0095-1137/82/110926-09$02.00/0
Copyright©1982,AmericanSocietyforMicrobiology
Development and Evaluation of
a
Biochemical Scheme for
Identification of Endocervical
Lactobacilli
JOANE. FAGNANT,CHRISTINEC. SANDERS,*AND W.EUGENE SANDERS, JR. DepartmentofMedicalMicrobiology, Creighton UniversitySchoolofMedicine, Omaha, Nebraska 68178
Received4June1982/Accepted 18 August 1982
A biochemical scheme for the species identification of endocervical lactobacilli wasdeveloped and evaluated with 10 isolates obtained from the AmericanType Culture Collection (ATCC) and 106 endocervical isolates obtained from women reportingto alocal venereal disease clinic and a localhospital clinic. The scheme consisted oftwo stages. Stage I included six tests and was tested and modified withresults obtainedwith ATCC strains. From the modifiedstage I,stage IIwas
developed. Tests to be performed in this stage were determined from expected characteristicsof lactobacilli. StageIIwas also tested with the ATCC strains. Of the106 endocervicalisolates,78(74%)wereidentified withthetwo-stage scheme asdevelopedwith the ATCC strains. Unexpectedresults were obtainedin one or both stages with the other 28 isolates. For 10 isolates, the final species identified werenot
previously
expected
tobe recovered. A "best-fit" methodwasusedtodetermine the most likely identification of the remaining 18 isolates. In a few instances, theuse ofathird stage was necessaryto reach an identification. The final identification scheme, although complicated in appearance, generated a
species identification with a total of 12 tests with a range of 7 to 10 tests per isolate.
Lactobacilli are
ubiquitous microorganisms
inhabiting
numerous ecological niches. There aremany reportsinthe literatureonthe identifi-cation oflactobacilli fromavariety
ofsources.Briggs (2), in 1953, was the first to attempt to
standardize the classification of lactobacilli
by
physiological
tests. Serology, colonial morphol-ogy,nutritionalrequirements,
andcultural,
cel-lular, and macromolecular characteristics have also been utilized in attempts to develop anidentification scheme(4, 6,7, 10, 11,13, 14, 16, 17).However, todate,thereisnowidely accept-ed schemefortheidentification ofspecies within thegenusLactobacillus. There are several rea-sonsfor the lack of asimple, broadly applicable identification scheme: (i) the particular species present vary widely from niche to niche; (ii) lactobacilli are usually considered to be non-pathogenic; (iii) the lactobacilli are relatively fastidiousorganisms.
The results of apreviousstudy indicated that endocervical lactobacilli may play a role in resistance of females to gonorrhea (15). We wishedto pursue these studies further to deter-mine whether anyparticular species were more important than others in this natural defense mechanism.However, there was no simple iden-tification scheme in the published literature. Therefore, the present study was designed to developand evaluate asimplebutreliable
identi-fication scheme for the lactobacilli
inhabiting
the femalegenital tract.MATERIALS ANDMETHODS
Strains. Stock strains obtainedfrom the American Type Culture Collection (ATCC), Rockville, Md., were(i) L. acidophilus ATCC 4356; (ii) L.leichmannii
ATCC 4797; (iii) L. lactis ATCC 12315; (iv) L. lactis ATCC8000;(v)L.brevis ATCC 14869; (vi) L.
fermen-tumATCC 14931; (vii)L.salivarius subsp. salivarius ATCC 11741;(viii)L.cellobiosus ATCC 11739; (ix) L. cellobiosus ATCC 11740; and (x) L. casei subsp. rhamnosus ATCC 7469. Atotal of 106 endocervical isolates were obtained from women reporting to a local venereal diseaseclinic and a local hospital clinic.
Endocervicalcultures.Endocervical specimens were obtained with a sterile calcium alginate swab (Calgi-swab, Inolex Corp.). The swab was immediately placed in 1 ml of eugonic broth (Difco Laboratories)
andblended in aVortex mixer for 3min. The surfaces ofa dextrose starch agar (Difco) plate supplemented with 5% sheep blood (Colorado Serum Co.) and a lactobacillus selective agar (LBS, BBL Microbiology Systems) plate were inoculated with swabs dipped into theeugonic broth. The plates were streaked for isola-tion in the four-quadrant fashion and incubated in a candlejar for 48 h at 37°C. Thus, only aerobic and facultatively anaerobic lactobacilli were examined.
Identification of lactobacilli. Gram-positive, cata-lase-negative rods forming small translucent colonies
onLBSwere identifiedpresumptively as lactobacilli.
To ensure a pure isolate, a single-colony cloning procedure was employed. This involved aspirating a
926
on February 7, 2020 by guest
http://jcm.asm.org/
IDENTIFICATION
TABLE 1. Stage I identification testsa Value
Group 1 test assigned for Group2 test apositive
testb
Acid fromglucose 1 Acid from lactose Gasfrom glucose 2 Acid from mannitol Acidfromamygdalin 4 Acidfrom galactose
a Sumof results from group 1 tests = first digit of stageIidentification value(0 to 7). Sum of results from group 2 tests = seconddigit of stage I identification value(0 to 7).
bAllnegative tests were assigned a value of zero.
colony into a sterile Pasteur pipette from the agar surface with theuseofadissectingmicroscope (30x). The colony wasplaced into broth and incubated for 24 to 48 h. The broth was subcultured again onto agar. Fromthis subculture,asingle colony was selected as described above, and thecloning procedure was
re-peated again. Further tests, including some orall of the following, were performed on each isolate for identification.
Testsfor the fermentation of glucose,lactose,
man-nitol, galactose, cellobiose, salicin, sorbitol, and xy-lose were performed in LBS broth base (14). The
carbohydrateswereaddedtotheLBS broth base ina
final concentration of 2%. For all except glucose,
sterile solutions of the carbohydrates wereadded to
thebroth base afterautoclaving. Thefinal volume of thesemediawas3ml. Glucose wasaddedtothe base beforeautoclaving,andthe mediaweredispensed in 8-mlaliquots. A Durham tube was inverted into each
glucose-containingmedium, and the mediumwasthen sterilized. All tubes in which the Durham tubes
con-tained bubbles after sterilizing were discarded. The final pHwas5.9.
Themedium foramygdalin utilization consisted of 1.5% neopeptone, 0.1% Tween80, 0.6% yeast extract, 0.15% agar, and 0.4% chlorophenol red (19).
Amygda-linwasaddedtothesterile medium inafinal
concen-trationof0.2%. The final pHwas7.0to7.2. The inoculum was prepared by suspending the growth (aftersingle-colony cloning)froma2-day
eu-gonicagarorblood agarplate in 9 mlofsaline. The salinewascentrifugedtoobtainapellet.The
superna-tant wasdiscarded,and thepelletwasresuspendedin the smallamount(ca. 0.2ml)of salineremainingin the tube. Each tubewasinoculated withone or twodrops
of this heavycellsuspension (ca. 109colony-forming units/ml). After inoculation of theglucosemedium,1
mlof melted sterileparaffin waslayeredoverthetop of the medium. Inoculated tubeswereincubated for14
daysin10%oCO2inairat37°Candwereexaminedon days 1, 2, 4,7, and14.Mostcolorchangesindicatinga
positivetestoccurredwithin2days.Inthe LBSbroth, acolor change from slightly green to lemon yellow
indicatedapositivereaction.Fortheamygdalintest,a colorchange from darkmaroon to scarlet red witha
yellowtingewas apositivereaction.
Thearginine hydrolysis medium consisted of 0.3% yeastextract,0.5% peptone, 1% sodium acetate, 0.3% glucose,0.1%Tween80,and0.5% salt solution(3, 14). The final pH was 7.4. The sterile
L-arginine-hydro-chloride was added in afinal concentration of 0.3%
after the mediumwassterilized.Tubeswere
inoculat-ed in the same manner as the LBS tubes and were
incubated for 10 days at37°C in10o C02 in air. A positivetestconsisted ofanorangecolor afterNessler
reagentwasadded (8).
The ability ofastraintogrowat45°Cwastestedin 5
mlofEugon broth. Inoculated tubeswereincubatedin acandle jarat45°C until growthwas apparent. The positive tubeswerethensubculturedtoensurethatno
contamination hadoccurred. All negative tubeswere
heldfor7days before being discarded.
RESULTS
Aschemewasdeveloped for the identification
ofspecies oflactobacilli reportedly isolated from humans (9). It was designed to identify these species withaminimum number of biochemical tests.
Stage I. Stage I was developed from those
characteristics of lactobacilli commonly
recov-eredfrom humansourcesand listed inBergey's
Manual ofDeterminative Bacteriology (9). A
minimum of sixtestsappearedtobe sufficientto
yield an identification of most species. These
were(i) acidfrom glucose, (ii)gasfrom glucose,
(iii) acid from amygdalin, (iv) acid from lactose, (v)acid from mannitol, and (vi) acid from galac-tose. Eachpositivetestwasassigned a
numeri-cal value (Table 1). Each negative test was
assignedavalueofzero.Thesumsof the values
of thefirstthreetestsand the second threetests
were determined, and these sums provided a
two-digit numeral. The two-digit stage I value
was used to provide an initial identification
(Table 2).
Thevalidity of this schemewasdetermined by
testing 10 reference strains of various speciesa
TABLE 2. Identification of lactobacilliby using
stageIvaluesa
Expected Otherpossible
stageI Mostlikely species species value
74, 75 L.cellobiosus
57 L. caseisubsp.rhamnosus
55 L.acidophilus L. leichmannii 54 L.jensenii
51 L.leichmannii 37 L. brevis
36 L. brevis
35 L.fermentum L. brevis
34 L.brevis 33 L. brevis 32 L.brevis
31 L. brevis 30 L.brevis
17 L.salivarius
15 L.lactis
aDerived from results of stageItests aspredicted
bycharacteristics listed inBergey'sManualof
Deter-minativeBacteriology(9).
VOL. 16,1982
on February 7, 2020 by guest
http://jcm.asm.org/
928 FAGNANT, SANDERS, AND SANDERS
TABLE 3. Stage Ivaluesmodified from results of multipletestswithnine reference strainstoinclude
inherently variabletests
Stage I Mostlikelyspecies Other possible
value species
74, 75 L.cellobiosus
57 L.casei subsp. rhamnosus L. cellobiosus 55 L.acidophilus L. leichmannii 54 L.jensenii
51 L.leichmannii
50 L.leichmannii
37 L.brevis L.cellobiosus 36 L.brevis
35 L.fermentum, L.brevis L.cellobiosus 34 L.brevis L.cellobiosus 33 L.brevis
32 L. brevis 31 L.brevis 30 L.brevis 17 L.salivarius
15 L.lactis L. acidophilus
minimum of four times. All but three strains generated the stage I value predicted by Ber-gey'sManualofDeterminativeBacteriologyfor
the species they represented in the majority (68
to 100%) of replicate tests. One strain, L. lactis (ATCC 12315),washighly variable in thestageI values it generated. In 68% of the replicatetests,
it generated a value correspondingto L. leich-mannii. It deviated from the expected results in fourmajortestsandwastherefore consideredto
be inappropriate for use as a control strainfor
this species. Neither L. cellobiosus control strain generated the stage I value expected. Amygdalinwasnotutilizedby either strain, thus generating a stage I value of 35 that
corre-sponded to an identification of L. fermentum.
AccordingtoBergey's Manual ofDeterminative Bacteriology, an additional distinguishing
char-acteristic betweenL.cellobiosus and L. fermen-tumis theability of the former butnotthe latter
toferment cellobiose. Since both ATCC strains
of L. cellobiosus fermented cellobiose, they
were considered to be valid control strains. Stage I was modified, however, to include L.
cellobiosus under "other possible species" for strainsgeneratingastage I value of 34 or35.
Further modifications to stage I were neces-sary when it became apparent that variable results were obtained in certain specific tests evenwhen theywere performed in anidentical
fashion with the samecontrol strain. The
vari-able tests were amygdalin for L. acidophilus,
lactose forL.leichmannii,andmannitolandgas
from glucose forL. cellobiosus. Results ofthese
tests varied from the expected in over 10% of
tests, indicating that the test was inherently
variable for the particular species. The interpre-tation ofstage I valueswasmodified toinclude values generated due to "inherently variable tests" (Table 3). Withthe modifications indicat-ed,stageI should becorrectatleast 90%of the time.
Stage II. From the modified stageI, a second stage oftests was developed to distinguish
be-tween those species giving the same stage I values and to verify the species suggested by
stageI. Asbefore, thetests tobe performed in
stage II were determined from characteristics
listedinBergey'sManualofDeterminative Bac-teriology and evaluated with the nine reference strains. The tests performed in stage II varied
dependinguponthestageIvalue and
represent-ed the minimum number of tests required to
differentiate the strain from all species consid-eredcapable of generating thestageI value.The
stage II scheme and expected results for each speciesare showninTable 4.
TABLE 4. StageIIidentification of lactobacilli
Stage I value StageII test(result) Identification
74, 75 Cellobiose (+) L. cellobiosus
57 Cellobiose(+); sorbitol (+) L. casei subsp. rhamnosus
Cellobiose (+); sorbitol (-) L. cellobiosus
55 Cellobiose(+);arginine (+) L. Ieichmannii
Cellobiose(+);arginine (-) L. acidophilus
54 Cellobiose (+) L.jensenii
51 Cellobiose(+);arginine (+) L.leichmannii
50 Cellobiose (+); arginine (+) L.Ieichmannii
37, 34 Cellobiose (+); growth at45C(-) L. cellobiosus Cellobiose(-);growth at 45°C (-) L. brevis
36 Cellobiose (-); growth at 45°C (-) L. brevis
35 Cellobiose (-); growth at 45°C (+) L. fermentum Cellobiose(-);growth at 45°C (-) L. brevis Cellobiose(+); growthat 45°C (-) L. cellobiosus
30-33 Cellobiose (-); growth at 45°C (-) L. brevis
17 Cellobiose (-); sorbitol (+) L. salivarius
15 Cellobiose(-) L. lactis
Cellobiose (+) L. acidophilus
on February 7, 2020 by guest
http://jcm.asm.org/
TABLE 5. Identificationof lactobacilli recovered on endocervical cultures
Stage I StageII
No.of
Value isolates No. of Identification
obtained generating Test (result) isolates
value giving result
75 5 Cellobiose (+) 5 L. cellobiosus
74 4 Cellobiose (+) 4 L. cellobiosus
55 23 Cellobiose(+); arginine (+) 6 L. leichmannii
Cellobiose(+); arginine (-) 17 L. acidophilus
54 24 Cellobiose(+) 24 L.jensenji
50 11 Cellobiose(+);arginine (+) 11 L. leichmannii 35 6 Cellobiose(-); growth at 45°C (+) 2 L. fermentum
Cellobiose(-); growth at 45°C (-) 4 L. brev'is
17 1 Cellobiose (-); sorbitol (+) 1 L. salivarius
15 4 Cellobiose (+) 4 L. acidophilus
When the control strains wereused to assess the validity of the stage II tests, all strains produced the expected results in over 80% of replicate runs ofeach test, with one exception. L. brevis fermented cellobiose in three offour tests. L. brevis and L. cellobiosus differed in theirability to ferment cellobiose and amygda-lin. The only situation in which it would be necessary to differentiate between these two organisms in stage I isin the instance ofa rare amygdalin-negative strain of L. cellobiosus. Therefore, the cellobiose test was left in the scheme with the assumption that the L. brevis control strainwas ararecellobiose-positive iso-late.
Use of the scheme for the identification of endocervical lactobacilli. Atotal of106isolates of lactobacilli were obtained from endocervical cultures from 72women. Thespecies of Lacto-bacillus to which a strain belonged could be determined for78
(74%)
ofthestrains (Table 5). Theremaining 28 isolates generated either unex-pected stage I values or unexpected stage II results or both. Modification and expansion of thetwo-stage scheme wasrequired
for identifi-cation ofthese 28 isolates.Ten isolates gave unexpected stage I values. A list ofspecies most likely to have given the unexpected stage I values was
prepared
by (i)
assumingthat thefirst digit ofthe stage I value wascorrectandselectingthe mostlikely species from Table 3 to generate the first number,(ii)
selecting the most likely species from Table 3that could generate the
unexpected
value ob-served by assuming only one ofthe six stage I tests was aberrant, and(iii)
selecting additional species listedinBergey's Manualof
Determina-tive
Bacteriology
that could havegenerated
the unexpected stage I value. These lastspecies
were not included in the schemeoriginally
asthere wasnoindication that they had ever been recovered from the human genital tract. After the most likely species had been determined
(from i to iii), stage II tests were devised to differentiateamongthem. Theisolatewasfinally identified by usinga "best-fit"analysis, i.e., for each speciesonthe list, the number of aberrant testresultswasdetermined, and the species with thefewest aberranttests wasselected. With this system, the most likely species for the 10 iso-lates weredetermined (Table 6).
A similar method of"best-fit" analysis was employed to determine the identification of the remaining18isolatesthat gave unexpected stage II results based ontheir stage I values. Criteria usedinselectingthemost likelyspecieswere:(i) anassumptionthat the stage Ivaluewas correct and the aberrant result was in stage II, (ii) an assumption that the stage II result was correct and the aberrant result was in stage I, and (iii) selection of additional species from Bergey's ManualofDeterminative Bacteriology. In many cases,application of these three criteria generat-edonlyonespecies for whichasingle character-istic was aberrant (Table 7). Thus, no further tests were performed as this species was the most likely. Inother instances, additional tests (Stage III) were necessary to determine the identification from among several possible spe-cies.
Finalidentificationscheme. Byusingresultsof testswith the 106 isolates, a final identification scheme was derived (Table 8). The scheme generated in two or three stages either
(i)
a"final identification" if the strain possessed characteristics expected ofa particular species (Table4), or
(ii)
a"mostlikely identification" if the strainpossessedaberrantcharacteristicsand a "best-fit" analysis wasapplied.
Thisscheme,
althoughcomplicated
in appearance,required
theuseofonlyatotal of 12 tests, witharange of 7 to10 tests per isolate.DISCUSSION
SinceDoderlein (5)firstdescribed the
lactoba-VOL.16, 1982
on February 7, 2020 by guest
http://jcm.asm.org/
Zs s -Ze
.0 .0
. .
+I+
C)C)C)
7;CZ)E~ CZ )C) C
___ _ _ __C) C)
-#: CS o U#C U)C; CU) _
C U t
~~~~~ ~ _ ~~~~ Q U _- _- ^ _-C.. .0~ C) C)U) o.0 >0 o0 on0 C-0 _ *_ *_U2 C..C)a
cO et _
eo eo
0 _ c
e= r: C.
00u:
0 0A
CiC 0U)
OC O C)
0
o o 0.0
L C)
°S-"
00 0o0 O
"Y EZ
Um
m Ucilli in thevagina in 1892, most early investiga-tions assumed that the lactobacilli recovered from the female genital tract were all members of one species, L. acidophilus. However, it has been shown subsequently that these organisms do notform ahomogeneousgroup(12). Thus, it was apparent that any study of the species of endocervical lactobacilli would require the use ofasimple but reliable identification scheme. A search of the literature revealed that no such scheme exists (6, 9, 18). Most identification schemes involved the use of a multitude of tests, some of which would require expensive equip-ment (9, 12, 14, 16-18). Thus, an identification scheme wasdeveloped in this studythat would allow maximum identification of lactobacilli with the use of aminimumnumberof biochemi-cal tests.
With the identification scheme developed for this study, most isolates (74%) of lactobacilli recoveredfrom thegenitaltractofwomen were readilyidentified. Theremaining26% were iden-tified by using a best-fit analysis. With some species, there were several strains that varied from the expected reactions according to Ber-gey's ManualofDeterminativeBacteriology (9). Forexample, only 82%of strains ofL. acidophi-lus fermentedamygdalin.This test wasfoundto be inherently variable for the species, as the ATCC type strain fermented amygdalin only 86% of the time. The same was true for the fermentationoflactosebyL.leichmannii.Three of five strains ofL. delbrueckii did not ferment glucose.Since we had noATCCtypestrain with which to compare these lactobacilli, the best-fit
analysis
was used todesignate
these as L. del-brueckii. Of the species of lactobacilli listed in Bergey'sManualofDeterminativeBacteriology (9), L. delbrueckii is the most asaccharolytic. Another aberrant result was found to occur in tests with all endocervical isolates of L. casei subsp. rhamnosus, but in no instance with the ATCC type of the species. No endocervical isolate fermented sorbitol. A stage I value of57 was the only value requiring not only a stage II but a stage III determination as well. Stage III consisted of salicin and xylose to distinguish among three possible species: (i) sorbitol-nega-tive L. caseisubsp. rhamnosus, (ii) L. cellobio-sus producing no gas from glucose, or (iii) L. xylosus. Since all strains fermented salicin but not xylose, the best-fit analysis designated the strains as sorbitol-negative L. casei subsp. rhamnosus. Thisdesignation was also supported bythe VPIidentificationschemefor the lactoba-cilli, which lists weak acid production from sorbitol for this species (18). Thus, from this analysis,itappears that the schemedeveloped in thisstudy permits arelativelyreliable identifica-tion of lactobacilli recovered from theendocer-o 0 Ct u C) 0 a)-cm , - 0 0o> C; CA aL) V) C) C' _ CU C) 0 CL) 0._C) V) 'Z. *_ 0 CT U) 4-C) CU 0-0 CZI Cy) 0. C-0 0 Q. X C) c CU s -0 *_ 0 c '-ce 0 : 0 CZ C)~ H
on February 7, 2020 by guest
http://jcm.asm.org/
cr C crcrn n
C
D D fD CD
z-r: -i
z-uq CD f Df DC
>
-x~~~~
+I)++ + m~~~~~~~Q
n- CD- 2 ww. 2w-w w
Sct-p: p: p: ct -F~CD C
OO1S S m O f~~~~~~~~~~~~D sm'aq m m m ct~~
tyy-- . - - . y
-IDENTIFICATION OF ENDOCERVICAL LACTOBACILLI 931 VOL. 16, 1982
CD
(,bD
CDx
C OO
33
ccim
C D _.Q
3 3:
CDx fi *_
"Cn
=r
CD _
;>
=
r
m <
.
X (3D
_.
>
n &'
°3--^.t
> a ^ 0
>
^ _.
_ Ce
_.
O
&' eCt Ut.
fi_.
_
5
-n3t
nt nt
n
n
C
U:
P:
gq fD
_
_
ND
U:
v
_
Ut. 5 rm
t'}t.
fD
n O
3
5 nt
S-U'_
m
eCt
_ < ;_
fD
U:
C,
3
Zt
t-
on February 7, 2020 by guest
http://jcm.asm.org/
932 FAGNANT, SANDERS,
C0
c
) .l .On
_ _00
-~~~~~~~~~~~~~~~~~0~~~~~~~~~~~~~~~IZIIZ
0 > ++ 0
vt~ a0 4) 0)
0) 00o0o .0 -- 0 )
-_ __ ) *i) ~ >> o. ~ .
oo~~ ~
~~~~~~
.t
..E
ES .os~~....~ % ~~ ~%oq..o.0 0
0 .~~~~~~L.14
U 0
06~~~~~~
0.~~~ 0 0 00 0 .00
.0
:~~~~~~
0 0 0 000~~ 00 01-&. .&. I. .000 0 -
-++++ + +I+I+I+ + +
U U
0 U00 U. U - W W
OOOO O O O O O O O O O . 0 0 0.
0)0 04 00.D.-00 O} O O O O O O O.0O O O 0- 0-. .
w ~ ~~~~~CK Ct Ct (UZ Ct Ct 0C 0 004 ) es 0s es N) 0ut 0 0 cn 0 _ _ C
C~~~~~~~~~~~~~~~~~~~~~0)~~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~C
00 000 000 0c rc acot0c oc r ~ ~ C
O0_00 00 000..0000000°°°°,>D
>
_I O
W VE t ^ O b ot ^ b(.)(..)Q(.)QQOQOQ;>0
S > ^ ^ ^ O O t <~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0
CQ~~~
+1on February 7, 2020 by guest
http://jcm.asm.org/
LACTOBACILLI
vix. Most deviations from Bergey's Manual of Determinative Bacteriology were due either to inherent variability intests or to the isolation of too few strains of aparticular species for accu-rateanalysis.
A comparison of results generated by this identification scheme with those from more complex schemes used by other investigators also helps to confirm its simplicity and reliabil-ity. In 1955, Davis (3) used an identification scheme employingover 30 tests to identify 473 strains oflactobacilli isolated from the human mouth. Of these 30 tests, 8 were used in the present scheme. A comparison of the results of testsperformed byDavis (3) that wereincluded in the scheme developed in the present study revealed excellent agreement between the two studies. The only disagreement with the charac-teristics listed by Davis was the ability of L. brevis to grow at45°C. This characteristic was not observed in the present study, nor was it expected to be observed according to Bergey's Manualof Determinative Bacteriology. Also in contrast to Bergey's Manual ofDeterminative Bacteriology, but inagreement with this study, Davis found only 80% of 146 strains of L. fermentum capable offermentinglactose. Thus, theresults obtained by Davis support the valid-ityoftheidentification scheme developed here. Therehas only beenone report in theliterature concerned solely with the identification of hu-man vaginal lactobacilli. Rogosa and Sharpe (12), in 1960, isolated 35 strains oflactobacilli from 21 women and studied 21 of the strains in detail.Rogosa and
Sharpe
found four species of lactobacilli to be present in the vagina: (i) L. acidophilus,14strains; (ii)L.caseisubsp. rham-nosus, 2 strains; (iii) L.fermentum, 4 strains; and (iv) L. cellobiosus, 1 strain. The reactions obtainedby Rogosaand Sharpeintheir identifi-cation scheme were identical to those obtained in thepresentinvestigation, withoneexception. Bothof their strains ofL. caseisubsp. rhamno-susfermented sorbitol. Nineteentests were per-formedbythoseinvestigators, of which11 wereutilizedin the present
investigation.
Thus, from these comparisons, it appears that thenewly
devised schemeprovides
a relatively reliable identification of endocervical lactobacilli.To establish highly reliable identification schemes, phenotypes generated with biochemi-cal testsmustbeconsideredincombinationwith genotypes determined in studies of DNA
con-tent and relatedness (1). London has recently reviewedthetaxonomic status of thelactobacilli based upon these and otherdiverse characteris-tics(6). However, the purpose of thisstudywas
not toestablish aprecise taxonomicdescription of the lactobacilli. Rather, it was to develop a
practical approachto the identification ofthose
species commonly recovered from the female genitaltract. Tothis end,the schemedeveloped has the advantage over previous schemes of using fewer and relatively simple tests. The reliabilityof theidentification generated by this scheme appears to be at least as good as more complex schemes (3, 12, 14, 18).
ACKNOWLEDGMENTS
This workwassupportedbyPublicHealth Servicegrant 5
R01-AI-11584 from the National Institute of Allergy and Infectious Diseases.
We thank the staffs of Equilibria Medical Center and CreightonHealthClinic OB-GYN for their assistance in this study.
LITERATURECITED
1. Brenner, D. J. 1980. Taxonomy, classification, and no-menclature ofbacteria, p. 1-6. In E. H. Lennette, A. Balows, W. J.Hausler,Jr., and J.P. Truant(ed.), Manual of clinical microbiology, 3rd ed. American Society for Microbiology,Washington, D.C.
2. Briggs, M. 1953. The classification of lactobacilli by means ofphysiological tests.J. Gen. Microbiol. 9:234-248.
3. Davis, G. H. G. 1955. The classification of lactobacilli fromthe human mouth.J. Gen. Microbiol. 13:481-493. 4. Davis, G. H. G., K. A. Bisset, and C. M. F. Hall. 1955.
Correlation between morphological and physiological characters intheclassification of members ofthegenus Lactobacillus.J.Gen. Microbiol. 13:68-71.
5. Doderlein,A. 1892. DasScheidensekret und seine Bedeu-tungfurdasPuerperalfieber. Leipzig.
6. London,J.1976.Theecologyandtaxonomicstatusof the lactobacilli. Annu.Rev. Microbiol. 30:279-301. 7. McDonald, I. J., and W. C. Frazier. 1951. Variation in
morphology of colonies of lactobacilli. J. Bacteriol. 61:627-637.
8. Niven, C.F., Jr.,K. L.Smiley,andJ.M. Sherman.1942. The hydrolysis of arginine by streptococci.J. Bacteriol. 43:651-660.
9. Rogosa,M.1974.GenusLactobacillusBeijerinck 1901,p. 576-593. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey'smanual ofdeterminative bacteriology, 8th ed. TheWilliams&WilkinsCo., Baltimore.
10. Rogosa,M., J. G.Franklin,and K. D.Perry.1961. Corre-lation of the vitamin requirements with cultural and biochemical characters of Lactobacillus spp. J. Gen. Microbiol.25:473-482.
11. Rogosa, M., and J.A. Mitchell. 1950. Induced colonial variation ofatotalpopulationamongcertainlactobacilli. J.Bacteriol. 59:303-308.
12. Rogosa, M.,and M. E.Sharpe.1960.Species differentia-tion of human vaginal lactobacilli. J. Gen. Microbiol. 23:197-201.
13. Rogosa,M.,R. P.Tittsler,and D. S.Geib. 1947. Correla-tionof vitaminrequirements and cultural and biochemical characteristics ofthegenus Lactobacillus. J. Bacteriol. 54:13-14.
14. Rogosa,M.,R. F.Wiseman,J.A.Mitchell,M. N. Disrae-ly, and A.J.Beaman.1953.Speciesdifferentiation oforal lactobacillifrommanincludingdescriptionsof Lactoba-cillus salivariusnov.spec. andLactobacillius cellobiosus nov. spec.J. Bacteriol.65:681-699.
15. Saigh, J. H.,C. C.Sanders,and W. E.Sanders,Jr.1978. Inhibition ofNeisseriagonorrhoeae byaerobicand facul-tativelyanaerobic componentsof theendocervical flora: evidence foraprotectiveeffect against infection. Infect. Immun. 19:704-710.
16. Schleifer, K.H., and 0. Kandler. 1972. Peptidoglycan typesofbacterialcell wallsandtheirtaxonomic implica-tions.Bacteriol. Rev.36:407-477.
VOL. 16, 1982
on February 7, 2020 by guest
http://jcm.asm.org/
17. Sharpe, M. E. 1955. Aserological classification of lacto-bacilli. J. Gen. Microbiol. 12:107-122.
18. VirginiaPolytechnic Institute Anaerobe Laboratory. 1977. Luctobacillus,p. 63-71. In L. V.Holdeman, E. P. Cato, and W. E.C. Moore(ed.),Anaerobe laboratorymanual.
4th ed., VirginiaPolytechnicInstitute and State Universi-ty,Blacksburg,Va.
19. Wheater, D. M. 1955. Thecharacteristics of Lactobacillus acidophillusandLactobacillitsbulgaricus.J. Gen. Micro-biol. 12:123-132.