The effects of stress hormones on growth of selected periodontitis related bacteria.

The focus of this study was to examine in vitro the effects of stress hormones (catecholamines: epinephrine, norepinephrine, dopamine and hydrocortisone: cortisol) on the growth of four anaerobic species of periodontitis-related bacteria (Fusobacterium nucleatum, Porphyromonas gingivalis, Prevotella intermedia and Tannerella forsythia) and one facultative anaerobic species (Eikenella corrodens). Bacterial growth was determined by two different methods: fluorescence in situ hybridization (FISH), and the viable count by culture method. To simulate stress, each single strain was grown in a special growth medium with three different concentrations of each hormone, using an anaerobic chamber at 37 °C. Growth of F. nucleatum increased in the presence of all stress hormones. Growth of P. gingivalis was not significantly influenced by any hormone. Growth of P. intermedia and E. corrodens was inhibited by almost all stress hormones tested. Both methods of analysis revealed that the highest concentrations of norepinephrine and cortisol increased the growth of T. forsythia. Different hormones have a different effect on the growth of periodontitis-related bacteria in vitro. It appears that bacterial viability is more strongly influenced than is bacterial metabolic activity. The growth of F. nucleatum particularly and partially of T. forsythia is increased by several stress hormones and may have an additional negative impact on periodontal disease.

The periodontopathogenic bacterium Eikenella corrodens produces an autoinducer-2-inactivating enzyme.

Eikenella corrodens produces autoinducer-2 (AI-2) in the mid log phase, and AI-2 activity decreases dramatically during the stationary phase. We investigated the mechanism underlying this decrease in AI-2 activity. To analyze the mechanism, we extracted and purified AI-2 from the supernatant of mid-log-phase culture. Simultaneously, the stationary-phase culture supernatant was fractionated by ammonium sulfate precipitation. On incubating purified AI-2 and 4-hydroxy-5-methyl-3(2H)-furanone (MHF) with each fraction, the 30% fraction decreased both AI-2 and MHF activities. The data suggest that AI-2 and MHF were rendered inactive in the same manner. Heat and/or trypsin treatment of the 30% fraction did not completely arrest AI-2 inactivation, suggesting that partially heat-stable proteins are involved in AI-2 inactivation. We observed that an enzyme converted MHF to another form. This suggests that E. corrodens produces an AI-2 inactivating enzyme, and that AI-2 can be degraded or modified by it.

Caries-related plaque microcosm biofilms developed in microplates.

In vivo dental plaque biofilms consist of complex communities of oral bacteria that are a challenge to replicate in vitro. The aim of this investigation was to establish human dental plaque microcosms in microplates to reflect conditions that are relevant to dental caries. Microcosm plaque biofilms were initiated from the saliva of two different donors, grown for up to 10 days in 24-welled microplates on Thermanox coverslips in various types of artificial saliva with and without sucrose, which were replaced daily. Microbiota composition of 40 species associated with oral health and dental caries was monitored in the plaques using Checkerboard DNA-DNA hybridization analysis. pH was measured as an indicator of cariogenic potential. The composition of the saliva inocula was different, and yielded plaque microcosms with different composition and growth responses to sucrose. Artificial saliva type and presence of sucrose, and the resulting growth and pH conditions, modified the growth of individual species and hence the ecological profile of the microplate plaques during development. Complex population shifts were observed during development, and older plaques comprised predominantly facultative anaerobic species. Sucrose supplementation limited the decline of Streptococci over time but did not increase the abundance of mutans Streptococci. Sucrose at 0.15% increased levels of caries-associated species including Lactobacillus fermentum, Lactobacillus acidophilus and Actinomyces gerensceriae; these were further increased with sucrose at 0.5%, in addition to Actinomyces israelii, Rothia dentocariosa and Capnocytophaga gingivalis. The microplate plaques demonstrated complex community dynamics that appeared to reflect the maturation of natural plaques, and sucrose induced a cariogenic plaque composition and pH.

Plasmid-mediated genomic recombination at the pilin gene locus enhances the N-acetyl-D-galactosamine-specific haemagglutination activity and the growth rate of Eikenella corrodens.

Eikenella corrodens belongs to a group of periodontopathogenic bacteria and forms unique corroding colonies on solid medium due to twitching motility. It is believed that an N-acetyl-D-galactosamine (GalNAc)-specific lectin on the cell surface contributes significantly to its pathogenicity and can be estimated by its haemagglutination (HA) activity. Recently, a plasmid, pMU1, from strain 1073 has been found; this plasmid affects pilus formation and colony morphology. To identify the gene involved in these phenomena, ORF 4 and ORFs 5-6 on pMU1 were separately subcloned into a shuttle vector, and the resultant plasmids were introduced into E. corrodens 23834. Transformants with the ORF 4 gene, which is identified to be a homologous gene of the type IV pilin gene-specific recombinase, lost their pilus structure and formed non-corroding colonies on a solid medium, whereas transformants with ORFs 5-6 exhibited the same phenotype as the host strain 23834. Southern analysis showed that the introduction of the ORF 4 gene into strain 23834 resulted in genomic recombination at the type IV pilin gene locus. The hybridization pattern of these transformants was similar to that of strain 1073. These results suggest that ORF 4 on pMU1 encodes a site-specific recombinase and causes genomic recombination of the type IV pilin gene locus. Furthermore, the introduction of ORF 4 into strain 23834 increased GalNAc-specific HA activity to a level equivalent to that of strain 1073. Although the morphological colony changes and loss of pilus structure are also observed in phase variation, genomic recombination of the type IV pilin gene locus did not occur in these variants. Moreover, an increase was not observed in the GalNAc-specific HA activity of these variants. These results suggested that the loss of pilus structure, the morphological change in colonies and the increase in HA activity due to plasmid pMU1 might be caused by a mechanism that differs from phase variation, such as a genomic recombination of the type IV pilin gene locus.

Involvement of N-acetyl-D-galactosamine-specific lectin in biofilm formation by the periodontopathogenic bacterium, Eikenella corrodens.

Eikenella corrodens is known not only as one of the periodontopathogenic bacteria but also as a pathogen associated with many infectious diseases of humans. Dental plaque is a complex biofilm community comprised of many bacterial species. E. corrodens has a lectin on its cell surface that is thought to be involved in its pathogenicity. In this study, we found that E. corrodens forms a biofilm on a polystyrene surface. A biofilm was formed at the bottom of the wells in microtiter plates after 24 h. Microcolonies were observed as the amount of biofilm became larger. When anaerobic respiration was repressed due to nitrate limitation, the biofilm formed only at the air-water interface. Strain 1073 and HU, which have higher lectin activity, formed a biofilm more effectively than other strains. Biofilm formation was repressed by the addition of N-acetyl-D-galactosamine. These results suggest that the lectin on the surface of E. corrodens might be involved in biofilm formation.

[Eikenella corrodens infective endocarditis]. / Endocardite infecciosa causada por Eikenella corrodens.

The HACEK microorganisms (Haemophilus spp, Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens, and Kingella kingae) account for 3% of the cases of endocarditis. They have the following similar clinical and microbiological properties: are Gram-negative bacilli, more easily isolated in aerobic media; their cultures require prolonged incubation time for growing (mean, 3.3 days); and may be considered part of normal flora of upper respiratory tract and oropharynx. The following characteristics have been identified in endocarditis caused by the HACEK microorganisms: insidious clinical findings; difficult diagnosis due to the fastidious nature of the microorganisms; and negative cultures. The Eikenella corrodens endocarditis was first described in 1972. That microorganism continues to be a rare etiological agent. We report the case of a female patient with native valve, who had Eikenella corrodens infective endocarditis.

Endocardite infecciosa causada por Eikenella corrodens / Eikenella corrodens infective endocarditis

Os microorganismos do grupo HACEK (Haemophilus spp, Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens e Kingella kingae) são responsáveis por 3 por cento dos casos de endocardites. Eles apresentam propriedades clínicas e microbiológicas semelhantes entre si: são bacilos gram-negativos, isolados mais facilmente em meios aeróbicos, suas culturas necessitam de tempo prolongado de incubacão para crescimento (média 3,3 dias) e podem ser considerados como parte da flora normal do trato respiratório superior e da orofaringe1,2. Algumas características foram identificadas nas endocardites por esses agentes, como o quadro clínico insidioso , diagnóstico difícil pela natureza fastidiosa e culturas negativas3,4. A endocardite por Eikenella corrodens foi descrita pela primeira vez em 1972(5) e continua sendo um agente etiológico raro. Relatamos o caso de uma paciente com valva nativa que apresentou endocardite infecciosa causada por Eikenella corrodens.

Isolation and characterization of a plasmid DNA from periodontopathogenic bacterium, Eikenella corrodens 1073, which affects pilus formation and colony morphology.

Eikenella corrodens (Ec) is one of a group of periodontopathogenic bacteria. A plasmid DNA (8.7 kb) isolated from Ec 1073 was designated pMU1. Agarose gel electrophoresis and Southern analysis suggested that pMU1-like plasmids were carried in 2 Ec strains, including 1073, with higher hemagglutination (HA) activity than other strains. We determined the nucleotide sequence of this plasmid and identified 7 ORFs. A homology search revealed that 4 ORFs of pMU1 were homologous to ORFs in pJTPS1, found in a spontaneous avirulent mutant of the phytopathogenic bacterium, Ralstonia solanacearum. pJTPS1 is a putative hypovirulent plasmid, which is thought to control the virulence of R. solanacearum. We also found the ORF to be homologous to the recombinase specific to the type IV pilin gene. We introduced a part of pMU1 into the Ec 23834 strain, which has a pilus structure on its cell surface and forms corroding colonies on solid medium. No pilus structure was observed on the surface of transformants, most of which formed non-corroding colonies. When such transformants (or Ec 1073) were cured of pMU1 with acridine orange, they remained non-foliated and non-corroding. The results suggest that pMU1 might irreversibly affect pilus formation and colony morphology, and might be involved in the pathogenicity and virulence of Ec.

Microbial profile on metallic and ceramic bracket materials.

The placement of orthodontic appliances creates a favorable environment for the accumulation of a microbiota and food residues, which, in time, may cause caries or exacerbate any pre-existing periodontal disease. The purpose of the present study was to compare the total bacterial counts present on metallic and ceramic orthodontic brackets in order to clarify which bracket type has a higher plaque retaining capacity and to determine the levels of Streptococcus mutans and Lactobacillus spp on both types of brackets. Thirty-two metallic brackets and 24 ceramic brackets were collected from orthodontic patients at the day of debonding. Two brackets were collected from each patient; one from a maxillary central incisor and another from a maxillary second premolar. Sixteen patients who used metallic brackets and 12 patients who used ceramic brackets were sampled. Bacterial populations were studied using "checkerboard" DNA-DNA hybridization, which uses DNA probes to identify species in complex microbial samples. The significance of differences between groups was determined using the Mann-Whitney U-test. Results showed no significant differences between metallic and ceramic brackets with respect to the caries-inducing S mutans and L acidophilus spp counts. Mean counts of 8 of 35 additional species differed significantly between metallic and ceramic brackets with no obvious pattern favoring one bracket type over the other. This study showed higher mean counts of Treponema denticola, Actinobacillus actinomycetemcomitans, Fusobacterium nucleatum ss vincentii, Streptococcus anginosus, and Eubacterium nodatum on metallic brackets while higher counts of Eikenella corrodens, Campylobacter showae, and Selenomonas noxia were found on ceramic brackets.

Contamination of interdental brushes by periodontopathogens.

Experimental studies have shown that intraoral transmission of bacteria can occur. Of course, the question arises as to how this transmission may happen. In this study, the contamination of interdental brushes by periodontopathogens is examined and compared to the microbial load of the periodontal pockets. In ten untreated chronic periodontitis patients, four interdental sites were professionally brushed with one interdental brush per patient. Subsequently, samples from the depths of the pockets (of the specific interdental sites) were obtained with paper-points. The interdental brush samples and the samples of the subgingival plaque, obtained by the pooled paper-points, were processed for dark-field microscopy examination as well as anaerobic culturing. The results showed that, although significant differences could be found between the brushes and paper-points with direct microscopy, the culturing did not uncover many differences. On the contrary, the detection frequencies of specific bacterial species were almost the same between the two. The total anaerobic colony-forming units (CFU), P. gingivalis, F. nucleatum, and E. corrodens found on the brushes showed a significant correlation with the subgingival plaque samples (P<0.005). These results suggest that, in untreated situations, interdental brushes are contaminated relatively easily by putative periodontopathogens in numbers comparable to their presence in periodontal pockets. This contamination could be a factor in the intraoral spread of bacteria.

Salivary microbiota levels in relation to periodontal status, experience of caries and miswak use in Sudanese adults.

OBJECTIVES: The purpose of the present investigation was to assess the salivary levels of 25 oral bacteria in relation to periodontal status and experience of caries, and to compare the levels of these bacteria between habitual miswak and toothbrush users in adult Sudanese subjects. MATERIAL AND METHODS: The study subjects consisted of 56 individuals with age range 19-53 years (mean 35.2 years) and included 30 miswak and 26 toothbrush users. The periodontal status and presence of dental caries were assessed clinically. Whole saliva was collected from all subjects, and the levels of 25 selected bacterial species in saliva were assessed by the checkerboard DNA-DNA hybridization method using whole genomic DNA probes. RESULTS: A high percentage of the subjects had detectable levels (> or = 105 bacterial cells) of several bacterial species in saliva. Between 12% and 16% of the subjects showed high (> or = 106 cells) salivary levels of the periodontitis-associated bacteria A. actinomycetemcomitans, P. melaninogenica, P. intermedia, C. rectus and E. corrodens, whereas only two (3.6%) and four (7.1%) subjects had high levels of P. gingivalis and F. nucleatum, respectively. There were no significant differences in the levels of all or most bacterial species by age group, gender or periodontal status. Presence of > or = 105 L. acidophilus bacterial cells in saliva was associated with high caries scores (p = 0.02). There were significantly higher levels of A. actinomycetemcomitans, P. melaninogenica, C. rectus, P. micros, V. parvula, S. mutans, S. anginosus, A. israelii, C. sputigena, and C. gingivalis, and significantly lower levels of P. intermedia, F. nucleatum, S. sputigena, E. corrodens, L. acidophilus, S. sanguis, S. salivarius, S. oralis, and S. mitis in the miswak than in the toothbrush group. CONCLUSIONS: : The findings suggest that miswak may have a selective inhibitory effect on the level of certain bacteria in saliva, particularly several oral streptococci species. This is the first report that the checkerboard DNA-DNA hybridization method can be useful for assessing the levels of a wide range of bacterial taxa in saliva.

The characterization of a (nutritionally important) proline iminopeptidase from Eikenella corrodens.

Eikenella corrodens generates energy primarily through the oxidative deamination of specific amino acids, a process that is coupled to dissimilatory nitrate reduction to nitrite. Cell yields resulting from chemostat-growth of the organism in simple, chemically defined media containing varying amounts of proline confirm that this amino acid is a likely source of energy for E. corrodens in the oral environment. The importance of proline in ATP generation by the organism is reflected in molar growth yields, which showed that biomass production per mole of this amino acid was significantly higher than that for other amino acids. The organism was found to express, constitutively, the enzyme proline iminopeptidase, which releases proline from the N-terminus of small peptides. The enzyme was partially purified and characterized and found to exist in the cytoplasm as a 35 kDa monomer. Inhibition studies showed that the enzyme, although classified as a serine protease, also appears to require thiol groups for activity, a finding which is consistent with previous reports. The enzyme obeyed Michaelis-Menten kinetics and was found to have a Km value of 0.223 mM for the substrate proline-p-nitroanilide.

The localization of periodontal-disease-associated bacteria in human periodontal pockets.

Some Gram-negative anaerobes are associated with the incidence and progression of periodontal disease. In periodontal pockets, however, the localization of those bacteria is unknown. We investigated the localization of 5 bacterial species in human periodontal pockets. Fifteen teeth with a part of periodontal pockets from 10 adult periodontitis patients were obtained, and the localization of bacteria was examined immunohistochemically. Positive reactions with anti-Prevotella nigrescens antibody were located at the epithelium-associated plaque area in the middle pocket zones. In the middle and deep pocket zones, Fusobacterium nucleatum and Treponema denticola were especially localized in the unattached plaque area, but Eikenella corrodens was observed in the tooth-attached plaque area. Actinobacillus actinomycetemcomitans, detected in 2 of 15 samples examined, was found in the unattached plaque area, in the middle pocket zone. The present findings indicated that the 5 bacterial species examined localized at distinct regions in human periodontal pockets.

Antimicrobial effect of acidified nitrite on periodontal bacteria.

The antimicrobial agent nitric oxide (NO) is formed in the mouth and its concentration is directly related to salivary nitrite, which in turn is related to dietary nitrate intake. The aim of this study was to determine whether nitrite under acidic conditions will have an inhibitory effect, possibly occurring through NO production, on the periodontal disease pathogens Fusobacterium nucleatum, Eikenella corrodens and Porphyromonas gingivalis. Whereas the growth of these organisms was inhibited by a more acid pH, the addition of nitrite caused a marked, further dose-dependent reduction in bacterial numbers after exposure. The ability of these bacteria to recover from nitrite exposure was also affected by pH and nitrite concentration. At acidity levels below pH 5.0, low concentrations of nitrite (0.2 mM) caused effective complete killing of the periodontal bacteria. Addition of sodium thiocyanate did not increase the bacteriostatic or bacteriocidal activity of acidified nitrite against any of the 3 bacteria. These results demonstrate the possibility that nitrite in saliva, under appropriate conditions, may have an effect on the growth and survival of the bacteria implicated in periodontal disease.

Relationship of periodontopathic bacteria with early-onset periodontitis in Down's syndrome.

BACKGROUND: Down's syndrome (DS) patients often develop severe early-onset marginal periodontitis in early adulthood; however, there is little information available on the microbiology of DS periodontitis. METHODS: Subgingival plaque specimens were taken from 67 DS young adults and 41 age-matched systemically healthy individuals with mental disabilities (MD). The prevalence of 10 possible periodontopathic bacterial species, Actinobacillus actinomycetemcomitans, Porphyromonas gingivalis, Bacteroides forsythus, Treponema denticola, Prevotella intermedia, Prevotella nigrescens, Capnocytophaga ochracea, Capnocytophaga sputigena, Campylobacter rectus, and Eikenella corrodens, were investigated in their subgingival plaque samples using a polymerase chain reaction method. The detection of P. gingivalis fimA genotypes was also performed in P. gingivalis-positive samples. RESULTS: Although DS subjects generally develop an earlier and more extensive periodontal breakdown than those with MD, no significant differences were observed in the bacterial profiles. The profiles of subjects with periodontitis were significant in DS, but not in MD. The prevalence of P. gingivalis, B. forsythus, and P. intermedia were significant in the DS periodontitis group, compared to DS gingivitis group. Moreover, the occurrence of P. gingivalis with the type II fimA gene was significantly related to periodontitis in both DS and MD, with odds ratios of 6.32 and 12.03, respectively. CONCLUSIONS: These results suggest that early-onset periodontitis in DS is mainly due to the more susceptible host for the causative microbial agents including P. gingivalis with type II fimA.

"Checkerboard" assessments of periodontal microbiota and serum antibody responses: a case-control study.

BACKGROUND: We explored the association between subgingival microbial profiles and serum IgG responses to periodontal microbiota in relation to clinical periodontal status. METHODS: One hundred thirty-one (131) periodontitis patients aged 29 to 74 years (mean 51.8) were age- and gender-matched with 74 periodontally intact controls (range 26 to 77, mean 49.3). Smoking habits and health history were recorded and assessments of plaque, bleeding on probing, probing depth, and attachment level were performed at 6 sites per tooth on all present teeth, excluding third molars. Subgingival plaque samples were obtained from each tooth in one upper and one lower quadrant (maximum 14 samples/subject; 2,440 samples total) and analyzed with respect to 19 species by means of whole genomic DNA probes. Serum IgG antibodies against the same 19 species were assessed by an immunoassay. RESULTS: Cases displayed an average of 22.7 teeth, 20.3 sites with probing depth > or =6 mm, and 18.9 sites with attachment loss > or =6 mm. Corresponding figures for controls were 27.1, 0.1, and 1.0, respectively. Heavy smoking was 3 times more frequent among cases than controls (32.1% versus 9.6%). Higher levels of Porphyromonas gingivalis, Porphyromonas endodontalis, Prevotella intermedia, Prevotella nigrescens, Prevotella melaninogenica, Bacteroides forsythus, Fusobacterium nucleatum, Treponema denticola, Eubacterium nodatum, Peptostreptococcus micros, and Campylobacter rectus were found in cases and higher levels of Eikenella corrodens, Veillonella parvula, and Actinomyces naeslundii in controls. Cases displayed higher IgG levels against P. gingivalis and Actinobacillus actinomycetemcomitans, while controls displayed higher levels against F. nucleatum, T. denticola, E. nodatum, and Capnocytophaga ochracea. Positive correlations between bacterial colonization and antibody responses were identified for 9 species in controls. In cases, however, statistically significant correlations were observed for only 3 species out of which only one was positive (V. parvula). Both bacterial levels and antibody responses declined in ages over 55 years. A logistic regression employing selected elements of bacterial colonization and antibody responses as independent variables resulted in 81.1% correct diagnosis, with sensitivity of 83.1%, specificity of 77.8%, positive predictability of 86%, and negative predictability of 73.7%. Smoking did not reach statistical significance in this model. CONCLUSION: A combined microbial colonization/antibody response profile can effectively discriminate between periodontitis patients and periodontally intact controls.

The prevalence of pathogenic periodontal microflora in healthy young adult smokers.

BACKGROUND: Smoking is a major risk factor in periodontitis, although the mechanisms of its effects are not well understood. The overall goal of this clinical study was to determine if smoking enhances the colonization of the oral cavity by pathogenic bacteria in a periodontitis-free population. The prevalence of Actinobacillus actinomycetemcomitans, Porphyromonas gingivalis, Prevotella intermedia, Fusobacterium nucleatum, Campylobacter rectus, Eikenella corrodens, Bacteroides forsythus, and Treponema denticola was investigated in 25 smokers and 25 non-smokers by using DNA probes. METHODS: The subjects were 21 to 35 years of age with a healthy periodontium or slight gingivitis and were systemically healthy. The test group included subjects who had a minimum of a 1.5 pack-year history of smoking, while the control subjects never smoked. Subgingival plaque samples were taken by paper point following the assessment of multiple clinical parameters. RESULTS: This investigation showed: 1) no statistically significant differences were noted in any clinical parameter measured between the groups; 2) of the 8 subjects who were infected by at least 1 tested pathogen, seven were smokers (P= 0.02); 3) infected smokers had a 15.7+/-3.5 pack-year history and smoked a mean of 27+/-5 cigarettes/day versus 4.4+/-0.8 pack years and 15+/-1 cigarettes/day for the non-infected smokers (P = 0.0001 and P = 0.004); and 4) smokers were 18 times more likely to exhibit the presence of pathogens than non-smokers. CONCLUSIONS: These data indicate that the prevalence of colonization of the sulcus by pathogenic bacterial species in periodontitis-free individuals is related to the quantity and duration of cigarette smoking.

Eikenella corrodens phase variation involves a posttranslational event in pilus formation.

The human pathogen Eikenella corrodens synthesizes type IV pili and exhibits a phase variation involving the irreversible transition from piliated to nonpiliated variants. On solid medium, piliated variants form small (S-phase), corroding colonies whereas nonpiliated variants form large (L-phase), noncorroding colonies. We are studying the molecular basis of this phase variation in the clinical isolate E. corrodens VA1. A genomic fragment encoding the major type IV pilin was cloned from the S-phase variant of strain VA1. Sequence analysis of the fragment revealed four tandemly arranged potential open reading frames (ORFs), designated pilA1, pilA2, pilB, and hagA. Both pilA1 and pilA2 predict a type IV pilin. The protein predicted by pilB shares sequence identity with the Dichelobacter nodosus FimB fimbrial assembly protein. The protein predicted by hagA resembles a hemagglutinin. The region containing these four ORFs was designated the pilA locus. DNA hybridization and sequence analysis showed that the pilA locus of an L-phase variant of strain VA1 was identical to that of the S-phase variant. An abundant pilA1 transcript initiating upstream of pilA1 and terminating at a predicted hairpin structure between pilA1 and pilA2 was detected by several assays, as was a less abundant read-through transcript encompassing pilA1, pilA2, and pilB. Transcription from the pilA locus was nearly indistinguishable between S- and L-phase variants. Electron microscopy and immunochemical analysis showed that S-phase variants synthesize, export, and assemble pilin into pili. In contrast, L-phase variants synthesize pilin but do not export and assemble it into pili. These data suggest that a posttranslational event, possibly involving an alteration in pilin export and assembly, is responsible for phase variation in E. corrodens.