1950

The Identification of Vaginal Lactobacillus Species and the Demographic

and Microbiologic Characteristics of Women Colonized by These Species
May A. D. Antonio,1 Stephen E. Hawes,3 1
Magee-Womens Research Institute and 2Department of Obstetrics,
and Sharon L. Hillier1,2 Gynecology, and Reproductive Sciences, University of Pittsburgh,
Pittsburgh, Pennsylvania; 3Human Papillomavirus Research Group,
University of Washington, Seattle

Lactobacillus acidophilus has been reported to be the predominant vaginal species. Vaginal
lactobacilli isolated from 215 sexually active women were identified using whole-chromosomal
DNA probes to 20 American Type Culture Collection Lactobacillus strains. Most women
were colonized by L. crispatus (32%), followed by L. jensenii (23%), a previously undescribed
species designated L. 1086V (15%), L. gasseri (5%), L. fermentum (0.3%), L. oris (0.3%), L.
reuteri (0.3%), L. ruminis (0.3%), and L. vaginalis (0.3%). H2O2 was produced by 95% of L.
crispatus and 94% of L. jensenii isolates, compared with only 9% of L. 1086V. Colonization

Downloaded from http://jid.oxfordjournals.org/ by guest on September 5, 2016

by L. crispatus or L. jensenii was positively associated with being white (P ! .001), age >20
years (P = .05), barrier contraceptive usage (P = .008 ), and lower frequency of bacterial va-
ginosis (P ! .001) and gonorrhea (P = .03 ). L. crispatus and L. jensenii, not L. acidophilus, are
the most common species of vaginal lactobacilli.

Establishing the identity of Lactobacillus species colonizing
the vagina of women is of importance, because clinical studies
have demonstrated an association between the presence of
H2O2-producing strains of Lactobacillus and a decreased prev-
alence of gonorrhea, bacterial vaginosis (BV) [1], and human
immunodeficiency virus (HIV) infection [25]. However, in
these studies lactobacilli have usually been identified only to
the genus level because of the technical difficulties in the species
identification of Lactobacillus.
The identity of the predominant Lactobacillus species colo-
nizing the vagina has been uncertain because of the unreliability
of classic identification methods, which employ sugar fermen-
tation and other phenotypic assays. From these methods, var-
ious lists of vaginal Lactobacillus species have been developed,
including any of the following: L. acidophilus, L. fermentum,
L. plantarum, L. brevis, L. jensenii, L. casei, L. cellobiosus, L.
leichmanii, L. delbrueckii, and L. salivarius [68]. Of all these
species, L. acidophilus has been the vaginal lactobacillus most
widely accepted to be predominant . However, the general use
Received 10 August 1998; revised 3 August 1999; electronically published
12 November 1999.
Presented in part: International Society for Sexually Transmitted Disease
Research Meeting, New Orleans, August 1995 (abstract 207).
Informed consent was obtained from each woman in a protocol approved
by the Human Subjects Committee at the University of Washington. In the
conduct of clinical research, human experimentation guidelines of the US
Department of Health and Human Services were followed.
Grant support: NIH (AI-31448, AI-38513).
Reprints or correspondence: Dr. Sharon L. Hillier, University of Pitts-
burgh, Dept. of Obstetrics, Gynecology, and Reproductive Sciences,
Magee-Womens Hospital, 300 Halket St., Pittsburgh, PA 15213-3180
(slh61@pitt.edu).
The Journal of Infectious Diseases 1999; 180:19506
q 1999 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/1999/18006-0025$02.00
of L. acidophilus to identify human oral, intestinal, and vaginal
Lactobacillus has been more a historic than a scientific desig-
nation because of the poor reliability of existing tests used to
differentiate Lactobacillus species [9]. Even 80 years ago, there
was uncertainty whether L. acidophilus characterized a group
of related species or a single group of organisms that under-
goes transformation [10].
Several investigators who were questioning the reliability and
reproducibility of classic identification methods for Lactoba-
cillus species sought more dependable identification protocols
[11, 12]. On the basis of DNA homology studies, the taxonomy
of lactobacilli has been under revision [12, 13]. Formerly, the
species group of L. acidophilus comprised 6 DNA homology
groups that could not be distinguished biochemically [12]. Two
of these homology groups are now L. crispatus and L. gasseri.
When DNA homology methods were used to evaluate the lac-
tobacilli from a group of 27 asymptomatic women, Giorgi et
al. [14] identified L. gasseri, L. jensenii, and L. crispatus, not
L. acidophilus, as the predominant vaginal Lactobacillus species
colonizing asymptomatic women.
The present study was undertaken to identify which species
of lactobacilli were present in a cross-sectional sample of 302
women, using DNA homology to American Type Culture Col-
lection (ATCC) strains of lactobacilli. H2O2 production and the
species specificity of this characteristic were determined. The
distribution of Lactobacillus species colonizing women was also
assessed demographically and microbiologically.
Methods
In this study, 319 women visiting an adolescent medicine clinic
and 2 sexually transmitted disease clinics in Seattle were enrolled.
A standardized questionnaire concerning demographic character-
JID 1999;180 (December) L. crispatus and L. jensenii
istics and contraceptive history was administered. Two vaginal
swabs were used to obtain samplings from the lateral vaginal wall.
One swab was rolled onto a slide for a vaginal smear, and the other
swab for culture was placed into an Amies transport medium
(MML Diagnostics, Troutdale, OR). The slide and the transport
medium were delivered to the research laboratory within 12 h.
Calcium alginate swabs or Dacron swabs were inserted into the
cervix to obtain material for gonococcal and chlamydial cultures.
Seventeen women were excluded because samples were not ob-
tained for Gram staining or culture, leaving 302 evaluable subjects.
Specimens for Chlamydia trachomatis cultures were transported
to the laboratory in 0.2 mL of sucrose-phosphate containing 2%
fetal calf serum, 100,000 U of penicillin, 1 mg/mL of gentamicin,
25 mg/mL of vancomycin, and 25 U/mL of nystatin. C. trachomatis
was cultured both in vials and in 96-well microtiter plates of cy-
cloheximide-treated McCoy cells. The cell layers were stained with
fluorescein-conjugated monoclonal antibodies to a species-specific
C. trachomatis antigen and were examined for inclusions at 4872
h with an epifluorescence microscope [15].
Kelloggs medium and either modified Thayer-Martin medium
or enriched chocolate agar were streaked for the isolation of Neis-
seria gonorrhoeae. Gonococci were identified by sugar utilization,
oxidase tests, and examination of Gram-stained preparations [16].
Slides were Gram stained and evaluated by use of the Nugent
criteria [17]. A score of 010 was assigned in light of the relative
proportions of large gram-positive rods (lactobacilli), small gram-
negative or gram-variable rods (Bacteroides, Prevotella, or Gard-
nerella species), and curved gram-variable rods (Mobiluncus spe-
cies). A score of 03 was interpreted as consistent with normal
flora, a score of 46 was considered consistent with intermediately
disturbed flora, and a score of 710 was considered consistent with
BV [17].
Vaginal swabs were removed from the transport medium and
used to inoculate Rogosa agar (Difco, Detroit), Columbia 5% sheep
blood agar, prereduced brucella agar with 5% sheep blood, vitamin
K, hemin, prereduced laked-blood kanamycin agar, and 2 human
blood bilayer Tween (HBT) agar plates (Prepared Media Labo-
ratories, Tualitin, OR). The Columbia agar and 1 HBT plate were
incubated at 367C in 5%7% CO2 for a minimum of 48 h. The
remaining plates were incubated within an anaerobic glove box at
367C for a minimum of 5 days. An A7B agar plate and broths for
isolation of Ureaplasma urealyticum and Mycoplasma hominis [18],
all prepared in-house, were also inoculated and incubated at 367C
in 5%7% CO2 for a minimum of 72 h. The broths were further
subcultured onto A7B agar.
Aerobic bacteria were first identified by use of Gram stain and
colony morphologies and catalase, and they were further identified
by use of gas chromatography and biochemical tests [19]. Anaer-
obic gram-negative rods were identified by use of Gram stain and
colony morphologies, pigment production on HBT, and their in-
ability to grow aerobically [19]. The mycoplasmas were identified
by their characteristic morphology on the A7B agar plate [18].
Lactobacilli were identified to the genus level by Gram stain and
colony morphologies, negative catalase test, and production of pre-
dominant lactic acid peak as assessed by gas chromatographic anal-
ysis [20]. All lactobacilli were tested for the production of H2O2 in
a qualitative assay on a tetramethylbenzidine (TMB) agar plate
[21]. After 48 h of incubation in an anaerobic glove box at
1951
367C377C, the agar plates were exposed to ambient air. The H2O2
that was formed reacted with the horseradish peroxidase in the
agar to oxidize the TMB, causing the colonies of lactobacilli to
turn blue. Lactobacillus isolates were stored at 2707C in litmus
milk until they were transferred to the Infectious Disease Labo-
ratory at the Magee-Womens Research Institute for DNA studies.
For the DNA studies, each Lactobacillus isolate was grown in
PYTSG broth (PY basal medium [20], 1% [wt/vol] dextrose, 1%
[wt/vol] soluble starch, and 0.02% [vol/vol] Tween 80) incubated
for 2448 h in 6% CO2 at 377C. Growth in broth was checked for
purity by plating a drop onto a Columbia 5% sheep blood agar
plate (Prepared Media Laboratories) and incubating as described
above. The following DNA isolation procedure was modified from
Luchansky et al. [22]. The cells were washed in TES buffer (50 mM
NaCI, 5 mM EDTA, 30 mM Tris, pH 8.0). The pellet was resus-
pended in lysis buffer (25% ultrapure sucrose, 50 mM Tris, 1 mM
Downloaded from http://jid.oxfordjournals.org/ by guest on September 5, 2016
EDTA, pH 8.0) containing lysozyme and incubated at 377C for at
least 1 h. About 35 U of RNase A (Sigma, St. Louis) was added
to the solution, which was then incubated at 607C for 30 min.
Sixteen units of a nonspecific protease type XIV from Streptomyces
griseus (Sigma) was added and was incubated at 377C for at least
1 h. After the addition of 1 mL of 0.25 M EDTA, pH 8.0, the
solution sat in ambient temperature for 5 min. A volume of 0.4
mL of 20% SDS was added and then incubated at 607C for 1 h.
The solution was further incubated at 607C for 30 min, with the
addition of 10 U of proteinase K (Sigma). The lysate was extracted
once with an equal volume of buffered phenol and once with an
equal volume of chloroform. A double volume of 95% ethanol was
added to precipitate the DNA. The DNA pellet was dissolved in
a nominal amount of TE (10 mM Tris, 1 mM EDTA, pH 8.0) and
stored at 47C. DNA concentration was determined using absorb-
ance readings at a wavelength of 260 nm. Gel electrophoresis and
ethidium bromide staining were done to evidence genomic DNA
isolation.
The Random Primed DNA Labeling Kit (Boehringer Mann-
heim, Indianapolis) was used to make whole-chromosomal probes
according to the manufacturers protocol. The specificity of this
method has been demonstrated for identification of Mobiluncus
[23] and Campylobacter [24] species. For each whole-chromosomal
probe, a 20-mL reaction volume (consisting of 20 mM each dCTP,
dGTP, and dTTP; 2 mL of a reaction mixture containing random
hexanucleotide primers; 25 ng denatured target genomic DNA; 2
U of Klenow enzyme; and 50 mCi [a32P] dATP [NEN Life Science
Products, Boston]) was incubated at 377C for at least 30 min. The
addition of 2 mL of 0.2 M EDTA stopped the reaction. Whole-
chromosomal probes were made from 20 ATCC Lactobacillus
strains: L. acidophilus 521 and 4356, L. brevis 11577, L. buchneri
11579, L. casei subspecies casei 4646, L. catenaformis 25536, L.
confusus 10811, L. crispatus 33197, L. delbrueckii subspecies del-
brueckii 9649, L. delbrueckii subspecies bulgaricus 11842, L. fer-
mentum 23271, L. gasseri 4963, L. jensenii 25258, L. minutus 33267,
L. oris 49062, L. parabuchneri 49374, L. rhamnosus 21052, L. rum-
inis 25644, L. salivarius subspecies salivarius 11741, and L. vaginalis
49540. A probe was also made to a previously unidentified species
of Lactobacillus derived from a vaginal specimen, designated
1086V. The estimated specific activity of each probe was 2 3 10 9
disintegrations/min/mg.
All DNA from patient isolates and ATCC Lactobacillus strains
1952 Antonio et al.
were slot-blotted onto nylon membranes of a Minifold II slot blot-
ter (Schleicher & Schuell, Keene, OH) according to the manufac-
turers instructions. In brief, to a 400-mL total volume of TE (pH
7.0) containing 5 mg of DNA, a 1 : 10 volume of 3 N NaOH was
added. The sample was incubated at 607C for 45 min. After the
sample was cooled to ambient temperature, an equal volume of 1
M ammonium acetate was added. The control filters consisted of
DNA from the ATCC strains listed above and from ATCC strains
L. acidophilus 4357, L. brevis 14869, L. buchneri 4005, L. casei
subspecies casei 393, L. delbrueckii subspecies lactis 4797 and
12315, L. fermentum 11739 and 14931, L. gasseri 9857, L. johnsonii
33200, L. minutus 33267, L. paracasei subspecies paracasei 27216,
L. plantarum 14917, L. rhamnosus 21052, L. ruminis 25644, and L.
salivarius subspecies salicinius 11742. The control filters were in-
cluded with each whole-chromosomal probe hybridization. About
100 unknown isolates were tested against probes to all of the control
strains. The specificity of the probe technique was excellent, with
none of the unknown isolates having homology to 11 of the control
species.
The wells of the slot blotter are arranged in a 3 3 24 fashion,
allowing 24 samples to be vertically arranged with 3 wells per
sample. The slot blotter was connected to a vacuum, and during
vacuuming, each sample was aliquoted into 3 wells. About 1.5 mg
of sample DNA was aliquoted into each slot. After vacuuming was
complete, the slot blotter was disassembled. Before removing the
membrane from the slot blotter template piece, we marked the
width of each well on the backside of the membrane, using a black
permanent ink marker. Each 24-sample membrane was given a
number and baked for 2 h at 807C. After being baked, each
membrane was cut into 3 strips and then cut again in half, dividing
the wells in 2, resulting in 6 strips, each with an identical sequence
of samples and filter number.
For each probe, sample and control strips were laid flat, not
overlapping one another, and placed in an 8 3 12 inch heat-seal-
able pouch (Kapak, Minneapolis). Prehybridization solution [25]
was added to equilibrate the nylon membranes. The heat-sealed
pouch was incubated at 427C in a rocking incubator for at least 1
h before the addition of whole-chromosomal probes. A corner of
the pouch was cut open, and a 20-mL volume of random
primedlabeled whole-chromosomal probe was added to the equil-
ibrated membranes.
After the addition of a probe, the membrane was incubated
overnight at 427C. Three 45-min washes in 0.53 standard saline
citrate (0.15 M NaCl, 0.015 M sodium citrate) and 0.1% SDS were
done at 657C. If more than the intended species turned positive on
the control filter of ATCC Lactobacillus strains, higher-stringency
washes were performed. A dark band on the autoradiograph was
read as positive if its intensity was similar to or greater than that
of the control positive band. Two individuals analyzed the auto-
radiographs; one was blinded to the identity of all unknown sam-
ples and control strains. The blinded individual did not work
with the processing of DNA samples onto the nylon membranes.
The location of a sample, unknown or control, was not revealed
to this individual. Agreement between the 2 readers was 100%,
suggesting that visual interpretation of a positive hybridization was
not subjective. Examples of 2 autoradiographs used in the DNA
hybridization to L. crispatus and L. gasseri whole-chromosomal
probes are shown in figure 1.
JID 1999;180 (December)
Pearson x2 tests were utilized to compare discrete variables with
respect to the presence of L. crispatus or L. jensenii. P values <.05
were considered statistically significant.
Results
Lactobacilli were recovered from 215 (71%) of the 302 women
(table 1). About one-third of the women were colonized by L.
crispatus, and one-fourth were colonized by L. jensenii. A Lac-
tobacillus strain we designated L. 1086V colonized 15% of the
women. L. gasseri was recovered from only 5% of the women,
and only 1 woman each was colonized by L. ruminis, L. reuteri,
L. fermentum, L. oris, or L. vaginalis. The following species
were not recovered from the vagina of any of the women in
Downloaded from http://jid.oxfordjournals.org/ by guest on September 5, 2016
this group: L. acidophilus, L. brevis, L. buchneri, L. casei, L.
catenaformis, L. confusus, L. delbrueckii, L. parabuchneri, L.
rhamnosus, and L. salivarius.
At least 2 different Lactobacillus species simultaneously col-
onized 25 of the women. Both L. crispatus and L. jensenii col-
onized 11 women, including 1 who was also colonized by L.
1086V. Both L. crispatus and L. 1086V were coisolated from 4
women, and both L. jensenii and L. 1086V were coisolated from
6 women. Other isolated pairs included L. jensenii and L. gas-
seri, L. crispatus and L. fermentum, L. 1086V and L. gasseri,
and L. jensenii with an isolate not homologous to any of the
probes used.
Nearly all of the L. crispatus and L. jensenii strains produced
H2O2 (95% and 94%, respectively; table 1). Only 14 L. gasseri
strains were identified, and 71% of them produced H2O2. In
contrast, only 9% of L. 1086V isolates produced H2O2.
The P values in tables 2 and 3 were obtained by comparing
women with L. crispatus or L. jensenii with women not colo-
nized by either species, including women lacking lactobacilli.
This comparison was chosen because nearly all of the L. cris-
patus and L. jensenii strains produced H2O2. Either L. crispatus
or L. jensenii colonized a total of 154 women (51%). There were
61 lactobacilli-colonized women (20%) who did not have either
species. In some instances, data were not available for some of
the women, resulting in a different denominator.
Demographic characteristics and birth control usage strati-
fied by species of lactobacilli are shown in table 2. Marital
status, oral contraception, or douching did not appear to have
any significant effect on the colonization of L. crispatus or L.
jensenii. However, women with L. crispatus or L. jensenii were
less likely to be !20 years old (37% vs. 49%; P = .05) and more
likely to be white (72% vs. 50%; P < .001) and users of barrier
contraceptives (41% vs. 27%; P = .008).
The microbiologic findings for the 302 women are shown in
table 3. Of the sexually transmitted pathogens, N. gonorrhoeae
and C. trachomatis, only N. gonorrhoeae infection was signif-
icantly decreased in women colonized by L. crispatus or L.
jensenii (1% vs. 5%; P = .03). Women with L. crispatus or L.
jensenii were also less likely to have BV (9% vs. 69%; P <
JID 1999;180 (December) L. crispatus and L. jensenii 1953

.001) and to have decreased colonization of BV-related species: Table 1. Lactobacillus species detected among 302 women with or
Gardnerella vaginalis (38% vs. 84%; P < .001); anaerobic non- without bacterial vaginosis (BV), and H2O2 production by the
lactobacilli.
pigmented gram-negative rods (38% vs. 77%; P < .001); an-
No. of H2O2
aerobic black-pigmented gram-negative rods (11% vs. 33%;
DNA homology group women colonized production BV present
P < .001); and Mycoplasma hominis (22% vs. 61%; P < .001). a
L. crispatus 96 (32) 91 (95) 9 (9)
Except for Escherichia coli, there was no significant decrease L. jensenii
a
69 (23) 65 (94) 5 (7)
in colonization of other vaginal bacterial species, such as Urea- L. gasseri 14 (5) 10 (7) 6 (43)
plasma urealyticum, Group B Streptococcus and Enterococcus, L. ruminis 1 (.3) 9 (0) 1 (100)
L. reuteri 1 (.3) 0 (0) 0 (0)
and Staphylococcus species. E. coli was less likely to be recov- L. fermentum 1 (.3) 0 (0) 0 (0)
ered from women with L. crispatus or L. jensenii (15% vs. 27%; L. oris 1 (.3) 0 (0) 0 (0)
P = .01). L. vaginalis 1 (.3) 0 (0) 0 (0)
Lactobacillus 1086V 44 (15) 4 (9) 16 (36)
No homology 13 (4) 4 (31) 6 (46)
No lactobacilli 87 (29) NA 73 (84)
Discussion
NOTE. Values are no. (%) of the 302 women. NA, not applicable.
a
L. crispatus and L. jensenii are the predominant vaginal Lac- Both L. crispatus and L. jensenii were recovered from 11 women, with 154

Downloaded from http://jid.oxfordjournals.org/ by guest on September 5, 2016

tobacillus species colonizing women of reproductive age. L. ac-
idophilus was not recovered from any of the women in this
study. This study confirms the DNA homology study of Giorgi
et al. [14], who did not identify L. acidophilus in a group of 27
asymptomatic women but did find strains homologous to L.
gasseri, L. jensenii, and L. crispatus. In light of the differences
in species identification by use of phenotypic and DNA-based
methods, we encourage the use of genomic-based methods for
identifying lactobacilli to the species level. The identification of
women being colonized by 1 or both of these species.
L. crispatus and L. jensenii allows us to focus on the potential
benefits of these species in the vagina.
To our knowledge, this is the first published report to support
the species specificity of H2O2 production by lactobacilli. Most
clinical isolates of L. crispatus and L. jensenii produce H2O2,
compared with only a few isolates of L. 1086V. McGroarty et
al. [26] and Eschenbach et al. [21] both reported that all of

Figure 1. Identical sample portions of slot blot hybridization autoradiographs, using whole-chromosomal probes made from American Type
Culture Collection (ATCC) Lactobacillus crispatus strain 33197 (A) and ATCC L. gasseri strain 4963 (B). Same samples were used and arranged
in a similar fashion for each autoradiograph shown. Columns ac include DNA samples of unknown vaginal lactobacilli. Column d consists of
DNA samples of ATCC Lactobacillus strains: L. crispatus 33197 (d1), L. vaginalis 49540 (d2), L. plantarum 14917 (d3), L. catenaformis 25536
(d4), L. buchneri 11579 (d5), L. buchneri 4005 (d6), L. gasseri 4963 (d7), and L. gasseri 9857 (d8). Specificity of each whole-chromosomal probe
was evidenced when DNA samples of ATCC strains (identical to probe species) were positive. In A, L. crispatus ATCC 33197 whole-chromosomal
probe hybridized only to L. crispatus ATCC 33197 sample DNA (d1) on control filter. Clinical vaginal lactobacilli homologous to probe were
identified at a1a4, a6, a7, b4, b5, b7, b8, c6, and c7. In B, L. gasseri ATCC 4963 whole-chromosomal probe hybridized only to DNA of ATCC
L. gasseri strains (d7, d8) on control filter. One unknown vaginal Lactobacillus isolate was homologous to probe (c4).
1954 Antonio et al. JID 1999;180 (December)

Table 2. Demographic characteristics and birth control usage of vaginalis, anaerobic gram-negative rods and M. hominis were
women colonized or not colonized by Lactobacillus species, stratified less frequently recovered from women colonized by these H202-
by species of lactobacilli.
producing lactobacilli, compared with women not colonized by
No. (%) of women with Lactobacillus species
these species. Klebanoff et al. [27] demonstrated in vitro that
L. crispatus Other None were H2O2 produced by vaginal lactobacilli was the source of the
a b c
Characteristic or L. jensenii species isolated P
cidal activity against G. vaginalis and Prevotella bivia (formerly
Age !20 years 57/154 (37) 31/60 (52) 41/87 (47) .05 Bacteroides bivius). This activity was observed either in the pres-
White race 111/154 (72) 31/60 (52) 42/87 (48) !.001
Unmarried 129/154 (84) 48/60 (80) 81/87 (93) .48 ence of a halide-peroxidase complex or alone.
Contraception An association between H2O2-producing lactobacilli and de-
Oral 29/138 (21) 16/54 (30) 8/78 (10) .52 creased diagnosis of BV also has been reported in pregnant
Barrier 57/138 (41) 16/54 (30) 19/78 (24) .008
Douche 12 times/ 5/146 (3) 5/58 (9) 6/86 (7) .12 women [28]. This is noteworthy, since BV has been linked to
month pregnancy complications, such as preterm birth [29] and his-
NOTE. Data are no. positive for characteristic/no. positive for lactobacilli tologic chorioamnionitis [30]. The vaginal colonization of L.
(%).
a
crispatus or L. jensenii during pregnancy may offer protection
Includes women colonized by L. crispatus and/or L. jensenii, 14 of whom
against these complications. Concentrations of >107 cfu of vag-

Downloaded from http://jid.oxfordjournals.org/ by guest on September 5, 2016

were also colonized by other species of lactobacilli.
b
Sixty-one women were colonized by species other than L. jensenii or L. inal lactobacilli per gram of vaginal fluid have been linked to
crispatus. Demographic data were not available for 1 woman.
c
a decreased incidence of preterm delivery [31].
Compares women with L. crispatus or L. jensenii with women not colonized
We do not know whether L. crispatus and L. jensenii com-
by these species, including women lacking lactobacilli.
plement one another. However, it is beneficial to know that
none of the women colonized by both of these species were
their identified L. jensenii isolates were H2O2 producers. Nagy
diagnosed with BV. In this study, 8% of the women were col-
et al. [8], however, reported H2O2 production by only 56% of
onized by 11 Lactobacillus species. The L. crispatus and L.
L. jensenii strains. Although some Lactobacillus species may be
jensenii combination accounted for nearly half of the Lacto-
biochemically identifiable, the discrepancy of the latter result
bacillus combinations found. In addition, Lactobacillus 1086V
may be due to the unreliability and irreproducibility of the
was coisolated with L. crispatus or L. jensenii in 40% of the
classic methods of identification between laboratories. In ad-
women colonized by a species combination. This prevalence is
dition, the percentage of L. acidophilus isolates producing H2O2
unexpected, contradicting the claim that no 2 women are col-
was reported to be 77% by McGroarty et al. [26], 43% by
onized by the same Lactobacillus species combination [32].
Eschenbach et al. [21], and 71% by Nagy et al. [8]. This disparity
There were significantly fewer diagnoses of gonorrheal in-
may result from the inability of biochemical assays to differ-
entiate between the species formerly belonging to the L. aci-
dophilus group. Table 3. Microbiologic characteristics of women colonized or not
colonized by lactobacilli.
In this study, an increased frequency of L. crispatus or L.
jensenii was significantly linked to a lower prevalence of BV. No. (%) of women with Lactobacillus
isolated
L. crispatus or L. jensenii colonized three-fourths of women
L. crispatus Other None
without BV, whereas only 12% of women with BV were col- a b
or L. jensenii species isolated
onized by either species (table 3). Similarly, Eschenbach et al. Characteristic (n = 154) (n = 60) (n = 87) P
c

[21] reported that L. acidophilus or L. jensenii colonized 64% Neisseria gonorrhoeae 1 (1) 1 (2) 6 (7) .03
of 28 uninfected women and only 7% of 67 BV-positive women. Chlamydia trachomatis 9 (6) 4 (7) 4 (5) .87
This differs from the study by Nagy et al. [8], which suggested Bacterial vaginosis 14 (9) 29/61 (48) 73 (84) !.001
Gardnerella vaginalis 59 (38) 47 (78) 77 (89) !.001
that there were no differences in the Lactobacillus species dis- Anaerobic gram-
tribution between asymptomatic women without BV and negative rods
women with BV. Nonpigmented 59 (38) 38 (63) 75 (86) !.001
Black-pigmented 17 (11) 14 (23) 34 (39) !.001
The clinical relevance of H202-producing isolates of lacto- Mycoplasma hominis 34 (22) 29 (48) 60 (69) !.001
bacilli was suggested by Hawes et al. [1] in a longitudinal study Ureaplasma urealyticum 91 (59) 32 (53) 59 (68) .67
of nonpregnant women. After an adjustment was made for Group B Streptococcus 27 (18) 17 (28) 19 (22) .15
Enterococcus species 20 (13) 4 (7) 20 (23) .43
douching and having multiple sex partners, it was shown that Staphylococcus species 5 (3) 4 (7) 7 (8) .11
nonpregnant women lacking H202-producing vaginal lactoba- Escherichia coli 23 (15) 13 (22) 26 (30) .01
cilli were twice as likely to develop BV than were women col- a
Includes women colonized by L. crispatus and/or L. jensenii, 14 of whom
onized by H202-producing lactobacilli. As the primary species were also colonized by other species of lactobacilli.
b
of lactobacilli that colonize the vagina and produce H2O2, L. There were 61 women colonized by species other than L. jensenii or L.
crispatus. Complete microbiologic data were not available for 1 woman.
crispatus and L. jensenii may protect a woman from developing c
Compares women with L. crispatus or L. jensenii with women not colonized
BV by inhibiting the growth of BV-related microorganisms. G. by these species, including women lacking lactobacilli.
JID 1999;180 (December) L. crispatus and L. jensenii
fection in women colonized by L. crispatus or L. jensenii, com-
pared with women not colonized by either species or lacking
lactobacilli. Previously, Saigh et al. [33], who were studying the
inhibition of N. gonorrhoeae by lactobacilli, reported that the
presence of inhibitory lactobacilli may be protective for women
with gonorrhea-infected partners, since fewer of these women
subsequently developed gonorrhea [33]. Unfortunately, it is un-
known whether the presence of vaginal lactobacilli benefits a
woman with a gonorrhea-infected partner by decreasing the
efficiency of transmission. Hawes et al. [1] found that women
colonized by H2O2-producing lactobacilli were less frequently
infected by N. gonorrhoeae than were women lacking H2O2-
producing lactobacilli. Zheng et al. [34] discovered that pro-
duction of lactic acid and a catalase inhibitor by lactobacilli
inhibits N. gonorrhoeae growth. It is unknown whether the
production of catalase inhibitor is universal among lactobacilli.
The absence of lactobacilli, as determined by Gram stain, is
positively associated with an increased prevalence of HIV [25].
Some authors have suggested that treatment of BV may reduce
HIV transmission, because women with abnormal Gram stain
scores had an increased risk of HIV, compared with women
with predominant vaginal lactobacilli [3]. One recent study of
women in Zimbabwe reported an increased rate of HIV sero-
conversion among pregnant women with BV [35], compared
with those without clinical evidence of BV. Since women col-
onized by L. crispatus and L. jensenii are less likely to have BV,
further studies are needed to evaluate whether these 2 Lacto-
bacillus species may be protective against HIV. Klebanoff et al.
[36] implicated H2O2, either alone or in combination with a
halide and a peroxidase, in the cidal activity against HIV. They
suggested that women lacking H2O2-producing lactobacilli may
be at greater risk of being infected with HIV.
The identification of L. crispatus and L. jensenii as the pre-
dominant vaginal lactobacilli narrows the list of potential pro-
biotic strains and focuses research to either species or to both.
These 2 species may offer protection against some vaginal in-
fections, such as BV and gonorrhea. These species inherently
meet 2 requirements for use in successful probiotic therapy:
They are endogenous vaginal lactobacilli, and they adhere to
vaginal epithelial cells [37]. In addition, these 2 species are
known to produce H2O2, a well-accepted mode of host defense.
Vaginal colonization of women with these species may be ad-
vantageous in the maintenance of a normal microflora and the
prevention of sexually transmitted diseases. Randomized, con-
trolled trials will be needed to test this hypothesis.
References
1. Hawes SE, Hillier SL, Benedetti J, et al. Hydrogen peroxideproducing lac-
tobacilli and acquisition of vaginal infections. J Infect Dis 1996; 174:
105863.
1955
2. Cohen CR, Duerr A, Pruithinithada H, et al. Bacterial vaginosis and HIV
seroprevalence among female commercial sex workers in Chiang Mai,
Thailand. AIDS 1995; 9:10937.
3. Sewankambo N, Gray RH, Wawer MJ, et al. HIV-1 infection associated with
abnormal vaginal flora morphology and bacterial vaginosis. Lancet
1997; 350:54650.
4. Taha ET, Gray RH, Kumwenda NI, et al. HIV infection and disturbances
of vaginal flora during pregnancy. J AIDS Hum Retrovirol 1999; 20:529.
5. Royce RA, Thorp J, Granados JL, Savitz DA. Bacterial vaginosis associated
with HIV infection in pregnant women from North Carolina. J AIDS
Hum Retrovirol 1999; 20:3826.
6. Rogosa M, Sharpe ME. Species differentiation of human vaginal lactobacilli.
J Gen Microbiol 1960; 23:197201.
7. Levison ME, Corman LC, Carrington ER, Kaye D. Quantitative microflora
of the vagina. Am J Obstet Gynecol 1977; 127:805.
8. Nagy E, Petterson M, Mardh PA. Antibiosis between bacteria isolated from
the vagina of women with and without signs of bacterial vaginosis. APMIS
1991; 99:73944.
Downloaded from http://jid.oxfordjournals.org/ by guest on September 5, 2016
9. Rogosa M, Wiseman RF, Mitchell JA, Disraely MN, Beaman AJ. Species
differentiation of oral lactobacilli from man including descriptions of Lac-
tobacillus salivarius nov spec and Lactobacillus cellobiosus nov spec. J
Bacteriol 1953; 65:68199.
10. Thomas S. Doderleins bacillus: Lactobacillus acidophilus. J Infect Dis
1928; 43:21827.
11. Stahl M, Molin G, Persson A, Ahrne S, Stahl S. Restriction endonuclease
patterns and multivariate analysis as a classification tool for Lactobacillus
spp. Int J Syst Bacteriol 1990; 40:18993.
12. Johnson JL, Phelps CF, Cummins CS, London J, Gasser F. Taxonomy of
the Lactobacillus acidophilus group. Int J Syst Bacteriol 1980; 30:5368.
13. Collins MD, Phillips BA, Zanoni P. 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 1989; 39:1058.
14. Giorgi A, Torriani S, Dellaglio F, Bo G, Stola E, Bernuzzi L. Identification
of vaginal lactobacilli from asymptomatic women. Microbiologica 1987;
10:37784.
15. Stamm WE, Tam M, Koester M, Cles L. Detection of Chlamydia trachomatis
inclusions in McCoy cell cultures with fluorescein-conjugated monoclonal
antibodies. J Clin Microbiol 1983; 17:6668.
16. Knapp JS, Holmes KK. Disseminated gonococcal infections caused by Neis-
seria gonorrhoeae with unique nutritional requirements. J Infect Dis
1975; 132:2048.
17. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial va-
ginosis is improved by a standardized method of Gram stain interpre-
tation. J Clin Microbiol 1991; 29:297301.
18. Krohn MA, Hillier SL, Lee ML, Rabe LK, Eschenbach DA. Vaginal Bac-
teroides species are associated with an increased rate of preterm delivery
among women in preterm labor. J Infect Dis 1991; 164:8893.
19. Marray PR, Baron EF, Pfaller MA, Tenover FC, Yolken RH, eds. Manual
of Clinical Microbiology. 6th ed. Washington, DC: American Society for
Microbiology, 1995.
20. Holdeman LV, Cato EP, Moore WEC, eds. Anaerobe laboratory manual.
4th ed. Blacksburg, VA: Anaerobe Laboratory Virginia Polytechnic In-
stitute and State University, 1977:1438.
21. Eschenbach DA, Davick PR, Williams BL, et al. Prevalence of hydrogen
peroxideproducing Lactobacillus species in normal women and women
with bacterial vaginosis. J Clin Microbiol 1989; 27:2516.
22. Luchansky JB, Tennant MC, Klaenhammer TR. Molecular cloning and
deoxyribonucleic acid polymorphisms in Lactobacillus acidophilus and
Lactobacillus gasseri. J Dairy Sci 1991; 74:3293302.
23. Roberts MC, Hillier SL, Schoenknecht FD, Holmes KK. Comparison of
Gram stain, DNA probe, and culture for the identification of species of
Mobiluncus in female genital specimens. J Infect Dis 1985; 152:747.
1956 Antonio et al.
24. Gebhart CJ, Murtaugh MP, Lin GF, Ward GE. Species-specific DNA probes
for Campylobacter species isolated from pigs with proliferative enteritis.
Vet Microbiol 1990; 24:36779.
25. Sambrook J, Fritsch EF, Maniatis T. Hybridization of radiolabeled probes
to immobilized nucleic acids. In Molecular cloning: a laboratory manual.
2d ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press,
1989:9.479.55.
26. McGroarty JA, Tomeczek L, Pond DG, Reid G, Bruce AW. Hydrogen per-
oxide production by Lactobacillus species: correlation with susceptibility
to the spermicidal compound nonoxynol-9. J Infect Dis 1992; 165:11424.
27. Klebanoff SJ, Hillier SL, Eschenbach DA, Waltersdorph AM. Control of
the microbial flora of the vagina by H2O2-generating lactobacilli. J Infect
Dis 1991; 164:94100.
28. Hillier SL, Krohn MA, Klebanoff SJ, Eschenbach DA. The relationship of
hydrogen peroxideproducing lactobacilli to bacterial vaginosis and gen-
ital microflora in pregnant women. Obstet Gynecol 1992; 36972.
29. Hillier SL, Nugent RP, Eschenbach DA, et al. Association between bacterial
vaginosis and preterm delivery of a low-birth-weight infant. N Engl J Med
1995; 333:173742.
30. Hillier SL, Martius J, Krohn M, Kiviat N, Holmes KK, Eschenbach DA.
A case-control study of chorioamnionic infection and histologic chorioam-
nionitis in prematurity. N Engl J Med 1988; 319:9728.
JID 1999;180 (December)
31. Krohn MA, Hillier SL, Lee ML, Rabe LK, Eschenbach DA. Vaginal Bac-
teroides species are associated with an increased rate of preterm delivery
among women in preterm labor. J Infect Dis 1991; 164:8893.
32. Redondo-Lopez V, Cook RL, Sobel JD. Emerging role of lactobacilli in the
control and maintenance of the vaginal bacterial microflora. Rev Infect
Dis 1990; 12:85672.
33. Saigh JH, Sanders CC, Sanders WE. Inhibition of Neisseria gonorrhoeae by
aerobic and facultatively anaerobic components of the endocervical flora:
evidence for a protective effect against infection. Infect Immun 1978; 19:
70410.
34. Zheng H, Alcorn TM, Cohen MS. Effects of H2O2-producing lactobacilli on
Neisseria gonorrhoeae growth and catalase activity. J Infect Dis 1994; 170:
120915.
35. Taha TE, Kumwenda NI, Liomba GN, et al. Bacterial vaginosis and dis-
turbances of vaginal flora: association with increased acquisition of HIV.
AIDS 1998; 12:1699706.
36. Klebanoff SJ, Coombs RW. Viricidal effect of Lactobacillus acidophilus on
human immunodeficiency virus type 1: possible role in heterosexual trans-
Downloaded from http://jid.oxfordjournals.org/ by guest on September 5, 2016
mission. J Exp Med 1991; 174:28992.
37. Wood JR, Sweet RL, Catena A, Hadley WK, Robbie M. In vitro adherence
of Lactobacillus species to vaginal epithelial cells. Am J Obstet Gynecol
1985; 153:7403.