Antibiotic susceptibility of members of the Lactobacillus acidophilus group using broth microdilution and molecular identification of their resistance determinants

International Journal of Food Microbiology 144 (2010) 81–87

Contents lists available at ScienceDirect

International Journal of Food Microbiology

j ourna l homepage: www.e lsev ie r.com/ locate / i j foodmicro

Antibiotic susceptibility of members of the Lactobacillus acidophilus group usingbroth microdilution and molecular identification of their resistance determinants

Sigrid Mayrhofer a, Angela H.A.M. van Hoek b,1, Christiane Mair a, Geert Huys c, Henk J.M. Aarts b,1,Wolfgang Kneifel a, Konrad J. Domig a,⁎a BOKU — University of Natural Resources and Life Sciences, Department of Food Sciences and Technology, Institute of Food Sciences; Muthgasse 18, A-1190 Vienna, Austriab RIKILT — Institute of Food Safety, Wageningen UR, Wageningen, The Netherlandsc BCCM/LMG Bacteria Collection, Laboratory of Microbiology, Faculty of Sciences, Ghent University; K.L. Ledeganckstraat 35, B–9000 Gent, Belgium

⁎ Corresponding author. Tel.: +43 1 47654 6750; faxE-mail address: [email protected] (K.J. Dom

1 Present address: Laboratory for Zoonoses andNational Institute for Public Health and the EnvironmBox 1, 3720 BA Bilthoven, The Netherlands.

0168-1605/$ – see front matter © 2010 Elsevier B.V. Aldoi:10.1016/j.ijfoodmicro.2010.08.024

a b s t r a c t

Article history:Received 3 May 2010Received in revised form 16 August 2010Accepted 26 August 2010

Keywords:Lactobacillus acidophilusAntibiotic susceptibilityMicrodilutionMicroarrayPCR

The range of antibiotic susceptibility to 13 antibiotics in 101 strains of the Lactobacillus acidophilus group wasexamined using the lactic acid bacteria susceptibility test medium (LSM) and broth microdilution.Additionally, microarray analysis and PCR were applied to identify resistance genes responsible for thedisplayed resistant phenotypes in a selection of strains. In general, narrow as well as broad unimodal andbimodal MIC distributions were observed for the Lactobacillus acidophilus group and the tested antimicrobialagents. Atypically resistant strains could be determined by visual inspection of the obtained MIC ranges forampicillin, chloramphenicol, clindamycin, erythromycin, quinupristin/dalfopristin, streptomycin and tetra-cycline. For most of these atypically resistant strains underlying resistance determinants were found. To ourknowledge erm(A) was detected in lactobacilli for the first time within this study. Data derived from thisstudy can be used as a basis for reviewing present microbiological breakpoints for categorization ofsusceptible and resistant strains within the Lactobacillus acidophilus group to assess the safety ofmicroorganisms intended for use in food and feed applications.

: +43 1 47654 6751.ig).Environmental Microbiology,ent (RIVM)—Division CIb, PO.

l rights reserved.

© 2010 Elsevier B.V. All rights reserved.

Introduction

Bacteria belonging to a healthy gastrointestinal microbiota andaffecting the microbial balance are among the first candidates in theselection of probiotics (Ishibashi and Yamazaki, 2001). Due to theirautochthonous presence in human and animal gut flora and purportedbenefits for gut function and health lactobacilli are of particularinterest (De Angelis et al., 2006). Especially the beneficial effects ofisolates of the Lactobacillus acidophilus group, e.g. L. acidophilus,L. amylovorus, L. crispatus, L. gallinarum, L. gasseri and L. johnsonii, inhuman and animal nutrition are well known (Bernardeau et al., 2006;Holzapfel, 2006).

In the last decade, a concern has arisen that microorganisms usedfor food and feed production can be vehicles for transmission ofantibiotic resistance genes (EFSA, 2008; Teuber et al., 1999).According to the Panel on Additives and Products or SubstancesUsed in Animal Feed (FEEDAP) of the European Food Safety Authority(EFSA), all bacteria intended for use as feed additives in Europe mustbe examined to ensure the susceptibility of the component strains to a

relevant range of antibiotics (EFSA, 2008). In contrast to bacteria usedas animal feed additives, there is no formal safety assessmentstandard with legal status for dietary supplements containingprobiotics (von Wright, 2005; Wassenaar and Klein, 2008), but itwas recommended that a consistent approach should be adopted forall microbial products entering the food chain (EC, 2001). Recently thequalified presumption of safety (QPS) status was launched by theEFSA (EFSA, 2007). Within this approach the absence of acquiredantibiotic resistance traits has to be confirmed for all strains of specieswith QPS status (EFSA, 2007).

To distinguish susceptible wild-type populations from subpopula-tions with acquired resistance mechanisms microbiological break-points are required (Turnidge and Paterson, 2007). A major obstaclefor proposing microbiological breakpoints for lactobacilli is that nogeneral agreement on reliable methods for Minimal InhibitionConcentration (MIC) determination has been reached. Consequently,a wide variety of different techniques have been reported. Theestablishment of breakpoints is clearly linked to the method used todetermine these data, and differences in the choice of media ormethods may have an impact on the MIC breakpoint values (Ferraro,2001). Accordingly, the breakpoints suggested by EFSAmay be seen asa pragmatic response, which are reviewed on a regular basis andmodified when necessary (EFSA, 2008).

The development of the lactic acid bacteria susceptibility testmedium (LSM; Klare et al., 2005) proved to be a first major step

82 S. Mayrhofer et al. / International Journal of Food Microbiology 144 (2010) 81–87

forward to establish a standardized method for lactobacilli. Mean-while, this medium has been frequently applied for antimicrobialsusceptibility testing of lactobacilli (Danielsen et al., 2008; Devirgiliiset al., 2008; Egervärn et al., 2007; Flórez et al., 2006; Huys et al., 2008;Klare et al., 2007; Korhonen et al., 2007; Mayrhofer et al., 2008;Rossetti et al., 2009). Largely based on the use of the LSM formulation,standard operating procedures for antimicrobial susceptibility testingof lactobacilli have been proposed within the 6th framework EUproject ACE-ART (Huys et al., 2010) and were recently made availableas ISO 10932/IDF 233 standard (ISO, 2010).

In the scope of these developments, the current study aimed todetermine the susceptibility of 101 strains belonging to theL. acidophilus group to 13 antibiotics proposed by EFSA at the time ofinvestigation (EFSA, 2005). Previously, the occurrence of antimicrobialresistance to seven antibiotics (i.e. ampicillin, clindamycin, erythro-mycin, gentamicin, streptomycin, tetracycline and vancomycin) inthese strains was determined using the Etest (Danielsen et al., 2008).Etest is a commercial, agar based diffusion method using plastic stripsimpregnated with an antibiotic concentration gradient. However,MICs for delineation of microbiological breakpoints should preferablybe determined using a referencemethodwith serial two-fold dilutionsin agar or broth such as the broth microdilution test (Vankerckhovenet al., 2008). Thus, in the present study MIC ranges for these strainswere established now including a larger set of antimicrobial agentsusing LSMmedium and the broth microdilution method following thenewest method developments. The obtained MIC distributions mayserve as a basis for the definition of microbiological breakpoints formembers of the L. acidophilus group. Additionally, molecular screeningfor a large number of antibiotic resistance (AR) genes by microarrayanalysis and PCR was performed to identify resistance determinantsresponsible for the resistant phenotypes displayed.

Materials and methods

Bacterial strains and growth conditions

A total of 101 strains of the L. acidophilus groupwere included in thisstudy belonging to the species: L. acidophilus (n=10), L. amylovorus(n=31), L. crispatus (n=7), L. gallinarum (n=7), L. gasseri (n=26) andL. johnsonii (n=20). The origin and identification of these isolates atspecies and strain level were previously described by Danielsen et al.(2008).

Strains were maintained in deMan Rogosa Sharpe broth (MRS,Oxoid, Hampshire, UK) containing 20% glycerol at a temperature of−80 °C. Before performing antimicrobial susceptibility testing, cultureswere streaked on LSM agar (Klare et al., 2005) consisting of 90%IsoSensitest broth (Oxoid), 10% MRS broth, supplemented with 15 g/Lbacteriological agar (Oxoid) and incubated for 24 h at 37 °C in ananaerobic chamber (80%N2, 10% CO2, 10% H2; Scholzen Technik, Kriens,Switzerland).

Antimicrobial susceptibility testing

Antimicrobial susceptibility testing was performed according tothe ISO 10932/IDF 233 standard (ISO, 2010)withminormodifications.Thus, LSM instead of MRS agar was used for the cultivation of thetested strains. Furthermore, the dilution of the antibiotic stocksolutions was done in LSM broth instead of water. Due to the use ofsingle strength instead of double strength LSM broth for the dilutionof the inocula, the same strength was finally achieved.

In short, ACE-ART VetMIC™ 96-well microtiter plates (NationalVeterinary Institute, Uppsala, Sweden) were used for determining theMICs (μg/mL) of ampicillin (0.12–8 μg/mL), clindamycin (0.12–8 μg/mL),erythromycin (0.12–16 μg/mL), gentamicin (0.5–32 μg/mL), streptomy-cin (2–256 μg/mL) and tetracycline (0.5–128 μg/mL). Additionally,susceptibilitywas tested separately against chloramphenicol, kanamycin,

linezolid, neomycin, quinupristin/dalfopristin, trimethoprim and vanco-mycin in microdilution assays in a concentration range of 0.12–128 μg/mL (Table 1). With the exception of quinupristin/dalfopristin(Sanalog, Kist, Germany) and linezolid (Pfizer, New York, USA), allantibiotics originated from Sigma-Aldrich (Saint Louis, Missouri, USA).For the production of the test plate, antibiotic stock solutions of5120 μg/mLwere prepared in sterilewater. To dissolve chloramphenicol,erythromycin and trimethoprim, 95% ethanol (chloramphenicol, eryth-romycin) or 0.05 MHCl (trimethoprim) was required in volumes as lowas possible. Consequently, stock solutions were diluted in LSM broth toobtain solutions with preliminary concentrations in the range of 0.25–256 μg/mL. Fifty microliters of these solutions were dispensed in eachwell of the microtiter plates.

Inocula of the strains were prepared by suspending colonies from24-h-incubated LSM agar plates in 5 mL 0.85% NaCl solution to aturbidity of McFarland standard 1. Subsequently, adjusted inoculawere diluted 1:1000 in LSM broth for inoculation of ACE-ARTVetMICTM microdilution plates by adding 100 μL of diluted inoculumto each well. The in-house prepared microdilution plates wereinoculated with 50 μL portions of a 1:500-diluted inoculum, resultingin a final antibiotic concentration of 0.12–128 μg/mL.

After incubating the plates under anaerobic conditions at 37 °C for48 h, MIC values were read as the lowest concentration of anantimicrobial agent at which visible growth was inhibited.

The accuracy of susceptibility testing was monitored by paralleluse of the quality control strain Enterococcus faecalis ATCC 29212.

Detection of antibiotic resistance genes

All strains phenotypically resistant to more than one of the testedantimicrobial agents were examined for the presence of AR genesapplying the oligonucleotide microarray assay described by van Hoeket al. (2005) and van Hoek and Aarts (2008).

In brief, oligonuclotide probes, 50–60 nucleotides in length, weredesigned representing AR genes belonging to the antibiotic classesdescribed in Table 2. Oligonucleotides were manufactured by Biolegio(Nijmegen, The Netherlands) and spotted on SCHOTT Nexterion Eslides (Isogen Life Science, IJsselstein, The Netherlands) using aMicroGrid II Microarrayer (BioRobotics Ltd., Cambridge, UK). Afterspotting, the microarrays were washed and blocked according to themanufacturer's instructions (SCHOTT JENAer GLAS GmbH, Jena,Germany).

Microarrays were prehybridized in hybridization buffer (5×SSC,0.2% SDS, 5×Denhardt's solution, 50% (v/v) formamide, 0.2 mg/mLdenatured herring sperm DNA) in a humidity chamber at 42 °C for atleast 4 h and subsequently rinsed. For hybridization, DNA of the testedstrains was isolated using the Wizard genomic DNA purification kit(Promega Benelux, Leiden, The Netherlands) and fluorescently labelledwith the cyanine dye Cy3 using the BioPrime DNA labelling system(Invitrogen BV, Breda, The Netherlands). Labelled DNAwas dissolved inhybridization buffer, denatured at 65 °C for 5 min and immediatelyplaced on ice. After applying the probes on themicroarray and coveringthe hybridization area with a cover slip, the slides were hybridizedovernight at 42 °C in a humidity chamber. Subsequently, slides werewashed and dried. Finally, they were scanned using a confocal laserscanner ScanArray ExpressHT (PerkinElmer Life). The software packageArrayVision (Imaging Research, London, Ontario, Canada) was used forthe analysis of the fluorescent signals.

Additionally, all strainswere examined for the presence of tet genesrelated to ribosomal protection proteins by PCR with the universalprimer set Ribo2-FW/Ribo2-RV (Aminov et al., 2001). Subsequently,positive strains were subjected to PCR amplification with the tet(W)-specific primers TetW-FW/TetW-RV (Aminov et al., 2001). PCRreactions were performed in a total volume of 25 μL containing0.5 μL of DNA, 1 μL of each primer (10 pmol/μL), 2.5 μL 10×PCR buffer(Finnzymes, Espoo, Finland), 0.5 μL dNTPs (10 mM), 0.25 μL DNA

Table 1MIC distribution of 13 antibiotics for species of the L. acidophilus group,b,c

Antibiotic Species Number of strains with indicated MIC (μg/mL):

aAMP L. acidophilus 9 1

L. amylovorus 1 4 17 7 1 1−

L. crispatus 6 1

L. gallinarum 5 2

L. gasseri 3 20 3

L. johnsonii 3 9 8

1 10 55 30 4 1−

aCHL L. acidophilus 8 2

L. amylovorus 3 27 1

L. crispatus 1 2 4

L. gallinarum 7

L. gasseri 2 16 8

L. johnsonii 3 12 2 3cat

All

All

1 10 74 13 3cat

aCLI L. acidophilus 2 2 1 3 2

L. amylovorus 2 8 10 4 2 1 4d, erm

L. crispatus 1 3 2 1

L. gallinarum 4 2 1

L. gasseri 3 2 3 2 16

L. johnsonii 5 2 1 5 3e, erm 4d, erm

All 3 17 20 13 7 12 21e, erm 8d, erm

aERY L. acidophilus 3 7

L. amylovorus 18 9 1erm 3erm

L. crispatus 6 1

L. gallinarum 5 2

L. gasseri 9 17

L. johnsonii 8 6e, erm 2 1erm 3erm

All 49 42e, erm 2 1erm 1erm 6erm

aGEN L. acidophilus 1 2 3 4

L.amylovorus 1 5 20 5

L. crispatus 2 4 1

L. gallinarum 2 1 4

L. gasseri 3 4 9 10

L. johnsonii 1 11 8

All 3 9 17 48 24aKAN L. acidophilus 1 1 1 2 5

L. amylovorus 1 2 5 22 1

L. crispatus 1 4 2

L. gallinarum 1 2 3 1

L. gasseri 1 3 10 12

L. johnsonii 2 10 8

All 3 1 5 14 54 24aLIN L. acidophilus 9 1

L. amylovorus 2 18 11

L. crispatus 4 3

L. gallinarum 5 2

L. gasseri 2 4 20

L. johnsonii 4 9 7

All 8 40 52 1aNEO L. acidophilus 1 3 1 4 1

L. amylovorus 1 3 5 15 7

L. crispatus 1 2 3 1

L. gallinarum 3 2 2

L. gasseri 2 2 10 11 1

L. johnsonii 3 6 10 1

All 1 3 3 14 16 34 28 2aQ/D L. acidophilus 2 2 4 2

L. amylovorus 7 17 4 2 1vat

L. crispatus 3 4

L. gallinarum 5 1 1

L. gasseri 11 14 1

L. johnsonii 3 9 6 2

All 20 35 32 11 2 1vat

aSTR L. acidophilus 3 6 1

L. amylovorus 2 8 9 2 4 3 3−

L. crispatus 1 2 4

L. gallinarum 2 1 1 3−

L. gasseri 1 2 18 3 2

L. johnsonii 5 9 4 1 1

All 4 20 44 13 6 5 3 6−

aTET L. acidophilus 2− 5− 1− 2−

L. amylovorus 2− 3− 5− 1− 2-/tet 4tet 6tet 8tet

L. crispatus 1− 5− 1tet

≤0.12 0.25 0.5 1 2 4 8 16 32 64 128 256 >256

(continued on next page)

83S. Mayrhofer et al. / International Journal of Food Microbiology 144 (2010) 81–87

Antibiotic Species Number of strains with indicated MIC (μg/mL):

L. gallinarum 4tet 3tet

L. gasseri 1− 1− 5− 16− 3−

L. johnsonii 1− 1− 3− 2− 3-/tet 4tet 6tet

All 3− 10− 15− 26− 4− 4-/tet 7-/tet 15tet 17tet

aTMP L. acidophilus 10

L. amylovorus 1 3 5 6

L. crispatus 1 6

L. gallinarum 1 1 5

L. gasseri 1 1 3

L. johnsonii 3 1 2

All 2 9 7 11aVAN L. acidophilus 7 3

L. amylovorus 1 28 2

L. crispatus 3 4

L. gallinarum 5 2

L. gasseri 23 3

L. johnsonii 2 18

All 1 45 52 3

16

21

14

72

≤0.12 0.25 0.5 1 2 4 8 16 32 64 128 256 >256

a Abbreviation of antimicrobials; AMP, ampicillin; CHL, chloramphenicol; CLI, clindamycin; ERY, erythromycin; GEN, gentamicin; KAN, kanamycin; LIN, linezolid; NEO,neomycin; Q/D, quinupristin/dalfopristin; STR, streptomycin; TET, tetracycline; TMP, trimethoprim; VAN, vancomycin.b Shaded areas show the range of dilutions tested for each antibiotic. MICs above the range exceed the highest concentration tested. The lowest MICs are equal to or lower than

the lowest concentration tested.c Atypical resistant strains deviating from the wild-type population (i.e. for ampicillin, chloramphenicol, erythromycin, quinupristin/dalfopristin), displaying very high

(N256 μg/mL) MIC values (i.e. for clindamycin, streptomycin) or belonging to the subpopulation at the high-end concentration range in case of bimodal MIC distributions (i.e. fortetracycline) are indicated in bold. Resistance genes (cat, erm, vat, tet) detected for these strains are typed in superscripts, whereas for not detected resistance mechanisms - isindicated.d Strains displaying clindamycin MIC values of N8 μg/mL by broth microdilution were found to be highly resistant (N 256 μg/mL) by Etest (Danielsen et al., 2008).e Resistance not phenotypically expressed in one strain harboring an erm gene.

Table 1 (continued)

84 S. Mayrhofer et al. / International Journal of Food Microbiology 144 (2010) 81–87

polymerase (2 U/μL, Dynazyme; Finnzymes, Espoo, Finland) and 19.25µl distilled sterile water. The following PCR program was used: 95 °C,5 min, then 35 cycles of 95 °C,60 s; 50 °C (Ribo2) or 64 °C (tet(W)),60 s; 72 °C, 60 s endingwith 72 °C, 8 min. The obtained PCR fragmentswere analyzed by electrophoreses on a 2% agarose gel, stained withethidium bromide and visualized with UV light.

Results

Antimicrobial susceptibility testing

The results of antimicrobial susceptibility testing of 101 L.acidophilus strains using broth microdilution are listed in Table 1. Ingeneral, relatively narrow unimodal MIC distributions at the low-endconcentration range, occasionally joined by a few atypical strains,were determined by visual inspection of the antimicrobial agentsampicillin (≤0.12–2 μg/mL), chloramphenicol (1–8 μg/mL), erythro-mycin (≤0.12–0.5 μg/mL), gentamicin (≤0.5–8 μg/mL), linezolid (1–8 μg/mL), quinupristin/dalfopristin (≤0.12–2 μg/mL) and vancomycin(0.25–2 μg/mL).

For kanamycin and neomycin, most species displayed broaderunimodal MIC distributions at the high-end concentration range, i.e.4–128 μg/mL and 1–128 μg/mL, respectively.

The unimodal MIC distributions of streptomycin for L. acidophilus,L. gasseri and L. johnsonii were also in a higher concentration range(≤2–64 μg/mL), whereas MIC values of strains of the speciesL. amylovorus, L. crispatus and L. gallinarumwere unevenly distributedat the whole concentration range tested.

Either uni or bimodal distributions were observed for tetracycline.In case of bimodal distributions (as detected for L. amylovorus, L.crispatus and L. johnsonii), two subpopulations could be differentiat-ed: onewith lowMICs (2–16 μg/mL) and another onewith higherMICvalues (32–N128 μg/mL). Concerning unimodal MIC distributions(obtained for the species L. acidophilus, L. gallinarum and L. gasseri),MIC values were either positioned at the low-end (1–16 μg/mL) orhigh-end (128–N128 μg/mL) concentration range.

With the exception of some atypical strains with MICs higher than8 μg/mL (broth microdilution) or 256 μg/mL (Etest; Danielsen et al.,

2008), the distribution of clindamycin MIC values of the investigatedL. acidophilus group strains was broad and positioned at the low-endconcentration range (≤0.12–8 μg/mL).

The trimethoprim MIC values were usually situated in higherconcentration ranges from 16 to N128 μg/mL.

Detection of antibiotic resistance genes

Strains deviating from the wild-type population (i.e. for ampicillin,chloramphenicol, erythromycin and quinupristin/dalfopristin), dis-playing very high (N256 μg/mL) MIC values (i.e. for clindamycin andstreptomycin) or belonging to the subpopulation at the high-endconcentration range in case of bimodal MIC distributions (i.e. fortetracycline) were considered as atypical resistant strains (Table 1). Inthis way, 44 of the 101 strains tested were considered to displayresistant phenotypes. Of these, 29 strains were resistant to one of theinvestigated antibiotics (i.e. 28 strains to tetracycline and one strain tostreptomycin), whereas 15 strains displayed atypical resistance tomore than one of the antimicrobial agents tested.

To identify resistance determinants responsible for the displayedresistant phenotypes, the 15 multi-resistant strains were screened forthe presence of AR genes using the oligonucleotide microarrayplatform. Additionally, PCR was performed to examine the occurrenceof tet genes in all strains resistant to tetracycline. The investigation ofthe 15 multi-resistant strains demonstrated the presence of vat(E) inthe quinupristin/dalfopristin-resistant L. amylovorus strain and a cat-TC like gene in all three chloramphenicol-resistant L. johnsonii strains(Table 1). An erm(A) determinant could be detected in three L.amylovorus strains with erythromycin MICs of 8 or higher than 16 μg/mL, whereas erm(B) was found in all four erythromycin-resistant L.johnsonii strains and one L. amylovorus strain with an erythromycinMIC value higher than 16 μg/mL. As erm genes also encode resistanceto lincosamide antibiotics, the clindamycin MIC values of the ermpositive strains were expectedly higher than 8 μg/mL by brothmicrodilution (this study) or 256 μg/mL by Etest (Danielsen et al.,2008). When investigating all tetracycline-resistant strains by micro-array or PCR, the tet(W) gene was determined in all strains displayingMIC values higher than 64 μg/mL. With the exception of one L.

Table 2Antibiotic resistance genes represented on the oligonucleotide microarray.

Class of antibiotic Genes

Aminoglycoside aac(3)-Ia; aac(3)-Ib; aac(3)-IIa; aac(3)-IIIb; aac(3)-IIIc; aac(3)-VII; aac(6')-aph(2''); aac(6')-Ib; aac(6')-Il; aac(6')-Iq; aac(6')-Iy;aac(6)-II; aac(6)-IIb; aac(6)-IIc; aacA1; aacA4; aacA5; aacA7; aacC1; aacC2; aacC3; aacC9; aadA1; aadA2; aadA6; aadA10; aadA11;aadA13; aadB; aadD; aadE (also referred to as ant(6)); aph(2')-Ib; aph(2')-Ic; aph(2')-Id; aph(3')-Id; aph(3')-IIa; aphA1; aphA1-IAB;aphA2; aphA3 (also referred to as aph(3′)-IIIa); aphA6; aphA7; aphE; nptII; sat2; sat3; sat4; strA; strB

β-lactamsa blaACC-01_03; blaACT-01,02; blaCARB-04_09; blaCMY-01,09_11,19; blaCTX-M-01 groupb; blaCTX-M-02 groupb; blaCTX-M-09 groupb; blaDHA-01_02;blaFOX; blaIMP; blaKPC; blaMIR; blaMOR; blaMOX-01; blaOXA-01_07,10,13_17,19_35,37,40,47_49,56,72_74,101,102; blaPER-01_03; blaPSE-01; blaROB-01; blaTEM;blaVIM-01_11,14; blaZEG-01

Chloramphenicol cat; cat(pC194); cat-TC; catII; catIII; catA1; catA3; catB; catB2; catB3; catB6; catB8; catB9; catD; catP; catQ; cmlA; cmlA1; cmlA4; cmlA5;cmlA6; cmlA7; cmlB; floR

Macrolides- Lincosamides-Streptogramins (MLS group)

ere(A); ere(A2); ere(B); erm(A); erm(B); erm(C); erm(D); erm(F); erm(FS); erm(FU); erm(G;) erm(GM); erm(GT); erm(J); erm(K);erm(Q); erm(T); mef(A); mef(E); mph(A); mph(BM); mph(C); mph(K); msr(A); msr(B); msr(SA); sat(A); sat(G); vat; vat(A); vat(B);vat(C); vat(D); vat(E); vat(E3_E8); vga; vga(A); vga(A)LC; vgb; vgb(B)

Sulfonamide sul1; sul2; sulATetracycline otr(A); otr(B); tet(30); tet(31); tet(32); tet(33); tet(34); tet(35); tet(36); tet(37); tet(A); tet(A(P)); tet(B); tet(B(P)); tet(C); tet(D);

tet(E); tet(G); tet(H); tet(J); tet(K); tet(L); tet(M); tet(O); tet(Q); tet(S); tet(T); tet(U); tet(V); tet(W); tet(X); tet(Y); tet(Z)Trimethoprim dfrA; dfrA1; dfrA2; dfrA3; dfrA5; dfrA6; dfrA7; dfrA8; dfrA9; dfrA10; dfrA12; dfrA14; dfrA15; dfrA16; dfrA17; dfrA19; dfrA21; dfrA22; dfrA23;

dfrB2; dfrB3; dfrC; dfrDVancomycin vanA; vanB; vanC1; vanC2/C3; vanD; vanE

a For example blaACC-01_03 means that oligonucleotide(s) representing blaACC-01, blaACC-02 and blaACC-03 are present on the microarray. The nomenclature for genes encoding forESBLs has been standardized by adding a number to the gene name (for review see Paterson and Bonomo, 2005).

b For genes included in these groups see van Hoek and Aarts, 2008.

85S. Mayrhofer et al. / International Journal of Food Microbiology 144 (2010) 81–87

johnsonii strain, tet(W) was also found in all strains with MICs of64 μg/mL, whereas the gene could only be detected in one out of fourL. acidophilus group strains with MICs of 32 μg/mL. All these tet(W)-harbouring strains were also positive for the presence of tet genesusing the universal primer set Ribo2 in a PCR, whereas no ribosomalprotection protein related genes were found in all other strains withtetracycline MICs lower than 32 μg/mL. Likewise, examination of oneampicillin- and six streptomycin-resistant strains for the presence ofresistance genes by microarray analysis did not allow the identifica-tion of the underlying resistance mechanism. In contrast, one L.johnsonii strain phenotypically susceptible to erythromycin (0.25 μg/mL) and clindamycin (8 μg/mL) seemed to harbour an erm(B)resistance determinant.

Discussion

In order to make a meaningful interpretation of the antimicrobialsusceptibility profiles of lactobacilli used in food and feed applications,microbiological breakpoints are needed for categorizing susceptible orresistant strains. As the establishment of breakpoints is clearly linkedto the method used, the antimicrobial susceptibility of a collection ofstrains belonging to the L. acidophilus group was tested following thenewest method developments.

According to the literature, lactobacilli are usually susceptible toampicillin, chloramphenicol, erythromycin, linezolid and quinupris-tin/dalfopristin (Ammor et al., 2007; Katla et al., 2001; Luh et al., 2000).Furthermore, homofermentative lactobacilli like species of theL. acidophilus group are also susceptible to vancomycin (EFSA, 2008).This is in good concordance with results of our study. However, in anumber of cases higher MIC values for ampicillin, chloramphenicol,erythromycin and quinupristin/dalfopristin indicated the presence ofstrains with acquired resistance among the collection under study.With the exception of one ampicillin-resistant strain, resistance genes,i.e. cat-TC, erm(A), erm(B) and vat(E), could be identified in all strainsdetermined as atypically resistant by visual inspection of thedistribution of MIC values. The resistance determinants cat-TC, erm(B) and vat(E) have already been described in lactobacilli (Ammor etal., 2007), but to our knowledge, this is the first report of erm(A) in thegenus Lactobacillus. Surprisingly, one L. johnsonii strain phenotypicallysusceptible to clindamycin and erythromycin seemed to harbour erm(B). Additional investigations have revealed that four tandem insertednucleotides in this resistance determinant probably result in a non-transcribable erm(B) (A. van Hoek, unpublished results). Thus, silent

genes, pseudogenes, partial or unexpressed genes may cause falsepositive results, leading to discordance between phenotypic andgenotypic resistance data (Sundsfjord et al., 2004; Woodford andSundsfjord, 2005). Furthermore, only known resistance determinantsor those with sufficient DNA homology can be screened for usinggenotypic analyses (Sundsfjord et al., 2004;Woodford and Sundsfjord,2005), explaining the missing underlying resistance mechanism incase of the phenotypically ampicillin-resistant strain. Nevertheless,phenotypic and genotypic resistance data were generally in goodagreement for thementioned antimicrobial agents as resistance genescould be determined in 92.3% of the strains showing phenotypicresistance. In a bimodal MIC distribution where resistant non-wildtype strains of a species population are well separated from thesusceptible wild-type, the determination of breakpoints is usuallyrather straightforward. In contrast, it is difficult to make appropriatebreakpoint determinations when the distribution is unimodal (Fer-raro, 2001) as in case of linezolid and vancomycin. Hence, Turnidgeand Paterson (2007) suggested setting the breakpoint one or twodilutions above the high-end of the wild-type MIC range. This wouldbe preferable as a ±1 dilution variation of MICs is intrinsic to thetesting system (CLSI, 2008).

Compared to gentamicin, which showed a unimodal MIC distribu-tion at the low-end concentration range, the other aminoglycosideantibiotics streptomycin, kanamycin and neomycin displayed broadMIC distributions at the high-end concentration range. Generally, thereduced susceptibility of lactobacilli to aminoglycosides is thought tooccur because of membrane impermeability (Elkins andMullis, 2004).Due to the better ability of gentamicin to cross the membrane (Elkinsand Mullis, 2004), a tendency of lower MICs for gentamicin is known(Danielsen and Wind, 2003; Egervärn et al., 2007). Additionally,interference by low test medium pH with aminoglycosides wasdescribed, reducing their antibacterial activity (EFSA, 2008). Thismay explain why no aminoglycoside resistance genes could bedetected by microarray analysis in the tested strains, including thosewith very high streptomycin MIC values. Also in other studiesinvestigating the antimicrobial resistance of lactobacilli (Klare et al.,2007; Ouoba et al., 2008), no aminoglycoside resistance determinantshave been identified. In contrast, Rojo-Bezares et al. (2006) andTenorio et al. (2001) found the genes aac(6′)-aph(2″), ant(6) and aph(3′)-IIIa whichwere also represented on themicroarray applied in ourstudy. However, the former authors stated that the presence of thedetected resistance genes was not associated with an increase in the“intrinsically high” MIC values.

86 S. Mayrhofer et al. / International Journal of Food Microbiology 144 (2010) 81–87

Acquired resistance to tetracycline is a well known phenomenonin strains of the L. acidophilus group (Cauwerts et al., 2006; Danielsenand Wind, 2003; Delgado et al., 2005; Klare et al., 2007; Korhonen etal., 2007). Thus, unimodal as well as bimodal distributions could bedetected for this antimicrobial agent and the species investigated. Thetet(W) gene, which has already been reported in strains of L. crispatus,L. johnsonii, L. paracasei and L. reuteri (Egervärn et al., 2009), wasfound in all strains belonging to the resistant subpopulations, exceptfor one L. amylovorus and three L. johnsonii strains with MICs of 32 or64 μg/mL. Interestingly, these strains were categorized as intermedi-ate based on the MIC values obtained by Etest (Danielsen et al., 2008).Although results of Etest and broth microdilution were generally ingood agreement (Mayrhofer et al., 2008), one of the other two L.amylovorus strains, additionally considered as intermediate byDanielsen et al. (2008), could be classified as resistant with a MICvalue of 32 μg/mL and a tet(W) genewithin our study. The other strainwas considered as susceptible displaying a MIC of 16 μg/mL but notlinked to an underlying ribosomal protection protein gene. As in thecase for susceptibility testing of clinical isolates, the definition of anintermediate or so-called indeterminate (Metzler and DeHaan, 1974)category could be meaningful as a buffer zone between resistant andsusceptible categories to prevent discrepancies in interpretations(CLSI, 2008). This would especially be necessary in case of bimodaldistributions with an overlap of wild-type and resistant MICs. On theother hand, one has to bear in mind that an unknown resistancemechanism could also be responsible for the increased MIC values ofthese strains.

With the exception of some atypically resistant strains harbouringan erm gene, broad MIC distributions at the low-end concentrationrange with no clear pattern could be observed for clindamycin. Hence,Dutta and Devriese (1984) noticed that the lincosamide antibioticsclindamycin and lincomycin are unique in that many bacterial specieshave two populations of naturally susceptible strains. Otherwise, abroad distribution of MICs might be the result of an accumulation ofdifferent resistance determinants (Phillips, 1998). In particular,resistance mechanisms involved in low-level resistance are difficultto detect by phenotypic methods (Sundsfjord et al., 2004) and in caseof mutations often more than one nucleotide polymorphism may berequired before resistance is expressed (Ferraro, 2001). Thus, thebroad MIC distribution concerning clindamycin could be due tomutations which are not detected by the oligonucleotide microarrayapplied. As an extended distribution range of MICs hampers thedetermination of breakpoints, further characterization studies arerequired.

All tested species of the L. acidophilus group either exhibitedunimodal or bimodal MIC distributions at the high-end concentrationrange for trimethoprim. However, none of the 23 trimethoprimrelated dfr genes, represented on the microarray (Table 2), could bedetected in the strains investigated. Concerning this antimicrobialagent, it is known that the presence of antagonistic components in theLSM medium complicates susceptibility testing (Danielsen et al.,2004; Klare et al., 2007). Thus, MIC testing of trimethoprim for lacticacid bacteria is no longer considered as relevant according to thecurrent technical guidance of EFSA (EFSA, 2008).

In conclusion, most phenotypic resistances observed from MICdeterminations could be confirmed by molecular methods. Thus, MICranges derived from this study could be used as a basis for reviewingpresent microbiological breakpoints. However, MIC determination ofmicroorganisms intended for use in food and feed applications shouldbe accompaniedby a broad screeningof potential resistancegenes usingmolecular tools, especially in case of the possible occurrence of silentresistance genes. To investigate the discordance between phenotypicand genotypic resistance data in more detail, further studies regardingthe molecular characterization of unknown atypical resistance patternsand the possible influence of the test medium on the results ofphenotypical methods are needed.

Acknowledgements

This study was performed as part of the EU project “ACE-ART”(CT-2003-506214) within the 6th framework. G. Huys is a postdoc-toral fellow of the Fund for Scientific Research — Flanders, Belgium(F.W.O.–Vlaanderen). Pfizer (New York, USA) is thanked for providinglinezolid.

References

Aminov, R.I., Garrigues-Jeanjean, N., Mackie, R.I., 2001. Molecular ecology oftetracycline resistance: development and validation of primers for detection oftetracycline resistance genes encoding ribosomal protection proteins. Appl.Environ. Microbiol. 67, 22–32.

Ammor, M.S., Flórez, A.B., Mayo, B., 2007. Antibiotic resistance in non-enterococcallactic acid bacteria and bifidobacteria. Food Microbiol. 24, 559–570.

Bernardeau, M., Guguen, M., Vernoux, J.P., 2006. Beneficial lactobacilli in food and feed:long-term use, biodiversity and proposals for specific and realistic safetyassessments. FEMS Microbiol. Rev. 30, 487–513.

Cauwerts, K., Pasmans, F., Devriese, L.A., Haesebrouck, F., Decostere, A., 2006. CloacalLactobacillus isolates from broilers often display resistance toward tetracyclineantibiotics. Microb. Drug Resist. 12, 284–288.

CLSI (Clinical and Laboratory Standards Institute), 2008. Development of in vitrosusceptibility testing criteria and quality control parameters; approved guideline.M23-A3 third edition. CLSI document, 28.

Danielsen, M., Wind, A., 2003. Susceptibility of Lactobacillus spp. to antimicrobialagents. Int. J. Food Microbiol. 82, 1–11.

Danielsen, M., Andersen, H.S., Wind, A., 2004. Use of folic acid casei medium revealstrimethoprim susceptibility of Lactobacillus species. Lett. Appl. Microbiol. 38,206–210.

Danielsen, M., Mayrhofer, S., Domig, K.J., Amtmann, E., Mayer, H.K., Flórez, A.B., Mayo,B., Korhonen, J., Tosi, L., 2008. Assessment of the antimicrobial wild-type minimuminhibitory concentration of species of the Lactobacillus delbrueckii group. DairyScience and Technology 88, 183–191.

De Angelis, M., Siragusa, S., Berloco, M., Caputo, L., Settanni, L., Alfonsi, G., Amerio, M.,Grandi, A., Ragni, A., Gobbetti, M., 2006. Selection of potential probiotic lactobacillifrom pig feces to be used as additives in pelleted feeding. Res. Microbiol. 157,792–801.

Delgado, S., Flórez, A.B., Mayo, B., 2005. Antibiotic susceptibility of Lactobacillus andBifidobacterium species from the human gastrointestinal tract. Curr. Microbiol. 50,202–207.

Devirgiliis, C., Caravelli, A., Coppola, D., Barile, S., Perozzi, G., 2008. Antibiotic resistanceand microbial composition along the manufacturing process of Mozarella di BufalaCampana. Int. J. Food Microbiol. 128, 378–384.

Dutta, G.N., Devriese, L.A., 1984. Observations on the in vitro sensitivity and resistanceof Gram positive intestinal bacteria of farm animals to growth promotingantimicrobial agents. J. Appl. Bacteriol. 56, 117–123.

EFSA, 2005. Opinion of the Scientific Panel on Additives and Products or SubstancesUsed in Animal Feed (FEEDAP) on the updating of the criteria used in assessment ofbacteria for resistance to antibiotics of human and veterinary importance. The EFSAJournal 223, 1–12.

EFSA, 2007. Opinion of the Scientific Committee on a request from EFSA on theintroduction of a Qualified Presumption of Safety (QPS) approach for assessment ofselected microorganisms referred to EFSA. The EFSA Journal 587, 1–16.

EFSA, 2008. Technical guidance prepared by the Panel on Additives and Products orSubstances in Animal Feed (FEEDAP) on the update of the criteria used in theassessment of bacterial resistance to antibiotics of human and veterinaryimportance. The EFSA Journal 732, 1–15.

Egervärn, M., Danielsen, M., Roos, S., Lindmark, H., Lindgren, S., 2007. Antibioticsusceptibility profiles of Lactobacillus reuteri and Lactobacillus fermentum. J. FoodProt. 70, 412–418.

Egervärn, M., Roos, S., Lindmark, H., 2009. Identification and characterization ofantibiotic resistance genes in Lactobacillus reuteri and Lactobacillus plantarum. J.Appl. Microbiol. 107, 1658–1668.

Elkins, C.A., Mullis, L.B., 2004. Bile-mediated aminoglycoside sensitivity in Lactobacillusspecies likely results from increased membrane permeability attributable to cholicacid. Appl. Environ. Microbiol. 70, 7200–7209.

European Commission (EC), 2001 updated 2003. Opinion of the Scientific Committeeon Animal Nutrition on the criteria for assessing the safety of micro-organismsresistant to antibiotics of human clinical and veterinary importance. [http://ec.europa.eu/food/fs/sc/scan/out108_en.pdf].

Ferraro, M.J., 2001. Should we reevaluate antibiotic breakpoints? Clin. Infect. Dis. 33,227S–229S.

Flórez, A.B., Egervärn, M., Danielsen, M., Tosi, L., Morelli, L., Lindgren, S., Mayo, B., 2006.Susceptibility of Lactobacillus plantarum strains to six antibiotics and definition ofnew susceptibility-resistance cutoff values. Microb. Drug Resist. 12, 252–256.

Holzapfel, W.H., 2006. Introduction to prebiotics and probiotics. In: Goktepe, I., Juneja,V.K., Ahmedna, M. (Eds.), Probiotics in food safety and human health. CRC Press,Boca Raton, pp. 1–33.

Huys, G., D'Haene, K., Danielsen, M., Mättö, J., Egervärn, M., Vandamme, P., 2008.Phenotypic and molecular assessment of antimicrobial resistance in Lactobacillusparacasei strains of food origin. J. Food Prot. 71, 339–344.

87S. Mayrhofer et al. / International Journal of Food Microbiology 144 (2010) 81–87

Huys, G., D'Haene, K., Cnockaert, M., Tosi, L., Danielsen, M., Flórez, A.B., Mättö, J.,Axelsson, L., Korhonen, J., Mayrhofer, S., Egervärn, M., Giacomini, M., Vandamme, P.,2010. Intra- and interlaboratory performances of two commercial antimicrobialsusceptibility testing methods for bifidobacteria and nonenterococcal lactic acidbacteria. Antimicrob. Agents Chemother. 54, 2567–2574.

Ishibashi, N., Yamazaki, S., 2001. Probiotics and safety. Am. J. Clin. Nutr. 73, 465S–470S.ISO (International Organization for Standardization), 2010. ISO 10932/IDF 233 standard

“Milk and milk products—Determination of the minimal inhibitory concentration(MIC) of antibiotics applicable to bifidobacteria and non-enterococcal lactic acidbacteria (LAB)”.

Katla, A.K., Kruse, H., Johnsen, G., Herikstad, H., 2001. Antimicrobial susceptibility ofstarter culture bacteria used in Norwegian dairy products. Int. J. Food Microbiol. 67,147–152.

Klare, I., Konstabel, C., Müller-Bertling, S., Reissbrodt, R., Huys, G., Vancanneyt, M.,Swings, J., Goosens, H., Witte, W., 2005. Evaluation of new broth media formicrodilution antibiotic susceptibility testing of lactobacilli, pediococci, lactococci,and bifidobacteria. Appl. Environ. Microbiol. 71, 8982–8986.

Klare, I., Konstabel, C., Werner, G., Huys, G., Vankerckhoven, V., Kahlmeter, G.,Hildebrandt, B., Müller-Bertling, S., Witte, W., Goossens, H., 2007. Antimicrobialsusceptibilities of Lactobacillus, Pediococcus and Lactococcus human isolates andcultures intended for probiotic or nutrional use. J. Antimicrob. Chemother. 59,900–912.

Korhonen, J.M., Sclivagnotis, Y., von Wright, A., 2007. Characterization of dominantcultivable lactobacilli and their antibiotic resistance profiles from faecal samples ofweaning piglets. J. Appl. Microbiol. 103, 2496–2503.

Luh, K.T., Hsueh, P.R., Teng, L.J., Pan, H.J., Chen, Y.C., Lu, J.J., Wu, J.J., Ho, S.W., 2000.Quinupristin–Dalfopristin resistance among Gram-positive bacteria in Taiwan.Antimicrob. Agents Chemother. 44, 3374–3380.

Mayrhofer, S., Domig, K.J., Mair, C., Zitz, U., Huys, G., Kneifel, W., 2008. Comparison ofbroth microdilution, Etest, and agar disk diffusion methods for antimicrobialsusceptibility testing of Lactobacillus acidophilus group members. Appl. Environ.Microbiol. 74, 3745–3748.

Metzler, C.M., DeHaan, R.M., 1974. Susceptibility tests of anaerobic bacteria: statisticaland clinical considerations. J. Infect. Dis. 130, 588–594.

Ouoba, L.I.I., Lei, V., Jensen, L.B., 2008. Resistance of potential probiotic lactic acidbacteria and bifidobacteria of African and European origin to antimicrobials:Determination and transferability of the resistance genes to other bacteria. Int. J.Food Microbiol. 121, 217–224.

Paterson, D.L., Bonomo, R.A., 2005. Extended-spectrum β-lactamases: a clinical update.Clin. Microbiol. Rev. 18, 657–686.

Phillips, I., 1998. The subtleties of antibiotic resistance. J. Antimicrob. Chemother. 42,5–12.

Rojo-Bezares, B., Sáenz, Y., Poeta, P., Zarazaga, M., Ruiz-Larrea, F., Torres, C., 2006.Assessment of antibiotic susceptibility within lactic acid bacteria strains isolatedfrom wine. Int. J. Food Microbiol. 111, 234–240.

Rossetti, L., Carminati, D., Zago, M., Giraffa, G., 2009. A Qualified Presumption of Safetyapproach for the safety assessment of Grana Padano whey starters. Int. J. FoodMicrobiol. 130, 70–73.

Sundsfjord, A., Simonsen, G.S., Haldorsen, B.C., Haaheim, H., Hjelmevoll, S.O., Littauer, P.,Dahl, K.H., 2004. Genetic methods for detection of antimicrobial resistance. APMIS:Acta Pathol. Microbiol. Immunol. Scand. 112, 815–837.

Tenorio, C., Zarazaga, M., Martinez, C., Torres, C., 2001. Bifunctional enzyme 6´-N-aminoglycoside acetyltransferase-2´´-O-aminoglycoside phosphotranferase in Lac-tobacillus and Pediococcus isolates of animal origin. J. Clin. Microbiol. 39, 824–825.

Teuber, M., Meile, L., Schwarz, F., 1999. Acquired antibiotic resistance in lactic acidbacteria from food. Antonie Leeuwenhoek 76, 115–137.

Turnidge, J., Paterson, D.L., 2007. Setting and revising antibacterial susceptibilitybreakpoints. Clin. Microbiol. Rev. 20, 391–408.

van Hoek, A.H.A.M., Aarts, H.J.M., 2008. Microarray-based detection of antibioticresistance genes in Salmonella. Food Anal. Methods 1, 95–108.

van Hoek, A.H.A.M., Scholtens, I.M.J., Cloeckaert, A., Aarts, H.J.M., 2005. Detection ofantibiotic resistance genes in different Salmonella serovars by oligonucleotidemicroarray analysis. J. Microbiol. Meth. 62, 13–23.

Vankerckhoven, V., Huys, G., Vancanneyt,M., Vael, C., Klare, I., Romond,M.B., Entenza, J.M.,Moreillon, P., Wind, R.D., Knol, J., Wiertz, E., Pot, B., Vaughan, E.E., Kahlmeter, G.,Goossens, H., 2008. Biosafety assessment of probiotics used for human consumption:recommendations from the EU-PROSAFE project. Trends Food Sci. Technol. 19,102–114.

von Wright, A., 2005. Regulating safety of probiotics — the European approach. Curr.Pharm. Des. 11, 17–23.

Wassenaar, T.M., Klein, G., 2008. Safety aspects and implications of regulation ofprobiotic bacteria in food and food supplements. J. Food Prot. 71, 1734–1741.

Woodford, N., Sundsfjord, A., 2005. Molecular detection of antibiotic resistance: whenand where? J. Antimicrob. Chemother. 56, 259–261.