ARTICLE IN PRESS

1075-9964/$ - se

doi:10.1016/j.an

�CorrespondE-mail addr

Anaerobe 13 (2007) 178–184

www.elsevier.com/locate/anaerobe

Antimicrobial susceptibility

Antimicrobial activity and protective properties of vaginal lactobacillifrom healthy Bulgarian women

Silviya P. Dimitonovaa,b, Svetla T. Danovaa,�, Julia P. Serkedjievac, Boris V. Bakalovb

aDepartment of Microbial Genetics, Institute of Microbiology ‘‘Stephan Angeloff’’, Bulgarian Academy of Sciences, 26, Acad. G. Bontchev Str.,

1113 Sofia, BulgariabDepartment of General and Applied Microbiology, Biological Faculty, Sofia University, 8, Dragan Tzankov Bvd., 1000 Sofia, BulgariacDepartment of Virology, Institute of Microbiology, Bulgarian Academy of Sciences, 26, Acad. G. Bontchev Str., 1113 Sofia, Bulgaria

Received 9 January 2007; received in revised form 13 August 2007; accepted 18 August 2007

Available online 29 August 2007

Abstract

The role of vaginal Lactobacillus as an efficient barrier against invading pathogens is of considerable interest. The purpose of the

present study was to assess in vitro the ability of 20 recently identified vaginal lactobacilli to protect the vagina. In order to evaluate their

significance, the antimicrobial, hemagglutination (HA) and aggregation (Agg) activities, as well as acid and hydrogen peroxide (H2O2)

production, were estimated. The cell-free cultures of eight strains showed a stable antimicrobial activity after elimination of the putative

effects of lactic acid and H2O2. Three of the isolated vaginal lactobacilli expressed a broad spectrum of anti-bacterial activity including

Gram-negative pathogens. Strains with anti-Gardnerella and anti-herpes simplex virus type 2 activities were found. All tested isolates

were H2O2 producers, actively acidifying the growth media to pH 3.9270.04, which is presumed to neutralize sexually transmitted

infection pathogens. The major part (75%) expressed an HA activity and different Agg phenotypes, estimated as important properties in

the competition with invading pathogens and in host defense. These results are encouraging and prompt further research of the

characterized active strains and their possible application in prophylaxis of vaginal disorders.

r 2007 Elsevier Ltd. All rights reserved.

Keywords: Vaginal lactobacilli; Antimicrobial activity; Aggregation phenotypes; Hemagglutination

1. Introduction

Recently, an increasing interest has developed inmicrobiota, that promote a woman’s health. In particular,Lactobacillus species, that commonly are present in thehuman vagina, have received considerable attention due totheir protective and probiotic properties [1]. Lactobacilliproduce acids, hydrogen peroxide, bacteriocins and bio-surfactants and thus confer protection of the host [2]. Thereis a good clinical evidence that the vaginal and urogenitalfloras play a central role in maintaining both the well-beingand the illnesses of women [3]. When the vaginallactobacilli are diminished or absent, other microorganismsmay grow excessively, causing disorders including bacterialvaginosis (BV), yeast vaginitis and sexually transmitted

e front matter r 2007 Elsevier Ltd. All rights reserved.

aerobe.2007.08.003

ing author. Tel.: +359 2 9793119; fax: +359 2 8700109.

ess: std@microbio.bas.bg (S.T. Danova).

diseases (STD) [4]. Furthermore, the use of living micro-organisms with probiotic properties to replace or augmentprotective populations is gradually achieving scientificacceptance.In addition, the increased risk of acute and recurrent

herpes simplex virus (HSV) infections associated with BVremains a serious problem. In this respect it is important toaddress the role of vaginal lactobacilli in the preventionand treatment of HSV infections.Good knowledge of the biology and the metabolic

activity of these bacteria is a starting point in advanceof therapy and prevention of urogenital infections. In ourrecent study, a laboratory collection of well-characterizedvaginal strains from healthy Bulgarian women has beencreated (unpublished data). The present study was under-taken to evaluate the ability of 20 vaginal Lactobacillus

isolates to act as natural antagonists, inhabiting vagina.The main goal was to define their spectrum of antagonistic

ARTICLE IN PRESSS.P. Dimitonova et al. / Anaerobe 13 (2007) 178–184 179

activity and to select the uropathogens-inhibiting strainswith putative ability to protect the vagina.

2. Materials and methods

2.1. Microorganisms, media and growth conditions

Twenty vaginal lactobacilli were included in this study.They were isolated from vaginal swab samples fromhealthy Bulgarian volunteers at childbearing age. As apart from our laboratory collection we have identified themby phenotypic and molecular methods (such as species-specific and Box-PCR, ARDRA and sequence analyses).Stock cultures were stored at �70 1C in MRS broth(Oxoid), supplemented with glycerol 20%v/v. The strainswere subsequently cultivated on MRS broth (pH 6.5), for24–48 h at 37 1C by anaerobiosis (BBLs GasPak Anaero-bic system, Baltimore, USA).

Different test cultures were used in the study. They weregrown in appropriate media and temperatures as it isshown in Table 1.

2.2. Antimicrobial activity

The inhibitory effect of vaginal strains on selected testmicroorganisms and reference lactobacilli was determinedby the well-diffusion and agar-spot methods as describedpreviously [5]. Normally, 1.5% agar was used. For overlay,0.75% soft agar inoculated with a reference culture wasprepared. Over-night (18 h) test cultures were diluted insaline and were inoculated in appropriate agar medium tofinal concentration of 109CFU/mL. In order to eliminatethe effect of lactic acid on the test organisms, the filteredsupernatants (0.45 mm filters, Millipore, USA) were ad-justed to pH 6.0 with 5M NaOH and were additionallytreated with catalase (1mg/mL, Merck) to exclude theinhibition due to hydrogen peroxide production. As acontrol, aliquots of MRS broth, treated as filtered super-natants, were used. After 24 h of incubation inhibition,zones were measured. All tests were performed in triplicate.The samples giving more than 10mm zone were reported aspositive.

2.3. Anti-viral activity test

Neutralized and acid (without pH correction) filteredsupernatants of selected 10 vaginal strains were used. Two-fold dilutions of the supernatants were made ex tempore

in cell culture medium. The replication of herpes simplexvirus type 2, strain Bja (HSV-2), infectious titer(1018 TCID50/mL) in Madin–Darby bovine kidney(MDBK) cells was used as a model system describedbefore [6]. The positive control was BVDU (E)-5-(2-bromovinyl)-(20-deoxyuridine) (Sigma). The cellulartoxicity test and cytopathic effect (CPE) reduction assaywere performed as described previously [7]. The dilutioncausing 50% CPE (50% toxic concentration, TC50) with

respect to cell control as well as the concentration reducingCPE by 50% (50% effective concentration, EC50) withrespect to virus control were estimated from graphic plots.The selectivity index (SI) was determined from the ratioTC50/EC50. SIX4 was considered to stand for a significantselective inhibition.

2.4. Agar plate method for H2O2 detection

The hydrogen peroxide production of vaginal isolateswas examined using a modification of the Eschenbachquantitative method [8]. Briefly, single colonies (in tripli-cate for each strain) were cultivated on freshly preparedMRS agar plates with 0.05mg/mL of 3,30,5,50-tetramethyl-benzidine (TMB, Sigma) and 0.1mg/mL horseradishperoxidase (Type I Sigma, 116U/mg) added. After 48 hincubation at 37 1C under 5% CO2, the plates were exposedto ambient air for 30min to allow the colony color changes.The horseradish peroxidase oxidized TMB in the presenceof H2O2 and formed a blue pigment in the H2O2-producingcolonies. Non-producers were colorless. The Lactobacillus

acidophilus ATCC 4356T was used as a positive control.

2.5. Acidification ability of vaginal isolates

A change in pH during Lactobacillus cultivation isconsidered as an indicator of their ability to acidify theculture media. Freshly thawed aliquots of vaginal isolatestocks were inoculated (1% inoculum) in MRS broth(pH 6.5) and grown for 24 h at 37 1C in anaerobicconditions. The pH was measured during the growingexponential phase of each strain with a pH meter(SensoDirect, Lovibond pH110, Germany) using a cali-brated glass electrode. The experiments were performed intriplicate and the mean DpH(pH0h�pH24h)7s.d. wascalculated.

2.6. Hemagglutination assay

The hemagglutination activity (HA) was studied by amodification of the microtiter-plate method describedpreviously [9]. Briefly, the lactobacilli were grown inMRS broth at 37 1C for 16–20 h. Washed bacterial cells,collected by centrifugation, (5000g, 10min), were sus-pended in phosphate buffered saline (PBS), pH 7.2, at aconcentration of optical density (OD) of 2.0 at 600 nm.Fifty mL of serially diluted (1:2) microbial suspensions weremixed with 50 mL of 1% suspension of sheep red blood cells(SRBC) in saline. After 4 h at room temperature, thehemagglutination titer (HT) was determined as �log2of the last dilution that causes complete agglutination.Zymosan (Sigma, 10mg/mL) was used as a positivecontrol.

ARTICLE IN PRESS

Table 1

Test microorganisms and culture conditions

Test microorganisms Strains and collectionsa Culture conditions Inhibited by the vaginal strains

Phylogenetically closely related

Lactobacillus acidophilus ATCC 4356T 37 1C, MRS 811, 822

Lactobacillus acidophilus BUCSAV 375 37 1C, MRS No inhibition

Lactobacillus bavaricus NCDO 2588 37 1C, MRS No inhibition

Lactobacillus brevis ATCC 13648 37 1C, MRS No inhibition

Lactobacillus coryniformis subsp. torquens ATCC 25600 37 1C, MRS 712, 822, 832

Lactobacillus fermentum ATCC 13557 37 1C, MRS No inhibition

Lactobacillus fructivorans NCDO 2167 37 1C, MRS No inhibition

Lactobacillus gasseri ATCC 9857 37 1C, MRS 303

Lactobacillus parabuchneri ATTC 49874 37 1C, MRS No inhibition

Lactobacillus ruminis ATCC 25644 37 1C, MRS No inhibition

Lactobacillus plantarum ATCC 14917 37 1C, MRS 303, 304

Lactobacillus plantarum ATCC 8014 37 1C, MRS No inhibition

Other Gram (+) and Gram (�) bacteriab

Bacillus subtilis ATCC 6633 37 1C, MPA

Bacillus megatherium ATCC 9885 37 1C, MPA

Listeria innocua F CIP T82 37 1C, BHI

Sarcina lutea ATCC 9341 37 1C, MPA

Staphylococcus aureus IMB 209 37 1C, blood agar

Staphylococcus aureus MSSA ATCC 25923 37 1C, blood agar

Staphylococcus aureus MRSAc ATCC 39592 37 1C, blood agar

Staphylococcus epidermidis ATCC 14990 37 1C, MPA

Staphylococcus saprophyticus ATCC 15305 37 1C, blood agar

Streptococcus pyogenes ATCC 12344 37 1C, blood agar

Alcaligenes faecalis ATCC 19018 37 1C, MPA

Enterobacter aerogenes NBIMCC 8110 37 1C, MPA

Escherichia coli HB IMB 101 37 1C, LB and MacConkey agar

Escherichia coli WF+ IMB 295 37 1C, LB and MacConkey agar

Escherichia coli NBIMCC 1438 37 1C, LB and MacConkey agar

Escherichia coli 2747d Clinical isolate 37 1C, MacConkey agar

Gardnerella vaginalis ATCC 14018 37 1C BHI+20% (v/v) horse fetal serum

Klebsiella pneumoniae ATCC 10031 37 1C, MPA

Klebsiella pneumoniae NCTC 9133 37 1C, MPA

Proteus vulgaris NBIMCC 1393 37 1C, MPA

Providencia rustigianii ATCC 33673 37 1C, MPA

Pseudomonas aeruginosa ATCC 25668 37 1C, MPA

Pseudomonas aeruginosa 2792 Clinical isolate 37 1C, MacConkey agar

Pseudomonas aeruginosae ATCC 27853 37 1C, MacConkey agar

Salmonella typhimurium IMB 4914 37 1C, MPA

Candida albicans IMB 562 28 1C, Sabouraud’s agar

Virus

Herpes simplex virus type 2 IMB-strain Bja 37 1C, DMEM+5% fetal BSA

aCollections: ATCC—American Type Culture Collection, Virginia, USA; BUSCAV—Institute of Biology, Czech Academy of Sciences, Prague, Czech

Republic; CIP—Collection of Bacteries de l’Institute Pasteur, Paris, France; IMB—Institute of Microbiology, Sofia, Bulgaria; NBIMCC—National Bank

for Industrial Microorganisms and Cell Cultures, Sofia, Bulgaria; NCDO—National Collection of Dairy Organisms, UK; NCFB—National Collection of

Food Bacteria, UK; NCTC—National Collection of Type Cultures, London, UK.bThe results from antagonistic activity are shown in Table 2.cA metacilin resistant strain.dA clinical isolate producing b-lactamases.eAntibiotic-multiresistant test culture (b-lactames, cephalosporins, nalidixic acid, nitrofurantoin, trimethoprim).

S.P. Dimitonova et al. / Anaerobe 13 (2007) 178–184180

2.7. Aggregation (Agg) studies

Exponential anaerobically grown cultures (24 h) of thevaginal strains in the MRS broth (pH 6.5) were used,according to the method of Kmet and Lucchini (1997) [10].Escherichia coli HB 101 cells, washed and re-suspended inPBS (pH 6.0), aggregated in the presence of supernatants(10%v/v) from Lactobacillus cultures in the Agg assay.

A control with pure MRS broth (10%v/v) was used.Co-aggregation (Co-Agg) of E. coli HB 101 cells(�108CFU/mL) with live water-washed Lactobacillus cells(�108CFU/mL) in PBS was examined by appropriatecontrols—Lactobacillus and E. coli suspensions alone inPBS. The auto-aggregation (A-Agg), Agg and Co-Aggwere scored as positive when clearly visible sand-likeparticles, formed by the aggregated cells, gravitated to the

ARTICLE IN PRESSS.P. Dimitonova et al. / Anaerobe 13 (2007) 178–184 181

bottom of the tubes within 2 h at room temperature,leaving clear supernatant fluid in comparison with corre-sponding controls. The Agg within first 15min wasconsidered as rapid (R).

3. Results

3.1. Antagonistic activity of vaginal lactobacilli

In this study, the inhibitory activity of vaginal lactoba-cilli, isolated from healthy Bulgarian women, was assessed.In vitro tests with pH neutralized and catalase-treatedsupernatants (Ns) from exponential Lactobacillus culturesshowed antimicrobial activity for eight out of 20 examinedstrains. They were determined as active because of theobserved ability to inhibit the growth of one or more targetstrains (Table 2). Two vaginal isolates Lactobacillus

fermentum 303 and 361 were active against Gardnerella

vaginalis, the microorganism usually associated with BV.Activity was also found against clinical isolates resistant tocommonly used antibiotics (Table 2). However, none of thestrains inhibited the growth of Candida albicans, Staphy-

lococcus saprophyticus, Staphylococcus epidermidis, Proteus

vulgaris and Salmonella typhimurium.When the antimicrobial tests were performed with acid

supernatants (As), 11 strains were determined as active(Table 2). The effect expressed as mm inhibition zone(mean of three assays) was higher in comparison withresults obtained with neutralized cell-free supernatants(Table 2). Likewise, all 20 isolates acidified actively the

Table 2

The inhibitory effect of selected vaginal strains on test cultures

Tests with Vaginal lactobacillia

301 302 303 304 361

As Ns As Ns As Ns As Ns As

Indicator strains

P. aeruginosa 27853 0 0 0 0 0 0 8 0 0

P. aeruginosa 2792 0 0 0 0 0 0 0 0 0

S. aureus MSSA 0 0 0 0 0 0 0 0 0

S. aureus MRSA 0 0 0 0 13 0 0 0 0

S. aureus 209 0 0 nd nd 0 0 0 0 0

St. pyogenes 15436 0 0 0 0 0 0 0 0 0

Ent. aerogenes 8110 0 0 nd nd 16 14 0 0 0

Kl. pneumoniae 10031 0 0 nd nd 17 14 0 0 0

Kl. pneumoniae 9133 0 0 nd nd 0 0 20 15 18

E. coli 2747 0 0 0 0 0 0 0 0 0

E. coli 1438 0 0 12 0 0 0 0 0 0

E. coli HB101 0 0 0 0 25 14 25 12 22

P. vulgaris 1393 0 0 0 0 0 0 13 0 0

B. subtilis 6633 0 0 0 0 0 0 0 0 0

B. megatherium 9885 12 0 0 0 12 0 12 0 12

Sarcina lutea 9341 0 0 0 0 0 0 0 0 0

G. vaginalis 14018 0 0 0 0 17 15 0 0 16

L. innocua F 82T 0 0 0 0 18 0 17 0 0

aInhibitory activity expressed in mm sterile zone (mean of three assays), As—

not determined.

growth media. The pH of the MRS broth (initially 6.5) waschanged to pH 3.9270.04. The acidification rates(DpH0–24 h) were calculated for each of the vaginal strains(Table 3). Two of them, identified as L. fermentum strains303 and 361, demonstrated ability to lower rapidly the pH(in the first 5 h of cultivation) by 0.5 and 1.2 units,respectively.Most of the tested 20 lactobacilli did not inhibit each

other, except strains 302, 303, 304, 361 and 811. The isolates302, 304, 361 and 811 were able to prevent the growth ofstrain 713. Also, strain 303 suppressed the growth of thestrains 304 and 364. An anti-bacterial activity against otherLactobacillus species was shown for strains 303, 304, 732,811, 822 and 832. The growth of all four referenceLactobacillus species was inhibited as well (Table 1).Cell-free filtrates of vaginal cultures were tested for their

cytotoxic effect on MDBK cells and their antiviralpotential towards the replication of HSV-2 in the samecell line. The cytotoxicity differed considerably. TC50’sranged from 0 (non-toxic) to the dilution 1:40. Themajority of the samples did not inhibit the replication ofHSV-2. Cultures of strains 361, 831 (neutralized to pH 6.0)and 301 (acid, pH 4.5) exhibited antiviral activity (SI were10, 5 and 9, respectively). It should be noted that the 3active extracts were not cell toxic.In a qualitative study of H2O2 production all 20 strains

and the positive control L. acidophilus ATCC 4356 gavevisible blue color change of the colonies within 30min afterexposure of the test plates to air [8]. Thus, they weredetermined as H2O2-positive strains.

362 364 711 712 713 732

Ns As Ns As Ns As Ns As Ns As Ns As Ns

0 11 9 10 0 0 0 9 0 11 0 0 0

0 0 0 0 0 0 0 0 0 13 0 0 0

0 0 0 0 0 0 0 8 0 8 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0

0 15 12 0 0 nd nd 0 0 0 0 nd nd

0 0 0 15 13 0 0 0 0 0 0 0 0

0 0 0 nd nd nd nd 0 0 nd nd nd nd

0 0 0 nd nd nd nd 0 0 nd nd 0 0

14 18 14 0 0 0 0 0 0 nd nd nd nd

0 0 0 0 0 0 0 10 0 8 0 0 0

0 0 0 0 0 0 0 12 11 12 11 0 0

0 28 0 14 0 0 0 11 0 20 12 18 12

0 0 0 14 0 0 0 18 0 18 0 0 0

0 0 0 0 0 0 0 12 0 0 0 0 0

0 12 0 13 0 15 0 18 0 17 0 14 0

0 0 0 0 0 0 0 13 0 12 0 12 0

14 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0

acid supernatant, Ns—neutralized and catalase-treated supernatant, nd—

ARTICLE IN PRESS

Table 3

Acidification ratio, aggregation and hemagglutination activity of vaginal lactobacilli isolated from healthy Bulgarian women

Vaginal strains DpH7s.d. A-Agga Agga Co-Agga HTb

L. fermentum 301 2.5870.07 7 � + 4

L. fermentum 302 2.3370.07 + + � 0

L. fermentum 303 2.670.03 + +R +R 3

L. fermentum 304 2.5970.03 7 +R +R 2

L. .fermentum 361 2.7970.02 + +R +R 2

L. fermentum 362 2.5870.06 + +R +R 1

L. fermentum 364 2.7870.03 + + + 3

Lactobacillus sp. 611 2.5670.05 7 + � 2

Lactobacillus sp. 612 2.6170.09 7 + � 2

Lactobacillus sp. 613 2.6170.04 7 + � 2

L. brevis 622 2.670.02 � � � 1

L. acidophilus sp. 711 2.6170.03 7 + + 0

L. acidophilus sp. 712 2.7970.02 � � � 4

L. acidophilus sp. 713 2.7970.01 � � � 4

L. acidophilus sp. 732 2.7970.02 � � � 5

Lactobacillus sp. 811 2.2770.15 7 + � 4

Lactobacillus sp. 821 2.0970.08 7 + � 2

L. paracasei 822 2.0970.1 � � 7 3

L. brevis 831 2.6270.07 � � � 0

L. salivarius 832 2.7770.06 7 � 7 0

Zymosan—a positive control in HA tests 5

a(+) Positive and (�) negative aggregation activity; a positive A-Agg: scored with a clear (+) or less clear supernatant (7) within 2 h; (+R)—rapid

aggregation and Co-Agg within first 15min.bHT—hemagglutination titer.

S.P. Dimitonova et al. / Anaerobe 13 (2007) 178–184182

3.2. HA and Agg activity

As a first step toward the screening of strains possessingcarbohydrate-binding activity, the HA of 20 strainswas examined. There was a difference in the HT of 16strains agglutinating the SRBC (Table 3). Only the strain713 showed HT, equal to the positive control zymosan(10mg/mL). Furthermore, an A-Agg activity was foundin 14 isolates out of the tested 20 Lactobacillus strains(Table 3). The addition of water-washed Lactobacillus cellsto their own supernatant culture fluid caused A-Agg,unambiguously detectable for five isolates (Table 3).Moreover, isolates 303, 304, 361 and 362 caused a rapidAgg within the first 15min as well as the Co-Agg withE. coli cells. The cell-free cultures of nine vaginal isolatesdid not cause a positive Agg reaction and the washed cellsof 11 out of the tested 20 Lactobacillus strains were not ableto co-aggregate the cells of E. coli HB 101 (Table 3).

4. Discussion

The importance of vaginal Lactobacillus as a barrier toinfections is of considerable interest. The experiments thatdemonstrate their protective role in a woman’s healthwould benefit from the development of simple methods,allowing their detection and identification. In the presentstudy, different in vitro methods were applied to character-ize the antagonistic properties of 20 vaginal strains. Theability of eight lactobacilli to inhibit the growth ofpotential human pathogens and other undesirable bacteriawas revealed. Three of them (strains 303, 304 and 713)

exerted a broad spectrum of activity, including Gram (�)bacteria and other Lactobacillus species. Thus, the activestrains could dominate in the vagina by suppressing closelyrelated and other species [11]. The effect of lactic acidbacteria in controlling the proliferation of pathogenicbacteria in the vagina is questionable [12]. Nonetheless,recent reports on the activity against Gram-negativebacteria [5,13–15] stimulated a renewed interest in studyingantimicrobials-producing strains and their interaction withvaginal pathogens. The capacity to produce microbocideswith different spectrum of activity was reported forLactobacillus rhamnosus, L. fermentum and Lactobacillus

salivarius subsp. salivarius strains and other lactobacilliisolated from vagina [16–18]. A limited data exist on theactivity exhibited by endogenous vaginal lactobacilliagainst G. vaginalis [19].Our results demonstrated also a selective antiviral effect

of two Lactobacillus strains. To our knowledge this is thefirst data for an anti-HSV-2 effect, shown by cell-freecultures of vaginal lactobacilli. A complete inactivation ofHSV-1 (strain F) and HSV-2 (strain G) was observed aftertreatment with the STV4 purified peptide, produced fromEnterococcus mundtii [20]. As a close relationship wasreported between the infections, caused by HSV-2 and BV[21], it could be speculated that, as a natural barrier againstinvading pathogens, the vaginal lactobacilli possibly reducethe susceptibility to HSV-2.The production of H2O2 could be an alternative effective

way by which these bacteria suppress the growth of othermicroorganisms [22]. In vitro H2O2-producing lactobacillihave bactericidal effect on several BV-associated and other

ARTICLE IN PRESSS.P. Dimitonova et al. / Anaerobe 13 (2007) 178–184 183

pathogens. It has been proposed that the productionof H2O2 may explain the success of lactobacilli as vaginalcolonizers [11]. While a very high percent of H2O2

producers (96–98%) was found among vaginal lactobacilli,isolated from healthy women of reproductive age [8,23], inthe cases with infections pertinent to preterm delivery, theirpercent was only 6–23% [10]. There are no data concerningthe predominance of these bacteria in healthy Bulgarianwomen. The fact that all of the 20 randomly selectedvaginal lactobacilli proved to be H2O2 producers deservesan attention and requires additional detailed study.

Hydrogen peroxide does not seem to be the sole activeinhibitory product in the studied group of vaginal strains.The detected antagonistic activity (especially for strains303, 361, 364, 712, 713 and 732), after elimination of theputative effects of lactic acid and H2O2, raised the questionfor a possible production of bacteriocin-like or low-molecular microbocides. The observed inhibitory effect ofthe vaginal cultures on Gram (�) human pathogens couldbe potentially useful, as they did not inhibit a large numberof other Lactobacillus species. The complex mechanism/s ofsuch antagonistic interactions of lactobacilli with otherbacteria has not been fully characterized. Further studiesare in progress to select some of the active strains aspotential bacteriocin producers after additional tests forcharacterization of the bactericidal substances. The anti-microbial effect of some strains in vivo may be completedwith the production of relevant concentration of lactic acidin the microenvironment, which inhibits the growth ofGram-negative pathogenic bacteria [24]. Thus, the reportedhigher antagonistic activity with non-neutralized super-natants also seemed to be important because the effectwould be exerted in the vagina. All strains selected for thisstudy were able to create an acidic environment, whichmight protect the area against the invasion of undesirablemicroorganisms [11,19]. They maintained a low pHthrough their fermenting activity. Our results confirmedthe ability of the vaginal lactobacilli to decrease pH to 4.0and less that corresponds well to the pH of a healthyvagina [25].

Cell adhesion of bacteria on vaginal tract is the first stepof infection and is affected by several factors. Thecarbohydrate-binding activity leading to hemagglutinationis closely related to the ability to adhere to intestinal andurogenital epithelial cells. The speculation is that vaginallactobacilli, like intestinal ones, can adhere to epithelialcells like lectins through interactions between the carbohy-drate-recognizing molecules on the bacterial cells’ surfaceand the glycoproteins on the host cells’ surface [9]. Thehemagglutination assay has been applied in the screeningof lectin or lectin-like components. Thus, the ability toagglutinate erythrocytes might be used as a model systemto predict the possible interaction between bacteria andtheir hosts. The HA activity of 15 strains (75%) has beenevaluated (Table 3). The effect appeared not to be speciesspecific. The obtained data are in accordance with theresults reported for Lactobacillus reuteri strains [26].

Moreover, some adhering Lactobacillus strains couldinhibit adhesion and internalization within the host’s cellsof uropathogenic E. coli [27] or other bacteria. Horosovaet al. (2006) [28] proved, by in vitro tests, that the adhesionof lactobacilli reduced the number of adhering pathogenicE. coli strains and was a pre-condition for the Lactobacillus

health benefits.We have applied a simple method to show the capacity

of vaginal isolates to exhibit various Agg phenotypes. Withthe exception of only three strains (isolates 712, 713 and732) the Agg assays showed a correlation between the anti-E. coli activity and the Agg ability expressed in vitro. Aggand Co-Agg activities have been actively studied forvaginal isolates [29]. The A-Agg (or formation of multi-cellular clumps between cells of the same strain), observedfor the major part of the tested strains, is one of theproposed mechanisms to explain the protective role oflactobacilli in the human vagina. This property related tothe adhesion ability to epithelial vaginal cells can force thelactobacilli to produce a biofilm on the vaginal epithelia,thus preventing the entry of pathogens. In addition, thepotentially probiotic lactobacilli would be exposed tovaginal environment with fluctuating conditions, such asdifferent pH values (pH 4.0–4.5 in normal women; pH5.0–6.0 in women with BV, pH close to 7.0 around themenstruation). Therefore, the aggregating microorganismswill survive and proliferate under conditions that promotethe approach of partner cells. The domination of lactoba-cilli in a healthy vaginal ecosystem could be achieved byvarious defense mechanisms. The in vitro tests used by uscould be useful in the pre-selection of candidate Lactoba-

cillus strains for vaginal probiotics. In vivo studies areneeded to confirm the significance of the H2O2 andbacteriocin-like production and other factors for theLactobacillus domination in the vaginal ecosystem.In conclusion, the maintenance or the reconstruction of

the normal composition of the vaginal microflora byapplying properly selected lactobacilli may be of prophy-lactic value in preventing various urogenital infections inwomen. More complex studies based on well-standardizedmethods of selected candidates for vaginal probiotics arenecessary, before entering the stage of clinical studies.

Acknowledgment

We would like to thank Prof. N. Ivanovska fromDepartment of Immunology, Institute of Microbiology,BAS, Sofia, for the support and valuable advices.

References

[1] Andreu A. Lactobacillus as a probiotic for preventing urogenital

infections. Rev Med Microbiol 2004;15:1–6.

[2] Reid G, Bruce AW. Urogenital infections in women: can probiotics

help? Postgrad Med J 2003;79:428–32.

[3] Hillier SL, Krohn MA, Klebanoff SJ, Eschenbach DA. The

relationship of hydrogen peroxide-producing lactobacilli to bacterial

ARTICLE IN PRESSS.P. Dimitonova et al. / Anaerobe 13 (2007) 178–184184

vaginosis and genital microflora in pregnant women. Obstet Gynecol

1992;79:369–73.

[4] Burton JP, Cadieux PA, Reid G. Improved understanding of the

bacterial vaginal microbiota of women before and after probiotic

instillation. Appl Environ Microbiol 2003;69:97–101.

[5] Ivanova I, Miteva V, Stefanova T, Pantev A, Budakov I, Danova S,

et al. Characterization of a bacteriocin produced by Streptococcus

thermophilus 81. Int J Food Microbiol 1998;42:147–58.

[6] Sokmen M, Serkedjieva J, Daferera D, Gulluce M, Polissiou M, Tepe

B, et al. In vitro antioxidant, antimicrobial, and antiviral activities of

the essential oil and various extracts from herbal parts and callus

cultures of Origanum acutidens. J Agric Food Chem 2004;52:3309–12.

[7] Serkedjieva J, Hay AJ. In vitro anti-influenza virus activity of a plant

preparation from Geranium sanguineum L. Antiviral Res 1998;37:

121–30.

[8] Eschenbach DA, Davick PR, Williams BL, Klebanoff SJ, Young-

Smith K, Critchlow CM, et al. Prevalence of hydrogen peroxide-

producing Lactobacillus species in normal women and women with

bacterial vaginosis. J Clin Microbiol 1989;27:251–6.

[9] Yamada M, Saito T, Toba T, Kitazawa H, Uemura J, Takatoshi I.

Hemagglutination activity of Lactobacillus acidophilus group lactic

acid bacteria. Biosci Biotechnol Biochem 1994;58:910–3.

[10] Kmet V, Lucchini F. Aggregation-promoting factor in human vaginal

Lactobacillus strains. FEMS Immunol Med Microbiol 1997;19:111–4.

[11] Aroutcheva A, Gariti D, Simon M, Shott S, Faro J, Simoes JA, et al.

Defense factors of vaginal lactobacilli. Am J Obstet Gynecol 2001;

185:375–9.

[12] Brink M, Todorov SD, Martin JH, Senekal M, Dicks LM. The effect

of prebiotics on production of antimicrobial compounds, resistance

to growth at low pH and in the presence of bile, and adhesion of

probiotic cells to intestinal mucus. J Appl Microbiol 2006;100.

[13] Kaewsrichan J, Peeyananjarassri K, Kongprasertkit J. Selection and

identification of anaerobic lactobacilli producing inhibitory com-

pounds against vaginal pathogens. FEMS Immunol Med Microbiol

2006;48:75–83.

[14] Todorov SD, Danova ST, Van Reenen CA, Meincken M, Dinkova

G, Ivanova IV, et al. Characterization of bacteriocin HV219,

produced by Lactococcus lactis subsp. lactis HV219 isolated from

human vaginal secretions. J Basic Microbiol 2006;46:226–38.

[15] Caridi A. Selection of Escherichia coli-inhibiting strains of Lactoba-

cillus paracasei subsp. paracasei. J Ind Microbiol Biotechnol 2002;29:

303–8.

[16] Ocana VS, Pesce De Ruiz Holgado AA, Nader-Macias ME.

Characterization of a bacteriocin-like substance produced by a

vaginal Lactobacillus salivarius strain. Appl Environ Microbiol

1999;65:5631–5.

[17] Reid G, Beuerman D, Heinemann C, Bruce AW. Probiotic

Lactobacillus dose required to restore and maintain a normal vaginal

flora. FEMS Immunol Med Microbiol 2001;32:37–41.

[18] Voravuthikunchai SP, Bilasoia S, Supamalaa O. Antagonistic activity

against pathogenic bacteria by human vaginal lactobacilli. Anaerobe

2006;12:221–6.

[19] Aroutcheva AA, Simoes JA, Faro S. Antimicrobial protein produced

by vaginal Lactobacillus acidophilus that inhibits Gardnerella vagina-

lis. Infect Dis Obstet Gynecol 2001;9:33–9.

[20] Todorov SD, Wachsman MB, Knoetze H, Meincken M, Dicks LM.

An antibacterial and antiviral peptide produced by Enterococcus

mundtii ST4V isolated from soya beans. Int J Antimicrob Agents

2005;25:508–13.

[21] Cherpes TL, Meyn LA, Krohn MA, Lurie JG, Hillier SL.

Association between acquisition of herpes simplex virus type 2 in

women and bacterial vaginosis. Clin Infect Dis 2003;37:319–25.

[22] Otero MC, Nader-Macias ME. Inhibition of Staphylococcus aureus

by H2O2-producing Lactobacillus gasseri isolated from the vaginal

tract of cattle. Anim Reprod Sci 2005.

[23] Rabe LK, Hillier SL. Optimization of media for detection of

hydrogen peroxide production by Lactobacillus species. J Clin

Microbiol 2003;41:3260–4.

[24] Alakomi HL, Skytta E, Saarela M, Mattila-Sandholm T, Latva-Kala

K, Helander IM. Lactic acid permeabilizes gram-negative bacteria by

disrupting the outer membrane. Appl Environ Microbiol

2000;66:2001–5.

[25] Schwiertz A, Taras D, Rusch K, Rusch V. Throwing the dice for the

diagnosis of vaginal complaints? Ann Clin Microbiol Antimicrob

2006;5:4.

[26] Mukai T, Kaneko S, Ohori H. Haemagglutination and glycolipid-

binding activities of Lactobacillus reuteri. Lett Appl Microbiol

1998;27:130–4.

[27] Atassi F, Brassart D, Grob P, Graf F, Servin AL. Vaginal

Lactobacillus isolates inhibit uropathogenic Escherichia coli. FEMS

Microbiol Lett 2006;257:132–8.

[28] Horosova K, Bujnakova D, Kmet V. Effect of lactobacilli on E. coli

adhesion to Caco-2 cells in vitro. Folia Microbiol (Praha)

2006;51:281–2.

[29] Juarez Tomas MS, Wiese B, Nader-Macias ME. Effects of

culture conditions on the growth and auto-aggregation ability of

vaginal Lactobacillus johnsonii CRL 1294. J Appl Microbiol 2005;99:

1383–91.