Introduction

Lactobacilli are Gram-positive bacteria of high biotechnological and natural significance. They populate nutrient-rich habitats associated with food, feed, plants, animals, and humans [15]. In latter, Lactobacillus spp. belong to the resident gut microbiota being responsible for many of its beneficial effects on human health [28]. As a result, lactobacilli are widely used as probiotics to maintain or to replenish the gut microbiota after antibiotic treatment [23].

With regard to antibiotic resistance of the lactobacilli, vancomycin and ciprofloxacin have low inhibitory effect among the majority of Lactobacillus species [23, 54]. Lactobacilli are generally susceptible to the inhibitors of protein synthesis, such as chloramphenicol, macrolides, lincosamides, and tetracycline, but their resistance to aminoglycosides is often high [1, 23, 27, 32]. Moreover, lactobacilli are usually sensitive to the cell wall-targeting β-lactams such as penicillin, but are more resistant to cephalosporins [1, 23]. Resistance to other antibiotics varies greatly among lactobacilli.

Antibiotic resistance of probiotic Lactobacillus strains is an essential property for their application to reinforce the concomitant antibiotic therapy. On the other hand, intake of bacteria with acquired antibiotic resistance poses the risk of its dissemination in the gastrointestinal microbiota and totally in the environment. Studying of antibiotic susceptibility pattern and resistance genes in Lactobacillus spp. is a comparably recent approach [30]. Up to now, the presence of several antibiotic resistance genes, both intrinsic and acquired, has been reported in Lactobacillus spp. For example, chloramphenicol-resistance cat gene has been found in L. reuteri [16] and L. plantarum [55]. Different erythromycin-resistance genes (ermB, ermA, ermC, ermT) and at least 11 tetracycline resistance genes (tetW, tetM, tetS, tetO, tetQ, tet36, tetZ, tetO/W/32/O/W/O, tetW/O, tetK, and tetL) have been detected to date in lactobacilli [38, 42, 57], among which tetM and ermB were suggested to be the most widely-distributed among Lactobacillus spp.[3, 4, 6, 14]. Some of them were found to be located on plasmids and in transposons and thus were considered acquired [14, 18, 42, 44, 47]. These findings evidence the view on the food and probiotic bacteria as reservoirs of antibiotic resistance genes and facilitate the revision of GRAS (generally recognized as safe) status of lactic acid bacteria (LAB).

However, there is still much to be explored about the problem of antibiotic resistance of lactobacilli. Assessment of antibiotic resistances among lactobacilli is confounded by the lack of standards for susceptibility testing, as well as susceptibility breakpoints for most antibiotics. Determination of antibiotic susceptibility patterns of a representative number of different strains from each species is necessary for working out of reliable criteria for the differentiation between the intrinsic and the acquired antibiotic resistance in probiotic bacteria. Although some effort has been made to this end, work has only been carried out for some antibiotics and particular Lactobacillus species, such as L. casei, L. acidophilus, L. reuteri, L. rhamnosus, L. delbrueckii, L. brevis, L. fermentum [2, 4, 20, 37, 42]. Besides, antibiotic resistance profile of Lactobacillus strains for practical application should be studied by both phenotypic and genotypic methods, because a susceptible phenotype may carry silent genes, which are detected by PCR-based molecular methods [54].

In the present study, antibiotic resistance pattern of 20 Lactobacillus strains was investigated through the disc diffusion method and microdilution method as well as molecular methods to check the presence of different antibiotic resistance genes.

Materials and Methods

Isolation of Bacteria and Growth Conditions

Lactobacilli were isolated from probiotics, commercial dairy products and fermented plant material and identified in our previous studies [5, 7]. L. brevis DSM 20,054, L. buchneri DSM 20,057, L. brevis ssp. gravesensis LMG 7934, and L. rhamnosus B-8238, obtained from Culture Collections, were used as reference strains (Table 1). The cultures preserved as glycerol stocks were activated by growing in sterile de Man, Rogosa, Sharpe (MRS) broth [12] under microaerophilic conditions at 37 °C for 24 h. The cultures thus activated were stored at 5 °C with fortnightly subculture in MRS broth. Working cultures were prepared in sterilized MRS broth using 1% inocula followed by incubation under microaerophilic conditions at 37 °C for 16–18 h.

Table 1 Bacterial strains used in this study

Antibiotic Susceptibility Testing

Antibiotic susceptibility was assessed by the disk diffusion method, as described earlier [5, 7]. In brief, all strains were diluted in 0.85% saline to obtain turbidity equivalent to McFarland scale 0.5 and aliquots were pour-plated on MRS agar plates. Antibiotic discs (Scientific Research Centre of Pharmacotherapy, Russia) were placed on the surface of inoculated plates. After 48 h incubation in anaerobic conditions (Anaerogas GasPak, NIKI MLT, Russia) at 37 °C, inhibition halos were measured in mm (means ± SD of 3 trials) and interpreted as susceptible (S), moderately susceptible (MS), or resistant (R) according to [9, 43], as indicated in Table S1.

The MIC values of cephalosporins were determined by the broth microdilution method in MRS broth in 96-well nontreated cell culture plates (Eppendorf). Cefazolin, ceftriaxone, and cefotaxime (all Sigma-Aldrich) were tested in concentration range of 0.5–1024 μg/ml obtained after a series of two-fold dilutions in MRS broth. Wells were inoculated with 200 μl of the bacterial culture (3 × 107 CFU/ml) and incubated at 37 °C. The MIC was read after 24 h of incubation as the lowest concentration of an antibiotic at which visible growth was inhibited.

Detection of Antibiotic Resistance Genes

Genomic DNA was extracted from lactobacilli cells and purified as described earlier [4]. Antibiotic resistance genes tested by PCR are indicated in Table S2. The PCR reaction was carried out in a total volume of 25 μl containing DNA template, 10 pmol of each primer (Table S2) [10, 16, 17, 21, 25, 27, 31, 40, 48, 50, 52, 59, 60], 1U Taq DNA polymerase, each of four dNTPs at a concentration of 200 μM, and PCR buffer containing Tris–HCl, KCl (NH4)2SO4, MgSO4, and Triton X-100. The amplification program was as following: initial denaturation step of 94 °C for 5 min, and then 35 cycles of 94 °C for 30 s, annealing temperature (Table S2) for 30 s, and a final extension step at 72 °C for 7 min. The obtained PCR fragments were analyzed by electrophoreses on a 1% agarose gel, stained with Midori Green DNA Stain (Nippon Genetics Europe, Germany) and visualized with UV light. The positive amplicons obtained in these PCRs were confirmed by sequencing. PCR products were purified with a GenJET Plasmid Miniprep Kit (Thermo Scientific, Lithuania) according to the instructions of the manufacturer. The purified products were sequenced with the forward and reverse primer (Evrogen JSC, Russian). The obtained primary nucleotide sequences were analyzed using the NCBI-BLAST algorithm and GenBank database.

Results

Phenotypic Resistance of Lactobacilli

In this study, 20 Lactobacillus strains were analyzed for the antibiotic susceptibility by disc diffusion method and were classified either as resistant (R), moderately susceptible (MS), or sensitive (S) based on zones of growth inhibition. All tested Lactobacillus strains were susceptible to ampicillin and chloramphenicol. Sensitivity to rifampicin and to the inhibitors of protein synthesis erythromycin, tetracycline, and clindamycin was also widely distributed among lactobacilli (Fig. 1b). Most of Lactobacillus strains were resistant to vancomycin (95% of strains), ciprofloxacin (95%) and aminoglycosides used in this work: amikacin (95%), kanamycin (100%), gentamicin (90%), streptomycin (85%). Noteworthily, that well-known probiotic strain L. planrarum 8PA3 possessed an antibiotic resistance profile not typical for Lactobacillus. For example, it revealed sensibility to vancomycin, while the vast majority of other tested lactobacilli were resistant to this antibiotic. Besides, unlike most lactobacilli, L. plantarum AG15 revealed rifampicin resistance.

Fig. 1
figure 1

Antibiotic resistance of lactobacilli. a Antibiotic resistance profiles exhibited by Lactobacillus strains. b Prevalence of the resistance to specific antibiotics among tested Lactobacillus strains

The broth microdilution method was used to characterize the resistance of lactobacilli to cephalosporins. Microbiological breakpoints categorizing Lactobacillus species as resistant to cefotaxime, ceftriaxone and cefazolin are not determined yet. Therefore, we assumed that strains with high MIC values (64 and 128 µg/ml) were more likely to be resistant while two L. plantarum isolates (strains FCa1L and S6) with MIC values 0.5 μg/mL for three tested cephalosporins were considered as sensitive to them.

Genotypic Resistance of Lactobacilli

The Lactobacillus strains were tested for the presence of antibiotic resistance genes by PCR. The results are presented in Table 2 and fig. S1. Four silage Lactobacillus isolates were positive for genes encoding erythromycin-resistance ermB and mefA. The tetK gene was detected in L. buchneri DSM 20057 and tetL was found in L. plantarum FCa1L and in two L. plantarum silage isolates S1 and AG15 (Fig. S1, b). Detection of tetK gene was confirmed by the results of NCBI BLAST search for the homologs of tetracycline-resistance protein of Staphylococcus aureus PM1 (YP_006958133.1) in the genome sequence of L. buchneri DSM 20057. As a result, we identified nine secondary transporters which belong to the major facilitator superfamily (MFS) proteins and facilitate the transport across bacterial membranes. Among them, the EmrB QacA subfamily drug resistance transporter (KRK67099.1) gave the highest query cover (99%) and shared 41% similarity and 24% identity with the typical tetracycline efflux protein TetK.

Table 2 Antibiotic resistance of Lactobacillus strains and detection of antibiotic resistance genes by polymerase chain reaction (PCR)

Other erythromycin and tetracycline-resistance determinants (ermA, mefE, tetM) were not detected in any strain.

The gene parC associated with resistance to ciprofloxacin was detected in L. plantarum FCa3L, L. brevis DSM 20,054, L. brevis ssp. gravesensis LMG 7934, and L. buchneri DSM 20,057. No PCR products were obtained for another ciprofloxacin resistance gene gyrA. In addition, the PCR analysis showed that none of tested Lactobacillus strains possessed aac(6′)-aph(2″), ant(6), aph(3′)-III, and ant(2″)-I genes, which encode enzymatic inactivation of aminoglycosides. Yet, streptomycin resistance genes aadA and aadE were found in L. rhamnosus B-8238, L. plantarum FCa3L, L. plantarum AG1, and L. plantarum AG10.

The vancomycin resistance gene vanX was detected in 15 Lactobacillus strains (Table 2, Fig. S1, c), while other genes of this cluster vanA and vanE were not revealed in any strain.

Sixteen tested Lactobacillus strains gave a 0.5 kbp band, presumably corresponding to blaTEM gene of cephalosporins resistance (Fig S1, a). The amplicons were sequenced and a 99% similarity with the blaTEM gene of Acinetobacter baumannii (GenBank Accession No. MK764360.1) and Escherichia sp. (GenBank Accession No. NG050218.1) was revealed by NCBI BLAST algorithm. The genes blaOXA-1 and blaSHV also related to cephalosporins resistance were less frequent in tested Lactobacillus strains and were detected in 3 and 4 strains, correspondingly.

The chloramphenicol resistance gene cat as well as the mecA gene were not detected in any strain.

Discussion

In the present study, we characterized phenotypic and genotypic antibiotic resistance profiles of 20 Lactobacillus strains, including 11 strains newly isolated from fermented plant material. In addition, four reference strains and five probiotic and dairy isolates were included in this study (Table 1). Irrespective of the origin, all the strains exhibited phenotypic resistance to a number of antibiotics revealing multidrug resistance pattern (Fig. 1, a). Moreover, all the strains carried at least one gene for antibiotic resistance (Tables 2, 3).

Table 3 Characterization of cephalosporin resistance of Lactobacillus strains

It is well known that lactobacilli are generally susceptible to antibiotics inhibiting nucleic acid synthesis (except for ciprofloxacin) and protein synthesis (except for aminoglycosides), and to cell wall synthesis inhibitors (except for vancomycin) [1, 11, 27, 34]. Indeed, in our study, 95% of tested Lactobacillus strains were susceptible to tetracycline, erythromycin, and rifampicin, and 90% of strains demonstrated sensitivity to clindamycin (Fig. 1b). Ampicillin and chloramphenicol were found to inhibit all bacteria used in this study (Fig. 1b). The latter notion coincided with the genomic program of tested lactobacilli. The chloramphenicol acetyl transferase gene cat which is associated with resistance to chloramphenicol was not detected in any of the strains. Gene mecA which encodes penicillin-binding protein 2A and confers resistance to penicillin-like antibiotics also was not found in tested Lactobacillus strains.

Tetracycline or erythromycin-resistant lactobacilli have been encountered in previous studies [4, 11, 21, 23, 42, 56]. Genes encoding these two resistances are often located on mobile genetic elements such as conjugative plasmids and transposons [3, 14, 18, 44, 47]. Therefore, detection of resistances to tetracycline and erythromycin in bacterial strains for food and agricultural applications always constitutes the risk of spread of antibiotic resistance genes in the environment. The most common determinants for resistance to tetracycline found in lactobacilli are genes tet (K, M, O, Q, S, W, 36), sometimes also present in combination [3, 23]. Among erythromycin-resistance genes, the ermB gene is the most frequently found among Lactobacillus spp.[42, 57].

In the present study, using PCR, we detected tetracycline efflux genes tetK (in L. buchneri DSM 20,057) and tetL (in L. plantarum FCa1L, L. plantarum S1, and L. plantarum AG15). The ermB gene, encoding 23S ribosomal rRNA methyltransferase, was found in three silage isolates, and macrolide efflux gene mefA was detected in another silage isolate L. plantarum AG15. All the strains which contained erythromycin and tetracycline-resistance genes displayed phenotypic susceptibility to these antibiotics. Similarly, in our previous investigation L. fermentum 5–1 sensitive to tetracycline was discovered to carry silent genes tetK and tetM, and L. fermentum 3–4 sensitive to erythromycin was positive for ermC [4]. These discrepancies between the resistance phenotype and genotype may be due to defective expression of resistance genes and have been described earlier by [41].

It is well documented that lactobacilli are usually resistant to aminoglycosides [23, 60]. This resistance is considered intrinsic and originates from the low impermeability of lactobacillar cell surface for aminoglycosides [9, 32, 34]. Yet, resistance genes for enzymatic modification of aminoglycosides such as aac(6′)-aph(2″), ant(6) and aph(3′)-III have been reported in several Lactobacillus spp.[50]. The vast majority of tested strains revealed phenotypic resistance to streptomycin, kanamycin, gentamycin, and amikacin (Table 2). The aminoglycoside adenyltransferase genes aadA and aadE which confer resistance to streptomycin were identified in four strains (Table 2, Fig S1, b), but other aminoglycoside-resistance genes (aac(6′)-aph(2″), aph(3′)-III, ant(6), ant(2″)-I) were not detected by PCR analysis in any of the tested lactobacilli. Thus, according to our results, among Lactobacillus spp. widely distributed resistance to aminoglycosides is likely to be intrinsic, though enzymatic inactivation of streptomycin is possible in some strains.

Lactobacilli have high natural resistance to vancomycin and ciprofloxacin. However, susceptibility to these antibiotics was shown to be species-dependent and varied several folds between species [11]. Hence, some resistant strains may harbor spontaneous mutations or acquired genes. In this work, we studied resistances to vancomycin and ciprofloxacin to find out their genetic determinants and assess their potential transferability.

We demonstrated that all tested Lactobacillus strains except for L. plantarum 8PA3 were resistant to vancomycin (Table 2). The vancomycin resistance has been reported to be intrinsic, chromosomally encoded and not inducible or transferable in lactobacilli [35, 51]. The most studied mechanism of vancomycin resistance in Lactobacillus spp. includes the replacement of the terminal D-alanine residue by d-lactate or d-serine in muramyl-pentapeptide molecule [13, 54]. Vancomycin has low-affinity binding to such altered peptidoglycan termini, and thus lactobacilli are generally resistant to vancomycin. The d-alanyl-d-alanine dipeptidase, a product of vanX gene, is critical for vancomycin resistance, because it prevents synthesis of the usual D-alanyl-D-alanine termini of the peptidoglycan precursor side chain. Homologs of vanX were found in genomes of sequenced Lactobacillus spp. (e.g. KRK67761.1 in L. buchneri DSM 20,057, ERK42887.1 in L. brevis DSM 20,054, ARW34669.1 in L. plantarum SRCM102022, according to NCBI GenBank database). Besides, in two strains of L. plantarum (LP1, LP2), the vanX gene was detected after sequencing and alignment [40]. In this study, PCR analysis revealed vanX gene in 15 tested Lactobacillus strains (Table 2, fig. S1, c). Guo et al. [24] also showed that vanX gene was frequent in lactobacilli. Among the genes of the vancomycin resistance cluster, only vanA gene is considered transferable via conjugation within the plasmid DNA [56, 58] or the conjugative transposon [26, 56]. In our work, no PCR products were amplified with vanA and vanE primer sets. Therefore. we conclude that tested Lactobacillus strains did not carry these resistance genes.

According to our results, all tested Lactobacillus strains showed resistance to ciprofloxacin, except for L. rhamnosus B-8238 which was moderately susceptible. Frequently encountered within the genus Lactobacillus resistance to ciprofloxacin has been earlier described by [11, 32, 39]. It is considered to arise from intrinsic characteristics, such as cell wall impermeability or efflux mechanism [27]. However, other mechanisms can be involved in the development of resistance to fluoroquinolones in Gram-positive bacteria. Some are the consequence of mutations involving genes encoding DNA gyrase and topoisomerase IV, essential type II topoisomerases necessary for DNA replication, chromosome segregation and DNA compaction in the cell [36, 49]. Here, we tested mutations in the quinolone resistance-determining region (QRDR) of the parC (topoisomerase IV) and gyrA (DNA gyrase) genes using PCR and consequent DNA sequencing of the obtained amplicons. The PCR fragments for the parC gene were obtained in four Lactobacillus strains and none of the strains possessed gyrA gene. Our data partly corroborate the results of [20, 27], which also demonstrated absence of typical mutations in the QRDR of gyrA or parC genes for ciprofloxacin resistance in Lactobacillus spp.

Regarding beta-lactams, lactobacilli are generally susceptible to penicillins, but more resistant or variable to cephalosporins [1, 53]. With few exceptions, Lactobacillus strains showed high MIC values of cefotaxime, ceftriaxone, and cefazolin (Table 3), as previously reported by [1, 22, 29]. Notably, among tested Lactobacillus strain sensitivity to cefotaxime was more frequent rather than to two other cephalosporins (Table 3). Indeed, cefotaxime has been shown to be the most active type I β-lactamase inhibitor, in comparison to the other β-lactam antibiotics [19]. The understanding of the mechanisms underlying resistance to cephalosporins is still very limited in lactobacilli. Although multiple beta-lactamases can be identified in the available genomic sequences of Lactobacillus spp., the presence of extended spectrum beta-lactamase (ESBL) in Gram-positive lactic acid bacteria remains obscure [33]. The major ESBL enzymes are TEM, SHV, CTX-M KPC, VIM, IMP, NDM-1, and OXA [8, 46]. Using PCR amplification and subsequent sequencing we identified the blaTEM gene in 80% of tested Lactobacillus trains, whereas blaSHV and blaOXA-1 genes were less frequent. To our knowledge, this is the first data on the detection of ESBL in lactobacilli. Although it is believed, that SHV enzymes confer much higher resistance than do TEM enzymes [45], according to our results TEM enzyme was the most frequent lactamase responsible for resistance to cephalosporins in lactobacilli.

Conclusions

Phenotypic multidrug resistance was revealed in all tested Lactobacillus strains. Studying of corresponding resistance genes showed that blaTEM (80% of strains) and vanX (75%) were the most frequently identified, and the occurrence of ermB (15%), mefA (5%), tetK (5%), tetL (15%), parC (20%), aadA (5%), aadE (15%), blaSHV (20%), and blaOXA-1 (15%) was less frequent. To our knowledge, genes for ESBLs were found in lactobacilli for the first time. Consideration should be given to the potentially transferable resistance determinants such as ermB, tetK, and tetL which were found in this work, fueling the debate about the safe use of lactobacilli in food and their potential to spread resistance in the environment. Future studies should be focused on horizontal transfer of detected resistance genes to other species.