System. Appl. Microbiol. 4, 123-131 (1983)

"The Institute of Physical and Chemical Research, Wako, Saitama 351, Japan2 Department Biomedical Science, Faculty of Agriculture, University of Tokyo, Bunkyo-ku,Tokyo 113, Japan

Modification of the Minitek System for Identification

of Lactobacilli

YOSHIMI BENN0 1 and TOMOTARI MITSUOKAl,2

Received August 10, 1982

Summary

Seventy-eight strains of 23 Lactobacillus species, including 41 reference strains and 37fresh isolates from faeces of men and pigs or caecal contents of mice, were used in com­paring the conventional tube test with the Minitek system (BBL) to evaluate the feasibilityof the latter to identify lactobacilli. In addition, the following parameters were tested:(i) the effect of variation in inoculum broth and cultural methods, (ii) the reproducibilityof reactions on each substrate. The Minitek system includes a new lactobacilli inoculumbroth (LIB) and three substrates: ribose, melezitose, and amygdalin. The overall correlationwas 84% by the anaerobic steel-wool jar method and 72.5% by overlaying sterile oil withthe Minitek inoculum broth (MIB) for the identification of anaerobes; whereas with LIB,the overall agreement was 95.4% by the anaerobic jar method and 84.4% by the oil pro­cedure. The agreement of the identification between the Minitek systems and conventionaltests was best (85.9%) with LIB by the anaerobic jar method and poorest with the MIBoverlaid with oil. In the reproducibility of reactions, 247 (17.2%) of a total of 1,440 setsof disks gave variable reactions with MIB when incubated under the anaerobic jar methodas compared to only 29 (2.0%) with LIB. The Minitek system for the identification of lac­tobacilli is more accurate and reliable when inoculated with LIB and incubated under theanaerobic jar method.

Key words: Minitek - Identification - Lactobacillus

Introduction

Recently, several miniaturized systems have been marketed to facilitate identifica­tion of Enterobacteriaceae and anaerobic bacteria. Minitek (BBL MicrobiologySystems, Division of Becton, Dickinson and Co., Cockeysville, Md, U.S.A.), basedupon substrate impregnated disks, is such a system. Many investigators have reportedthat the Minitek system is reliable for rapid characterization of Enterobacteriaceae(Hansen et al., 1974; Kiehn et al., 1974; Finklea et al., 1976) and anaerobic bacteria

124 Y.Benno and T.Mitsuoka

(Starr et al., 1973; Moore et al., 1975; Hansen and Stewart, 1976; Stargel et al., 1976;Essers and Haralambie, 1977; Hanson et al., 1979). Gilliland and Speck (1977)reported the Minitek system to maintain a satisfactory performance level in thecharacterization of lactobacilli by employing the anaerobic GasPak (BBL) jar.

The present study was carried out to determine the accuracy of the results obtainedwith the Minitek system when different inoculum broths and anaerobic incubationprocedures were employed, and the reproducibility of the identification of Lacto­bacillus species by the Minitek system.

Materials and Methods

Substrates

The Minitek and conventional systems were compared with respect to the fermentationof twenty-four different substrates: arabinose, xylose, rhamnose, ribose, glucose, mannose,fructose, galactose, sucrose, maltose, cellobiose, lactose, trehalose, melibiose, raffinose,melezitose, starch, glycerol, mannitol, sorbitol, inositol, esculin, salicin, and amygdalin.Disks for fermentation of ribose, melezitose, and amygdalin were prepared in our labora­tory according to previously described procedures (Gilliland and Speck, 1977).

Organisms

The seventy-eight strains of 23 Lactobacillus species used in this study are listed inTable 1. Fresh isolates were obtained from the faeces of 5 healthy adults and five healthypigs or from the caecal contents of 3 mice, using a modified Lactobacillus selective (LBS)medium (Mitsuoka et al., 1965). All strains were maintained on the pre-reduced Briggsliver agar slants with Na2C03-C02 buffer (Mitsuoka et al., 1965) and stored at 4°C.

Conventional system

Conventional tube tests for the cultural and biochemical characterization of the Lacto­bacillus strains were carried out according to the previously reported method (Mitsuoka,1969). The acid and gas formation and growth at 15°C were repeatedly determined afterincubation periods of 3 days and 7 days.

Minitek system

All strains tested were streaked on Briggs liver agar (Mitsuoka, 1969) to obtain singlecolonies. The Briggs liver agar plates were incubated at 37°C in an anaerobic steel-wooljar (Parker, 1955) filled with 100% CO 2, After incubation for 2 days the cultures wereexamined for purity with a microscope.

Comparison of the Minitek system and the conventional tube test applying a modifiedlactobacilli inoculum broth (LIB)

(i) Four to six colonies from Briggs liver agar plates were suspended in two screw-cappedvials with the Minitek inoculum broth (MIB) for the characterization of anaerobic bac­teria supplied by the manufaturer in screw-capped vials. All cell suspensions were adjustedto the optical density of a McFarland no. 5 nephelometric standard. These suspensionswere used to inoculate two sets of Minitek plates according to the manufacturer's direc­tions. The inoculated substrate disks in one set of Minitek plates for each culture wereoverlaid with 0.1 ml of sterile mineral oil, and the plates were incubated in a humidor for48 h at 37°C (Minitek-oil method). The remaining set of plates for each culture with nooverlay of mineral oil was placed in an anaerobic steel-wool jar filled with 100% CO2andincubated for 48 h at 37°C (Minitek-jar method).

Minitek for Lactobacilli 125

Table 1. Seventy-eight strains of Lactobacillus species tested in this study

Lactobacillus species No. of No. of fresh isolate fromreference Human Pig Mousestrain (5)a (5) (3)

I. acidophilus 5 3 11 8"I. alimentarius" b 1"I. batatus" 1I. brevis 2I. buchneri 2I. bulgaricus 1I. casei subsp. casei 2I. casei subsp. rhamnosus 1I. cellobiosus 1I. confusus 1I. coryniformis 1I. curvatus 1I. delbruekii 1I. fermentum 4 1 10 3"I. frigidus" 1I. fructivorans 1I. helveticus 4I. hilgardii 1I. lactis 2I. leichmanii 1I. plantarum 3I. salivarius subsp. salicinius 1I. salivarius subsp. salivarius 1 1 1I. viridescens 1

Total 40 5 22 11

a Number of samples used for the isolation of lactobacilli.b Names in quotation marks are not contained in the Approved Lists of Bacterial Names

(Skerman et al., 1980).

(ii) A single colony from Briggs liver agar plates was transferred to Briggs liver brothcontaining some marble chips (Mitsuoka et al., 1965). All further subcultures were ob­tained from this broth. After boiling the medium was aseptically supplemented with 0.05 mlof ascrobic acid-cysteine solution (Mitsuoka, 1969) per 3 ml, giving the final concentrationof 0.34 g ascorbic acid/rnl and 0.01 g L-cysteine HCI· H,O Iml, immediately before use.After aerobic incubation of the broth, inoculated with the test strain for 24 h at 37°C,the culture was centrifuged for 10 min at 1800 x g and washed with sterile saline solutionsupplemented with 1 g L-cysteine HCI· H 20 II and 1 g sodium thioglycolate/l (Mitsuoka,1969). The tube with the saline solution was again centrifuged and the cells were used tomake inoculum with the modified lactobacilli inoculum broth (LIB). The turbidity of theinoculum was adjusted with LIB to approximate a McFarland no. 5 nephelometric standard.The modified LIB contained Bact-liver (Difco) infusion (5.5 gil); Proteose peptone no. 3(Difco) (10 gil); Trypticase (BBL) (5 gil); Yeast extract (Difco) (3 gil); Tween 80 (1 rnljl};and Solution B (Mitsuoka et al., 1965), (5 mlrl). The pH was adjusted to 6.5 before auto­claving. Excessive aeration of the LIB was avoided by boiling before preparing suspen-

126 Y.Benno and T.Mitsuoka

sions. Two sets of Minitek plates were prepared with LIB instead of MIB. One set ofMinitek plates was overlaid with sterile mineral oil, the other one was placed in theanaerobic steel-wool jar. Incubation was carried out for 48 h at 37°C. The reactionswere recorded after 0.05 ml of bromocresol purple solution (0.2 gil H20 , pH 7.2) wasadded to each fermentation well. Only a clearly yellow color was considered positive;any shade of orange was considered weak; purple color was considered negative. Theesculin reaction was considered positive if a brown or brown-black color developed. Thesetests were performed in duplicate.

The reproducibility of the results obtained with Minitek disks applying either MIB orLIB as substrates under the anaerobic steel-wool jar condition for 48 h at 37°C was con­firmed with the repeated six tests.

Results

Tables 2 and 3 show the results obtained with 78 Lactobacillus strains whenassayed with MIB and LIB as substrates and incubated either in anaerobic jars orunder the overlaid oil condition recommended by the supplier. By the Minitek-oilmethod, an overall agreement of 72.5% was obtained with the MIB, whereas 84%with the LIB. Both the suspension media gave marked differences from the conven-

Table 2. Agreement of results obtained with the Minitek-oil method and conventionalsystems with two different suspension media

.- --~----------

Minltek + MInit ek - Tests In agreementSubstrate Co n v e n t i o n a l - jJ& Conventional + (%) (%)

MIG LIB MIB LIB MIB LIB_._-----~----~---_.

Arabinose 0 0 11. 8 0 88. 100Xylose 0 5. 5. 9 0 94. 94. 4Rhamnose 5. 0 11. 8 11. 82. 88. 9Ribose

a0 5. 52. 9 22. 47. 72. 2

Glucose 0 0 23. 5 0 76. 100Mannose 5. 5. 35. 3 0 58. 94.Fructose 0 5. 29. 4 0 70. 6 94.Galactose 0 0 35. 3 27, 64. 7 72.Sucrose 0 0 52. 9 16. 47. 1 83.Ma Ito s e 0 0 35. 3 11. 1 64. 7 88.Ce l l o b i o s e 5. 9 5. 6 23. 5 22. 2 70. 6 72.Lactose 0 5. 6 57. 1 38. 9 52. 9 55.Trehalose 5. 11. 1 23. 5 11. 1 70. 6 77.Me I I b lOS e 0 5. 6 35. 3 11. 1 64. 7 83.RaffInose 0 0 23. 5 11. 1 76. 5 88.Melezitose

a0 11. 11.8 0 88. 2 88.

Starch 11.8 0 0 0 88. 2 100Glycerol 0 5. 5. 9 0 94. 1 94. 1Mannitol 5. 9 11. 5. 9 5. 9 88. 2 83. 0Sorb I tol 5. 9 11. 1 17. 6 16. 7 76. 5 72. 2Inositol 0 5. 9 0 0 100 94. 1Es c u l i n 5. 9 11. 1 35. 3 11. 1 58. 8 77. 8Sa I i c I n 5. 9 5. 9 41.2 16. 7 52. 9 77. 4Amygda I In a

5. 9 5. 9 29. 4 33. 3 64. 7 60. 8

Total 2. 7 4. 9 24. 8 11. 1 72. 5 84. 0

a This disk prepared according to Gilliland and Speck 1977.

Mi nitek for Lactobacill i 127

T able 3. Agreement of res ults ob tain ed wit h the Minitek-jar method and con venti on alsystem s with rwo different suspension media

Sub st r ateMl ni te k +Co nve nt Iona l

Ar a b i no s eXy los eRb amn o s eR,b ose'Gluco seMa nn o seF r u ctoseGala c tos eSuc roseMa I t o s eC e l lob io s eL a c t o s eT rehaloseMe l ib lO S eRa ffinoseMe 1e ZIt a s e'S ta r c hGl y ce ro lMann Ito lS or b i to lI n os it o IEs cu l inS ~ I I r j nAmygd a I I n '

T o t al

MIB

oo2oo2.uooo2.2.2.oo2. 95 . 7o2 . 9oo2.2 . 92 . 9

1. 4

LIB

2. 21. 5I. 52. 9no2. 2] . 5on. 72. 9O. 7I. 52. 24. 4oO. 7o2. 2oo2. 95. I1.5

I. 5

MIB

8 . 611. 4

5. 75 I. 4o

20 .5 .

25 .22.17.

5. 714. 3

2 . 917 . I14. 3

5 . 7o

11. 414 . 320 . 0o

10. 022 . 922 . 9

14 .

LIB

oO. 71. 55. 8oo2. 95. 1O. 7O. 7O. 73. 6I. 58 . 06. 6O. 73 . 61. 55. 82. 9o8. 07 . :36. 6

3.

MIB

9 I. 48 8 . 69 I. 448. 6

1no77 .194 . 374 . 377 . 182 . 99 1 . 48 2. 894 . 282 . 985. 79 I. 494. 388 . 682 . 880. 0

1008 7 . 174. 274. 2

84 4

LIB

97 . 897 . 897 . 09 I. 3

10 0100

94.93.99.98.96 .9 5 .9 7 . 089. 889 .099. 395. 798. 592 . 097 . 1

10 089 . 18 7 . 69 I. 9

95

& This disk prepared accord ing to Gilliland and Speck 1977.

tional tests with regard to the results on lactose, amygdalin, ribose, salicin, esculin,trehalose and sorbitol.

On the other hand, the percentage of agreement was higher when the Minitek-jarmethod was used (Table 3). An overall agreement of 84.4% was obtained with MIB,whereas use of LIB resulted in 95.4% agreement. Results of 15 of 24 substrateswith MIB agreed with conventional tests less than 90%. In cont rast, when LIB wasused, this method gives over 95% agreement with the results of conventional testsin fifteen substrates, but less than 90% agreement in 4 of all the substrates.

T able 4. Agreement of identi fication of seventy-eight strai ns of Lactobacillus spp. by twoMinitek systems

Determ inati on M initek-oi l methodMIB LIB

M ini tek-jar methodMIB LIB

Agreement (%) withconventional systems

42.3 69.2 73.1 85.9

128 Y.Benno and T.Mitsuoka

Table 4 shows the comparison of the identification percentages in various systems.The best agreement (85.9%) was obtained with LIB by the anaerobic jar method.Minitek-oil method with MIB showed the least agreement (42.3%). In some substratesof the Minitek system, certain strains gave negative results in spite of the positivereactions by the conventional tube tests. Acid production from melibiose, givingslow positive reaction with I. acidopbilus, was difficult to detect. Some strains ofI. cellobiosus, I. coniusus, I. [ermentum, "I. [rigidus" and I. lactis had negativeresults with glucosides (Table 5).

Table 5. Difference of slow fermentation between the Minitek and conventional systems

Species SubstrateRefer­encea

ReactionConven- Minitek-tional oilTest MIB LIB

Minitek­jar

MIB LIB

Liacidopbilus ATCC 4356 Melibiose +S +SL.acidophilus 1-85 Melibiose +S +SL. brevis ATCC 327 Fructose +S +

Maltose +S +SMannitol +S/- +S

L. brevis 8215 Galactose +S (+)L. casei ATCC 392 Sucrose +S/+ +L. casei subsp. rhamnosus Sucrose +/+S +

ATCC 7469L. cellobiosus ATCC 11738 Amygdalin +S (+)S

Esculin +S +sSalicin +S +S

L. confusus DSM 20194 Amygdalin NT (+)SL.fermentum ATCC 12315 Salicin -/+S (+)S"L.frigidus" DSM 20173 Cellobiose NT +S

Salicin NT (+)SL. helveticus ATCC 10797 Maltose +S (+)L.lactis ATCC 10697 Salicin (+)S (+)

++

++

++

(+)

+++

++

a The results of reaction given by Mitsuoka 1969.Symbols: +, reaction positive; (+), weak reaction; -, negative reaction; NT, not

tested. S, slow reaction (positive after 3 days incubation; / = or.

The reproducibility of each biochemical test in representative strains is shown inTable 6. Of a total of 1,440 sets of 6 disks, 247 (17.2%) of the sets gave variablereactions with MIB applying the anaerobic jar method, whereas 29 (2.0%) gavevariable reactions with LIB. Results of 13 of 24 substrates with LIB were 100%reproducible. The remaining 11 biochemical reactions varied in their reproducibilityfrom 85 to 98%.

Discussion

The agreement between the Minitek systems and conventional tests was best(95.4%) with LIB under the anaerobic jar method and poorest with the MIB over-

Minitek for Lactobacilli 129

Table 6. Variability of the Minitek jar method using six disks for each test

Substrate Number of testsperformed showing showing

varia bility" variability"with MIB with LIB

Arabinose 60 5 3Xylose 60 0 1Rhamnose 60 12 0Ribose 60 24 2Glucose 60 14 0Mannose 60 12 0Fructose 60 12 1Galactose 60 0 1Sucrose 60 16 0Maltose 60 11 0Cellobiose 60 6 0Lactose 60 14 2Trehalose 60 7 0Melibiose 60 20 9Raffinose 60 18 3Melezitose 60 12 0Starch 60 0 0Glycerol 60 0 0Mannitol 60 14 4Sorbitol 60 8 0Inositol 60 0 0Esculin 60 12 2Salicin 60 18 0Amygdalin 60 12 1

Total 1440 247 29Percentage 17.2 2.0

a Six disks per set.b Number of tests in which the reaction of one of the six disks differed from that of the

other five.

laid with oil. Results with mel ibiose, raffinos e, escu lin and salicin agreed with con­ventional test resu lts less th an 90%, even if th e LIB was used in combination withthe anae ro bic steel-wool jar meth od. These substrates were essential for the identi­ficat ion of lactobacilli. Gilliland and Speck (1977) have recently repo rte d th at mele­zitose, xylose , ara binose, mannose, and raffinose gave less th an 90% agreementwh en MRS medium as the inoculum was used in combina tio n with th e GasPakmethod. In our study, however, results with these subst ra tes, except for raffinose,exceeded 97% agreement with conventional results, when inoc ulated with LIB andincubated under th e anae ro bic jar method.

It has been reported tha t ara binose and xylose for cha rac ter izing of a few anaerobicbacteria were giving frequent false-negat ive reacti ons with slow ly growing ba cteri a(Hansen et al., 1976). Most of pr oblem substrates in the present work depended

9 System atic and Applie d Microbiology, Vol. 4, No . 1

130 Y.Benno and T.Mitsuoka

upon slow reactions but not the slowly growing bacteria (Table 5). Therefore, itmight be necessary for all organisms to incubate for longer periods (48 + x h).However, in order to eliminate the problem disks it is necessary to prepare a suitablescheme for the identification of lactobacilli by the Minitek method.

It has been suggested that variables, such as inoculum size, incubation time andthe technician's interpretation, might influence the results of micromethods (Butleret aI., 1975). The use of a large inoculum size (McFarland no. 5 nephelomeric stan­dard) and the new suspension medium (LIB) are necessary to improve the agreementbetween the Minitek system and conventional tube tests. A sufficient growth oflactobacilli strains was obtained by using the Briggs liver broth. To remove the fairamount of glucose remaining in the growth culture, cell suspensions from the brothwere centrifuged and washed with the same volume of sterile saline solution sup­plemented with 1 g i-cysteine Hel . H20/l and 1 g sodium thioglycolate/l,

All substrates containing indicators were found to be almost completely reducedafter the anaerobic incubation. The addition of a drop of fresh indicator to all fer­mentation wells after a 48 h incubation greatly facilitated the interpretation of reac­tions. This procedure is currently recommended by the manufacturer.

Based upon our experience with the new Minitek lactobacilli identification systemwe conclude that this system, when inoculated with the newly described lactobacilliinoculum (LIB) broth and incubated under the anaerobic jar method, to be a rapid,useful and acceptable substitute for the more time-consuming conventional tubetests of carbohydrate fermentation.

Acknowledgments. We are grateful to Nobue Siragami for her technical assistance.Kazunori Sugita and Seiji Fukushima of Becton Dickinson Overseas Inc., Tokyo generouslyprovided most of the Minitek equipment and substrates used in this investigation. We alsothank Melvin R. Smith for his assistance in manuscript preparation.

References

Butler, D.A., Lobregat, C.M., Gavan, T.L.: Reproducibility of the Analytab (API20E)system. J. Clin. Microbiol. 2, 322-326 (1975)

Essers, L., Haralambie, E.: Erfahrungen mit dem API 20A-Test System bei der Identifizie­rung von Anaerobiern aus der taglichen Routine. Zbl. Bakt. Hyg., I.Abt. Orig. A 238,394-401 (1977)

Finklea, P.]., Cole, M.A., Sodeman, T.M.: Clinical evaluation of the Minitek differentialsystem for identification of Enterobacteriaceae. J. Clin. Microbiol. 4, 400-404 (1976)

Gilliland, S.E., Speck, M. L.: Use of the Minitek system for characterizing lactobacilli.Appl. Environ. Microbiol. 30, 1289-1292 (1977)

Hansen, S.L., Hardesty, D. R., Myers, B.M.: Evaluation of the BBL Minitek system forthe identification of Enterobacteriaceae. Appl,Microbial. 28, 798-801 (1974)

Hansen, S.L., Stewart, B.].: Comparison of API and Minitek to center for disease controlmethods for the biochemical characterization of anaerobes. J. Clin, Microbial. 4, 227­231 (1976)

Hanson, C. W., Cassorla, R., Martin, W.].: API and Minitek systems in identification ofclinical isolates of anaerobic gram-negative bacilli and Clostridium species. J. Clin.Microbiol. 10, 14-18 (1979)

Kiehn, T. E., Brennan, K., Ellner, P. D.: Evaluation of the Minitek system for identificationof Enterobacteriaceae. Appl. Microbiol. 28, 668-671 (1974)

Minitek for Lactobacilli 131

Mitsuoka, T.: Vergleichende Untersuchungen iiber Laktobazillen aus den Faeces von Men­schen, Schweinen und Hiihnern. Zbl. Bakt., I. Abt. Orig. 210,32-51 (1969)

Mitsuoka, T., Sega, T., Yamamoto, S.: Eine verbesserte Methodik der qualitativen undquantitativen Analyse der Darmflora von Menschen und Tieren. Zbl. Bakt., I. Abt. Orig.195,455-469 (1965)

Moore, H. B., Sutter, V.L., Finegold, S.M.: Comparison of three procedures for bio­chemical testing of anaerobic bacteria. J. Clin. Microbiol. 1, 15-24 (1975)

Parker, C.A.: Anaerobiosis with iron wool. Aus. J. expo BioI. Med. Sci. 33,33-38 (1955)Sherman, V.B.D., McGowan, V., Sneath, P.H.A.: Approved lists of bacterial names. Int.

J. System. Bact. 30, 225-420 (1980)Stargel, M.D., Thomson, F.S., Phillips, S.E., Lombard, G.L., Dowell ir., V.R.: Modifica­

tion of the Minitek miniaturized differentiation system for characterization of anaerobicbacteria. J. Clin. Microbiol. 3,291-301 (1976)

Starr, S.E., Thompson, F.S., Dowell jr., V.R., Balows, A.: Micromethod system for identi­fication of anaerobic bacteria. Appl. Microbiol. 25, 713-717 (1973)

Dr. Yoshimi Benno, The Institute of Physical and Chemical Research, Wako, Saitama 351,Japan