DNA Fingerprinting of Lactobacillus crispatus Strain CTV-05 by ...

JOURNAL OF CLINICAL MICROBIOLOGY, May 2003, p. 1881–1887 0095-1137/03/$08.00⫹0 DOI: 10.1128/JCM.41.5.1881–1887.2003 Copyright © 2003, American Society for Microbiology. All Rights Reserved.

Vol. 41, No. 5

DNA Fingerprinting of Lactobacillus crispatus Strain CTV-05 by Repetitive Element Sequence-Based PCR Analysis in a Pilot Study of Vaginal Colonization May A. D. Antonio1 and Sharon L. Hillier1,2* Magee-Womens Research Institute1 and Department of Obstetrics, Gynecology, and Reproductive Sciences, University of Pittsburgh,2 Pittsburgh, Pennsylvania Received 9 October 2002/Returned for modification 15 December 2002/Accepted 14 February 2003

Lactobacillus crispatus is one of the predominant hydrogen peroxide (H2O2)-producing species found in the vagina and is under development as a probiotic for the treatment of bacterial vaginosis. In this study, we assessed whether DNA fingerprinting by repetitive element sequence-based PCR (rep-PCR) can be used to distinguish the capsule strain of L. crispatus (CTV-05) from other endogenous strains as well as other species of vaginal lactobacilli. Vaginal and rectal lactobacilli were identified to the species level by using wholechromosome probe DNA hybridization. The DNAs from L. crispatus, L. jensenii, L. gasseri, and an as-yetunnamed H2O2-negative Lactobacillus species designated 1086V were subjected to rep-PCR. The results of gel electrophoresis and ethidium bromide staining of the DNA fingerprints obtained were compared. L. crispatus CTV-05 had a unique DNA fingerprint compared to all other lactobacilli. DNA fingerprints for 27 production lots of L. crispatus sampled from 1994 through 2001 were identical to that of the original strain isolated in 1993, suggesting strain stability. In a pilot study of nine women, this DNA fingerprinting method distinguished CTV-05 from other endogenous vaginal lactobacilli prior to and after vaginal capsule use. rep-PCR DNA fingerprinting is useful for strain typing and for evaluating longitudinal loss or acquisition of vaginal lactobacilli used as probiotics. vaginal flora (33). However, commercially available products sold as dietary supplements do not always contain the Lactobacillus species advertised on the label and do not always contain H2O2-producing lactobacilli (15). Furthermore, compared to vaginal lactobacilli, yogurt-derived lactobacilli poorly adhere to vaginal epithelial cells in vitro (44). Studies evaluating the efficacy of some probiotic products containing lactobacilli remain inconclusive because the strain identification techniques have been inadequate, relying on classic phenotypic identification methods, such as sugar fermentation, to distinguish lactobacilli isolated before and after product use (7, 36). In one study, efficacy was revealed by comparing the clinical cure rate for BV in a group using the Lactobacillus product to that in a placebo group (12), while in another study, Nugent scoring for BV Gram staining was used to assess the outcomes for the treatment groups (32). Only one study to date has provided a DNA-based strain tracking method; in that study, random amplified polymorphic DNA-PCR analysis (8) was used to compare endogenous vaginal Lactobacillus strains to the probiotic strain. L. crispatus strain CTV-05, a vaginally derived H2O2-producing strain, is a probiotic that is being evaluated for the treatment and prevention of BV. The vehicle for the probiotic is a gelatin capsule which is inserted vaginally. The purpose of this study was to evaluate whether repetitive element sequence-based PCR (rep-PCR), a DNA fingerprinting technique, can be used to distinguish probiotic L. crispatus strain CTV-05 from other L. crispatus strains and other endogenous vaginal Lactobacillus species. rep-PCR, which makes use of repetitive sequences dispersed throughout a bacterial genome for the direct amplification of genomic DNA, generates bacterial species- or strain-specific fingerprint patterns (42).

Lactobacilli, the predominant group of microorganisms of the vaginal flora of healthy women, function as endogenous microbicides through the production of lactic acid, which acidifies the vagina, and hydrogen peroxide (H2O2), which reacts with myeloperoxidase to form reactive molecules toxic to human immunodeficiency virus (HIV) and other pathogens (18, 19). Bacterial vaginosis (BV), which is characterized microbiologically by the reduction or absence of lactobacilli, has been associated with an increased prevalence of HIV in several cross-sectional studies (2, 6, 17, 22, 25, 34, 35, 39) and an increased acquisition of HIV in one longitudinal study (40). Further, BV has been linked with increased viral shedding among HIV-infected women (5). In a longitudinal study, women without vaginal lactobacilli had an increased risk of HIV and gonorrhea acquisition (21), suggesting that a lack of vaginal lactobacilli may predispose women to sexually transmitted diseases, including HIV. Further, the absence of H2O2producing lactobacilli has been associated with an increased acquisition of BV (13, 21). Lactobacillus crispatus and L. jensenii, both H2O2-producing species (1), are the predominant vaginal Lactobacillus species colonizing women without BV in Europe (11), the United States (1, 41), and Japan (38). Lactobacillus-containing probiotic products have been proposed for the treatment of vaginal infections (3, 24, 29, 30, 37). Oral administration of a probiotic containing L. rhamnosus GR-1 and L. fermentum RC-14 once or twice daily for 28 days has been correlated with healthy * Corresponding author. Mailing address: University of Pittsburgh, Department of Obstetrics, Gynecology, and Reproductive Sciences, Magee-Womens Hospital, 300 Halket St., Pittsburgh, PA 15213-3180. Phone: (412) 641-6435. Fax: (412) 641-1133. E-mail: slh6⫹@pitt.edu. 1881

1882

ANTONIO AND HILLIER MATERIALS AND METHODS

Sources of strains. A total of 67 strains of L. crispatus, 41 strains of L. gasseri, 32 strains of L. jensenii, and 28 strains of an as-yet-unnamed species designated Lactobacillus sp. strain 1086V (1) were chosen and subjected to rep-PCR for comparison of their DNA fingerprints to the vaginal capsule L. crispatus CTV-05 DNA fingerprint. These lactobacilli were from a subset of previously identified lactobacilli isolated from vaginal and rectal cultures for 319 women visiting an adolescent medicine clinic and two sexually transmitted disease clinics in Seattle, Wash. (1). These species were chosen because of their vaginal prevalence (1). The DNA fingerprints for 24 Lactobacillus strains obtained from the American Type Culture Collection (ATCC) and representing 17 different Lactobacillus species of human origins were evaluated: L. acidophilus 4356, 4357, and 521; L. brevis 11577 and 14869; L. buchneri 11579 and 4005; L. casei subsp. casei 393; L. crispatus 33197; L. fermentum 23271; L. gasseri 9857 and 4963; L. jensenii 25258; L. johnsonii 33200; L. oris 49062; L. parabuchneri 49374; L. paracasei subsp. paracasei 27216; L. reuteri 23272; L. rhamnosus 7469 and 21052; L. ruminis 25644 and 27780; L. salivarius subsp. salivarius 11741; and L. vaginalis 49540. In addition, six other ATCC strains, L. alimentarius 29643, L. delbrueckii subsp. bulgaricus 11842, L. delbrueckii subsp. delbrueckii 9649, L. delbrueckii subsp. lactis 12315 and 4797, and L. plantarum 14917, were tested for comparison to strain CTV-05. Lactobacillus identification. Lactobacilli were identified to the genus level by Gram staining, colony morphology, negative catalase test, and production of lactic acid, as assessed by gas chromatographic analysis (14). H2O2 production was detected by inoculating tetramethylbenzidine plates with lactobacilli (29); the plates were incubated under anaerobic conditions for 48 h and then exposed to ambient air. Lactobacillus strains were preserved in approximately 400 ␮l of reconstituted litmus milk (Becton Dickinson Microbiology Systems, Cockeysville, Md.) and frozen at ⫺70°C. The strains were transported to Magee-Womens Research Institute, Pittsburgh, Pa., where the DNA fingerprinting studies were performed. Each vial of litmus milk was thawed at room temperature, and the contents were used to inoculate Columbia 5% sheep blood agar (BA) plates (PML Microbiologicals, Mississauga, Ontario, Canada). The plates were incubated at 36°C in 5 to 6% CO2 for a minimum of 24 h. Each strain was passed at least one more time to ensure purity and vigor prior to experimental procedures and restocking. Genomic DNA isolation and whole-chromosomal probe DNA hybridization, described previously (1), were performed to identify lactobacilli to the species level. Briefly, this procedure involved bacterial lysis with various enzymes (see below), phenol-chloroform DNA extraction, ethanol precipitation for DNA purification, and slot blot DNA hybridization with whole-chromosome probes. Small-volume DNA extraction for rep-PCR. The previously described DNA protocol (1) was modified to prepare DNA samples for rep-PCR. This modified protocol provided smaller volumes for easier manipulation of a large number of samples and avoidance of organic solvents. A 1.5-ml microcentrifuge tube containing 200 ␮l of lysis buffer (25% ultrapure sucrose, 50 mM Tris, 1 mM EDTA [pH 8.0]) was inoculated with a loopful of bacteria grown on BA plates. The loopful contained approximately 108 to 109 CFU. The sample was stored at ⫺20°C until needed for extraction. The following enzyme reaction mixtures were incubated at 60°C for at least 1 h. First, 75 ␮l of protease from Streptomyces griseus (1.0 mg/ml; Sigma, St. Louis, Mo.) was added. Then, 5 ␮l of RNase A (10 mg/ml; Boehringer Mannheim Biochemicals, Indianapolis, Ind.) was added. Finally, 5 ␮l of proteinase K (20 mg/ml; Sigma) was added. The following DNA extraction procedure avoided the use of organic solvents (26). A volume of 20% lauryl sulfate (Sigma) was added to the sample for a final concentration of 0.34% and then mixed. Then, 5 M NaCl was added for a final concentration of 1.8 M. The sample was vortexed and then centrifuged at 10,000 ⫻ g for 1 min. A double volume of 95% ethanol was added. The sample was again vortexed and then centrifuged at 60,000 ⫻ g for 1 min. After the supernatant was decanted, the pellet was washed with 70% ethanol. The sample was centrifuged at 10,000 ⫻ g for 1 min. After the ethanol was decanted, the pellet was air dried. The pellet was resuspended in 100 ␮l of 10 mM Tris–1 mM EDTA (pH 8.0). The DNA sample was stored at ⫺20°C until needed for repPCR processing. rep-PCR. The rep-PCR protocol used was based on the method of Versalovic et al. (42). The two repetitive extragenic palindromic (REP) consensus primers used for the PCR at 50 pmol each were REP1R-Dt (3⬘-CGGNCTACNGCNG CNIII-5⬘) and REP2-Dt (3⬘-CATCCGGNCTATTCNGCN-5⬘), where N is A, G, C, and T (GibcoBRL, Life Technologies, Gaithersburg, Md.). Of the available REP primers, these two were successfully applied to other gram-positive micro-

J. CLIN. MICROBIOL. organisms (43). The 25-␮l PCR volume also consisted of 10% dimethyl sulfoxide; 1.2 mM each dATP, dGTP, dCTP, and dTTP; 7 mM MgCl2; 2.5 U of recombinant Taq DNA polymerase (GibcoBRL); and PCR buffer. Either 100 ng of phenol-chloroform-extracted DNA sample or 7 ␮l of small-volume DNA extraction sample was used. The samples were placed in a GeneAmp PCR system 9600 thermocycler (Perkin-Elmer Cetus, Norwalk, Conn.). A modification to the method of Versalovic et al. (42) was the addition of ramp times. The PCR program started with an initial denaturation at 95°C for 7 min. This step was followed by 32 cycles of a 50-s ramp to 90°C for 30 s of denaturation, a 62-s ramp to 40°C for 1 min of annealing, and a 50-s ramp to 65°C for 8 min of extension. This step was followed by a final extension at 65°C for 16 min. The samples were stored at 4°C. An 8-␮l volume of each sample was electrophoresed in 0.8% SeaKem GTG agarose (FMC BioProducts, Rockland, Maine) in Tris-acetate-EDTA running buffer. Each gel was run with a 1-kb DNA ladder. The gels were stained with ethidium bromide and photographed. The photographs were viewed for the intense bands observed at 1,500, 1,650, and 2,400 bp in the capsule strain L. crispatus CTV-05. When these three major bands were present, two additional minor bands (1,950 bp and a doublet at 2,100 bp) were identified. Identity to L. crispatus CTV-05 required the presence of all five bands. Positive identification of L. crispatus CTV-05 was reported only when two independent readers agreed. Samples yielding discrepant rep-PCR interpretations were retested. Discrepant results were unusual, with only 3 of 380 consecutively evaluated Lactobacillus strains yielding different interpretations between two readers. The reproducibility of the rep-PCR method was also quite high in that the fingerprint patterns from the ATCC type strains run on different dates by two different technicians had the same major bands. Gelatin capsules containing L. crispatus CTV-05. Samples from various lots of gelatin capsules containing 108 CFU of CTV-05 were manufactured either at Chrisope Technologies, Lake Charles, La., or Gynelogix, Incorporated, Louisville, Colo. The capsules were sent to the laboratory for species identification and fingerprinting. Chrisope Technologies also supplied the gelatin capsules used in the human pilot study. The capsules were dissolved in a 1-ml solution of peptone yeast extract broth basal medium, 1% (wt/vol) dextrose, 1% (wt/vol) soluble starch, and 0.02% (vol/vol) Tween 80 (14). A drop of the broth was plated on a BA plate to check for purity. DNA was extracted from the lactobacilli and subjected to rep-PCR. The rep-PCR fingerprints were compared to that of original strain CTV-05. Pilot study. Nine female volunteers 18 to 40 years old were recruited. The Human Subjects Committee at the University of Washington approved the protocol and consent documents. None of the women had clinical signs or symptoms of BV. Women who were currently menstruating, pregnant, or using systemic antibiotics were excluded. Vaginal smear samples were collected at study entry and at the two follow-up visits. These smear samples were Gram stained and evaluated with Nugent criteria (27). Each woman inserted one gelatin capsule containing 108 CFU of L. crispatus CTV-05 twice daily for 3 days. Vaginal swab specimens were also collected at baseline and at 4 to 6 days and 9 to 11 days after enrollment. The specimens were used to inoculate one BA plate, one Rogosa agar (Difco Laboratories, Detroit, Mich.) plate, and two human blood bilayer Tween agar (PML Microbiologicals) plates. One of the human blood bilayer Tween agar plates and the BA plate were incubated at 36°C with 5 to 6% CO2, and the remaining plates were incubated at 36°C in an anaerobic chamber. Lactobacilli were identified to the genus level as described above. The isolates were stored at ⫺70°C in reconstituted litmus milk until they were transported to the infectious disease laboratory at the Magee-Womens Research Institute. Lactobacilli were identified from eight of the nine women at baseline and from all women at the follow-up visits. From the 27 visits, 66 Lactobacillus strains were recovered. All lactobacilli were tested for homology to L. crispatus by using whole-chromosomal probes as previously described (1). rep-PCR was performed for identification to the strain level.

RESULTS DNA fingerprints of L. crispatus strains were most distinguishable from each other in the 1,500- to 2,400-bp range (Fig. 1A). In this range, the fingerprint pattern of L. crispatus CTV-05 showed intense bands at approximately 1,500, 1,650, and 2,400 bp. Another band appeared at approximately 1,900 bp, and a doublet appeared at 2,100 bp (Fig. 1A, lane 2). In addition to the strains recovered from the nine women and

VOL. 41, 2003

rep-PCR DNA FINGERPRINTING OF L. CRISPATUS CTV-05

1883

FIG. 1. rep-PCR DNA fingerprints generated from Lactobacillus ATCC strains and vaginal isolates. The vaginal isolates were previously identified to the species level by using whole-chromosome probes produced from Lactobacillus ATCC strains. Each gel is representative of isolates having DNA homology to one Lactobacillus species. Clinical isolates within a group of Lactobacillus species were from different women. L. crispatus, L. jensenii, Lactobacillus sp. strain 1086V, and L. gasseri were chosen because of their vaginal prevalence. Lane 1 of each gel contains the 1-kb DNA size ladder. (A) Isolates homologous to the whole-chromosome probe of L. crispatus ATCC 33197. L. crispatus CTV-05 and ATCC 33197 are shown in lanes 2 and 3, respectively. Lanes 4 to 11 show the rep-PCR patterns for L. crispatus isolates from eight women. (B) Isolates homologous to the whole-chromosome probe of L. jensenii ATCC 25258. ATCC 25258 was run in lane 2. Lanes 3 to 12 show the rep-PCR patterns for 10 L. jensenii isolates, all from different women. (C) Isolates homologous to the Lactobacillus sp. strain 1086V whole-chromosome probe. The original Lactobacillus sp. strain 1086V is shown in lane 2. Lanes 3 to 12 show the rep-PCR patterns for Lactobacillus sp. strain 1086V-like isolates from nine women. (D) Isolates homologous to the whole-chromosome probe of L. gasseri ATCC 4963. L. gasseri ATCC 4963 and 9857 are shown in lanes 2 and 3, respectively. The rep-PCR DNA patterns for L. gasseri isolates from eight women are shown in lanes 4 to 11.

shown in Fig. 1A, 58 additional strains of L. crispatus evaluated by rep-PCR did not show identity to L. crispatus CTV-05. This result suggested that the L. crispatus strain chosen for development as a probiotic had a distinct fingerprint pattern not observed among 67 other vaginal L. crispatus strains. DNA fingerprints of other endogenous vaginal Lactobacillus species, L. jensenii, L. gasseri, and Lactobacillus sp. strain 1086V (Fig. 1B to D) as well as 24 Lactobacillus ATCC strains (data not shown) were distinguishable from that of L. crispatus CTV-05. Typical fingerprints of isolates homologous to the L. jensenii whole-chromosomal probe are shown in Fig. 1B. Less heterogeneity was observed among a total of 32 L. jensenii strains than among the L. crispatus wild-type strains. Fingerprints of strains homologous to Lactobacillus sp. strain 1086V are shown in Fig. 1C. A total of 28 isolates of this H2O2negative vaginal strain showed different rep-PCR fingerprints. Additionally, L. gasseri ATCC 4963 and 9857 and 41 clinical isolates having DNA homology to these two ATCC strains were distinguishable by rep-PCR fingerprinting. DNA fingerprints of 11 L. gasseri clinical isolates and the ATCC strains are

shown in Fig. 1D. A total of 24 ATCC strains, 17 of which are human associated, and 168 clinical isolates were evaluated by rep-PCR, and none showed identity to L. crispatus CTV-05. Various lots of L. crispatus CTV-05 produced between 1994 and 2001 had rep-PCR DNA fingerprints similar to that of the original strain (Fig. 2). This result suggested stability of the probiotic strain and reproducibility of the protocol. Changes in production processes for L. crispatus CTV-05 did not affect the genetic fingerprint of the microorganism. In order to evaluate whether the use of L. crispatus capsules intravaginally resulted in colonization, nine women without genital infections were recruited for a pilot study. Women used the capsules intravaginally twice daily for 3 days. None of the nine women had strains of lactobacilli with DNA fingerprints similar to that of L. crispatus CTV-05 at baseline (Table 1). Although four of the nine women were colonized by H2O2producing L. crispatus at baseline, none of the nine women had strains of lactobacilli with DNA fingerprints similar to that of L. crispatus CTV-05. At the first follow-up visit, at 4 to 6 days after enrollment, all five of the women who were not colonized

1884

ANTONIO AND HILLIER

FIG. 2. rep-PCR DNA fingerprints generated from different production lots of L. crispatus CTV-05 gelatin capsules produced from 1995 to 2001(lane 3, LB 045—January 1995; lane 4, LB 096—September 1996; lane 5, LB 112—June 1997; lane 6, LB 124—May 1998; and lane 7, LB 227—2001). The original strain CTV-05, isolated in 1993, was run in lane 2, and a 1-kb DNA size ladder was run in lane 1.

by L. crispatus at baseline were positive for L. crispatus CTV05. At the second follow-up visit, at 9 to 11 days, four of the five women remained vaginally colonized by the capsule strain. Of the four women already colonized by H2O2-producing L. crispatus at baseline, two became colonized by the capsule strain, but only at the second follow-up visit. Thus, vaginal colonization by the suppository strain was somewhat less successful among women already colonized by H2O2-producing L. crispatus. Representative gels are shown for two of the women in the pilot study in Fig. 3. For one woman, the L. crispatus strain colonizing the vagina at baseline (Fig. 3A, lane 3) persisted over all three visits (lanes 4 to 6, 8, and 9). Nevertheless, L. crispatus CTV-05 established colonization, although it was not detected by culturing until the third visit (Fig. 3A, lanes 11 and 12). Interestingly, this woman also acquired colonization by L. jensenii (Fig. 3A, lane 7), which persisted at the third visit (lane 10). A second woman was initially colonized by an H2O2producing strain (Fig. 3B, lanes 3 and 4) and remained colonized by that strain throughout the study (lanes 5 and 10). L. crispatus CTV-05 established colonization at the first follow-up visit (Fig. 3B, lane 7) and persisted through the third visit (lanes 8 and 9). An H2O2-negative strain was also detected at the second (Fig. 3B, lane 6) and third (lane 11) visits. These data suggest that colonization by L. crispatus CTV-05 does not necessarily result in displacement of other colonizing strains. DISCUSSION In this study, genomic fingerprinting based on rep-PCR was used to distinguish a probiotic strain, L. crispatus CTV-05, from other vaginal lactobacilli and 24 ATCC strains. The pilot study demonstrated successful colonization by L. crispatus CTV-05 in seven of nine women. Both women who were not

J. CLIN. MICROBIOL.

successfully colonized by L. crispatus CTV-05 were colonized by H2O2-producing L. crispatus at baseline, suggesting that colonization by an exogenous strain of L. crispatus may be less successful in persons already having predominant H2O2-producing lactobacilli. Nevertheless, colonization by the probiotic strain did not result in displacement of other endogenous lactobacilli. This study demonstrates the feasibility of colonizing the vagina with a probiotic strain of L. crispatus and the value of rep-PCR for fingerprinting of the probiotic strain. In a review of the use of probiotic lactobacilli, McGroarty addressed the need for DNA fingerprinting techniques for determination of whether a probiotic strain of lactobacilli colonizes the vaginal microflora (23). It is also necessary to differentiate the probiotic strain from endogenous lactobacilli colonizing the vagina. Although two studies reported the results of small clinical trials of a product containing a Lactobacillus strain of human origin, the authors did not present DNA fingerprinting data to support colonization by the strain of Lactobacillus found in the product (7, 12). Instead, evaluations were based on the number of cases of genital infections detected within a treatment group before and after use of the Lactobacillus preparations (7, 12). In most instances, classic biochemical or phenotypic tests were used to identify lactobacilli found before and after product use (12, 36). Because of the lack of specificity of these phenotypic tests, genotypic techniques are now more frequently used, even for identification to the species level (4, 16). In one recent study, a ribotyping method (33) was used to identify L. rhamnosus GR-1 and L. fermentum RC-14 in a clinical trial. However, the ribotypes of endogenous vaginal lactobacilli were not compared to the ribotypes of L. rhamnosus GR-1 and L. fermentum RC-14. According to Zhong et al., ribotyping appears to distinguish isolates only at the species level (46). More recently, Gardiner et al. used random amplified polymorphic DNA-PCR to track three probiotic strains (8). However, it was necessary for them to implement pulsedfield gel electrophoresis to distinguish their L. rhamnosus probiotic strain from closely related strains of the species. In our pilot study of an L. crispatus vaginal capsule, women were monitored for only 11 days. Vaginal insertion of CTV-05 gelatin capsules was done twice daily for only 3 days, so there

TABLE 1. Detection of L. crispatus strain CTV-05 by rep-PCR DNA fingerprinting at various times rep-PCR CTV-05 resulta at: Subject

101 102 103 104 105 106 107 108 109 a

Baseline

⫺ ⫺ ⫺ ⫺b ⫺ ⫺ ⫺b ⫺b ⫺b

Follow-up at days: 4–6

9–11

⫹ ⫹ ⫹ ⫺b ⫹ ⫹ ⫺b ⫺b ⫺b

⫹ ⫹ ⫹ ⫺b ⫹ ⫺ ⫹b ⫺b ⫹

⫹, positive for strain CTV-05; ⫺, negative for strain CTV-05. Positive for an H2O2-producing L. crispatus strain which had a DNA fingerprint different from that of L. crispatus strain CTV-05. b

VOL. 41, 2003


1885

FIG. 3. rep-PCR DNA fingerprints of lactobacilli isolated from two women in the L. crispatus CTV-05 pilot study. b, baseline visit; f1 and f2, follow-up visits 1 and 2, respectively. Each lane represents a different colony type of lactobacilli. For each gel, lane 1 is a 1-kb DNA size ladder and lane 2 is a control DNA fingerprint of L. crispatus CTV-05. (A) Patient 107. An H2O2-producing strain of L. crispatus was present at the baseline visit (lane 3) and at the first (lanes 4 to 6 and lane 8) and second (lane 9) follow-up visits. L. crispatus CTV-05 was not detected at the first follow-up visit but was detected at the third visit (lanes 11 and 12), suggesting that strain CTV-05 was present at lower levels than endogenous L. crispatus strains at the first follow-up visit. A third Lactobacillus strain having a fingerprint closely resembling the DNA fingerprints of strains homologous to L. jensenii ATCC 25258 (Fig. 1B, lane 2) was also detected at the second (lane 7) and third (lane 10) visits. (B) Patient 102. An H2O2-producing Lactobacillus strain was detected at the baseline visit (lanes 3 and 4) and at the second (lane 5) and third (lane 10) visits. Although the two H2O2-producing colony types from the baseline visit differed phenotypically, they had identical rep-PCR DNA fingerprints. Two additional strains, an H2O2-negative strain (lane 6) and L. crispatus strain CTV-05 (lane 7), were detected at the first follow-up visit. All three strains were detected at the second follow-up visit (lanes 8 to 11).

was only about a 1-week interval between the last capsule use and the follow-up visit. Vaginal colonization by CTV-05 was detected within 4 to 6 days. This time frame of colonization detection was also found in the study of the L. rhamnosus GR-1 or L. fermentum RC-14 probiotic (8). The vaginal capsule regimen of 3 days of use seems more amenable than either a regimen involving the oral intake of 3 ml of a probiotic suspension of either L. rhamnosus GR-1 or L. fermentum RC-14 twice daily for 14 days (33) or a 28-day regimen of a daily oral intake of one or two capsules of various combinations of L. rhamnosus GR-1 and L. fermentum RC-14 (32). This pilot study suggests that a short duration of vaginal capsule use may be adequate to establish colonization, although the number of subjects tested was limited and the follow-up period was short. Our decision to use rep-PCR DNA fingerprinting was based on the reported reliability of the method to type strains within a gram-positive species and to generate reproducible fingerprinting patterns (43). rep-PCR has been applied to the Lactobacillus genus at the species level. However, different sets of rep-PCR primers have been used (10). An alternative strain typing technique considered was pulsed-field gel electrophoresis, which has been shown to be more costly, labor-intensive, and time-consuming than rep-PCR (20, 28). rep-PCR DNA fingerprinting has been shown to be a useful tool for epidemiological typing (9, 28, 43). For example, repPCR DNA fingerprinting showed that strains of penicillinresistant Streptococcus pnuemoniae classified by capsular serotypes were genetically heterogeneous (43). Using rep-PCR, we

were able to determine whether different colony types had the same fingerprints. L. crispatus was chosen in this pilot study because this species has been reported to be prevalent among women with Lactobacillus-predominant vaginal microflora in North America, Europe, and Asia (1, 11, 38). Colonization by H2O2-producing Lactobacillus species has been associated with fewer diagnoses of gonorrhea (1, 13) and with a decreased acquisition of HIV (21). The production of lactic acid and catalase inhibitors by H2O2-producing lactobacilli was found to have cidal activity against Neisseria gonorrhoeae in vitro (45). Thus, the establishment of a Lactobacillus-predominant vaginal flora through the use of vaginal probiotic strains may have broad implications for public health in populations at high risk of sexually transmitted infections, including HIV. Vaginal colonization by lactobacilli is dynamic, and loss of colonization by H2O2-producing lactobacilli has been associated with frequent sexual intercourse and antibiotic usage (41). Lactobacillus species that produce H2O2, such as L. crispatus and L. jensenii, were more likely to persist over time in the vagina than H2O2-negative strains. Further, only one in three women who were not initially colonized by H2O2-producing lactobacilli became naturally colonized by H2O2-producing Lactobacillus species over the 8-month follow-up period in that study. The authors suggested that the introduction of exogenous H2O2-producing lactobacilli may be the best option to help normalize the vaginal flora of many women. If vaginal colonization is the intended consequence of probiotic use, selection of appropriate strains is critically impor-

1886

ANTONIO AND HILLIER

tant. Previous studies of Lactobacillus products for vaginal use reported the use of L. fermentum and L. rhamnosus (32, 33), both infrequent colonizers of the vagina (1, 11, 38). Although Wood et al. demonstrated that lactobacilli contained in yogurt did not adhere well to vaginal epithelial cells in vitro (44), dairy strains, such as L. acidophilus NCFM (31), are still being tested as potential vaginal probiotics. If colonization is desirable, then the use of strains capable of binding sites on vaginal epithelial cells is desirable. This pilot study demonstrated that L. crispatus CTV-05 is able to persist for several days in the vaginal ecosystem of reproductive-age women and that the rep-PCR method can be used to fingerprint the probiotic strain. A larger study population will be needed to garner more information about the rate of colonization of strain CTV-05, especially among women with BV.

J. CLIN. MICROBIOL.

17.

18. 19. 20.

21.

22.

ACKNOWLEDGMENTS This work was supported by Public Health Service grants 1U19 AI 38513, 1UO1 AI 31448, and 1UO1 AI 47785. REFERENCES 1. Antonio, M. A. D., S. E. Hawes, and S. L. Hillier. 1999. The identification of vaginal Lactobacillus species and the demographic and microbiologic characteristics of women colonized by these species. J. Infect. Dis. 180:1950– 1956. 2. Cohen, C. R., A. Duerr, N. Pruithithada, S. Rugpao, S. Hillier, P. Garcia, and K. Nelson. 1995. Bacterial vaginosis and HIV-seroprevalence among female commercial sex workers in Chiang Mai, Thailand. AIDS 9:1093– 1097. 3. Collins, J. K., G. Thornton, and G. O. Sullivan. 1998. Selection of probiotic strains for human applications. Int. Dairy J. 8:487–490. 4. Collins, M. D., B. A. Phillips, and P. Zanoni. 1989. Deoxyribonucleic acid homology studies of Lactobacillus casei, Lactobacillus paracasei sp. nov., subsp. paracasei and subsp. tolerans, and Lactobacillus rhamnosus sp. nov., comb. nov. Int. J. Syst. Bacteriol. 39:105–108. 5. Cu-Uvin, S., J. W. Hogan, A. M. Caliendo, J. Harwell, K. H. Mayer, and C. C. J. Carpenter. 2001. Association between bacterial vaginosis and expression of human immunodeficiency virus type 1 RNA in the female genital tract. Clin. Infect. Dis. 33:894–896. 6. Fonck, K., R. Kaul, F. Keli, J. J. Bwayo, E. N. Ngugi, S. Moses, and M. Temmerman. 2001. Sexually transmitted infections and vaginal douching in a population of female sex workers in Nairobi, Kenya. Sex. Transm. Infect. 77:271–275. 7. Friedlander, A., M. M. Drucker, and A. Schachter. 1986. Lactobacillus acidophilus and vitamin B complex in the treatment of vaginal infection. Panminerva Med. 28:51–53. 8. Gardiner, G. E., C. Heinemann, A. W. Bruce, D. Beuerman, and G. Reid. 2002. Persistence of Lactobacillus fermentum GR-1 but not L. rhamnosus GG in the human vagina as demonstrated by random amplified polymorphic DNA. Clin. Diagn. Lab. Immunol. 9:92–96. 9. Georghiou, P. R., A. M. Doggett, M. A. Kielhofner, J. E. Stout, D. A. Watson, J. R. Lupski, and R. J. Hamill. 1994. Molecular fingerprinting of Legionella species by repetitive element PCR. J. Clin. Microbiol. 32:2989–2994. 10. Gevers, D., G. Huys, and J. Swings. 2001. Applicapility of rep-PCR fingerprinting for identification of Lactobacillus species. FEMS Microbiol. Lett. 205:31–36. 11. Giorgi, A., S. Torriani, F. Dellaglio, G. Bo, E. Stola, and L. Bernuzzi. 1987. Identification of vaginal lactobacilli from asymptomatic women. Microbiologica 10:377–384. 12. Hallen, A., C. Jarstrand, and C. Pahlson. 1992. Treatment of bacterial vaginosis with lactobacilli. Sex. Transm. Dis. 19:146–148. 13. Hawes, S. E., S. L. Hillier, J. Benedetti, C. E. Stevens, L. A. Koutsky, P. Wolner-Hanssen, and K. K. Holmes. 1996. Hydrogen peroxide-producing lactobacilli and acquistion of vaginal infections. J. Infect. Dis. 174:1058– 1063. 14. Holdeman, L. V., E. P. Cato, and W. E. C. Moore (ed.). 1977. Anaerobe laboratory manual, 4th ed., p. 143–148. Virginia Polytechnic Institute and State University, Blacksburg. 15. Hughes, V. L., and S. L. Hillier. 1990. Microbiologic characteristics of Lactobacillus products used for colonization of the vagina. Obstet. Gynecol. 75: 244–248. 16. Johnson, J. L., C. F. Phelps, C. S. Cummins, J. London, and F. Gasser. 1980.

23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

35.

36.

37. 38.

39.

40.

41.

Taxonomy of the Lactobacillus acidophilus group. Int. J. Syst. Bacteriol. 30: 53–68. Kapiga, S. H., N. E. Sam, J. F. Shao, B. Renjifo, E. J. Masenga, I. E. Kiwelu, R. Manongi, W. Fawzi, and M. Essex. 2002. HIV-1 epidemic among female bar and hotel workers in northern Tanzania: risk factors and opportunities for prevention. J. Acquir. Immune Defic. Syndr. 29:409–417. Klebanoff, S. J., and R. W. Coombs. 1991. Viricidal effect of Lactobacillus acidophilus on human immunodeficiency virus type 1: possible role in heterosexual transmission. J. Exp. Med. 174:289–292. Klebanoff, S. J., S. L. Hillier, D. A. Eschenbach, and A. M. Waltersdorph. 1991. Control of the microbial flora of the vagina by H2O2-generating lactobacilli. J. Infect. Dis. 164:94–100. Louie, M., P. Jayaratne, I. Luchsinger, J. Devenish, J. Yao, W. Schlech, and A. Simor. 1996. Comparison of ribotyping, arbitrarily primed PCR, and pulsed-field gel electrophoresis for molecular typing of Listeria monocytogenes. J. Clin. Microbiol. 34:15–19. Martin, H. L., Jr., B. A. Richardson, P. M. Nyange, L. Lavreys, S. L. Hillier, B. Chohan, K. Mandaliya, J. O. Ndinya-Achola, J. Bwayo, and J. Kreiss. 1999. Vaginal lactobacilli, microbial flora, and risk of human immunodeficiency virus type 1 and sexually transmitted disease acquisition. J. Infect. Dis. 180:1863–1868. Mbizvo, E. M., S. E. Msuya, B. Stray-Pedersen, J. Sundby, M. Z. Chirenje, and A. Hussain. 2001. HIV seroprevalence and its associations with the other reproductive tract infections in asymptomatic women in Harare, Zimbabwe. Int. J. STD AIDS 12:524–531. McGroarty, J. A. 1993. Probiotic use of lactobacilli in the human female urogenital tract. FEMS Immun. Med. Microbiol. 6:251–264. McLean, N. W., and I. J. Rosenstein. 2000. Characterisation and selection of a Lactobacillus species to re-colonise the vagina of women with recurrent bacterial vaginosis. J. Med. Microbiol. 49:543–552. Moodley, P., C. Connolly, and A. W. Sturm. 2002. Interrelationships among human immunodeficiency virus type 1 infection, bacterial vaginosis, trichomoniasis, and the presence of yeasts. J. Infect. Dis. 185:69–73. Nath, K., and J. Galdi. 1995. Rapid salt-based mini-scale Gardnerella vaginalis DNA isolation procedure. BioTechniques 19:738–740. Nugent, R. P., M. A. Krohn, and S. L. Hillier. 1991. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of Gram stain interpretation. J. Clin. Microbiol. 29:297–301. Poh, C. L., V. Ramachandran, and J. W. Tapsall. 1996. Genetic diversity of Neisseria gonorrhoeae IB-2 and IB-6 isolates revealed by whole-cell repetitive element sequence-based PCR. J. Clin. Microbiol. 34:292–295. Rabe, L. K., and S. L. Hillier. Optimization of media for detection of hydrogen peroxide (H2O2) production by Lactobacillus species. J. Clin. Microbiol., in press. Reid, G. 1999. The scientific basis for probiotic strains of Lactobacillus. Appl. Environ. Microbiol. 65:3763–3766. Reid, G. 2000. In vitro testing of Lactobacillus acidophilus NCFM™ as a possible probiotic for the urogenital tract. Int. Dairy J. 10:415–419. Reid, G., D. Beuerman, C. Heinemann, and A. W. Bruce. 2001. Probiotic Lactobacillus dose required to restore and maintain a normal vaginal flora. FEMS Immun. Med. Microbiol. 32:37–41. Reid, G., A. W. Bruce, N. Fraser, C. Heinemann, J. Owen, and B. Henning. 2001. Oral probiotics can resolve urogenital infections. FEMS Immun. Med. Microbiol. 30:49–52. Royce, R. A., J. Thorp, J. L. Granados, and D. A. Savitz. 1999. Bacterial vaginosis associated with HIV infection in pregnant women from North Carolina. J. Acquir. Immune Defic. Syndr. Hum. Retrovirol. 20: 382–386. Sewankambo, N., R. H. Gray, M. J. Wawer, L. Paxton, D. McNairn, F. Wabwire-Mangen, D. Serwadda, C. Li, N. Kiwanuka, S. L. Hillier, L. Rabe, C. A. Gaydos, and T. C. Quinn. 1997. HIV-1 infection associated with abnormal vaginal flora morphology and bacterial vaginosis. Lancet 350: 546–550. Shalev, E., Battino, S., Weiner, E., Colodner, R., and Y. Keness. 1996. Ingestion of yogurt containing Lactobacillus acidophilus compared with pasteurized yogurt as prophylaxis for recurrent candidal vaginitis and bacterial vaginosis. Arch. Fam. Med. 5:593–596. Sieber, R., and U.-T. Dietz. 1998. Lactobacillus acidophilus and yogurt in the prevention and therapy of bacterial vaginosis. Int. Dairy J. 8:599–607. Song, Y., N. Kato, Y. Matsumiya, C. Liu, H. Kato, and K. Watanabe. 1999. Identification of and hydrogen peroxide production by fecal and vaginal lactobacilli isolated from Japanese women and newborn infants. J. Clin. Microbiol. 37:3062–3064. Taha, E. T., R. H. Gray, N. I. Kumwenda, D. R. Hoover, L. A. R. Mtimavalye, G. N. Liomba, J. D. Chiphangwi, G. A. Dallabetta, and P. G. Miotti. 1999. HIV infection and disturbances of vaginal flora during pregnancy. J. Acquir. Immune Defic. Syndr. Hum. Retrovirol. 20:52–59. Taha, E. T., D. R. Hoover, G. A. Dallabetta, N. I. Kumwenda, L. A. R. Mtimvalye, L.-P. Yang, G. N. Liomba, R. L. Broadhead, J. D. Chiphangwi, and P. G. Miotti. 1998. Bacterial vaginosis and disturbances of vaginal flora: association with increased acquisition of HIV. AIDS 12:1699–1706. Vallor, A. C., M. A. D. Antonio, S. E. Hawes, and S. L. Hillier. 2001. Factors

VOL. 41, 2003


associated with acquisition of, or persistent colonization by, vaginal lactobacilli: role of hydrogen peroxide production. J. Infect. Dis. 184:1431–1436. 42. Versalovic, J., T. Koeuth, and J. R. Lupski. 1991. Distribution of repetitive DNA sequences in eubacteria and application to fingerprinting of bacterial genomes. Nucleic Acids Res. 19:6823–6831. 43. Versalovic, J., V. Kapur, E. O. Mason, U. Shah, T. Koeuth, J. R. Lupski, and J. M. Musser. 1993. Penicillin-resistant Streptococcus pneumoniae strains recovered in Houston: identification and molecular characterization of multiple clones. J. Infect. Dis. 167:850–856.

1887

44. Wood, J. R., R. L. Sweet, A. Catena, W. K. Hadley, and M. Robbie. 1985. In vitro adherence of Lactobacillus species to vaginal epithelial cells. Am. J. Obstet. Gynecol. 153:740–743. 45. Zheng, H. Y., T. M. Alcorn, and M. S. Cohen. 1994. Effects of H2O2producing lactobacilli on Neisseria gonorrhoeae growth and catalase activity. J. Infect. Dis. 170:1209–1215. 46. Zhong, W., K. Millsap, H. Bialkowska-Hobrzanska, and G. Reid. 1998. Differentiation of Lactobacillus species by molecular typing. Appl. Environ. Microbiol. 64:2418–2423.