G. Huys 1, K. D Haene 1 and J. Swings 1,2 1 Laboratory of Microbiology, 2 BCCMä/LMG Bacteria Collection, Faculty of Sciences, Ghent University,

Transcription

1 Letters in Applied Microbiology, 3, Influence of the culture medium on antibiotic susceptibility testing of food-associated lactic acid bacteria with the agar overlay disc diffusion method G. Huys 1, K. D Haene 1 and J. Swings 1, 1 Laboratory of Microbiology, BCCMä/LMG Bacteria Collection, Faculty of Sciences, Ghent University, Ghent, Belgium 97/1: received 11 October 1, revised February and accepted February G. H U Y S, K. D H A E N E A N D J. S W I N G S.. Aims: To investigate the influence of the culture medium on antibiotic susceptibility testing of food-associated lactic acid bacteria (LAB) with the agar overlay disc diffusion (DD) method. Method: The antibiotic resistance profile of 39 food-associated lactobacilli and enterococci was determined with the agar overlay DD method using a defined medium (i.e. Iso-sensitest agar; ISA) or an undefined medium (i.e. de Man, Rogosa, Sharpe or MRS agar). Results: The study revealed that ampicillin discs and, although to a lesser extent, also tetracycline discs consistently produced larger zones on MRS medium compared to ISA medium. For the antibiotics gentamicin, bacitracin and erythromycin, the radius of the inhibition zones produced on MRS medium was significantly smaller in relation to ISA. For categorizing LAB isolates into resistant, intermediate and susceptible groups, it was demonstrated that major errors can occur in determining bacitracin and gentamicin resistance if MRS medium instead of ISA medium is used. On the other hand, the performance of both media was found to be equivalent for testing tetracycline resistance. Conclusions, Significance and Impact of the Study: Despite the fact that MRS medium generally supports the growth of lactic acid bacteria much better than the nutrient-poor ISA medium, the present study clearly demonstrates that both media are not compatible in susceptibility testing against various classes of antibiotics. These results may stimulate future discussions on a generally recommended DD method for susceptibility testing of food LAB strains. INTRODUCTION Until recently, antibiotic susceptibility testing of food and food-associated lactic acid bacteria (LAB) was only performed sporadically when an association with food-borne infection was suspected. However, current ecological and epidemiological insights have led to speculations that antibiotic-challenged environments linked to the production of dairy and fermented meat products may yield antibioticresistant LAB. Due to their capability to survive and grow in a variety of food matrices as well as in the human colon, it has been suggested that subpopulations of certain LAB taxa Correspondence to: G. Huys, Laboratory of Microbiology, Faculty of Sciences, Ghent University, K.L. Ledeganckstr. 35, B-9 Gent, Belgium ( geert.huys@rug.ac.be). can act as reservoirs of transferable antibiotic resistance throughout the human compartment (Teuber et al. 1999; Gevers et al. ). In this regard, routine antibiotic susceptibility testing of food-associated LAB may be advisable in a number of instances e.g. for checking the biosafety of potentially probiotic isolates and genetically modified strains used in biotechnology. Several methods have been reported for antimicrobial susceptibility testing of LAB food isolates, including disc diffusion (Sozzi and Smiley 19; Orberg and Sandine 195; Perreten et al. 199), E-test (Katla et al. 1), broth dilution (Perreten et al. 1997) and agar dilution (Butaye et al. ). Of these techniques, the disc diffusion (DD) method is most widely used because of its high degree of reliability towards the standardization of the antibiotic concentration and its relative ease of performance. The ª The Society for Applied Microbiology

2 DISC DIFFUSION TESTING OF LAB 3 primary role of the culture medium in the DD technique is to supply an optimal nutritional environment to support growth of the test organism. In addition, it should also provide a suitable gel matrix to allow reproducible and uniform diffusion of the antibiotic agent hereby minimizing possible chemical interactions between (un)defined medium components and the antibiotic gradient. The DD test media recommended by the National Committee for Clinical Laboratory Standards (NCCLS 1999) and by the British Society for Antimicrobial Chemotherapy (BSAC, Andrews and the BSAC Working Party on Susceptibility Testing 1), i.e. Mueller Hinton agar and Iso-Sensitest Agar (ISA), respectively, are believed to exert only limited antagonistic effects against the majority of test reagents. On the other hand, however, practical experiences with these media seem to point out that they do not support growth of a wide range of nonclinical LAB isolates. Previously, the use of MRS (de Man, Rogosa, Sharpe) medium in the agar overlay DD method has been described for testing drug resistance in animal isolates of Lactobacillus (Vescovo et al. 19) and in potentially probiotic LAB isolates (Charteris et al. 199a, b). This medium was specially designed to encourage growth of most LAB and is widely used for the cultivation of food-associated LAB. In addition, the agar overlay is thought to create an additional microaerophilic effect during the incubation of test plates. To our knowledge, however, very little information has so far been generated on the performance of MRS medium in DD susceptibility testing against the validated NCCLS or BSAC media. In the current study, the MRS medium was compared with the closely defined ISA medium for use in agar overlay DD testing of food-associated LAB. Using the inhibition zone diameter as a quantitative performance parameter, both media were evaluated for six antibacterial agents using a collection of 39 LAB isolates from food origin consisting of lactobacilli and enterococci. MATERIALS AND METHODS Bacterial strains A total of 39 LAB food isolates belonging to the genera Lactobacillus and Enterococcus were included in this study and encompassed the following species: Lact. brevis (n ¼ 9), Lact. parabuchneri (n ¼ 3), Lact. paracasei (n ¼ 3), Lact. bifermentans (n ¼ 1), Lact. buchneri (n ¼ 1), Lact. kefiri (n ¼ 1), Lact. paraplantarum (n ¼ 1), Ent. faecalis (n ¼ 17) and Ent. faecium (n ¼ 3). These strains were selected from the BCCMä/LMG Bacteria Collection, Ghent University, Belgium ( htm) and from the catalogue of enterococci of the FAIR-E collection (Project EU-FAIR-CT97-37; Franz et al. 1) on the basis that growth was observed on both MRS agar (Æ% [w/v], CM31; Oxoid) and on Iso-Sensitest Agar (ISA) (3Æ1% [w/v]; CM71; Oxoid) at C. Agar overlay disc diffusion testing Susceptibility testing was based on the agar overlay DD method described by Charteris et al. (199a) with slight modifications. From fresh overnight grown cultures, a few colonies were suspended in 5Æ% (w/v) MRS broth (CM359; Oxoid Ltd, Basingstoke, Hampshire, England) and incubated overnight under microaerophilic conditions at C. The inoculum density of the liquid cultures was adjusted to O.D. 59 of Æ1 ± Æ using a Vitalab 1 spectrophotometer (Vital Scientific, Dieren, The Netherlands). Depending on the type of medium, Æ ml of the adjusted inoculum was mixed with ml sloppy MRS medium or sloppy ISA medium, both containing Æ5% Bacteriological Agar no. 1 (CM; Oxoid) cooled to 5 C. Petri dishes (1 cm) containing either 15 ml of MRS medium or ISA medium were overlaid with Æ ml of sloppy MRS medium or ISA culture, respectively, and allowed to solidify for min at room temperature. Antibiotic discs were dispensed onto the overlay plates using the ST9 Disc Dispenser (Oxoid) and all plates were incubated under microaerophilic conditions at C for h. The following antibiotic discs (Oxoid) were tested: ampicillin (AMP1), tetracycline (TE3), erythromycin (E15), bacitracin (B1), rifampicin (RD3) and gentamicin (CN1). Inhibition zones were measured using digital callipers (Mauser digital, Ludwigsburg, Germany). RESULTS AND DISCUSSION Throughout the history of antibiotic DD testing various studies have been performed on the effects of the growth medium and its constituents on the measurement and interpretation of the resulting inhibition zones (Cotter and Adley, 1; Jensen et al. 199; Toama et al. 197; Dornbusch 19; Milne et al. 1993). However, as most of these reports focus on clinical isolates of well-recognized Gram-positive and Gram-negative pathogens, there remains a gap of knowledge regarding the antibiogram determination of commensal microorganisms such as food-associated LAB. For this purpose, the current study compared the performance of the complex and nutrient rich MRS medium vs. the defined ISA medium in agar overlay DD testing against six antibiotics, encompassing inhibitors of cell wall synthesis (i.e. ampicillin and bacitracin), of protein synthesis (i.e. tetracycline, gentamycin and erythromycin) and of nucleic acid synthesis (i.e. rifampicin). Preliminary growth tests on 7 LAB isolates encompassing the genera Lactobacillus, Enterococcus, Lactococcus, Pediococcus and Streptococcus ª The Society for Applied Microbiology, Letters in Applied Microbiology, 3,

3 G. HUYS ET AL. (a) Tetracycline (TE3) (b) Ampicillin (AMP1) Difference [(zone on ISA) (zone on MRS)] (c) Gentamicin (CN1) (d) Bacitracin (B1) (e) 1 Erythromycin (E15) (f) 1 Rifampicin (RD3) Inhibition zone on MRS agar (mm) Fig. 1 Polynomial regression expressing the relationship between inhibition zones of individual strains obtained on MRS and ISA medium for the antibiotics tetracycline (a), ampicillin (b), gentamicin (c), bacitracin (d), erythromycin (e) and rifampicin (f). If two or more strains showed the same zone size on MRS medium, values shown are the average mean values ± standard deviation. Breakpoints for classification of isolates as resistant (R), intermediate (I), or susceptible (S) proposed by Charteris et al. (199a) are indicated by vertical broken lines if applicable in the indicated zone range for the corresponding antibiotic disc load. d, data point contributed by Lactobacillus; r, data point contributed by Enterococcus; j, data point contributed by Lactobacillus and Enterococcus. showed that > % of the tested isolates subcultured on MRS medium did not display growth on ISA medium after a 7-h incubation at C under microaerophilic conditions. Supplementation of ISA medium with glucose and adaptation of the acidity to ph Æ did not significantly reduce this percentage (G. Huys, unpublished data). As ª The Society for Applied Microbiology, Letters in Applied Microbiology, 3,

4 DISC DIFFUSION TESTING OF LAB 5 such, this finding suggests that the ISA medium cannot be generally recommended for routine use in DD susceptibility testing of LAB. Subsequently, a group of 39 Enterococcus and Lactobacillus strains that displayed good growth on both media was selected for comparative testing of six agents. Towards quantitative analysis of inhibition zone diameters, we first set out to test the reproducibility of the agar overlay DD method by determining antibiograms in duplicate on MRS medium starting from separate pure cultures and using different batches of medium and antibiotic discs. For 39 LAB isolates tested for antibiotics, a mean standard deviation of 1Æ mm with a maximum variation of 3 mm was obtained for inhibition zones between the two determinations when aberrant zones were ignored (data not shown). As the zone radius of control strains is usually allowed to vary with to 3 mm in routine DD testing (NCCLS, BSAC), it can thus be concluded that the agar overlay DD method yields results with a comparable reproducibility. The effects of the choice of growth medium on the zones in the agar overlay DD method were found to be largely dependent on the antibiotic compound tested. However, there was a general tendency observed with all disc types that the proportional lack of congruence between zone reading on MRS medium and on ISA medium increased with the zone diameter. This indicates that the lower concentrations in the antibiotic diffusion gradient seem to be more affected by the medium compared with that of the higher concentrations close to the edge of the antibiotic disc. Considering the mean deviation of 1Æ mm, it was found that inhibition zones observed with tetracycline disc testing (Fig. 1a) showed a good overall correspondence between the data obtained with MRS medium and ISA medium data in the range of 9 mm. In this range, the MRS and ISA media can be considered equivalent whereas in the range higher than mm, zones on MRS medium displayed a higher radius compared to the corresponding diameter on ISA medium. Comparison of ampicillin test results (Fig. 1b) showed larger inhibition zones on MRS medium along the entire zone range tested, with a highly pronounced lack of correspondence between both media for zones exceeding a radius of mm. Within the limits of the range available for interpretation, it can thus be predicted that ampicillin inhibition zones will be consistently larger on MRS medium than on ISA medium for both enterococci and lactococci. An opposite effect was observed with the results of gentamicin (Fig. 1c), bacitracin (Fig. 1d) and erythromycin (Fig. 1e) testing. Within the concentration range tested for these antibiotics, the data indicate that the zone radius on ISA medium will be higher than that obtained on MRS medium. The results obtained for rifampicin testing (Fig. 1f) did not allow a clear interpretation due to the high strain-to-strain variation and the different genus-specific resistance profile exhibited by Enterococcus and Lactobacillus strains towards this antibiotic. For a number of antibiotics tested in this study, it can be concluded that the choice of culture medium will significantly affect the classification of LAB strains into susceptible, intermediate and resistant categories. However, there are as yet no guidelines available on the interpretation of DD susceptibility tests for commensal or food bacteria. For example, Charteris et al. (199a) in their study on Lactobacillus spp. used the interpretative standards that were originally designed for testing clinical isolates. Although this is not an ideal comparison since antibiotic breakpoints cannot be standardized amongst clinical and nonclinical microorganisms, projection of the guidelines of Charteris et al. (199a) to the present data indicate that, for gentamicin and bacitracin testing, major errors may occur in the categorization of strains when using different test media in the overlay DD method. LAB isolates that would otherwise be classified as resistant to these two antibiotics on MRS medium would be considered susceptible on ISA medium (Fig. 1c,d). In contrast, interpretation of the tetracycline data does not seem to be influenced by the choice of medium (Fig. 1a). Recently, Cotter and Adley 1) subjected a collection of clinical enterococci to a comparative evaluation of six national committee standardized DD procedures which differed in various parameters including the choice of the test medium, inoculum density, antibiotic disc content and interpretative breakpoint criteria. Coinciding with our findings, the latter authors noticed interpretative errors for testing gentamicin resistance when using different standardized procedures and likewise found that the classification of tetracycline-resistant strains was uniform for all procedures used. In conclusion, the present study clearly demonstrates that MRS medium is not compatible to the ISA medium for use in susceptibility testing against various classes of antibiotics, excluding tetracycline. To our knowledge, no general consensus has been reached on the DD method recommended for susceptibility testing of LAB. However, as research into the role of commensal and nonpathogenic LAB in the global ecology of antimicrobial resistance determinants is currently being stimulated by global initiatives like the Reservoirs of Antibiotic Resistance (ROAR) project ( it is the authors opinion that such studies will find high benefit from the coordination of existing and newly developed susceptibility testing methods for this group of bacteria. ACKNOWLEDGEMENTS This work was supported by the Fund for Scientific Research Flanders (Belgium) (F.W.O.-Vlaanderen) (contract G.39.1). G.H. is a postdoctoral fellow of the Fund for Scientific Research Flanders (Belgium) (F.W.O.- Vlaanderen). ª The Society for Applied Microbiology, Letters in Applied Microbiology, 3,

5 G. HUYS ET AL. REFERENCES Andrews, J.M. and the BSAC Working Party on Susceptibility Testing (1) BSAC standardized disc susceptibility testing method. Journal of Antimicrobial Chemotherapy, Butaye, P., Van Damme, K., Devrieze, L.A., Van Damme, L., Baele, M., Lauwers, S. and Haesebrouck, F. () In vitro susceptibility of Enterococcus faecium isolated from food to growth-promoting and therapeutic antibiotics. International Journal of Food Microbiology 5, Charteris, W.P., Kelly, P.M., Morelli, L. and Collins, J.K. (199a) Antibiotic susceptibility of potentially probiotic Lactobacillus species. Journal of Food Protection 1, Charteris, W.P., Kelly, P.M., Morelli, L. and Collins, J.K. (199b) Antibiotic susceptibility of potentially probiotic Bifidobacterium isolates form the human gastrointestinal tract. Letters in Applied Microbiology, Cotter, G. and Adley, C.C. (1) Comparison and evaluation of antimicrobial susceptibility testing of enterococci performed in accordance with six national committee standardized disk diffusion procedures. Journal of Clinical Microbiology 39, Dornbusch, K. (19) Influence of medium composition on bacterial susceptibility testing to gentamicin and netilmicin. Scandinavian Journal of Infectious Diseases (Suppl.), 53. Franz, C.M.A.P., Muscholl-Silberhorn, A.B., Yousif, N.M.K., Vancanneyt, M., Swings, J. and Holzapfel, W.H. (1) Incidence of virulence factors and antibiotic resistance among enterococci isolated from food. Applied and Environmental Microbiology 7, Gevers, D., Huys, G., Devlieghere, F., Uyttendaele, M., Debevere, J. and Swings, J. () Isolation and identification of tetracycline resistant lactic acid bacteria from pre-packed sliced meat products. Systematic and Applied Microbiology 3, 79. Jensen, K.T., Schonheyder, H., Gottschau, A., Thomsen, V.F. (199) Impact of the agar medium and disc type on disc diffusion susceptibility testing against teicoplanin and vancomycin. APMIS 1, 9 1. Katla, A.-K., Kruse, H., Johnsen, G. and Herikstad, H. (1) Antimicrobial susceptibility of starter culture bacteria used in Norwegian dairy products. International Journal of Food Microbiology 7, Milne, L.M., Crow, M.R., Emptage, A.G. and Selkon, J.B. (1993) Effects of culture media on detection of methicillin resistance in Staphylococcus aureus and coagulase negative staphylococci by disc diffusion methods. Journal of Clinical Pathology, National Committee for Clinical Laboratory Standards (NCCLS) (1999) Performance Standards for Antimicrobial Susceptibility Testing, ninth informational Supplement. M1 S9, Vol. 19, no. 1. Orberg, P.K. and Sandine, W.E. (195) Survey of antimicrobial resistance in lactic streptococci. Applied and Environmental Microbiology 9, Perreten, V., Kollöffel, B. and Teuber, M. (1997) Conjugal transfer of the Tn91-like transposon TnFO1 from Enterococcus faecalis isolated from cheese to other Gram-positive bacteria. Systematic and Applied Microbiology, 7 3. Perreten, V., Giampa, N., Schuler-Schmid, U. and Teuber, M. (199) Antibiotic resistance genes in coagulase-negative staphylococci isolated from food. Systematic and Applied Microbiology 1, Sozzi, T. and Smiley, M.B. (19) Antibiotic resistances of yoghurt starter cultures Streptococcus thermophilus and Lactobacillus bulgaricus. Applied and Environmental Microbiology, 5. Teuber, M., Meile, L. and Schwarz, F. (1999) Acquired antibiotic resistance in lactic acid bacteria from food. Antonie Van Leeuwenhoek 7, Toama, M.A., Issa, A.A. and Ashour, M.S. (197) Effect of agar percentage, agar thickness, and medium constituents on antibiotics assay by disc diffusion method. Pharmazie 33, 1 1. Vescovo, M., Morelli, L. and Bottazzi, V. (19) Drug resistance plasmids in Lactobacillus acidophilus and Lactobacillus reuteri. Applied and Environmental Microbiology 3, 5 5. ª The Society for Applied Microbiology, Letters in Applied Microbiology, 3,