ALTHOUGH BACTERIAL VAGINOSIS (BV) is the most prevalent cause of symptomatic vaginal discharge in the United States and has been implicated as a cause of preterm birth and acquisition of HIV, its cause remains unknown. 1–3 The microbiologic changes that are characteristic of this syndrome suggest that it may be caused by an “overgrowth” of anaerobic organisms, trigger unknown, though its epidemiology strongly suggests a sexual association. 4
Mobiluncus spp are among the anaerobic organisms most highly associated with BV. These fastidious curved rods are only rarely cultured from vaginal specimens of women without BV, yet are highly predictive of BV if found on Gram stain or wet-preparation examination of vaginal secretions. 5 Although these organisms are often suggested to be a marker of BV, their role in the etiology of this syndrome, if any, remains unknown. Because they are often difficult to isolate from clinical specimens and because the actual sensitivity of Grain stain detection is unknown, we used a highly sensitive polymerase chain reaction (PCR) methodology to compare the prevalence of Mobiluncus in women with and without evidence of BV.
Vaginal swab samples were obtained from women attending the Jefferson County Department of Health (JCDH) Sexually Transmitted Disease Clinic who participated in studies of the treatment and natural history of BV (i.e., the effect of treatment of asymptomatic BV and treatment of concurrent BV and cervicitis) and from women without evidence of infection. Approval for the study was obtained from the Institutional Review Boards of the University of Alabama at Birmingham and the JCDH. The presence of BV was determined at enrollment using the clinical criteria of Amsel et al. 6 Specimens for PCR analysis were vaginal wash samples obtained by gently agitating 5 ml sterile saline within the vaginal vault. Vaginal swab specimens were obtained from women without evidence of BV. Gram stain smears, which were prepared directly from vaginal swab specimens, were available for both groups.
Vaginal wash specimens were centrifuged for 10 minutes at 2000 × g. The supernatant was decanted and the pellet was resuspended in 1 ml RPMI media and frozen at −20 °C before testing. Vaginal swab specimens were agitated in 1 ml sterile water and pelleted in the same fashion. The DNA extraction was based on a modified version of Kasper 7 for bacteria. Archived samples were thawed and washed in 1 ml phosphate-buffered saline (pH 7.4), the pellet was resuspended in 600 μl lysis buffer (l M Tris, 0.5 M EDTA, 10% glucose, 2 mg/ml lysozyme) heat activated at 80 °C for 5 minutes then cooled to room temperature. Then, 3 μl 0.5 mg/ml RNase (Promega, Madison, WI) was added to the sample and digested for 1 hour at 37 °C, and proteins were precipitated with 0.2N NaOH, 1% sodium dodecyl sulfate, and 5 M potassium acetate (pH 4.8). The DNA was extracted with 600 μl isopropanol and washed with 600 μl 70% ethanol. The DNA pellet was rehydrated in 50 μl to 100 μl 10 mM Tris (pH 7.4)/1 mM EDTA (pH 8.0) and heat activated at 65 °C for 1 hour. Confirmation of DNA (1 μl/sample) was determined by electrophoresis on a 0.8% agarose/0.5 μg/ml ethidium bromide gel, viewed with an ultraviolet light box, and photographed with 667 Polaroid film.
Polymerase Chain Reaction Conditions
Two separate primer pairs were used to determine the presence of Mobiluncus spp and M curtisii. The primer sets used were Mob-AS- 5′CGCAGAAACACAGGATTGCA3′ and Mob-S- 5′GTGAACTCCTTTTTCTCGTGAA3′ for Mobiluncus spp and Mob-V3- 5′GCCAGCCTTCGGGGTGGTGT3′ and Mob-V4- 5′TCACGAGTCCCCGGCCGAACC3′ for M curtisi. 8,9 No specific primers were available for M mulieris. The primer concentration used for all primer sets was 10 pmol/μl. The PCR amplification mix was 2.5 mM each 5 μl 10X PCR buffer and 4 μl dNTP, 0.5 μl each primer pair (Gibco BRL, Frederick, MD), 5 U/μl 0.5 μl ampi-Taq gold polymerase (PE Applied Biosystems, Branchburg, NJ), 5 μl 2 mM dUTP, 0.5 μl AMPErase UNG (PE Applied Biosystems), 0.5 DNA product (5–10 ng/μl), and 34.5 μl sterile water, thereby yielding a 50-μl volume per reaction. To prevent false priming and carryover contaminants, ampi-Taq gold and UNG were used in all PCR reactions. Before the main thermocycling program, the UNG was activated a 50 °C for 2 minutes, followed by a 5-minute predenaturation cycle run at 94 °C before initiating the main PCR cycling program. The cycling program consisted of 35 cycles (denaturation cycle, 30 seconds at 94 °C; annealing cycle, 30 seconds at 60 °C; extension cycle, 45 seconds at 72 °C). After the main cycle program was complete, an additional extension step was added for 7 minutes at 72 °C. The PCR samples were then cooled to 4 °C for 30 to 60 minutes. The PCR products (5 μl/sample) were electrophoresed on a 1% agarose/0.5 μg/ml ethidium bromide gel, viewed on an ultraviolet light box, and photographed. Fluorescence of the Mobiluncus spp 450 bp and M curtisii 100 bp was considered a positive result. Positive and negative controls were run in all PCR batch runs. The negative controls were a Lactobacillus jensenni (ATCC strain #25258), the Master Mix solution with primers but no DNA, and human genomic DNA. These controls were used to rule out the presence of false priming for the samples taken during menses. The positive control was M curtisii (ATCC strain #35241).
DNA Sequencing of PCR Products
Various PCR products from this study were sequenced to determine the specificity of the primer pairs and the reproducibility of their selective priming ability. A simple digestion technique using enzymes to eliminate excess nucleotides and primers was used to prepare the PCR product for sequencing analysis as follows: l μl PCR product was digested with 0.4 μl 10 U/μl Exonuclease I (Amersham, Arlington Heights, IL) and 2 μl 1 U/μl shrimp alkaline phosphatase (Amersham) at 37 °C for 1 hour and 15 minutes at 72 °C, then cooled at 4 °C for 30 minutes. 10 The digested product was then sequenced with the Perkin-Elmer ABI Sequencing Program (PE Applied Biosystems). The sequence results of the PCR products were compared, and a BLAST search was done in the National Institutes of Health genebank.
All vaginal slides were obtained from the lateral wall, and were stained and graded using the method of Nugent et al. 11 In this method, the quantity of three different bacterial morphotypes (large, gram-positive rods representing Lactobacilli, small grain variable rods representing Gardnerella/Prevotella, and curved rods representing Mobiluncus spp) are scored. Composite scores of 0 to 3 represent normal flora, 4 to 6 are intermediate flora, and 7 to 10 indicate BV.
Statistical comparisons were made using Epi Info. 12 The Fisher exact test or the chi-square test was used to compare categorical variables.
Samples were available for 74 women, 58 of whom were enrolled in studies of BV, and 16 whom were without evidence of infection. Among women with BV, 84.5% had a positive PCR result for Mobiluncus, compared with 38% of women from the control group (P < 0.001). The PCR result for M curtisii was positive in 32 of 49 women (65.3%) in the BV group with Mobiluncus, compared with one of six women (16.6%) in the control group (P = 0.03). Gram stain scores indicative of BV were present in 41 of 49 (75%) specimens from women with a positive PCR result for Mobiluncus spp versus 6 of 19 (32%) women with negative PCR results (P = 0.002). The mean Gram stain score for women with a positive Mobiluncus PCR test result was 7.22 versus 3.42 for those with a negative result (P = 0.05).
Mobiluncus morphotypes were detected in the vaginal Gram stains of 23 of the 49 specimens from for Mobiluncus-positive BV patients (initial visits) by PCR and in none of the specimens that were negative by PCR analysis. Thus, the sensitivity and specificity of Gram stain for detection of Mobiluncus compared with PCR in a high-prevalence population was 46.9% and 100%, respectively. In the control group, a low-prevalence population for BV, and one in which Mobiluncus spp other than curtisii was the predominant species, the sensitivity of the Gram stain for the detection of Mobiluncus was 0% (zero of six samples).
Follow-up samples, which were obtained 4 weeks after receiving either metronidazole gel or placebo vaginal gel, were available for 19 of the BV patients. Among those patients who received placebo medication, there was a concordance rate of 92.8% (13 of 14 patients) for Mobiluncus PCR at the initial and follow-up visits. The species detected was the same in all cases. The one discrepant patient had no evidence of BV at the follow-up visit either by Amsel criteria or by Nugent Gram stain (Gram stain score, 2). Among those women who were treated with metronidazole gel, there was a concordance rate of 60% (three of five patients) for both the Mobiluncus and M curtisii PCR results at the initial and follow-up visits (P = 0.04 for the prevalence of Mobiluncus by PCR inpatients who received metronidazole versus placebo). Table 1 shows the PCR, Amsel, and Gram stain results for the women who received metronidazole. Two subjects with M curtisii at the initial visit who had resolution of BV after antibiotic therapy showed no evidence of Mobiluncus by PCR analysis at the second visit. One patient with persistence of M curtisii also had a persistence of BV; however, a patient with BV and Mobiluncus (not curtisii) had resolution of BV despite the persistence of the organism. The final patient had no evidence of Mobiluncus at either visit. The high concordance rate at the initial and follow-up visits in the patients who received placebo reinforces the validity of the PCR assay. Although the numbers are small, the association of persistence of M curtisii with lack of cure in women who received metronidazole is intriguing.
The PCR products from seven samples were sequenced, five of which were products identified as M curtisii by PCR, and two were identified as Mobiluncus (not curtisii). Genebank sequences showed 100% concordance with the M curtisii products and revealed that the two samples identified as Mobiluncus (not curtisii) were consistent with M mulieris.
Bacterial vaginosis, the most prevalent cause of vaginitis, remains poorly understood. Issues related to its diagnosis and treatment are hampered by a lack of understanding about the precise etiology of this infection. Mobiluncus spp are fastidious anaerobic bacteria that are highly specific, though not sensitive, markers for BV. Although other bacteria that are associated with BV are found in low concentrations in the vaginal fluid of women with Lactobacillus-predominant vaginal flora, Mobiluncus is rarely isolated from these cultures. 13,14 It is possible, however, that this organism is present in low concentrations, and is not easily cultivated because of its fastidious nature. In studies where careful anaerobic techniques have been used, Mobiluncus has been detected in 65% to 85% of women with clinically evident BV and in less than 5% of uninfected women. 13,14 The sensitivity of detection of Mobiluncus on Grain stain has been previously shown to be 84% compared with culture and whole-chromosomal DNA probe. 15
There are two known species of Mobiluncus, M curtisii and M mulieris.16 Additionally, two unique antigenic subgroups, one for each of the species, have been described. 17Mobiluncus curtisii appears as short, curved, gram-positive rods, whereas M mulieris is longer, curved, and gram-negative in appearance. Of the two species, M curtisii may be the more pathogenic. 18
Mobiluncus has been isolated from breast and umbilical abscesses, blood cultures, 19–22 and from the chorioamnionic membranes of placentas from women with preterm delivery and histologic evidence of chorioamnionitis. 23 In a Ugandan study of sexually transmitted disease control as a means to decrease HIV transmission, an increased frequency of HIV infection was associated with abnormal vaginal flora, especially among women with “severe BV” (i.e., those women with Mobiluncus morphotypes present on vaginal Gram stain). 24 Moi et al 18 studied the prevalence of antibody to Mobiluncus among women with and without BV, and found that women with BV had significantly higher titers than those without BV; titers to M mulieris were significantly lower than those to M curtisii. In a study of pregnant women, 75% had antibodies to M curtisii, whereas antibodies to M mulieris were lacking. 25 Serologic data such as these and isolation of M curtisii from postoperative and extragenital infections suggest that M curtisii may be more virulent than M mulieris. 18
Using PCR for Mobiluncus we found significant differences in the prevalence of this organism between women with BV and a control group of women without BV, and found significant associations between a positive PCR result for Mobiluncus and higher Nugent Gram stain scores. The prevalence of Mobiluncus by PCR analysis among women with BV was similar to that which has been reported using microbiologic culture techniques. 15 The prevalence was higher than anticipated among women in the control group, which may be due to the ability of PCR to detect low concentrations of organisms such as may be present in healthy women. Repeat samples showed persistence of the organism among women who had not received treatment for BV. Perhaps the most interesting finding was the distribution of species between the BV and control groups. Although Mobiluncus was detected by PCR in 38% of women in the control group, only one of these women had a PCR result indicative of M curtisii, suggesting that M mulieris was the predominant species among these healthy women. Although the numbers were small, analysis of the persistence of Mobiluncus after therapy for BV also suggested a difference between the two species (Table 1). Persistence of M curtisii was associated with treatment failure, whereas M mulieris persisted in a patient in whom BV was cured. These findings lend support to the hypothesis that M curtisii is the more virulent of the two species, and suggests a possible role for this organism in the pathogenesis of BV.
A limitation of this study was that we were unable to collect vaginal washes from the control group of women, and instead relied on vaginal swab specimens. Although it is possible that PCR using vaginal washes is more sensitive than that that using vaginal swabs, the higher than anticipated prevalence of Mobiluncus in the control group suggests that this may not be a factor.
The sensitivity of the Grain stain in our study compared with PCR for the detection of Mobiluncus was low, especially in the control group. This finding could be due to low concentrations of organisms combined with the fact that M mulieris is often more difficult to view than M curtisii. The specificity was 100% in that none of the samples in which Mobiluncus was seen on the Gram stain was negative by PCR.
In summary, detection of Mobiluncus by PCR analysis appears to be sensitive and specific. Advantages also include ease of the method compared with culture, and its ability to detect low concentrations of organisms. Our study indicates that this organism may be more common among healthy women than previously thought, but that there may be real differences in the prevalence of the two species between women with and without BV. The use of PCR for the detection of Mobiluncus spp among various cohorts of women and in longitudinal studies may provide data to elucidate the role of this organism, if any, in the pathogenesis of BV.
1. Eschenbach DA. The role of bacterial vaginosis as a cause of amniotic fluid infection, chorioamnionitis and prematurity: a review. Arch Gynecol Obstet 1993; 16: S282–S287.
2. Schwebke JR. Bacterial vaginosis: more questions than answers. Genitourin Med 1977; 73: 333–334.
3. Taha TE, Hoover DR, Dallabetta GA, et al. Bacterial vaginosis and disturbances of vaginal flora: association with increased acquisition of HIV. AIDS 1998; 12: 1699–1706.
4. Lossick JG. The descriptive epidemiology of vaginal trichomoniasis and bacterial vaginosis. In: Horowitz BJ, Mardh PA, eds. Vaginitis and Vaginosis. New York: Wiley-Liss, 1991: 77–84.
5. Thomason JL, Schreckenberger PC, Spellacy WN, Riff LJ, Le Beau LJ. Clinical and microbiological characterization of patients with non-specific vaginosis associated with mobile, curved anaerobic rods. J Infect Dis 1984; 149: 801–809.
6. Amsel R, Totten PA, Spiegel CA, Chen KCS, Eschenbach D, Holmes KK. Non-specific vaginitis: diagnostic and microbial and epidemiological associations. Am J Med 1983; 74: 14–22.
7. Shen AL, Porter TD, Wilson TE, Kasper GB. Structural analysis of the FMV binding domain of NADPH-Cytochrome P-450 oxidoreductase by site-directed mutagenesis. J Biol Chem 1989; 264: 7584–7589.
8. Tiveljung A, Backstrom J, Forsum U, Monstein HJ. Broad-range PCR amplification, DNA sequence analysis reveals variable motifs in 16S. rRNA genes of Mobiluncus
species. APMIS 1995; 103: 755–763.
9. Tiveljung A, Forsum U, Monstein HJ. Classification of the genes Mobiluncus
based on comparative partial 16S rRNA gene analysis. Int J Syst Bacteriol 1996; 46: 332–336.
10. Werle E, Schneider C, Renner M, Volker M, Fiehn W. Convenient single-step, one tube purification of PCR products for direct sequencing. Nucleic Acids Res 1994; 22: 4354–4355.
11. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of Gram stain interpretation. J Clin Microbiol 1991; 29: 297–301.
12. Dean A, Dean J, Columbeer D. Epi Info, Version 6. A word processing database and statistics program for epidemiology on microcomputers. Atlanta: Centers for Disease Control and Prevention, 1994.
13. Roberts MC, Hillier SL, Schoenknecht FC, Holmes KK. Comparison of grain stain, DNA probe and culture for the identification of species of Mobiluncus
in female genital specimens. J Infect Dis 1985; 152: 74–77.
14. Holst E. Bacterial vaginosis: clinical and microbiological findings. In: Horowitz BJ, Mardh PA, eds. Vaginitis and Vaginosis. New York: Wiley-Liss, 1991: 115–120.
15. Hillier SL, Critchlow CW, Stevens CE, et al. Microbiological, epidemiological and clinical correlates of vaginal colonization by Mobiluncus
species. Genitourin Med 1991; 67: 26–31.
16. Spiegel CA, Roberts M. Mobiluncus
gen. nov. Mobiluncus curtisii
sp. nov. Mobiluncus curtisii
subsp. nov., and Mobilincus mulieris
sp. nov., curved rods from human vagina. Int J Syst 1984; 34: 177–184.
17. Schwebke JR, Lukehart SA, Roberts MC, Hillier SL. Identification of two new antigenic subgroups within the genus Mobiluncus
. J Clin Microbiol 1991; 29: 2204–2208.
18. Moi H, Fredlund E, Tomquist E, Danielsson D. Mobiluncus
species in bacterial vaginosis: aspects of pathogenesis. APMIS 1991; 99: 1049–1054.
19. Glupezynski Y, Labbe M, Crockaert F, Pepersack F, Van Der Auwvera P, Yourassowsky E. Isolation of Mobiluncus
in four cases of extragenital infections in adult women. Eur J Clin Microbiol 1984; 3: 433–435.
20. Weinbren MJ, Perinpanayagan RM, Malnick H, Ormerod F. Mobiluncus
spp. pathogenic role in non-puerperal breast abscess. J Clin Pathol 1986; 39: 342–343.
21. Gomez-Garces JL, Balas D, Merino MT, Alos JI. Mobiluncus curtisii
bacteremia following septic abortion. Clin Infect Dis 1994; 19: 1166–1167.
22. Edmiston CE, Krepel CJ, Walker AP. Recovery of Mobiluncus curtisii
from mixed non-puerperal breast abscess. Eur J Clin Microbiol Infect Dis 1989; 8: 315–316.
23. Hillier SL, Martius J, Krohn M, Kiviat N, Holmes KK, Eschenbach DA. A case-control study of chorioamnionic infection and histologic chorioamnionitis in prematurity. N Engl J Med 1988; 319: 972–978.
24. Sewankambo N, Gray R, Waiver MJ, et al. HIV-1 infection associated with abnormal vaginal flora morphology and bacterial vaginosis. Lancet 1997; 350: 546–550.
25. Schwebke JR, Morgan SC, Hillier SL. Humoral antibody to Mobiluncus curtisii
, a potential serological marker for bacterial vaginosis. Clin Diagn Lab Immunol 1996; 3: 567–569.