) using primers (5′-GAAGAGTTTGATCMTGGCTCAGATTG-3′ and 5′-TACGGYTACCTTGTTACGACTTCAC-3′) slightly modified from the universal primers 27F and 1492R (Frank et al., 2008). PCR products were subjected to DNA sequencing by a dye-terminator method and sequencing reads were analyzed by the BLAST search (https://blast.ncbi.nlm.nih.gov/). Bacterial species were identified if samples showed the lowest expectation (E) value (i.e., the highest probability) in the list of BLAST results. Most of E values were 0, whereas the highest was 3 × 10−66, i.e., were sufficiently low to identify bacterial species.
Genotyping of fimA and mfa1
FimA genotypes were determined by PCR, sequencing, and BLAST analysis. Plaque samples in RTF or water were directly used as PCR templates. Primers for PCR (5′-AGCTTGTAACAAAGACAACGAGGCAG-3′ and 5′-GAGAATGAATACGGGGAGTGGAGCG-3′) were designed for common fimA regions based on fimA sequencing data for 84 P. gingivalis strains (Nagano et al., 2013). PCR-amplified fragments of a predicted size (around 1.2 kb) were sequenced by the dye-terminator method and the fimA genotype was determined by BLAST analysis.
Mfa1 genotypes were determined by PCR using two primer sets (5′-GAGCATTGCTCTCATTGGGCTTTG-3′ and 5′-CATCAGAAAAGGCAGCGTAAGCTG-3′, and 5′-GAGCATTGCTCTCATTGGGCTTTG-3′ and 5′-TTAGGTATTGGCGACGTTCTCCTTG-3′), which yielded mfa53 and mfa70 fragments of 410 and 830 bp, respectively.
The data were expressed as the mean ± SEM. Differences between groups were analyzed by the nonparametric Kruskal–Wallis H test, and were considered statistically significant at P < 0.01. The Chi square test is used to determine if there is a relationship in the genotype distribution (P < 0.01).
Isolation of P. gingivalis
The first 73 dental plaque samples were collected in RTF to isolate P. gingivalis by a culture method. Although black-pigmented colonies were obtained from the majority of samples, 16S rDNA sequencing analysis showed that they were mostly formed by Prevotella species, and the isolation rate of P. gingivalis was only 5.5%. Therefore, we decided to examine fimbrial genotypes by direct PCR; in addition, the collection solution was changed to water, did not affect experimental results, but slightly improved detectability.
Distribution of fimA and mfa1 genotypes
The distribution of fimbrial fimA and mfa1 genotypes is summarized in Table 1. Among the 100 samples, both fimbrial types were detected in 63 and a single type in 15 samples, whereas 22 had no fimbrial genes. FimA genotype II was the most prevalent and detected in 33 of the 63 samples positive for both fimbrial genes (52.4%), followed by genotypes Ib and I detected in 27.0% and 11.1% samples, respectively, whereas the frequency of the other fimA genotypes was low. Although there was no statistically significant difference in combination of the fimA and mfa1 genotypes, the following tendency was observed. The mfa53 and mfa70 genotypes were detected at comparable frequencies (44.4% and 55.6%, respectively) and each of them showed almost the same frequency in samples positive for fimA genotypes I, Ib, and IV. However, the prevalence of mfa70 was 1.75 times higher than that of mfa53 in genotype-II positive samples, whereas only mfa53 was detected in the two genotype III-positive samples, and no mfa1 genes were found in genotype V-positive samples.
|I + Ib||12||12||(4)||24 (38.1)|
|Total (%)||28 (44.4)||35 (55.6)||−||63 (100)|
Genes encoding both fimbrial types were determined in 63 of 100 samples. Samples marked “undetermined” were not included in the total numbers.
Relationship between clinical parameters and fimbrial genotypes
We also examined the association of the fimbrial genotypes with clinical characteristics of periodontitis (maximal and mean PD values, and BOP, PISA, and PESA values) (Table 2). However, no statistically significant differences in periodontitis severity were observed depending on the fimbrial genotypes.
|fimA||mfa1||Females (n)||Males (n)||Age (years)||PD (mm)||BOP (%)||PISA (mm2)||PESA (mm2)|
|Untyped||22||9||60.5 ± 2.5||3.22 ± 0.16||8.94 ± 0.51||44.2 ± 6.7||703 ± 135||1,640 ± 114|
|I + Ib||53||5||3||51.5 ± 4.1||2.91 ± 0.16||7.88 ± 0.61||38.9 ± 10.7||462 ± 136||1,436 ± 168|
|70||9||2||59.1 ± 3.8||3.73 ± 0.34||9.40 ± 0.85||73.1 ± 13.0||1,016 ± 244||1,951 ± 191|
|II||53||6||4||62.1 ± 3.1||3.21 ± 0.37||8.60 ± 0.92||50.7 ± 18.8||868 ± 407||1,752 ± 296|
|70||8||10||59.6 ± 2.0||3.50 ± 0.17||9.67 ± 0.48||49.3 ± 9.2||889 ± 155||1,839 ± 156|
In this study, we first attempted to isolate P. gingivalis from dental plaque samples by a culture method, because we thought that analysis of chromosomal DNA purified from isolated bacterial clones by PCR would provide unequivocal fimbrial genotyping results. However, P. gingivalis was rarely isolated by the culture method. On the other hand, direct PCR detected either fimA or mfa1 in 78% samples, indicating that P. gingivalis was present with high frequency in patients receiving periodontal maintenance therapy, although its proportion among dental plaque bacteria was low.
FimA genotypes have been determined by PCR using genotype-specific primers (Amano et al., 2004; Nakagawa et al., 2002b); in addition, restriction enzyme digestion is used to discriminate genotypes I and Ib (Nakagawa et al., 2002b), which, however, may not be necessary for the analysis of the entire fimA gene, because genotypes I and Ib cannot be clearly discriminated (Fig. S1 and Nagano et al., 2013). Furthermore, immunological analysis did not detect any differences in antigenicity between FimA I and Ib fimbriae (Nagano et al., 2013; Nakagawa et al., 2002b). Therefore, we do not discuss differences between genotypes I and Ib here. In contrast, genotype II (and possibly IV) could be further divided into two or more groups (Fig. S1 and Nagano et al., 2013). Regarding mfa1, two genotypes are currently known: mfa53 and mfa70. However, in 12% of fimA-positive specimens, mfa1 was not detected, suggesting that existence of additional mfa1 genotypes. Therefore, reclassification of fimA and mfa1 genotypes would be needed in the future.
In this study, we observed that in samples positive for fimA genotype II, mfa70 genotype was detected 1.75 times more frequently compared to mfa53, and in the previous study, where we analyzed 84 P. gingivalis strains stocked in our laboratory, the frequency of mfa70 detection among fimA II strains was 3.6 times higher than that of mfa53 (Nagano et al., 2015). These findings indicate that mfa70 is the major mfa1 genotype in P. gingivalis strains positive for fimA II. On the other hand, in this study, the two mfa1 genotypes had almost the same detection rate in samples positive for fimA I (including I and Ib), whereas our previous results indicate that mfa70 detection frequency was 2.3 times higher compared to that of mfa53 in fimA-I strains (Nagano et al., 2015). Among fimA IV-positive samples, the detection rate of each mfa1 genotype was the same in this study, and in our previous study, mfa53 and mfa70 genotypes were detected in four and two samples, respectively (Nagano et al., 2015). Taken together, these results suggest that strains with fimA genotypes I and IV tend to have either the same frequency of mfa1 genotypes or slightly higher prevalence of mfa70. Although we found only two genotype III-positive samples in this study, both had mfa53, which was consistent with our earlier findings that 12 out of 13 genotype-III strains carried mfa53 (Nagano et al., 2015). In this study, mfa1 was not detected in genotype V-positive samples, which, however, were all found mfa53-positive in our previous study (Nagano et al., 2015). These results indicate that genotype-III and -V strains almost exclusively carry mfa53. Thus, there is a tendency for correlation between the two fimbrial types in P. gingivalis: fimA II strains preferably carry mfa70, whereas fimA I/IV strains may have both mfa1 genotypes in equal proportions, and fimA III/V strains mostly carry mfa53. However, the reason for such correlations is unknown because there is a wide distance between the two genetic loci. There are polymorphisms in other P. gingivalis genes (Enersen, 2011). Thus, the ragA gene, which encodes a major outer membrane protein and is located downstream of the mfa1 gene, exhibits four genetic variants (Hall et al., 2005; Liu et al., 2013); besides, genetic variability has also been reported for capsular antigens (Laine, Appelmelk & Van Winkelhoff, 1996; Laine, Appelmelk & Van Winkelhoff, 1997). In future studies, it will be interesting to find out whether these genetic polymorphisms are correlated with those in the fimA and mfa1 genes.
We did not observe statistically significant associations between clinical parameters of periodontitis and the distribution of fimbrial genotypes. However, there was a time lapse between periodontal examination and sample collection, and it was possible that P. gingivalis clones were replaced during that interval; still, the chances for such clonal change may be low because it is reported that P. gingivalis showed high clonal stability (Valenza et al., 2009; Van Winkelhoff, Rijnsburger & Van Der Velden, 2008). In addition, we would like to note that most similar studies have the same problems which are inherent to this type of clinical research, because generally the treatment for chronic periodontitis takes a long time. Therefore, it is necessary to develop an appropriate study design for examining the relationship between bacterial genotypes and clinical symptoms of periodontitis.
There was a tendency in the distribution of fimbrial genotypes fimA (I–V) and mfa1 (mfa53 and mfa70) among patients with periodontitis. However, we did not observe any associations between fimbrial genotypes and the severity of the disease.
Fig. S1. The phylogenetic tree of the fimA gene.
This is a modification of the phylogenetic tree from our previous study (Nagano et al., 2013); fimA genotypes and the number of strains analyzed (in parenthesis) are indicated at each node. The areas of genotypes I and Ib were not circled because the division between these genotypes is unclear. On the other hand, genotypes II and IV are obviously divided into two or more groups.