Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Unravelling the Bacterial Vaginosis-Associated Biofilm: A Multiplex Gardnerella vaginalis and Atopobium vaginae Fluorescence In Situ Hybridization Assay Using Peptide Nucleic Acid Probes

  • Liselotte Hardy ,

    Affiliations Unit of Epidemiology and Control of HIV/STD, Department of Public Health, Institute of Tropical Medicine, Antwerp, Belgium, Laboratory Bacteriology Research, Faculty of Medicine & Health Sciences, University of Ghent, Ghent, Belgium, STI Reference Laboratory, Department of Clinical Sciences, Institute of Tropical Medicine, Antwerp, Belgium

  • Vicky Jespers,

    Affiliation Unit of Epidemiology and Control of HIV/STD, Department of Public Health, Institute of Tropical Medicine, Antwerp, Belgium

  • Nassira Dahchour,

    Affiliations STI Reference Laboratory, Department of Clinical Sciences, Institute of Tropical Medicine, Antwerp, Belgium, Plantijn Hogeschool, Antwerp, Belgium

  • Lambert Mwambarangwe,

    Affiliation Rinda Ubuzima, Kigali, Rwanda

  • Viateur Musengamana,

    Affiliation Rinda Ubuzima, Kigali, Rwanda

  • Mario Vaneechoutte,

    Affiliation Laboratory Bacteriology Research, Faculty of Medicine & Health Sciences, University of Ghent, Ghent, Belgium

  • Tania Crucitti

    Affiliation STI Reference Laboratory, Department of Clinical Sciences, Institute of Tropical Medicine, Antwerp, Belgium


Bacterial vaginosis (BV), a condition defined by increased vaginal discharge without significant inflammation, is characterized by a change in the bacterial composition of the vagina. Lactobacillus spp., associated with a healthy vaginal microbiome, are outnumbered by BV-associated organisms. These bacteria could form a polymicrobial biofilm which allows them to persist in spite of antibiotic treatment. In this study, we examined the presence of Gardnerella vaginalis and Atopobium vaginae in vaginal biofilms using Peptide Nucleic Acid (PNA) probes targeting these bacteria. For this purpose, we developed three new PNA probes for A. vaginae. The most specific A. vaginae probe, AtoITM1, was selected and then used in an assay with two existing probes, Gard162 and BacUni-1, to evaluate multiplex FISH on clinical samples. Using quantitative polymerase chain reaction (qPCR) as the gold standard, we demonstrated a sensitivity of 66.7% (95% confidence interval: 54.5% - 77.1%) and a specificity of 89.4% (95% confidence interval: 76.1% - 96%) of the new AtoITM1 probe. FISH enabled us to show the presence of a polymicrobial biofilm in bacterial vaginosis, in which Atopobium vaginae is part of a Gardnerella vaginalis-dominated biofilm. We showed that the presence of this biofilm is associated with high bacterial loads of A. vaginae and G. vaginalis.


Bacterial vaginosis (BV), a condition characterized by increased vaginal discharge without significant inflammation, is highly prevalent in women of reproductive age. It increases the risk for acquisition and transmission of sexually transmitted infections, including HIV, and is associated with preterm birth in pregnant women [12]. BV is a dysbiotic condition of unknown etiology and is characterized by a change in the microbial composition of the vagina. Lactobacillus spp., associated with a healthy vaginal microbiome, are outnumbered by an array of BV-associated organisms including Gardnerella vaginalis [13]. However, several studies suggest that the mere presence of G. vaginalis is not sufficient for the diagnosis of BV. Indeed, G. vaginalis is also present in 50% to 70% of women without BV according to Nugent score [46]. G. vaginalis expresses various virulence factors such as vaginolysin [7] and sialidase [8]. It can also produce a biofilm [9], thereby increasing its tolerance to lactic acid and hydrogen peroxide produced by lactobacilli [10,11] and to antimicrobial treatment [12,13]. Furthermore, it has been suggested that its adherence and biofilm-forming capacities allow G. vaginalis to initiate the colonization and scaffolding of the vaginal epithelium to which other species can attach subsequently [14,15].

As was first shown a decade ago, Atopobium vaginae is one of the many other species that are characteristic of BV [1620]. In one study, A. vaginae was detected in 80% of samples testing positive for G. vaginalis and made up 40% of the total biofilm mass dominated by G. vaginalis [9]. This association was confirmed in a study by Bradshaw et al. [21]: 93% of samples containing A. vaginae also contained G. vaginalis, whereas only 10% tested positive for G. vaginalis when A. vaginae was absent [22]. In contrast to G. vaginalis, A. vaginae is rarely part of the healthy vaginal microbiome and is considered a more specific marker of BV than G. vaginalis [17,21,23].

It is postulated that a biofilm provides bacteria with a competitive advantage over planktonic bacteria and that polymicrobial biofilms may offer additional advantages over single-species biofilms. Mechanisms that have been described in previous studies include metabolic cooperation, increased resistance to antibiotics or host immune responses [24] and an enlarged gene pool with more efficient sharing of genetic material compared to mono-species biofilms [25]. Polymicrobial coexistence is the dominant form in environmental biofilms, but is also prominent in the human body [24]. A well-known example is dental plaque: anaerobic bacteria, which are sensitive to oxygen, can survive and persist under the aerobic conditions in the oral cavity due to the consummation of oxygen by aerobic bacteria in the dental biofilm [26].

On the basis of these previous findings, we hypothesize that a polymicrobial biofilm consisting of A. vaginae and G. vaginalis and other bacteria not discussed in this study may serve as a marker of BV. Thus, better visualization of the structure of vaginal biofilms and identification of the bacterial components of the biofilm may contribute to better understanding of BV. To study the role of A. vaginae and G. vaginalis in BV, we designed and evaluated the performance of fluorescence in situ hybridization (FISH) with peptide nucleic acid (PNA) probes for A. vaginae and G. vaginalis. Three new PNA probes were designed for A. vaginae. The most specific A. vaginae probe was selected, and subsequently used together with a PNA probe that had been described for the detection of G. vaginalis and a positive control probe that detects a broad range of bacteria in order to evaluate the multiplex FISH on clinical samples.

Materials and Methods

Design of PNA probes

PNA probes targeting the bacterial 16S rRNA were synthesized by Panagene (Daejeon, South Korea). A fluorescent label was attached using a double 8-amino-3,6-dioxaoctanoic acid (AEEA) linker at the N terminus. We evaluated the performance of three new PNA probes for A. vaginae: Ato291, previously described as a DNA FISH probe [9,27] and AtoITM1 and AtoITM2, which we developed using the Applied Biosystems PNA designer software (, based on sequences of species-specific PCR primers from Burton [28] for AtoITM1 and from Fredricks [29] for AtoITM2. The probes met the following criteria: purine content was limited to less than 60%; a maximum of four purines in a purine-stretch and a maximum of three guanines in a guanine-stretch were allowed; and self-complementarity was absent, considering that PNA/PNA interactions are stronger than PNA/DNA interactions. For the detection of G. vaginalis, a previously described probe, Gard162 [30,31], was used. The broad-range bacterial probe, BacUni-1, previously designed [32] as a modified version of the broad-range eubacterial DNA probe [33], was used as a positive control. The probe specifications are listed in Table 1.

Bacterial culture techniques for evaluation performance PNA probes

The performance of the PNA probes was evaluated using clinical isolates, obtained from the collections of the Institute of Tropical Medicine (ITM) and Ghent University. The five most frequently detected Lactobacillus species [3,34], representing the non-BV microbiome, were chosen as a negative control to assure that the probes would not cross-hybridize with the normal microbiome (Table 2). Furthermore we selected the most frequent BV-associated bacteria (8 A. vaginae strains, 5 G. vaginalis strains) as a negative control for G. vaginalis and A. vaginae respectively in addition to 2 Prevotella melaninogenica strains (Table 2). This small test panel is a limitation of the study and more expansive testing will be required to assure that there is no cross-reactivity with other bacteria. The strains from frozen stocks in skimmed milk (-80°C) were cultured on Columbia agar base (Becton Dickinson Biosciences, Erembodegem, Belgium) + 5% horse blood and grown under anaerobic conditions (10% hydrogen, 10% carbon dioxide and 80% nitrogen), using an anaerobic incubator (Whitley DG250) at 37°C for 48 h and bacteria were streaked onto fresh plates every 48–72 hours. To determine the limit of detection, FISH was performed using the AtoITM1, Gard162 and BacUni-1 probe on serial tenfold dilutions using the fixative used throughout the study: Carnoy solution (6:3:1, ethanol:chloroform:glacial acetic acid [35,36]). Carnoy solution was chosen as a fixative solution because of its proven effectiveness in the stabilization of and minimal shrinkage in tissue structure [36,37]. The concentration of the dilutions was estimated based on the turbidity of the sample compared to McFarland Standards (Bio-Merieux SA, Marcy l’Etoile, France) and ranged from 1.2x109 to 1.2x103 cells per ml. The suspensions were vortexed briefly and 5 μl was spotted into the hybridization chamber; a 5 mm² area marked with a PAP Pen (Sigma Aldrich, St. Louis, USA) that creates a hydrophobic border on a Superfrost Plus slide (Menzel-Gläser, Braunschweig, Germany). The spotted suspensions were dried before performing FISH. Experiments were performed in duplicate.

Table 2. Specificity testing in duplicate of PNA probes using cultured bacteria.

The signal was considered positive if it had a positive counterpart in the DAPI stain and displayed a positive signal simultaneously with the broad-range probe. The signal was considered negative if no signal was seen with the species-specific probe.

Clinical samples

Ethics statement.

Vaginal samples were collected from 119 women participating in a clinical trial in Rwanda studying the vaginal microbiome and acceptability of a contraceptive ring (S1 Protocol) (the ‘Ring Plus’ study, identifier NCT01796613) (data analysis on-going) [38]. Participants were between 18 and 35 years old and provided written informed consent for participation in the study. The Ring Plus study and consent procedure were approved by the Rwanda National Ethics Committee, Rwanda; the Institutional Review Board of the ITM Belgium; and the ethics committee of the University Teaching Hospital in Antwerp, Belgium.

Vaginal sample collection and preparation.

Vaginal sampling was carried out by the study clinician as part of the study procedures. Two Copan flocked swabs (Copan, Brescia, Italy) and one cotton swab were brushed against the lateral walls of the vagina. The cotton swab was immediately rolled on a Superfrost Plus slide (Menzel-Gläser) which was heat-fixed by passing twice through a flame. The Superfrost Plus (Menzel-Gläser) slides were stored for maximum six months and shipped to ITM at room temperature and fixed for a minimum of 12 hours at ITM, submerged in Carnoy solution [35,36]. The Copan flocked swabs were eluated by vortexing each swab for at least 15 seconds in 1.2 ml of diluted phosphate buffered saline (PBS) (pH 7.4 - 1:9, PBS:saline). The two eluates were combined and divided into three aliquots, which were stored at -80°C. The swab eluates were shipped frozen (-191°C) in a dry shipper to the ITM to determine the total bacterial load of A. vaginae and G. vaginalis by means of quantitative real-time polymerase chain reaction (qPCR).

Urine sample collection and preparation.

According to an earlier described procedure [39], first-void urine was collected by the participants and 2 ml was transferred immediately to a 15 ml tube containing 2 ml of Carnoy solution. The sample was fixed overnight and after centrifugation (10 minutes at 3200 g), the supernatant was decanted and the pellet was treated for a second time with 0.75 μl of Carnoy solution. The samples were stored between 2–8°C and shipped at room temperature to the ITM. Before applying FISH, the urine samples were vortexed briefly and 5 μl was spotted into the hybridization chamber on a Superfrost Plus slide (Menzel-Gläser).

Quantitative PCR for quantification of bacteria in vaginal samples.

DNA was extracted from 250 μl of the vaginal swab eluate using the Abbott m2000sp automated extraction platform (Abbott, Maidenhead, UK), according to the manufacturer’s instructions. The volume of 200 μl DNA extract was stored at –80°C until testing. qPCR was performed for each bacteria species separately, to avoid competition between the primers. The 25 μl PCR mixture contained 12.5 μl Rotor-Gene SYBR Green RT-PCR Master mix (Qiagen, Venlo, the Netherlands), 5 μl DNA extract, 0.5 μM of A. vaginae or 1 μM of G. vaginalis forward and reverse primers (Integrated DNA Technologies, Leuven, Belgium) and RNase-free water provided with the Rotor-Gene SYBR Green PCR kit. The primers for A. vaginae and G. vaginalis and the amplification reactions (Rotor Gene Q MDx 5 plex, Qiagen) have been described before [34].

Quantification was done using standard curves, constructed using DNA extracts from A. vaginae (CCUG 38953T) and G. vaginalis (LMG 7832T), grown at 35°C ± 2°C on Columbia agar base (Becton Dickinson) + 5% horse blood, under anaerobic conditions. DNA concentrations were determined using NanoDrop (Thermo Fisher Scientific, Erembodegem, Belgium) and the number of genomes was calculated using the described genome sizes and G+C content of the strains. A total of six tenfold dilutions of the DNA stocks were prepared in high performance liquid chromatography (HPLC) grade water. Both the standard curve and samples were run in duplicate. The bacterial load was expressed as genome equivalents (geq)/ml.

PNA FISH procedure

Multiplex hybridization was performed on a Superfrost Plus slide (Menzel-Gläser) in a 5 mm² quadrant hybridization area marked with a PAP pen (Sigma Aldrich, St. Louis, USA), a liquid-repellent slide marker. The slide was covered with a cover slip after addition of a hybridization buffer that contained 200 nM of each probe: species-specific probes for A. vaginae (AtoITM1 or AtoITM2 or Ato291) and G. vaginalis (Gard162), and the broad-range BacUni-1 probe. The hybridization solution consisted of 10% (wt/vol) dextran sulphate (Sigma Aldrich), 10 mM NaCl (Merck KGaA, Darmstadt, Germany), 2% (vol/vol) formamide (Merck KGaA), 0.1% (wt/vol) sodium pyrophosphate (Sigma Aldrich), 0.2% (wt/vol) polyvinylpyrrolidone (Sigma Aldrich), 0.2% (wt/vol) Ficoll (Sigma Aldrich), 5 mM disodium EDTA (Merck KGaA), 0.1% (vol/vol) Triton X-100 (Acros Organics, Geel, Belgium) and 50 mM Tris-HCl at pH 7.5 (Sigma Aldrich).

The slides were incubated in a hybridization oven (Shake ‘N Bake, Boekel Scientific, Feasterville, Pennsylvania) in humid conditions, which were achieved by adding a small tray of water, at 60°C for 60 minutes. After the slides were rinsed with double-distilled (dd) H20, they were immersed in a washing solution containing 5 mM Tris base, 15 mM NaCl and 0.1% (vol/vol) Triton X-100 (at pH 10) for 15 min at 60°C on the rocking shelves of the hybridization oven. After this washing step, the slides were rinsed again with ddH20 and air-dried in the dark at room temperature. Subsequently, the slides were counterstained with 6-diamidine-2-phenylindole dihydrochloride (DAPI) (Serva, Heidelberg, Germany), a DNA-intercalating agent that stains the chromosomes of both prokaryotic and eukaryotic cells, for 5 minutes at room temperature in the dark and rinsed with ddH20. Before imaging, the slides were air-dried at room temperature in the dark.

Assessment of reproducibility of FISH

The inter-run repeatability of FISH was evaluated by comparing the outcomes of two independent FISH runs for a subset of the samples. The second hybridization was performed 6 months later on a subset of 15% of the samples (N = 17), which reflects one FISH run. We selected 17 samples showing variable results for the species-specific signal and positive for the broad-range probe. A new hybridization spot was drawn and fresh hybridization and washing buffer was used on the exact same slides used in the first FISH run. Visual inspection by confocal microscopy was performed by the same microscopist.

Microscopic analysis of hybridized samples

The hybridized samples were stored in the dark at room temperature for a maximum of one week before microscopic observation using laser scanning confocal microscopy (LSM700, Zeiss, Oberkochen, Germany). The microscope operates with four stable, solid-state lasers at wavelengths of 405 to 639 nm, and is therefore able to detect all three fluorescently labelled probes and the DAPI stain at once in one hybridization chamber. The sample was first scanned at 100x magnification (objective: EC Plan-Neofluar 10x/0.30 Ph1 M27), before individual bacteria were identified at 400x magnification (objective: Plan-Apochromat 40x/1.3 Oil Ph3 M27). Separate scattered bacterial cells were defined as dispersed bacteria (Fig 1A). Aggregates of bacterial cells attached to the vaginal epithelial cells were defined as adherent bacteria forming a biofilm (Fig 1B). The species-specific signal was considered positive only if it had a positive counterpart in the DAPI stain and if it displayed a positive signal simultaneously with the broad-range probe.

Fig 1. Dispersed bacteria versus biofilm.

Confocal laser scanning images with 400x magnification of G. vaginalis biofilm in 2 vaginal slides (A and B) in a superimposed image: vaginal epithelial cells DAPI in blue and G. vaginalis specific PNA-probe Gard162 with Alexa Fluor 647 in red. A: vaginal sample with dispersed bacteria; B: vaginal sample with bacteria in biofilm.

Statistical analysis

The specificity and sensitivity of FISH for vaginal slides was compared with the qPCR as a reference method. Bacterial counts were log 10 transformed before analysis. Data analysis was done using STATA13. The p-values reported for associations between the presence of bacterial species/biofilm and bacterial loads from the qPCR results originate from the non-parametric Kruskal-Wallis equality-of-populations rank test.


PNA FISH probe performance on bacterial strains

The three probes specific for A. vaginae clearly hybridized with all eight A. vaginae strains tested (Table 2). However, Ato291 showed cross-hybridization with three G. vaginalis strains, with all Lactobacillus species and with one Prevotella strain. The newly developed AtoITM2 probe cross-hybridized with three G. vaginalis strains. Only AtoITM1 performed without false positive results and was selected for further evaluation on the clinical samples. The Gard162 probe was able to identify all five G. vaginalis test strains and showed no cross-hybridization with any of the other 15 species tested (Table 2). All bacterial strains tested hybridized with the broad-range BacUni-1 probe. According to the FISH results of the serial dilutions, the limit of detection for AtoITM1, Gard162 and BacUni-1 probes was 1.2 x 105 cells per ml.

Detection of A. vaginae and G. vaginalis in clinical samples by PNA FISH

In a small pilot study, a total of 10 paired vaginal slides and urine samples were tested to evaluate the suitability of each type of sample for FISH analysis. Six out of 10 urine samples could not be assessed due to the low presence of vaginal epithelial cells in the urine, whereas this problem was experienced in only 2 vaginal samples. Therefore, it was decided to continue analyses on vaginal slides only.

Using the AtoITM1 PNA-probe, A. vaginae was visualized as dispersed entities, without the presence of biofilm, in 27/119 (22.7%) of the samples. A. vaginae biofilm was present in 26/119 (21.8%) samples. A. vaginae FISH was negative in the remaining 66/119 (55.5%) samples. PNA-FISH using Gard162 detected dispersed-only G. vaginalis in 31/119 (26%) samples, G. vaginalis biofilm in 58/119 (48.7%) samples (e.g., Fig 2) and 30/119 (25.3%) samples were negative for G. vaginalis. Of the 89 G. vaginalis FISH-positive samples (dispersed or biofilm), 36 samples (41%) were negative for A. vaginae. However, all samples with A. vaginae biofilm showed a G. vaginalis biofilm as well (e.g., Figs 3, 4 and 5).

Fig 2. Distribution of samples according to FISH.

Aggregates of bacterial cells attached to the vaginal epithelial cells, were defined as biofilm. Separate scattered bacterial cells, without the presence of biofilm, were defined as dispersed only bacteria.

Fig 3. G. vaginalis biofilm.

Montage of confocal laser scanning images with 400x magnification of G. vaginalis biofilm, negative for A. vaginae, in 4 vaginal samples (A-D) in a superimposed image: vaginal epithelial cells DAPI in blue and G. vaginalis specific PNA-probe Gard162 with Alexa Fluor 647 in red. For clarity we omitted the BacUni-1 plane; the bacteria that did not hybridize with Gard162 are visible in DAPI blue.

Fig 4. Superimposed image of polymicrobial biofilm of A. vaginae and G. vaginalis.

Montage of confocal laser scanning images with 400x magnification of polymicrobial biofilm in 6 vaginal samples (A-F) in a superimposed image: vaginal epithelial cells DAPI in blue, G. vaginalis specific PNA-probe Gard162 with Alexa Fluor 647 in red and A. vaginae specific PNA-probe AtoITM1 with Alexa Fluor 488 in green. For clarity we omitted the BacUni-1 plane; the bacteria that are not bound by the specific probes are visible in DAPI blue.

Fig 5. Polymicrobial biofilm of A. vaginae and G. vaginalis in different panes.

Confocal laser scanning image with 400 x magnification of polymicrobial biofilm in different panes, A: vaginal epithelial cells DAPI in blue, B: all bacteria, BacUni-1 PNA-probe with Alexa Fluor 555 in yellow, C: A. vaginae specific PNA-probe AtoITM1 with Alexa Fluor 488 in green, D: G. vaginalis specific PNA-probe Gard162 with Alexa Fluor 647 in red (superimposed image can be seen in Fig 3A).

Characterization of vaginal samples by qPCR

A total of 119 vaginal samples were available for qPCR analysis. A. vaginae was present in 72 (60%) of the samples with a mean log of 7.55 ± 1.34 geq/ml. G. vaginalis was detected in 95 (80%) of the samples with a mean log of 7.38 ± 1.11 geq/ml.

Performance of probes in vaginal samples

Quantitative PCR was used as the reference method for detection and quantification of A. vaginae and G. vaginalis. The signal of the species-specific probes was only considered positive if a positive counterpart was seen in the DAPI stain and with the universal BacUni-1 probe. When assessing the results with FISH probe AtoITM1 against the qPCR outcomes for A. vaginae for 119 vaginal samples, FISH results were false negative for 24 samples and false positive for 5 samples, resulting in a sensitivity of 66.7% (95% confidence interval (CI): 54.5% - 77.1%) and a specificity of 89.4% (95% CI: 76.1% - 96%) (Table 3). The mean log for the true positive samples (positive with qPCR and FISH) was 7.73 geq/ml, as compared to a mean log of 7.19 geq/ml for the false negative FISH results (p = 0.399). For Gard162, the G. vaginalis probe, 13 FISH results were false negative and six false positive. The sensitivity was 86.3% (95% CI: 77.4% - 92.2%) and the specificity 75.0% (95% CI: 52.9% - 89.4%) (Table 3). The mean log for the true positive results for G. vaginalis was 7.61 geq/ml compared to a mean log of 5.94 geq/ml for the false negative results (p<0.001).

Table 3. Performance of A. vaginae (AtoITM1) and G. vaginalis (Gard162) PNA probes, compared to qPCR results, for 119 vaginal slides.

Assessment of the repeatability was done using 17 samples. After the first hybridization, all samples showed a signal for the BacUni-1 probe, 5 and 9 samples out of 17 for the AtoITM1 and Gard162 probe respectively. The results of the second FISH with the BacUni-1 and Gard162 probe were in full agreement with the first run. For the AtoITM1 probe, only one sample had a different result in the second run (negative at first, but positive in the second run).

The presence of biofilm related to bacterial loads

The probability of detecting bacteria in a biofilm with FISH was higher when high (>106 geq/ml) bacterial loads for G. vaginalis (p<0.001) and A. vaginae (p<0.001) were present. The mean log of both species was highest when A. vaginae was part of the biofilm, compared to a biofilm of G. vaginalis only. A. vaginae biofilm was always observed together with G. vaginalis (Table 4).

Table 4. Presence of A. vaginae and G. vaginalis, as assessed by FISH, in relation to A. vaginae and G. vaginalis loads as determined by qPCR for 119 vaginal samples.


We set out to evaluate the performance of PNA FISH for the investigation of the vaginal polymicrobial biofilm consisting of G. vaginalis and A. vaginae. For this purpose, we evaluated three newly designed A. vaginae PNA probes for their specificity and applied the most specific one, AtoITM1, on a range of fixed vaginal slides together with an already existing G. vaginalis and broad-range PNA probe.

Hybridization-based techniques such as FISH have been used in various disciplines, such as cytogenetics and microbiology, to detect the presence or absence of nucleic acid sequences. Detection of DNA and RNA is generally done using DNA probes but the use of PNA probes is increasing. PNA molecules have a neutral backbone giving them a significant advantage in low ionic-strength conditions compared to DNA probes [40]. Low ionic-strength conditions prevent the complementary genomic sequences from reannealing when performing the FISH procedures; they facilitate denaturation of RNA secondary structures and favors hybridization of the PNA probes with nucleic acids. In combination with the superior penetration of PNA probes through the cell wall and hydrophobic bilayer of the target organism [41], PNA FISH is a fast, simple and robust assay. We compared DNA and PNA probes (data not reported) and can confirm that PNA FISH is faster and more robust than DNA FISH. In the current study, PNA FISH proved to be highly efficient for the identification and visualization of the spatial arrangement of A. vaginae and G. vaginalis in the BV-associated biofilms. Moreover, PNA FISH showed excellent inter-assay repeatability for all three probes used.

PNA FISH probe performance on bacterial strains

For the design of the PNA probes, we opted for Alexa fluorochromes, which have similar spectral properties as other fluorochromes, such as cyanine dyes, but are brighter and more resistant to photo bleaching [42].

PNA probes that specifically target A. vaginae have not been described before. The only probe thus far reported was a DNA probe (Ato291) [27]. The probe was designed to detect bacteria belonging to the Atopobium cluster in fecal samples. However, in silico evaluation of the specificity of the Ato291 probe showed cross-hybridization with other bacteria belonging to the Coriobacteriaceae, a family of Actinobacteria, to which A. vaginae belongs. This is not surprising, since the probe was originally designed on the basis of sequences of Coriobacteriaceae strains isolated from feces and clinical material. The probe has been used for the detection of A. vaginae in vaginal samples by Swidsinski et al. [9], but their findings have not yet been confirmed by other groups. In our experiments, using a PNA equivalent of the Ato291 probe, we showed a low specificity of the Ato291 probe on vaginal clinical isolates. The Ato291 probe cross-hybridized with three out of five G. vaginalis strains and all five Lactobacillus species. Therefore, we designed two new probes for A. vaginae targeting the 16S rRNA-gene, based on published PCR primers [28,29] and we adjusted the sequences to fit the requirements for PNA probes. One of the new probes, AtoITM1, did not cross-react to any of the tested species and was further used for detection of A. vaginae in vaginal slides.

Gard162 is the first PNA FISH probe designed specifically for G. vaginalis and has extensively been tested by Machado et al. on a variety of cultured bacterial strains and clinical samples [30,31]. Using this probe, we obtained clear hybridization for all G. vaginalis isolates tested and observed no cross-reaction with strains of the other species tested, confirming the findings of Machado et al. [30].

PNA FISH probe performance on clinical samples

Vaginal slides proved to be a valid sample type for imaging of the composition of the vaginal microbiome, if processed directly after sampling, as shown by Peltroche-Llacsahuanga et al. [43]. Collection of a vaginal swab is an easy and cheap sampling method, with a minimal burden on the study participant or patient. After heat fixation, we recorded that the slides can be stored at room temperature for up to at least six months and can be easily transported. A high quality vaginal sample can be obtained by thinly rolling the swab onto the slide. A thick vaginal ‘smear’ on the contrary where the material is smeared onto the slide is not useful for FISH and microscopic visualization.

The probes were also applied to 119 vaginal slides from women for whom the bacterial loads of G. vaginalis and A. vaginae had been quantified by qPCR. qPCR was used in this study as the reference method to evaluate the performance of the FISH probes, although comparison of these methods is subject to some limitations. qPCR is highly sensitive and was performed on a homogenized DNA extract representing half of the full sample. FISH however was carried out on 0.5 mm² of a vaginal slide, which could be heterogeneous. This is inevitably an underrepresentation of the vaginal sample. Both techniques were also performed using two different vaginal swabs, but the first collected specimen was used to prepare the slide as per study protocol.

After hybridization for 60 minutes and washing for 15 minutes, both at 60°C, the AtoITM1 probe gave only five false positive results, compared to qPCR results, resulting in a specificity of 90% for this set of samples However, 24 samples that were positive according to the qPCR were not detected by FISH, which gives a sensitivity of 67%. These results are comparable to the sensitivity observed for vaginal samples using FISH for detection of Group B Streptococcus [43]. The authors of this study obtained a higher sensitivity after extraction of the swabs by centrifugation of the swab head in a NaCl solution [43], but this method would probably disrupt the epithelial biofilm and thus prevent us from investigating the organization of the bacterial biofilm.

The relatively low sensitivity of the A. vaginae FISH assay cannot be explained by the bacterial load as measured by qPCR; the mean log of the true positive and false negative samples was not significantly different (log 7.73 and 7.19 geq/ml respectively). One possible explanation could be the typical structure of a biofilm, whereby an oxygen gradient exists from the top to the center of the biofilm [44]. Anaerobic bacteria such as A. vaginae are possibly found more embedded in the biofilm, to take advantage of the anaerobicity. It could be that the PNA probes are not able to fully penetrate into the inner parts of the biofilm; or that if the PNA probes do penetrate, the fluorescence could be masked and not detected due to low resolution of the equipment.

For G. vaginalis, 6 false positive results and 13 false negative results were found compared to 95 positive and 24 negative samples according to qPCR. This implies a sensitivity of 86% and specificity 75% for the detection of G. vaginalis by the Gard162 probe using our FISH protocol for this set of samples. This is lower than reported by Machado et al. [30], who, using the same probe, showed a full agreement between qPCR and FISH results for 13 vaginal samples. We were not able to elucidate why these discordant results were obtained.

Bacterial loads and the presence of a biofilm

Our study shows that higher bacterial loads of G. vaginalis and A. vaginae, as detected by qPCR, are associated with a higher probability of presence of a bacterial biofilm. Both bacterial species are important constituents of the vaginal epithelial biofilm [9,45]. No samples contained A. vaginae in the absence of G. vaginalis, but almost half of the G. vaginalis-positive samples did not contain A. vaginae according to FISH results. Both bacteria were seen in a dispersed and an adherent state, but A. vaginae was always accompanied by G. vaginalis. The mere presence of A. vaginae did not simply predispose to a polymicrobial biofilm, but when A. vaginae was part of the biofilm, compared to a biofilm of only G. vaginalis, both bacterial species were present in higher concentrations.

We hypothesize that G. vaginalis is one of the main initiators of a vaginal biofilm, when it is present in high amounts. This vaginal biofilm creates a favorable environment for anaerobic bacteria, such as A. vaginae. One reason for the appearance of A. vaginae may be the presence of an oxygen gradient within the biofilm. By embedding itself within the biofilm, A. vaginae can take advantage of the anaerobicity, proliferates and exists in a mutualistic relationship with G. vaginalis.


Our study confirms that PNA FISH is a valuable tool for detecting and visualizing biofilm-associated organisms in vaginal slides. This study describes the design and evaluation of a new PNA probe, AtoITM1, which can be included in multiplex FISH in BV biofilm research. Using the new probe, we have demonstrated the presence of a polymicrobial biofilm, with A. vaginae taking part in a G. vaginalis dominated biofilm.

Supporting Information

S1 Protocol. Final Approved Protocol for "The Ring Plus project: Safety and acceptability of vaginal rings that protect women from unintented pregenancy" version 2.0, 16 April 2013.



We are grateful to the participants of this study, the study staff of Rinda Ubuzima and the STI reference laboratory team of ITM, Antwerp.

Author Contributions

Conceived and designed the experiments: LH VJ TC. Performed the experiments: LH ND LM VM. Analyzed the data: VJ LH. Contributed reagents/materials/analysis tools: MV TC. Wrote the paper: LH VJ ND LM VM MV TC.


  1. 1. Marrazzo JM, Martin DH, Watts DH, Schulte J, Sobel JD, Hillier SL, et al. (2010) Bacterial vaginosis: identifying research gaps proceedings of a workshop sponsored by DHHS/NIH/NIAID. Sex Transm Dis 37: 732–744. pmid:21068695
  2. 2. Spiegel CA (1991) Bacterial vaginosis. Clin Microbiol Rev 4: 485–502. pmid:1747864
  3. 3. van de Wijgert JH, Borgdorff H, Verhelst R, Crucitti T, Francis S, Verstraelen H, et al. (2014) The vaginal microbiota: what have we learned after a decade of molecular characterization? PLoS One 9: e105998. ; PONE-D-14-20877 [pii]. pmid:25148517
  4. 4. Fredricks DN, Fiedler TL, Thomas KK, Oakley BB, Marrazzo JM (2007) Targeted PCR for detection of vaginal bacteria associated with bacterial vaginosis. J Clin Microbiol 45: 3270–3276. JCM.01272-07 [pii]; pmid:17687006
  5. 5. Nugent RP, Krohn MA, Hillier SL (1991) Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 29: 297–301. pmid:1706728
  6. 6. Srinivasan S, Hoffman NG, Morgan MT, Matsen FA, Fiedler TL, Hall RW, et al. (2012) Bacterial communities in women with bacterial vaginosis: high resolution phylogenetic analyses reveal relationships of microbiota to clinical criteria. PLoS One 7: e37818. ; PONE-D-12-04403 [pii]. pmid:22719852
  7. 7. Gelber SE, Aguilar JL, Lewis KL, Ratner AJ (2008) Functional and phylogenetic characterization of Vaginolysin, the human-specific cytolysin from Gardnerella vaginalis. J Bacteriol 190: 3896–3903. JB.01965-07 [pii]; pmid:18390664
  8. 8. Lopes dos Santos Santiago G, Deschaght P, El Aila N, Kiama TN, Verstraelen H, Jefferson KK, et al. (2011) Gardnerella vaginalis comprises three distinct genotypes of which only two produce sialidase. Am J Obstet Gynecol 204: 450–457. S0002-9378(11)00019-6 [pii];
  9. 9. Swidsinski A, Mendling W, Loening-Baucke V, Ladhoff A, Swidsinski S, Hale LP, et al. (2005) Adherent biofilms in bacterial vaginosis. Obstet Gynecol 106: 1013–1023. 106/5/1013 [pii]; pmid:16260520
  10. 10. Patterson JL, Girerd PH, Karjane NW, Jefferson KK (2007) Effect of biofilm phenotype on resistance of Gardnerella vaginalis to hydrogen peroxide and lactic acid. Am J Obstet Gynecol 197: 170–177. S0002-9378(07)00263-3 [pii]; pmid:17689638
  11. 11. Schwebke JR, Muzny CA, Josey WE (2014) Role of Gardnerella vaginalis in the pathogenesis of bacterial vaginosis: a conceptual model. J Infect Dis 210: 338–343. jiu089 [pii]; pmid:24511102
  12. 12. Olson ME, Ceri H, Morck DW, Buret AG, Read RR (2002) Biofilm bacteria: formation and comparative susceptibility to antibiotics. Can J Vet Res 66: 86–92. pmid:11989739
  13. 13. Van Acker H, Van Dijck P, Coenye T (2014) Molecular mechanisms of antimicrobial tolerance and resistance in bacterial and fungal biofilms. Trends Microbiol 22: 326–333. S0966-842X(14)00024-9 [pii]; pmid:24598086
  14. 14. Patterson JL, Stull-Lane A, Girerd PH, Jefferson KK (2010) Analysis of adherence, biofilm formation and cytotoxicity suggests a greater virulence potential of Gardnerella vaginalis relative to other bacterial-vaginosis-associated anaerobes. Microbiology 156: 392–399. mic.0.034280–0 [pii]; pmid:19910411
  15. 15. Verstraelen H, Swidsinski A (2013) The biofilm in bacterial vaginosis: implications for epidemiology, diagnosis and treatment. Curr Opin Infect Dis 26: 86–89. pmid:23221767
  16. 16. De Backer E, Verhelst R, Verstraelen H, Claeys G, Verschraegen G, Temmerman M, et al. (2006) Antibiotic susceptibility of Atopobium vaginae. BMC Infect Dis 6: 51. 1471-2334-6-51 [pii]; pmid:16542416
  17. 17. Ferris MJ, Masztal A, Aldridge KE, Fortenberry JD, Fidel PL Jr, Martin DH (2004) Association of Atopobium vaginae, a recently described metronidazole resistant anaerobe, with bacterial vaginosis. BMC Infect Dis 4: 5. ; 1471-2334-4-5 [pii]. pmid:15018635
  18. 18. Fredricks DN, Fiedler TL, Marrazzo JM (2005) Molecular identification of bacteria associated with bacterial vaginosis. N Engl J Med 353: 1899–1911. 353/18/1899 [pii]; pmid:16267321
  19. 19. Menard JP, Fenollar F, Henry M, Bretelle F, Raoult D (2008) Molecular quantification of Gardnerella vaginalis and Atopobium vaginae loads to predict bacterial vaginosis. Clin Infect Dis 47: 33–43. pmid:18513147
  20. 20. Verhelst R, Verstraelen H, Claeys G, Verschraegen G, Delanghe J, Van Simaey L, et al. (2004) Cloning of 16S rRNA genes amplified from normal and disturbed vaginal microflora suggests a strong association between Atopobium vaginae, Gardnerella vaginalis and bacterial vaginosis. BMC Microbiol 4: 16. ; 1471-2180-4-16 [pii]. pmid:15102329
  21. 21. Bradshaw CS, Tabrizi SN, Fairley CK, Morton AN, Rudland E, Garland SM (2006) The association of Atopobium vaginae and Gardnerella vaginalis with bacterial vaginosis and recurrence after oral metronidazole therapy. J Infect Dis 194: 828–836. JID36460 [pii]; pmid:16941351
  22. 22. Cools P, Verstraelen H, Vaneechoutte M, Verhelst R (2011) Atopobium. In: Liu Dongyou, editors. Molecular detection of human bacterial pathogens. CRC Press. pp. 31–43.
  23. 23. Verstraelen H, Verhelst R, Claeys G, Temmerman M, Vaneechoutte M (2004) Culture-independent analysis of vaginal microflora: the unrecognized association of Atopobium vaginae with bacterial vaginosis. Am J Obstet Gynecol 191: 1130–1132. S0002937804003990 [pii]; pmid:15507931
  24. 24. Elias S, Banin E (2012) Multi-species biofilms: living with friendly neighbors. FEMS Microbiol Rev 36: 990–1004. pmid:22229800
  25. 25. Wolcott R, Costerton JW, Raoult D, Cutler SJ (2013) The polymicrobial nature of biofilm infection. Clin Microbiol Infect 19: 107–112. pmid:22925473
  26. 26. Bradshaw DJ, Marsh PD, Watson GK, Allison C (1998) Role of Fusobacterium nucleatum and coaggregation in anaerobe survival in planktonic and biofilm oral microbial communities during aeration. Infect Immun 66: 4729–4732. pmid:9746571
  27. 27. Harmsen HJ, Wildeboer-Veloo AC, Grijpstra J, Knol J, Degener JE, Welling GW (2000) Development of 16S rRNA-based probes for the Coriobacterium group and the Atopobium cluster and their application for enumeration of Coriobacteriaceae in human feces from volunteers of different age groups. Appl Environ Microbiol 66: 4523–4527. pmid:11010909
  28. 28. Burton JP, Devillard E, Cadieux PA, Hammond JA, Reid G (2004) Detection of Atopobium vaginae in postmenopausal women by cultivation-independent methods warrants further investigation. J Clin Microbiol 42: 1829–1831. pmid:15071062
  29. 29. Fredricks DN, Fiedler TL, Thomas KK, Mitchell CM, Marrazzo JM (2009) Changes in vaginal bacterial concentrations with intravaginal metronidazole therapy for bacterial vaginosis as assessed by quantitative PCR. J Clin Microbiol 47: 721–726. JCM.01384-08 [pii]; pmid:19144794
  30. 30. Machado A, Almeida C, Salgueiro D, Henriques A, Vaneechoutte M, Haesebrouck F, et al. (2013) Fluorescence in situ Hybridization method using Peptide Nucleic Acid probes for rapid detection of Lactobacillus and Gardnerella spp. BMC Microbiol 13: 82. 1471-2180-13-82 [pii]; pmid:23586331
  31. 31. Machado A, Castro J, Cereija T, Almeida C, Cerca N (2015) Diagnosis of bacterial vaginosis by a new multiplex peptide nucleic acid fluorescence in situ hybridization method. PeerJ 3: e780. ; 780 [pii]. pmid:25737820
  32. 32. Perry-O'Keefe H, Stender H, Broomer A, Oliveira K, Coull J, Hyldig-Nielsen JJ (2001) Filter-based PNA in situ hybridization for rapid detection, identification and enumeration of specific micro-organisms. J Appl Microbiol 90: 180–189. jam1230 [pii]. pmid:11168720
  33. 33. Amann RI, Binder BJ, Olson RJ, Chisholm SW, Devereux R, Stahl DA (1990) Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl Environ Microbiol 56: 1919–1925. pmid:2200342
  34. 34. Jespers V, Menten J, Smet H, Poradosu S, Abdellati S, Verhelst R, et al. (2012) Quantification of bacterial species of the vaginal microbiome in different groups of women, using nucleic acid amplification tests. BMC Microbiol 12: 83. 1471-2180-12-83 [pii]; pmid:22647069
  35. 35. Matsuo K, Ota H, Akamatsu T, Sugiyama A, Katsuyama T (1997) Histochemistry of the surface mucous gel layer of the human colon. Gut 40: 782–789. pmid:9245933
  36. 36. Puchtler H, Waldrop FS, Conner HM, Terry MS (1968) Carnoy Fixation—Practical and Theoretical Considerations. Histochemie 16: 361–&. pmid:4179106
  37. 37. Shamsi M, Abdali K, Montazer NR, Kumar PV, Tabatabaee HR (2008) Comparison of Carnoy's solution and 96% ethyl alcohol fixation in bloody Pap smears. Acta Cytologica 52: 187–190. pmid:18499991
  38. 38. Schurmans C, De Baetselier I, Kestelyn E, Jespers V, Delvaux T, Agaba SK, et al. (2015) The ring plus project: safety and acceptability of vaginal rings that protect women from unintended pregnancy. BMC Public Health 15: 348. ; 10.1186/s12889-015-1680-y [pii]. pmid:25880636
  39. 39. Swidsinski A, Doerffel Y, Loening-Baucke V, Swidsinski S, Verstraelen H, Vaneechoutte M, et al. (2010) Gardnerella biofilm involves females and males and is transmitted sexually. Gynecol Obstet Invest 70: 256–263. 000314015 [pii]; pmid:21051845
  40. 40. Egholm M, Buchardt O, Christensen L, Behrens C, Freier SM, Driver DA, et al. (1993) PNA hybridizes to complementary oligonucleotides obeying the Watson-Crick hydrogen-bonding rules. Nature 365: 566–568. pmid:7692304
  41. 41. Williams B, Stender H, Coull JM (2002) PNA fluorescent in situ hybridization for rapid microbiology and cytogenetic analysis. Methods Mol Biol 208: 181–193. 1-59259-290-2-181 [pii]; pmid:12229288
  42. 42. Panchuk-Voloshina N, Haugland RP, Bishop-Stewart J, Bhalgat MK, Millard PJ, Mao F, et al. (1999) Alexa dyes, a series of new fluorescent dyes that yield exceptionally bright, photostable conjugates. J Histochem Cytochem 47: 1179–1188. pmid:10449539
  43. 43. Peltroche-Llacsahuanga H, Fiandaca MJ, von OS, Lutticken R, Haase G (2010) Rapid detection of Streptococcus agalactiae from swabs by peptide nucleic acid fluorescence in situ hybridization. J Med Microbiol 59: 179–184. jmm.0.013995–0 [pii]; pmid:19833782
  44. 44. Costerton W, Veeh R, Shirtliff M, Pasmore M, Post C, Ehrlich G (2003) The application of biofilm science to the study and control of chronic bacterial infections. J Clin Invest 112: 1466–1477. ; 112/10/1466 [pii]. pmid:14617746
  45. 45. Polatti F (2012) Bacterial vaginosis, Atopobium vaginae and nifuratel. Curr Clin Pharmacol 7: 36–40. BSP/CCP/E-Pub/0062 [pii]. pmid:22082330