Introduction

Lactic acid bacteria (LAB) have a long history of safe use as fermenting natural products and probiotics intended for health benefits and have acquired the “Generally recognized as safe” (GRAS) status [32]. Since the introduction of antibiotics more than 50 years ago, emergence of resistant microorganisms has become a major threat to public health [34]. Antibiotic resistance is well studied and documented for human pathogenic species [32]. However, since the last decade, researchers have also focused on characterizing antibiotic resistance in LAB [4]. LAB can serve as reservoir for antibiotic resistance genes and transfer it to other microorganisms including pathogens [23, 43]. This situation has become more hazardous by the overuse and misuse of antibiotics in feed, agriculture, and veterinary applications [2, 42]. European Food Safety Authority (EFSA) recommends that bacterial strains harboring transferable antibiotic resistance genes should not be used in animal feeds, fermented and probiotic foods for human [14].

Lactobacilli are dominant bacteria in several fermented foods such as meat and dairy products, which interact with gut microflora on ingestion [2]. Lactobacilli are generally intrinsically resistant for quinolones, trimethoprim, and sulphonamides, and susceptible to all protein synthesis inhibitors except aminoglycosides. Resistance of many species of lactobacilli except L. delbrueckii subsp. bulgaricus, L. acidophilus, L. johnsonii, and L. crispatus to glycopeptides is also considered intrinsic [8, 32]. Intrinsic resistance is not horizontally transferable as it is chromosomally encoded and related to the general physiology or anatomy of an organism. Acquired resistance is horizontally transferable, which emerges from genetic changes by mutations or acquisition of genetic elements (plasmids or transposons) most probably by conjugation [2, 4, 32]. Acquired antibiotic resistance in lactobacilli has been reported frequently in recent years from fermented milk products [10, 11, 29, 35, 41], pickles, and meat but there are no reports from China in this regard.

The objective of this study was a safety assessment of bacterial isolates from starter cultures and naturally fermented foods by phenotypic screening, polymerase chain reaction (PCR) and filter mating. This study provides data on the current resistance situation and insinuate for further monitoring and development of regulations for antibiotic resistance in food associated LAB in China.

Materials and Methods

Isolation of LAB and Growth Conditions

Bacterial strains were isolated from commercial and traditionally fermented food samples including Yogurt (n = 13), Whey or fermented milk (n = 10), Jiang shui (n = 3), and fermented vegetables (n = 15). Jiang shui is a traditional Chinese drink, made by fermenting a mixture of vegetables. Food samples obtained from local markets were homogenized and serial dilutions were plated on selective media and incubated at 37°C for 48 h. Lactobacilli were selected on MRS agar (Oxoid) plates incubated in anaerobic conditions (Anaerogen, Oxoid), while streptococci were isolated on M17 (Oxoid) agar plates in aerobic environment. After the incubation, distinguished colonies were selected, purified, and stored in MRS broth with 15% (w/v) glycerol at −20°C. Enterococcus faecalis ATCC 29212 and E. faecium ATCC 35667 were used as control strains in antibiotic sensitivity testing. Prior to all experiment bacterial isolates were subcultured at least twice.

Identification of LAB

Preliminary identification of isolates was performed by gram staining and catalase test. DNA was extracted from bacterial isolates using a commercial kit (TianGen, Beijing, China) according to the manufacturer’s instructions. Lactobacillus genus was identified by PCR using genus specific primers XB5 and LbLMA1-R. For the identification of species, amplification of partial 16S rRNA gene was performed by universal primers (Table 1) in an automated thermocycler (ABI 2720 USA) using the following program: initial denaturation at 95°C for 5 min; followed by 35 cycles of 95°C for 45 s, 55°C for 45 s, 72°C for 1 min, and a final extension at 72°C for 7 min. Amplified products were resolved by 1.5% (w/v) agarose gel electrophoresis and visualized by ethidium bromide staining (0.5 μg/ml) under UV transillumination. PCR products (nearly 1,380 bp) were purified and sequenced directly by Beijing genomics institute (BGI) by using the same universal primers. LAB isolates were identified to the species level by comparing their sequences with those in the NCBI databases by BLAST. All sequences were submitted to GenBank and accession numbers were obtained.

Table 1 Nucleotide sequences of primer sets used for PCR in this study
Full size table

Antibiotic Resistance

Minimum inhibitory concentrations (MICs) of 14 antibiotics were determined by broth microdilution method using LAB susceptibility test medium (LSM) [30], a mixed formulation containing Iso-Sensitest broth (90%) and MRS broth (10%). Penicillin G, ampicillin, erythromycin, clindamycin, tetracycline, cefsulodin, bacitracin, and ciprofloxacin were tested in concentration range of (0.032–64 mg/l) while, gentamicin, kanamycin, fusidic acid, streptomycin, chloramphenicol, nalidixic acid, and vancomycin in concentration range of (0.125–256 mg/l). In brief, 50 μl of antibiotics, doubly diluted in LSM broth were inoculated with 100 μl of test strain inoculum. Inoculum of the test strain was prepared by suspending colonies from LSM agar plates in 5 ml 0.85% NaCl solution to a turbidity of McFarland 1 standard and subsequently diluting them 1:1000 in LSM broth. Plates were incubated under anaerobic conditions at 37°C for 48 h. MICs were read as the lowest concentration of an antimicrobial agent at which visible growth was inhibited. Breakpoints were adopted from EFSA [15]. Breakpoints for antibiotic not covered by EFSA were adopted from Danielsen and Wind [10].

Amplification of Antibiotic Resistance Genes

Antibiotic resistance genes for β-lactam (bla), erythromycin [erm(A), erm(B), and mef(A)], tetracycline [tet(M), tet(O), tet(Q), tet(S), tet(W), tet(K), and tet(L)], gentamycin [aac(6′)-aph(2′′)], and chloramphenicol (cat) were amplified by PCR using primers given in Table 1. The reaction mixtures (50 μl) contained 25 pmol of each primer, 1.5, 2.0, 2.5, or 3.0 mM MgCl2 depending on experiment, 50–100 ng bacterial DNA, and 2.5 U of Taq DNA Polymerase. DNA fragments were amplified in a thermal cycler up to 40 cycles by using annealing temperatures given in Table 1. Positive and negative controls form our lab were used for all PCR reactions. PCR products were subjected to agarose gel electrophoresis in TAE buffer. Gels were stained with ethidium bromide (5 μg/ml) and visualized under UV transillumination.

Transfer of Antibiotic Resistance

Transfer of antibiotic resistance was analyzed by filter mating experiments as described by Gevers et al [24] with slight modifications. Enterococcus faecalis 181 and E. facium 258 were used as the recipient, which were sensitive to erythromycin and tetracycline, and resistant to rifampicin and fusidic acid. The tet(M), and erm(B) positive strains from the study were used as donors. Equal volumes (1 ml each) of donor and recipient strain, at exponential growth were mixed, filtered through a sterile 0.45 μm-pore-size nitrocellulose membrane filter (Millipore, USA), and placed on MRS agar plates. After incubation at 37°C for 24 h, cells were resuspended in phosphate buffer saline and serial dilutions were spread on selective medium. Following incubation for 48 h, transconjugants were selected and analyzed by antimicrobial susceptibility. The transfer of resistant gene tet(M) and erm(B) was confirmed in transconjugants by the detection of tet(M) and erm(B) by PCR. The number of transconjugants per recipient cell was given as conjugation frequencies.

The GenBank accession numbers of sequences reported in this article are mainly from HQ293024–HQ293117and HQ111074–HQ111078.

Results

Isolation and Identification of LAB

A total of 22 LAB, including, L. delbrueckii subsp. bulgaricus (n = 6), L. acidophilus (n = 6), L. rhamnosus (n = 1), L. kefiri (n = 1), and S. thermophilus (n = 8) were isolated from commercial Yogurts. 62 LAB strains belonging to species L. delbrueckii subsp. bulgaricus (n = 7), L. acidophilus (1), L. casei (n = 5), L. rhamnosus (3), L. fermentum (5), L. plantarum (7), L. brevis (6), L. salivarius (5), L. parabuchneri (12), L. alimentarius (2), L. anamalis (2), L. vaginalis (2), L. parabrevis (1), L. coryniformis (1), and S. thermophilus (n = 3) were isolated from traditional products of whey, yogurt, pickle, and Jiang shui. Species were identified by partially sequencing their 16S rRNA gene and all the sequences were submitted to GenBank. All the commercial products had the probiotic or starter strains of LAB as per label, except one yogurt sample that had L. kefiri instead of L. acidophilus as indicated on product label.

Phenotypic Profile of Antimicrobial Resistances

Table 2 indicates the number of strains that were subjected to antibiotic susceptibility, 90% MIC, range, and number of resistant strains for each species. All strains were susceptible to ampicillin, bacitracin, and cefsulodin and resistant to nalidixic acid and kanamycin. All the species except, L. bulgaricus and L. acidophilus and S. thermophilus were intrinsically resistant to vancomycin. Two strains, one each of L. salivarius and S. thermophilus species were resistant to penicillin G. High percentage of strains were resistant to gentamycin (65%) and ciprofloxacin (85%) while, a low percentage was resistant to streptomycin (7%) and chloramphenicol (12%) in species dependant manner. Moderately higher percentage of strains had acquired resistance for erythromycin (11%), clindamycin (6%), and tetracycline (17%). All erythromycin (n = 9) and tetracycline (n = 14) resistant strains originated from traditional products, except tetracycline resistant L. kefiri NWL78 and S. thermophilus NWL93, which were from two different commercially produced probiotic yogurts.

Table 2 Minimum inhibitory concentration (MIC) values of selected antibiotics against bacterial species
Full size table

Antibiotic Resistance Genes

Antibiotic resistance genes were detected by PCR from resistant strains and results are presented in Table 3. Erythromycin resistance gene erm(B) was found from eight lactobacilli strains including L. fermentum NWL24 and NWL26, L. vaginalis NWL 35 and NWL43, L. plantarum NWL22, L. acidophilus NWL23, L. salivarius NWL33, and L. animalis NWL39. One strain of S. thermophilus (NWL02), isolated from a traditionally fermented yogurt also had erm(B) gene. All the strains with atypical MIC to erythromycin were positive for erm(B), while none of the susceptible strains were positive for erm(B). Other erythromycin resistant determinants [erm(A) and mef(A)] were not detected from any strain.

Table 3 Characterization of 21 strains of LAB with acquired antibiotic resistances for erythromycin and tetracycline
Full size table

Out of 14 lactobacilli strains resistant to tetracycline, eight strains were positive for tet(M) gene only, two strains had tet(M) and tet(S) together, and two strains harbored only tet(S) gene. Two strains, resistant to tetracycline were negative for any of the tested tet resistance gene. None of the tet(W), tet(O), tet(Q), tet(K), and tet(L) genes were present in any strain. All tet gene positive strains were isolated from traditional sources except L. kefiri NWL78, which was isolated from a probiotic yogurt. One strain of each of L. plantarum (NWL16) and L. salivarius (NWL33) had erm(B) and tet(M) together. Penicillin, gentamycin, chloramphenicol, and streptomycin resistant strains were negative for any of the tested genes. Vancomycin resistance was intrinsic, and strains did not harbor acquired resistance gene van(B) (data not shown).

Transfer of Antibiotic Resistance Genes

LAB strains positive for erm(B) and tet(M) were used as donors, while E. faecalis 181 and E. facium258 were used as recipient strains in filter mating experiments. Results showed that erm(B) from L. fermentum NWL24 and L. salivarius NWL33 and tet(M) gene from L. plantarum NWL22 and L. brevis NWL59 were successfully transferred to E. faecalis. All transconjugants were resistant to erythromycin (MIC > 64) or tetracycline (MIC > 128) and positive for tet(M) or erm(B) genes. Transfer rate in filter mating experiments ranged from 2.9 × 10−6 to 1.39 × 10−5 for different combinations (Table 4).

Table 4 In vitro transfer of erm(B) and tet(M) from lactobacilli strains to E. faecalis 181
Full size table

Discussion

Since 1960s, the emergence of drug resistant pathogens and lack of newly developed antibiotics is a major concern for health professionals and researchers. Magnitude of problem increases significantly by the possibility of horizontal gene transfer, misuse, and overuse of antibiotics in human medicine, veterinary, agriculture, and food products. Food chain is considered as a major route of transmission of antibiotic resistant genes between animal and human populations [32]. LAB, a dominant part of microflora of fermented foods, can exchange genes to enhance their survival in antibiotic-containing environments and may transfer it to other commensal bacteria or pathogens in the intestine, on mucosal surfaces or in food [39, 43]. LAB with acquired resistance should be discouraged from entering into human food chain and a careful monitoring and regulation must be implied to check their antibiotic resistance profiles.

A wide spread susceptibility toward the inhibitors of cell wall synthesis, such as penicillin and ampicillin have been observed in many species of lactobacilli from different sources, such as cheese [4], probiotics or fermented foods [6], and human intestine [7, 11]. All of the lactobacilli from commercial or traditionally fermented foods in this study were sensitive to penicillin and ampicillin, except one strain of L. salivarius NWL31 (MIC > 64) from pickles that was resistant to penicillin. Resistance to penicillin has been reported in some species of lactobacilli from European probiotic products [41], from sausages [23] and cheese [19].

Susceptibility of all LAB strains to ampicillin, cefsulodin, and bacitracin, and intrinsic resistance to nalidixic acid, vancomycin, and kanamycin was in accordance with previous studies [10, 31, 32, 45]. Resistance of Lactobacillus species to vancomycin is considered as intrinsic [2], except for L. delbrueckii subsp. bulgaricus, L. acidophilus, L. johnsonii, and L. crispatus [4, 7].

Acquired resistance to erythromycin is generally because of erm(B) which has been reported previously in lactobacilli from different sources [2, 4, 5, 11, 18, 21, 25, 26, 31, 32]. Other resistant determinants [erm(A), erm(C), and erm(T)] are rarely detected from LAB, including lactobacilli and S. themophillus. Tetracycline resistance is also an acquired resistance in lactobacilli and S. thermophilus. Many tetracycline resistant determinants [tet(M), tet(O), tet(S), and tet(W) or the efflux proteins tet(K) and tet(L)] have been reported [26, 32]. The authors identified tet(M) in L. plantarum, L. salivarius, L. animalis, and L. brevis from different sources. The tet(M) has also been reported previously, in L. plantarum [17, 23] and L. salivarius from meat products [3]. The authors also identified, tet(S) gene in L. kefiri (NWL78) and L. brevis (NWL51 and 59). To the authors’ knowledge, the presence of tet(S) gene in L. kefiri or L. brevis has not been described previously; however, it has been detected in members of L. plantarum [27]. Other tetracycline resistant genes [tet(O), tet(Q), and tet(W)] have also been identified in lactobacilli such as tet(W) from a probiotic L. reuteri [29], but we did not find any of these genes in this study.

Although MIC can be considered as a “true” measure, however, it can have low reproducibility between labs. Recently, LAB susceptibility medium (LSM) [30] and microbiological breakpoints have been suggested by different researchers [10, 15, 16, 20, 31, 33]. Here, the authors propose tetracycline breakpoint for L.parabuchneri, isolated from fermented vegetables as 16 μg/ml. MIC of tetracycline for all the L. parabuchneri isolates (n = 12) from different sources, was from 16 to 32. Lactobacillus parabuchneri can be isolated frequently from fermented vegetables and pickles [9], from where resistances for different antibiotics have been described in other species of lactobacilli [36, 46]. MICs of bacitracin ranged from 0.063 to 16 in this study but none of the strain was resistant by the breakpoints given by Danielsen and Wind [10].

There are only few reports on conjugative transfer of naturally occurring antibiotic resistance determinants including erm(B) and tet(M) from lactobacilli to other bacteria in vitro [24, 35] or in gnotobiotic rats [28]. Feld et al. [18] successfully transferred an erythromycin resistant plasmid (pLFE1) in vitro and in vivo, from L. plantarum to a broad host-range including L. rhamnosus, L. lactis, E. faecalis, and L. monocytogenes. Hummel et al. [26] found erm(B) gene involved in erythromycin resistance from L. salivarius strain (BFE 7441), which was on genomic DNA. Recently, a transposons Tn916 containing tet(M) gene was transferred from L. paracasei to E. faecalis strain JH2-2 in mating experiments at low conjugation frequency [12]. Erythromycin resistance determinant Erm(B) from L. fermentum NWL24 and L. salivarius NWL33, and tet(M) gene from L. plantarum NWL22 and L. brevis NWL59 were successfully transferred to E. faecalis 181. Plasmid free recipient strains were characterized in the lab in another study. The authors also implied broth mating and filter separating techniques but conjugal transfer of erm(B) and tet(M) was not successful by these methods (data not shown), which further strengthen the notion that a close contact on a solid surface is required for better conjugal transfer in LAB [24].

The authors conclude that the prevalence of antibiotic resistant LAB in Chinese food chain is high and care should be taken whenever introducing any of these strains in functional foods intentionally. Although probiotic products and starter strains rarely had acquired antibiotic resistance, the presence of tet(S) gene in L. kefiri NWL78, isolated from a probiotic yogurt, draws the attention for a strict monitoring and regulation.