OUP user menu

Detection of bacterial vaginosis-related organisms by real-time PCR for Lactobacilli, Gardnerella vaginalis and Mycoplasma hominis

M. Reza Zariffard, Mohammed Saifuddin, Beverly E. Sha, Gregory T. Spear
DOI: http://dx.doi.org/10.1111/j.1574-695X.2002.tb00634.x 277-281 First published online: 1 December 2002


The aim of this study was to determine the feasibility of detecting bacterial vaginosis (BV)-related organisms in stored genital tract specimens using real-time PCR. Frozen cervicovaginal lavage (CVL) samples from 21 women were analyzed by real-time PCR for the numbers of Mycoplasma hominis, Gardnerella vaginalis and lactobacilli. Lactobacilli organisms were detected in all CVL samples, G. vaginalis was detected in all but one sample, while M. hominis was detected in only six samples. Using the Amsel criteria to define BV, the samples from women with BV had significantly higher numbers of G. vaginalis organisms than samples from women without BV (P=0.004). In contrast, the number of lactobacilli organisms in BV samples was significantly lower (P=0.013). The number of M. hominis organisms was not significantly different between BV-positive and BV-negative samples. A striking relationship was observed where most of the samples contained high numbers of either lactobacilli or G. vaginalis but not both. These results show that it is possible to determine the presence of BV-related organisms in stored genital tract samples by PCR, suggesting that this could be developed into an objective method that could be useful for certain applications.

1 Introduction

Bacterial vaginosis (BV) is a common disorder that involves an alteration in the genital tract microflora, where predominant lactobacilli are replaced by Gardnerella vaginalis and anaerobic bacteria [13]. Mycoplasma hominis is also commonly found in BV and this organism has been hypothesized to participate in the pathology of BV [4]. BV is the most common cause of abnormal vaginal discharge in women of childbearing age and has been implicated as a cause of pelvic inflammatory disease, preterm delivery of low birth-weight infants and amnionitis [57]. It has been reported in several studies that women with BV have a higher incidence of HIV infection and also that the prevalence of HIV infection correlated with increasing severity of BV [811].

Clinically, BV can be identified by the presence of at least three of the four Amsel criteria: an elevated vaginal pH, an increased vaginal discharge, the presence of clue cells and an amine odor after the addition of potassium hydroxide [12]. Another method used to detect BV is Gram staining of vaginal smears which scores genital tract flora from 0–10 depending on the numbers of three bacterial morphotypes observed on the slide [13]. A commercially available DNA-hybridization test for G. vaginalis has also been shown to have utility in detecting BV in patients [14].

If developed, PCR-based assays to detect BV-related organisms could be useful in several settings. For example, PCR-detection could be used in research to study the influence of specific organisms on either the development of BV or the pathologic outcomes of BV. Also, PCR can potentially detect and quantify genital tract organisms such as mycoplasmas that are not detected by Gram stain, Amsel test or the commercial test. This could be useful since M. hominis in the genital tract has been suggested to be a contributor to the pathology of BV [4]. Also, a PCR method could be used on stored genital tract samples in situations such as clinical trials where women can no longer be evaluated by Gram staining or clinical criteria. Finally, since it is possible to develop PCR assays that are quantitative for the numbers of specific types of organisms in the genital tract, a PCR-based assay could be less subjective in diagnosing BV than some of the available tests. In this study, we determined the feasibility of using real-time PCR to detect and quantify BV-related organisms in the genital tract of women using stored genital tract samples. The organisms detected were G. vaginalis, lactobacilli and M. hominis. The results of the PCRs were compared with the clinical diagnosis of BV.

2 Materials and methods

2.1 Bacteria

M. hominis, Gardnerella vaginalis, Lactobacillus crispatus, Lactobacillus jenseneii and Escherichia coli were grown in cultures for use as PCR standards. Values for the genomic size of each organism were obtained from the National Center of Biotechnology Information web site (National Institutes of Health, Bethesda, MD, USA). M. hominis (American Type Culture Collection, ATCC, Rockville, MD, USA, cat. #23114) was first cultured in the triphasic culture system (Irvine Scientific, Santa Ana, CA, USA, Cat. #D303–012) for 24–72 h. The culture was then expanded at 37°C for 24–48 h by adding 10 ml of this liquid culture into one liter of supplemented media (ATCC Mycoplasma Medium cat. #14268). G. vaginalis (ATCC #14018) was grown for 48 h on Stacker plate HBT bilayer agar plates (Fisher Scientific, Cockeysville, MD, USA, #297884) at 37°C in 5% CO2. The bacterial lawn was resuspended in 1 ml of normal saline and was expanded by culturing in 15 ml of brain–heart infusion broth. L. crispatus (ATCC #33820) and L. jenseneii (ATCC #25258) were cultured in 10 ml MRS broth (Difco #0038) for 24 h at 37°C under anaerobic conditions with 5% CO2.

2.2 CVL samples

Cervicovaginal lavage (CVL) samples were collected from volunteers enrolled in the Chicago Consortium of the Women's Interagency HIV Study (WIHS) in accordance with the human investigation committee. WIHS is a prospective epidemiological and natural history study of HIV-infected and high risk uninfected women. The women included in this study were evaluated at Rush-Presbyterian-St. Luke's Medical center in Chicago, IL from 1996 to 1999. Informed consent was obtained from all participants. Lavage was performed by irrigating the cervix and vagina sequentially with 10 ml of sterile normal saline. To minimize possible sampling errors, all samples in this study were collected by the same individual. The cells in the CVL sample were pelleted by centrifugation and the pellets frozen.

2.3 DNA Isolation

Bacteria in cultures or CVL samples were centrifuged for 30 min at 13,000×g and treated with 200–500 µl of lysis buffer (8% sucrose, 50 mM NaCl, 20 mM Tris–HCl pH 8.0, 1 mM EDTA, 1 mg ml−1 Lysosyme; Sigma cat. #P-5147) for 20 min at room temperature. RNase (10 µl of RNase at 500 µg ml−1; Cat. #1119915, Roche, Indianapolis, IN, USA) was added to each tube and incubated at 37°C for 15 min. Bacteria were further disrupted by incubation with 100 µl of 10% SDS for 30 min at 37°C. Protein was removed by mixing lysates for 30 s with 1 ml of DNAzole (cat. #10503–027 Gibco BRL, Grand Island, NY, USA) and centrifuging for 5 min. DNA was precipitated by incubation with two volumes of absolute ethanol overnight at −70°C and the precipitate was washed with 70% ethanol. Linear acrylamide 15 µg/µl (cat. #9520 Ambion, Austin, TX, USA) was added as a carrier during precipitation [15]. Ultraviolet spectrophotometry was used to determine DNA concentrations.

2.4 PCR Primers

The oligonucleotide primers for M. hominis 16S rRNA gene were used as published [16] (F- RNAH1 5′-CAATGGCTAATGCCGGATACGC-3′; R- RNAH2 5′-GGTACCGTCAGTCTGCAAT-3′). For G. vaginalis, primers were designed from the 16S rRNA gene with the Primer designer program (Clone Manager 4 Sci-ed, Durham, NC) (F- GV1 5′-TTACTGGTGTATCACTGTAAGG-3′; R- GV3 5′-CCGTCACAGGCTGAACAGT-3′). The Lactobacillus primer sequences were selected according to parameters defined by Primer Express™ Software (Applied Biosystems, Foster City, CA, USA) from regions of 16S rRNA that were homologous between L. jenseneii and L. crispatus (F- LBF 5′-ATGGAAGAACACCAGTGGCG-3′; R- LBR 5′-CAGCACTGAGAGGCGGAAAC-3′). These primers detect both organisms (not shown).

All three primer sets lacked homology with non-target bacteria as determined by searching the Gene Bank database using Blast (National Center of Biotechnology Information, National Institutes of Health, Bethesda, MD, USA). All primers were synthesized and column purified by Integrated DNA Technologies (Coralville, IA, USA). Each primer set was used for amplification of the appropriate bacteria in a conventional PCR and the amplified products were confirmed to be the correct size on agarose gels.

2.5 Real-time PCR amplification

For real-time PCR, syber green PCR core reagents (cat. #4304886, PE Biosystems, Wamington, UK) and optical tubes were used (cat. #4316567, Applied Biosystems, Foster City, CA, USA). Reagents were used as follows for all organisms: 50 ng of DNA isolated from CVL (unknowns), 2.5 mM MgCl2, 0.25 mM dNTPs, 0.05 µM primers (except for lactobacilli, 0.15 µM primers) 0.25 U/µl Ung, and 0.02 U/µl of Taq gold polymerase. Amplification was carried out using a gene amp 5700 thermocycler (Applied Biosystems) with a thermocycle profile as follows: stage 1, 50°C (2 min); stage 2, 95°C (10 min) and stage 3 consisting of 50 cycles of 95°C (1 min), 62°C (1 min) and 72°C (1 min) for M. hominis; 94°C (45 s), 55°C (45 s) and 72°C (45 s) for G. vaginalis; and 94°C (45 s), 50°C (45 s) and 72°C (1 min) for lactobacilli. All three primer sets resulted in amplification with DNA from the appropriate type of bacteria but showed no cross reactivity to the other two organisms or E. coli (not shown). To further evaluate the specificity of the primer sets, they were tested against DNA that was isolated from several additional organisms that can be found in the genital tract, including Peptostreptococcus anaerobius (ATCC #27337), Propionibacterium acnes (ATCC #11827), Actinomyces pyogenes (ATCC #49698), Moraxella (Branhamella) catarrhalis (ATCC #25240), Neisseria mucosa (ATCC #19695), Acinetobacter sp. (ATCC #9957), Streptococcus sanguinis (ATCC #10556), Enterococcus faecalis (ATCC #29212), Corynebacterium pseudodiphtheriticum (ATCC #10701), Pseudomonas aeruginosa (ATCC #27853), Candida albicans (ATCC #14053), Streptococcus (ATCC #12386), Torulopsis glabrata (ATCC #15126) and Oligella urethralis (ATCC #17960). For each organism, 0.3 ng of DNA was added to amplification tubes. This amount of DNA corresponds to one million organisms if the genome is similar in size to an average genital tract lactobacilli. Each of the above organisms were negative for amplification when tested against the three primer sets. In order to quantify the number of organisms in CVL samples, each thermocycler run contained seven standards consisting of 102 to 107 copies of DNA from the appropriate bacterium to generate a standard curve. A representative standard curve is shown in Fig. 1. The number of organisms in 50 ng of DNA isolated from CVL samples was then determined based on the standards. This number was then multiplied by the total DNA isolated from each sample and divided by 50 ng to give the total number of organisms estimated to be in the CVL sample. CVL sample DNA that resulted in values higher than the standard curve was diluted and rerun in subsequent thermocycler runs. The number of organisms in CVL samples was averaged from two separate thermocycler runs with the values between the runs on average agreeing within 30%.

Figure 1

Real-time PCR standard curve for M. hominis. The number of M. hominis organisms (102 to 107) added to each PCR tube is plotted versus the thermocycler run at which each sample was detected as having a positive signal. Duplicate PCR tubes for each target DNA amount were run in each assay. Similar results were obtained in 10 independent thermocycler runs.

3 Results

DNA was extracted from a total of 21 CVL samples and the number of organisms determined by real-time PCR. Lactobacilli organisms were detected in all of the CVL samples (range from 5.9×106 to 3.2×1010 organisms per CVL, Table 1). G. vaginalis organisms were detected in all but one of the samples (range from less than 104 to 1.2×1011). In contrast, only six of the samples had detectable numbers of M. hominis organisms (range from <104 to 7.5×107).

View this table:
Table 1

Number of organisms based on clinical diagnosis of BV

OrganismTotal (n=21)BV+ (n=5)BV− (n=16)P value
Lactobacillus3.4×108 (6×106–2.7×1010)8.5×106 (6×106–2.7×108)1.1×109 (1×107–2.7×1010)0.013
G. vaginalis4.2×108 (<104–1.2×1011)1.3×1010 (7×109–1.2×1011)3.75×107 (<104–1.3×1010)0.004
M. hominis<104 (<104–7.5×107)<104 (<104–7.2×105)<104 (<104–7.5×107)0.321
  • Mann–Whitney U-test.

  • Median (range) number of organisms per CVL sample.

Using the Amsel clinical criteria, sixteen of the CVL samples were obtained from women who were negative for BV, while five samples were from women with BV. The number of lactobacilli was significantly higher (P=0.013, Mann–Whitney U-test) in the BV-negative group (median number 1.1×109) than in the BV-positive group (median number 8.5×106, Table 1). In contrast, the number of G. vaginalis organisms was significantly higher (P=0.004) in the BV-positive group (median number 1.3×1010) than in the BV-negative group (median number 5.4×107). Only two of the BV-positive women were positive for M. hominis and the number of M. hominis organisms was not significantly different between the two groups.

A plot of the number of CVL lactobacilli vs. the number of CVL G. vaginalis organisms revealed a striking pattern (Fig. 2). The group of eight samples with >109G. vaginalis organisms contained all five BV-positive samples and all had low numbers of lactobacilli. Conversely, the samples with >109 lactobacilli were all BV-negative and had low numbers of G. vaginalis. Interestingly, the sample with the second-highest level of G. vaginalis was from a woman that was not diagnosed as having BV, suggesting that PCR may be more sensitive in picking up BV than the Amsel test.

Figure 2

Relationship of lactobacilli and G. vaginalis in CVL samples. The total numbers of lactobacilli and G. vaginalis for each CVL are plotted. Samples from the BV-positive women are shown as filled circles.

Since all of the above samples were from HIV-seropositive women, 10 additional CVL samples obtained from HIV-seronegative BV-negative women were assessed for the three organisms by PCR. The number of lactobacilli in the seropositive BV-negative (median number 1.1×109, Table 1) and seronegative BV-negative (median number 1.5×109) groups were similar (P=0.17, Mann–Whitney U-test). Similarly, M. hominis numbers were not different in the two groups since it was not detected in most women. However, the numbers of G. vaginalis organisms were significantly different in the two groups (P=0.01), with the seronegative group having a median of <104 organisms per CVL.

4 Discussion

BV is a common condition that has important clinical consequences but has a highly variable makeup as far as the types and numbers of bacteria present. This study sought to determine if PCR can be used to detect and quantify several of the genital tract bacteria that are positively or negatively associated with BV. The results show that real-time PCR can be used to detect both G. vaginalis, M. hominis and lactobacilli in genital tract samples. The results also show that samples from women with BV that was diagnosed clinically have significantly higher numbers of G. vaginalis, but significantly lower numbers of lactobacilli. Finally, there was a noticeable pattern where low numbers of lactobacilli were found in samples with high numbers of G. vaginalis and conversely, low numbers of G. vaginalis organisms were seen in samples that had high numbers of lactobacilli. M. hominis was detected at a lower frequency, and no statistically significant relationship with clinical BV was found.

A PCR method has several potential advantages in studying the biology of BV over the two most commonly used methods for detecting BV, the Amsel criteria and the Gram stain [12,13]. PCR is highly sensitive and can be optimized to pick up very low numbers of bacteria. However, this study and other studies [1,3] show that the number of either lactobacilli or G. vaginalis organisms are very high in the genital tract and therefore the high level of sensitivity of PCR is not needed to diagnose BV. However, the ability of real-time PCR to quantify the number of DNA copies over several logs is very useful since this means that preparation of multiple dilutions of sample for PCR amplification is not needed but means that this may be useful for quantifying the severity of BV. The Nugent test has also proved useful for quantifying the severity of BV. However, the Nugent test cannot identify organisms such as M. hominis that do not show up on Gram stains. Another feature of PCR is that it quantifies DNA rather than viable organisms. This is a potential advantage in that it can detect organisms in archived genital tract samples that were collected under conditions that were not optimized for organism viability.

In several previous studies, the number of G. vaginalis and lactobacilli in genital tract secretions were determined by counting the number of colony-forming organisms. For example, in four studies the mean number of G. vaginalis organisms per milliliter of vaginal fluid ranged from 107.6 to 108.3[1,1719]. Although the total amount of vaginal fluid in the genital tract is difficult to measure, it would appear that the current study measured a somewhat higher number of G. vaginalis organisms, since the median number in BV-positive samples per CVL was about 1010. There are many factors that could affect the numbers of bacteria arrived at in the above studies, such as differences in the patient populations (for example, CVL could harvest more bacteria than a swab), PCR may detect more organisms than culture methods since not all bacteria are viable and capable of growing into colonies, and counting colonies could underestimate the numbers of organisms such as G. vaginalis, since many organisms are associated with one clue cell, which would result in a single colony.

While the results of this study show that it is possible to detect and quantify the numbers of BV-related organisms in the genital tract, several important variables will have to be addressed before this type of test could be used to diagnose BV. For example, sampling error could significantly affect the numbers of organisms. Thus, if during collection of the 10 ml CVL, several ml were missed, this would result in a reduction of the number of organisms detected by PCR. To avoid these types of problems it may be useful to express the data as a ratio of lactobacilli to G. vaginalis or to run additional PCR to quantify all bacteria and express each bacteria as a relative number of the entire bacterial load.

In conclusion, this study indicates that it is feasible to use PCR to detect and quantify the numbers of BV-related organisms.


This work was supported by National Institutes of Health Grant P01HD40539. Samples and clinical data were provided by the Women's Interagency HIV Study (WIHS) Collaborative Study Group with centers (Principal Investigators) at New York City/Bronx Consortium (Kathryn Anastos), Brooklyn, NY (Howard Minkoff), Washington DC Metropolitan Consortium (Mary Young), the Connie Wofsy Study Consortium of Northern California (Ruth Greenblatt, Herminia Palacio), Los Angeles County/Southern California Consortium (Alexandra Levine), Chicago Consortium (Mardge Cohen), Data Coordinating Center (Alvaro Muñoz, Stephen J. Gange). The WIHS is funded by the National Institute of Allergy and Infectious Diseases, with supplemental funding from the National Cancer Institute, the National Institute of Child Health and Human Development, the National Institute on Drug Abuse, the National Institute of Dental Research, the Agency for Health Care Policy and Research, and the Centers for Disease Control and Prevention (U01-AI-35004, U01-AI-31834, U01-AI-34994, AI-34989, U01-HD-32632 (NICHD), U01-AI-34993, U01-AI-42590).


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
View Abstract