Antibacterial effect of cell-free supernatant fraction from Lactobacillus paracasei CH88 against Gardnerella vaginalis

Bacterial vaginosis (BV) is the most common vaginal infection in reproductive women, which is characterized by depleted level of lactic acid bacteria and overgrowth of anaerobes such as Gardnerella vaginalis spp. Lactic acid bacteria have been known to be beneficial for amelioration of BV, since they produce antimicrobial substances against G. vaginalis spp. The objectives of this study were to characterize different fractions of cell-free supernatant of Lactobacillus paracasei CH88 (LCFS) and investigate antibacterial activity of the LCFS fractions against G. vaginalis in-vitro and in-vivo. Antibacterial activity of the LCFS was stable during thermal treatment up to 120 °C for 30 min and maintained at pH ranging from 3.0 to 13.0 except pH 5.0. Fraction below 3 kDa of the LCFS partially lost its antibacterial activity after treatment with proteolytic enzymes. Precipitated protein fraction below 3 kDa of the LCFS (< 3 kDa LCFSP) inhibited the growth and biofilm formation of G. vaginalis. Treatment of L. paracasei CH88 or the < 3 kDa LCFSP attenuated G. vaginalis-induced BV in mice by inhibiting the growth of G. vaginalis, reducing exfoliation of vaginal epithelial cells, and regulating immune response. These results suggest that L. paracasei CH88 may have potential in ameliorating G. vaginalis-induced BV.

www.nature.com/scientificreports/ Intravaginal administration of L. rhamnosus GR-1 and L. fermentum RC-14 to the patients with BV suppressed recurrence of BV 19 . Intravaginal administration of L. crispatus CTV-05 to the patients with BV resulted in this bacterium colonizing the vagina, consequently restoring healthy vaginal microbiota in the patients 20 .
Although some reproductive women maintain a healthy vaginal environment without lactobacilli 21 , lactobacilli are still considered as the most common bacteria in vaginal tract of healthy reproductive-age women and play an important role in maintaining healthy vaginal mucosa 22 . Lactobacilli have been known to have protective effects against BV via production of H 2 O 2 and lactic acid [23][24][25] . Moreover, they can produce antimicrobial compounds that inhibit growth of pathogens in the urogenital ecosystem 26 . Bacteriocin produced by L. acidophilus 160 had inhibitory activity against G. vaginalis 27 .
We found that cell-free supernatant of L. paracasei CH88 possessed antibacterial activity against G. vaginalis. Moreover, cell-free supernatant of L. paracasei CH88 that was adjusted to pH 7.0 (LCFS) also showed the antibacterial activity against G. vaginalis.
Thus, we hypothesized that the antibacterial activity of the LCFS (pH 7.0) may be attributed to antibacterial compounds other than lactic acid. The aim of this study was to characterize and fractionate the LCFS and to determine antibacterial effects of the LCFS fraction against G. vaginalis in-vitro and ameliorative effect on G. vaginalis-induced BV in mice.
Preparation of the LCFS. Culture broth of L. paracasei CH88 was centrifuged at 10,000 × g for 30 min at 4 °C. After centrifugation, the supernatant was adjusted to pH 7.0 with 5 M NaOH and filtered using a sterilized bottle-top vacuum filter (0.22 μm; Merck, Darmstadt, Germany). The filtrate (LCFS) was stored at − 80 °C for further experiments. Prior to every experiment, we confirmed that no colony was observed when the LCFS was spread and incubated on MRS agar plate at 37 °C for 24 h.

Characterization of the LCFS
Thermal treatment and pH adjustment of the LCFS. To determine thermal stability on antibacterial activity of the LCFS against G. vaginalis, G. vaginalis was inoculated at 1% (v/v) in a 96-well plate with the BHIS broth containing 5% (v/v) LCFS heated for 30 min at 60, 80, 100, or 120 °C. G. vaginalis was inoculated in the BHIS broth containing 5% (v/v) MRS broth as control. The plate was incubated at 37 °C for 24 h and its optical density (OD) was measured at 600 nm using a microplate reader (SpectraMax 190, Molecular Devices, San Jose, CA, USA) under anaerobic condition (85% N 2 , 10% H 2 , and 5% CO 2 ).
To evaluate the effect of pH on antibacterial activity of the LCFS against G. vaginalis, pH of the LCFS was adjusted to 3.0, 5.0, 7.0, 9.0, 11.0, or 13.0 with 5 M NaOH or 5 M HCl. G. vaginalis was inoculated at 1% (v/v) in a 96-well plate with the BHIS broth containing 5% (v/v) LCFS with different pH. Antibacterial activity of each sample was examined as described above.

Size-exclusion filtration of the LCFS.
To determine molecular weight of the active fraction, Amicon Ultra-15 Centrifugal Filter (Merck) with molecular weight cut-offs (MWCO) of 3, 10, 30, 50, and 100 kDa was used. Fractions of over 100 kDa, 50-100 kDa, 30-50 kDa, 10-30 kDa, 3-10 kDa, and below 3 kDa were obtained according to the manufacturer's instruction. G. vaginalis was inoculated at 1% (v/v) in a 96-well plate with the BHIS broth added with 5% (v/v) LCFS fractions with different molecular weights. The plate was incubated at 37 °C for 24 h and its OD was measured at 600 nm under anaerobic condition (85% N 2 , 10% H 2 , and 5% CO 2 ). The fraction below 3 kDa of the LCFS (< 3 kDa LCFS), which had antibacterial activity, was used for further study.
Protein precipitation and plate count. The LCFS was mixed with ammonium sulfate (70%, (w/v); Samchun Pure Chemicals, Pyeongtaek, Korea), followed by stirring the mixture for 12 h at 4 °C. The mixture was centrifuged at 10,000 × g at 4 °C for 1 h. Precipitated proteins were suspended with phosphate buffered saline (PBS) and then dialyzed for 48 h using Pur-A-Lyzer Mega Dialysis Kit (Sigma-Aldrich Co.) with a MWCO of 1 kDa in distilled water at 4 °C. Collected proteins were filtered using a 3 kDa MWCO Amicon Ultra Centrifugal Filter to obtain protein fraction over 3 kDa of the LCFS (> 3 kDa LCFSP) and protein fraction below 3 kDa of the LCFS (< 3 kDa LCFSP). The fractionated protein fractions were lyophilized and suspended with PBS at a concentration of 1 mg/mL. G. vaginalis (1.1 × 10 6 colony forming unit (CFU)/mL) was inoculated in a 24-well plate with the BHIS broth, in which each 200 μL protein fraction was added. G. vaginalis was inoculated in the BHIS broth, in which 200 μL precipitated protein fraction (1 mg/mL) of MRS broth was added as control. The plate was incubated for 24 h at 37 °C under anaerobic condition (85% N 2 , 10% H 2 , and 5% CO 2 ). After incubation, the suspension was spread on tryptic soy broth (TSB; BD) supplemented with 1.5% (w/v) bacto agar (BD), 5% (v/v) horse blood (Synergy Innovation Co., Ltd., Seongnam, Korea), 0.02% (v/v) menadione solution (BDH Chemicals Ltd., Poole, UK), and 1% (v/v) hemin solution (Sigma-Aldrich Co. BV was induced in accordance to the method of Jang et al. 29 with minor modifications. After 1-week acclimatization, mice were divided into five groups with 6 mice each and all treatments were executed for 10 days. β-Estradiol 3-benzoate (0.5 mg; Sigma-Aldrich Co.) was diluted in 100 μL filter-sterilized sesame oil (Sigma-Aldrich Co.). The β-estradiol 3-benzoate solution was injected intraperitoneally 72 h before inducing BV and administered weekly thereafter 30 . On the day of BV induced, G. vaginalis (1.1 × 10 8 CFU/20 μL PBS) was inoculated intravaginally. The control group was intravaginally treated with 20 μL PBS instead of G. vaginalis suspension. Each day for 6 days after infection, the control and G. vaginalis-infected group were intravaginally treated with 20 μL PBS. 100 μg metronidazole (Sigma-Aldrich Co.) was diluted in 20 μL sterilized PBS and injected intravaginally each day for 6 days after infection. L. paracasei CH88 (1.1 × 10 10 CFU/20 μL PBS) was intravaginally injected each day for 6 days after infection. The lyophilized < 3 kDa LCFSP was diluted in sterilized PBS at a concentration of 1 mg/mL, 20 μL of which was then intravaginally injected each day for 6 days after infection. All mice were anesthetized with 4% isoflurane before intravaginal injection. Mice were euthanized with CO 2 after 24 h of final administration. Vagina was washed with 100 μL sterilized PBS to obtain vaginal fluid and then excised. The excised vagina was fixed in 10% neutral buffered formalin solution (Sigma-Aldrich Co.) for histological examination and the other was stored at − 80 °C for enzyme-linked immunosorbent assay (ELISA) analysis.
Homogenates were centrifuged at 13,000 × g at 4 °C for 30 min and cytokine levels of the supernatants were examined by measuring OD at 450 nm.
Statistical analysis. All statistical analyses were conducted using IBM SPSS Statistics 26.0 (Chicago, IL, USA). Results were represented as means and standard errors of the means (SEM). One-way analysis of variance (ANOVA) with Dunnett's multiple comparison test (p < 0.05) or Duncan's multiple range test (p < 0.05), and Kruskal-Wallis test with Dunn's test (adjusted p < 0.05) were used to determine statistical significance between the groups.

Results
The viability of G. vaginalis was significantly reduced after 24 h incubation with the LCFS (Fig. 1a). Inhibitory activities of the LCFS against the growth of G. vaginalis depended on pH (Fig. 1b). The LCFS with pH 5.0 had the least activity. However, the growth curve of G. vaginalis treated with the LCFS with pH 5.0 exhibited delayed exponential phase compared to the control group, suggesting that the LCFS with pH 5.0 still had inhibitory effect ( Supplementary Fig. 1b). The most active inhibition was observed in the LCFS with pH 11.0 and 13.0. These results suggest that antibacterial compound or compounds might be active potently at pH 11.0 and 13.0. Among the fractions filtered with different MWCO membranes (< 3, 3-10, 10-30, 30-50, 50-100, and > 100 kDa), the < 3 kDa LCFS had the most antibacterial activity against G. vaginalis (Fig. 2). Furthermore, the < 3 kDa LCFS also delayed the exponential phase of G. vaginalis (Supplementary Fig. 2).
The < 3 kDa LCFS treated with protease, trypsin, and α-chymotrypsin showed a tendency to lose their antibacterial activity against G. vaginalis after 24 h incubation (Fig. 3). Among them, the < 3 kDa LCFS treated with trypsin or protease showed higher OD values in exponential phase than the others (Supplementary Fig. 3).
We investigated whether protein precipitates from the LCFS, including the > 3 kDa LCFSP and the < 3 kDa LCFSP, had antibacterial activity against G. vaginalis. The CFU of G. vaginalis in the < 3 kDa LCFSP-treated group was significantly lower than the control and > 3 kDa LCFSP-treated groups (p < 0.05) (Fig. 4a). Both of the LCFSP reduced the biofilm formation of G. vaginalis compared to the control group (Fig. 4b). Especially, the < 3 kDa LCFSP significantly more inhibited the biofilm formation than the > 3 kDa LCFSP (p < 0.05).
The CFU of G. vaginalis were significantly reduced in vaginal fluid of the mice to which L. paracasei CH88, < 3 kDa LCFSP, or metronidazole was intravaginally administered (Fig. 5). Vaginal tissue of the G. vaginalis group showed increased exfoliation of vaginal epithelial cells and thickness of transitional epithelium in H&E www.nature.com/scientificreports/ staining (Fig. 6). On the other hand, the L. paracasei CH88 or < 3 kDa LCFSP group showed less vaginal epithelial cell exfoliation and thinner transitional epithelium than the G. vaginalis group. Intravaginal treatment of L. paracasei CH88 or the < 3 kDa LCFSP tended to reduce the levels of IL-1β, IL-6, and TNF-α in vaginal tissue of the mice (Fig. 7). IL-10 production was higher in the CH88 group than in the G. vaginalis and metronidazole groups.  www.nature.com/scientificreports/

Discussion
Antibacterial effect of cell-free supernatant fractions of L. paracasei CH88 against G. vaginalis was investigated and characterized in-vitro and in-vivo. The LCFS still had antibacterial activity after thermal treatment at the various temperatures, indicating that antibacterial compound or compounds in the LCFS were stably active even when treated at 120 °C for 30 min (Supplementary Fig. 1a). Bacteriocin fraction from L. paracasei SD1 cell-free supernatant lost their activity completely after thermal treatment at 120 °C for 20 min 32 . However, L. paracasei HD1-7 cell-free supernatant maintained their activity after thermal treatment at 121 °C for 20 min 33 . Class II bacteriocin, one of food preservatives, is known to be stable during thermal food processing 34 . The LCFS showed the highest antibacterial activity against G. vaginalis at pH 11.0 and pH 13.0 in this study. Antibacterial activity of L. paracasei ZFM54 bacteriocin was significantly reduced at pH 5.0 [e], which was consistent with our result. However, bacteriocin fraction from L. paracasei SD1 cell-free supernatant showed the highest antibacterial activity at pH 5.0 and 6.0 32 . The most active inhibition of L. paracasei HD1-7 cell-free supernatant was observed at pH 3.0, but its antibacterial activity was inactivated at pH 9.0 33 . The < 3 kDa LCFS had the most antibacterial activity against G. vaginalis in this study, suggesting that the antibacterial compound or compounds in the LCFS pass through 3 kDa MWCO membrane. Antibacterial compounds below 10 kDa, such as class II bacteriocins and bacteriocin from Lactobacillus spp., have been reported to easily permeate cell wall causing cell wall leakage [35][36][37] , leading to irreversible alteration of the cellular membrane.
After proteolytic enzyme treatment on the < 3 kDa LCFS, antibacterial activity against G. vaginalis was reduced. These results suggest that antibacterial compound or compounds in the < 3 kDa LCFS might be mainly composed of proteins and easily hydrolyzed by trypsin or protease, which prevents its accumulation in body unlike antibiotics 38 .
Among the LCFSP, < 3 kDa LCFSP had antibacterial activity in this study. Numerous studies identified inhibitory activity of protein precipitates secreted by various bacteria. Protein precipitates from cell-free supernatant of L. acidophilus KS400, Bacillus amyloliquefaciens, and B. coagulans had inhibitory activities against the growth of G. vaginalis and their molecular weights ranged from 3 kDa to 7.5 kDa [39][40][41] . In this study, antibiofilm compound or compounds in the LCFSP mainly existed in the < 3 kDa LCFSP. G. vaginalis develops a dense biofilm adherent to vaginal epithelium in the women who suffer from BV 42 . Biofilm offers G. vaginalis a favorable environment providing nutrients and protecting it from antimicrobial agents 43 . Therefore, preventing the biofilm formation of G. vaginalis might be a primary strategy to treat BV and to prevent its recurrence 43 . However, it was reported that antibiotics such as metronidazole and tobramycin have no significant disruptive effect on G. vaginalis biofilm in-vitro 44 . Saunders et al. reported that L. reuteri RC-14 and L. rhamnosus GR-1 could reduce G. vaginalis biofilm, suggesting that biosurfactants produced by these lactobacilli might contribute to inhibiting biofilm formation 45 . It    29,31 . Jang et al. reported that intravaginal administration of cell-free supernatant from L. fermentum SNUV175 and L. crispatus SNUV220 had ameliorative effects on vulvovaginal candidiasis in murine model by reducing the CFU of Candida albicans, which is also known as a representative pathogenic yeast causing vaginal disorder 47 . Lactobacilli have been known to be able to inhibit the growth of G. vaginalis by producing antimicrobial compounds as well as lactic acid 23,24 . Since pH of the < 3 kDa LCFSP was set at 7.0, it could be presumed that antibacterial activity of L. paracasei CH88 against G. vaginalis might be mainly attributed to unidentified compounds in the < 3 kDa LCFSP. Intravaginal treatment of G. vaginalis increased thickness of vaginal transitional epithelium in this study. Increased thickness of transitional epithelium in vagina was reported to be related to epithelial proliferation, which results in cell exfoliation 48 . Exfoliated vaginal epithelial cells covered with adherent anaerobic bacteria have been commonly observed in vaginal fluid of the women who have BV, but not in the women who do not 49 . Our results suggest that L. paracasei CH88 or the < 3 kDa LCFSP reduced G. vaginalis-induced vaginal epithelial cell exfoliation. Intravaginal inoculation of G. vaginalis was reported to induce immune response, resulting in elevating the levels of IL-1β, IL-6, IL-8, and IL-10 in cervicovaginal fluid 50 . L. paracasei CH88 and < 3 kDa LCFSP-treated group showed lower IL-1β, IL-6, TNF-α level compared to G. vaginalis group. Previous studies also reported that intravaginal and oral administration of L. johnsonii HY7042, L. rhamnosus HN001, and L. acidophilus GLa-14 tended to reduce the levels of IL-1β, IL-6, and TNF-α, while increasing IL-10 production in mice with G. vaginalis-induced BV 29,31 , which were consistent with our results.

Conclusion
The antibacterial effect of L. paracasei CH88 may be due to unidentified proteins in the fraction below 3 kDa of L. paracasei CH88 cell-free supernatant. L. paracasei CH88 or < 3 kDa LCFSP could ameliorate BV via not only inhibiting the growth of G. vaginalis and its biofilm formation, but also reducing exfoliation of vaginal cells and production of pro-inflammatory cytokines. However, this study has some limitations that in-vitro and in-vivo antibacterial experiments against F. vaginae and P. bivia should be followed since BV is a polymicrobial infection. Moreover, safety assessment must be followed to apply this treatment to humans.

Data availability
The data presented in this study are available on reasonable request and for non-commercial purposes.