To identify the resistance determinants of tetracycline-resistant strains, resaerchers analyzed the association between phenotypically resistant strains and their genotypes (
Table 1). Among the three tetracycline-resistant strains of
L. paracasei, the strain FCQHC12L3 was characterized by the presence of tet(M). However, no tetracycline resistance gene was found in the remaining two phenotypically resistant strains. Among the 17 tetracycline-resistant strains of
L. plantarum, six were characterized by the presence of tet(M), while QHLJZD13-L6 was characterized by the presence of tet(S), and no related resistance genes were detected in 10 strains. Among the 22 tetracycline-resistant strains of
L. reuteri, two strains were characterized by the presence of tet (M) and tet(L), whereas two strains were characterized by the presence of tet (W/N/W) and tet(L). Three strains and eleven strains harbored only tet(M) or tet(W/N/W), respectively. Furthermore, strain FYNLJ83L8 was characterized by the presence of tet(45), and three strains were not associated with resistance genes. Across the 15 tetracycline-resistant strains of
L. johnsonii, the gene tet(W/N/W) was found in 11 resistant strains. In addition, FHNXY70M2 was characterized by the presence of tet(W/N/W) and tet (L), and the remaining three strains did not display resistance-related genes. Of the 18 resistant strains of
L. crispatus, tet (M) was found in two resistant strains, and tet(W/N/W) was found in four resistant strains. Ten strains were characterized by the presence of tet(W/N/W) and tet(L), and the strain FHNXY70M14 was characterized by the presence of tet(M), tet(W/N/W), and tet(L). However, no resistance-related genes were detected in strain FHNXY56M7. No tetracycline resistance genes were found in any of the 32 tetracycline-resistant strains of
L. (para)gasseri and one tetracycline-resistant strain of
L. rhamnosus. In particular, two strains (
L. plantarum RS41-7 and
L. gasseri FHNFQ34_L1) harbored a tet(M) gene but were sensitive to tetracycline. Furthermore, the tet(M) sequence of the two strains had obvious deletions in the functional sites, resulting in the loss of resistance function (
Figure 2). Therefore, in this study, we did not classify these two strains as carriers of the gene tet(M). In brief, through genotypic and phenotypic association analysis, these five tetracycline resistance genes could explain the resistance phenotypes of 33% (1/3) of
L. paracasei, 41% (7/17) of
L. plantarum, 80% (12/15) of
L. johnsonii, 94% (17/18) of
L. crispatus, and 86% (19/22) of
L. reuteri strains. However, resistance phenotypes of 48% (52/108) of the resistant strains could not be explained based on their genotypes (
Figure 3).
Figure 2. Multi-sequence alignment of tet(M) sequences of Lactiplantibacillus plantarum RS41-7 and Lactobacillus gasseri FHNFQ34_L1 (the bottom two sequences) with the tet(M) sequences of other strains in this study.
Figure 3. MIC distribution of 478 lactic acid bacteria strains with or without the tetracycline resistance gene. Blue: strains without resistance gene. Pink: strains with the tetracycline resistance gene. Black dotted line: cutoff value established by EFSA. Red dotted line: the new cutoff value established by this work.
Table 1. Association between tetracycline-resistant strains and tetracycline resistance genes in eight lactic acid bacterial species.
Tetracycline Resistance Genes |
Number of Tetracycline-Resistant Strains |
tet(M) |
3. Definition of New Susceptibility–Resistance Cutoff Values
The results of genotype-phenotype association analysis revealed that the genetic basis for the resistance of 48% (52/108) of the strains with tetracycline resistance phenotype could not be determined, including that of two strains of
L. paracasei, one strain of
L. rhamnosus, 10 strains of
L. plantarum, one strain of
L. crispatus, three strains of
L. reuteri, three strains of
L. johnsonii, and 32 strains of
L. (para)gasseri. Lactobacillus (para)gasseri has a high drug resistance rate; however, the resistance determinants of all phenotypically resistant strains could not be identified. The cutoff value based on the fermentation type may not be applicable to all species of LAB. Thus, it is recommended to develop MCOFFs at the species level.
Lactobacillus (para)gasseri,
L. crispatus, and
L. johnsonii did not have species-specific cutoff values; therefore, we statistically analyzed the MIC frequency distribution of these species of
Lactobacillus to establish species-specific TMCOFFs, which could better distinguish the resistant strains from the sensitive strains without acquired ARGs.
Lacticaseibacillus paracasei,
L. rhamnosus,
L. plantarum, and
L. reuteri had MCOFFs at the species level; however, 10% of
L. plantarum had resistance phenotype but no resistance determinants. Therefore,researchers reformulated the cutoff value of
L. plantarum to determine whether the strains containing resistance genes could be better distinguished from sensitive strains. Based on the MIC distribution of tetracycline of
L.
(para)gasseri,
L. crispatus,
L. johnsonii, and
L. plantarum in this study, two different statistical approaches (Turnidge and Kronvall) and a “visual estimation” approach (eyeball method) were used to determine the new susceptibility–resistance cutoff values (
Table 2).
Table 2. Comparison of tentative microbiological cutoff values (TMCOFFs) for tetracycline calculated using two statistical methods and the eyeball method.
|
|
TMCOFFs Obtained Using the Indicated Method (%) a |
L. paracasei | (1), L. plantarum (6), L. reuteri (3), L.crispatus (2) |
tet(W/N/W) |
L. reuteri (11), L.johnsonii (11), L.crispatus (4) |
tet(S) |
L. plantarum (1) |
tet(45) |
L. reuteri (1) |
tet(M) and tet(L) |
L. reuteri (2), L.crispatus (10) |
Based on the new cutoff values, all
L. (para)gasseri strains were classified as sensitive, and the strains of
L. crispatus containing resistance genes were distinguished from the sensitive strains without acquired ARGs. One strain belonging to
L. johnsonii was not associated with the tetracycline resistance gene and was classified as phenotypically resistant. However, the MIC for this strain was found to be equivalent to that for another
L. johnsonii strain containing the tetracycline resistance gene. Accordingly, it contained potential tetracycline resistance genes. Therefore, the new cutoff value could completely distinguish the strains containing resistance genes from sensitive strains in
L. johnsonii. The cutoff value of tetracycline for
L. plantarum was newly formulated as 64 μg/mL, which is the same as that formulated by Flórez et al.
[17]. However, the resistance determinants of six strains of
L. plantarum still need to be further explored.
4. Prevalence and Distribution of Tetracycline Resistance Genes in LAB
To explore the distribution and prevalence of these five tetracycline resistance genes in LAB, this study performed statistical analysis on the detection of tetracycline resistance genes in eight species of LAB (
Table 3). The most widely distributed tetracycline resistance gene in LAB was tet(M), which was detected in four LAB species, including
L. paracasei,
L. plantarum,
L. reuteri, and
L. crispatus. The genes tet(W/N/W) and tet(L) were detected in three species of LAB, including
L. reuteri,
L. johnsonii, and
L. crispatus. The tet(S) gene was detected in only one strain of
L. plantarum, and the tet (45) gene was found in only one strain of
L. reuteri. The tet(W/N/W) gene was the most frequently detected tetracycline resistance gene in LAB; it was detected in 30 strains of LAB in this study. The genes tet(M), tet(L), tet(S), and tet(45) had detection frequencies of 26, 16, 1, and 1, respectively. Most types of the tetracycline resistance genes were detected in
L. reuteri, including tet(M), tet(W/N/W), tet(L), and tet(45). Three of the tetracycline resistance genes were detected in
L. crispatus, namely tet(M), tet(W/N/W), and tet(L). Two genes, tet(W/N/W) and tet(L), were detected in
L. johnsonii. Two tetracycline resistance genes, tet(M) and tet(S), were found in
L. plantarum. Among
L. paracasei strains, only one harbored the tet(M) gene. No tetracycline resistance genes were found in any of the
L. rhamnosus and
L. (para)gasseri strains in this study. Herein, the tetracycline resistance gene was detected in 12% (56/478) of the strains. A total of 26 LAB strains contained only the tet(W/N/W) gene, while 12 LAB strains contained only the tet(M) gene. Interestingly, the gene tet(L) did not appear alone in LAB but was always detected together with other tetracycline resistance genes. The genes tet(M) and tet(L) were detected together in 12 strains of LAB. The genes tet(W/N/W) and tet(L) were detected concurrently in three strains of LAB. Three tetracycline resistance genes, tet(M), tet(W/N/W), and tet(L), were detected simultaneously in one
L. crispatus strain. One strain of
L. plantarum contained only the tet(S) gene, and one
L. reuteri strain contained only the tet(45) gene.
Table 3. The detailed distribution of the detected tetracycline resistance genes in different lactic acid bacterial species.
Species |
Total Strain Number |
TETR |
tet(M) |
tet(W/N/W) |
tet(L) |
tet(S) |
tet(45) |
Species |
EFSA Cut Off |
Method of Turnidge et al. b |
Method of Kronvall |
Eyeball Method |
Median for the Method |
L. (para)gasseri |
4 (68%) |
16 (100%) |
256 (100%) |
16 (100%) |
16 (100%) |
L. paracasei |
116 |
3 |
1 |
0 |
L. johnsonii |
4 (17%) |
32 (38%) |
16 (38%) |
16 (38%) |
16 (38%) |
0 |
0 |
0 |
L. rhamnosus |
68 |
1 |
0 |
0 |
0 |
0 |
0 |
L. crispatus5 |
13 |
4 |
0 |
1 |
L. johnsonii |
18 |
tet(W/N/W) and tet(L) |
L. reuteri (2), L.johnsonii (1) |
4 (40%) |
8 (50%) |
16 (50%) |
16 (50%) |
16 (50%) |
L. plantarum |
99 |
17 |
6 |
0 |
0 |
1 |
0 |
L. plantarum |
15 |
0 |
tet(M), tet(W/N/W), and tet(L) |
L.crispatus (1) |
32 (83%) |
64 (87%) |
64 (87%) |
64 (87%) |
64 (87%) |
L. reuteri |
47 |
22 |
12 |
1 |
0 |
0 |
L. crispatus |
30 |
18 |
13 |
5 |
11 |
0 |
0 |
L.(para)gasseri |
100 |
32 |
0 |
0 |
0 |
0 |
0 |
No tetracycline resistance genes |
L. paracasei (2), L. rhamnosus (1), L. plantarum (10), L. reuteri (3), L.johnsonii (3), L.crispatus (1), L. (para)gasseri (32) |
Total |
478 |
108 |
25 |
30 |
16 |
1 |
1 |
The genes of tet(M) and tet(W/N/W) are the most widely distributed tetracycline resistance genes in LAB. In order to further explore their phylogenetic relationship, multiple sequence alignments of a total of 25 tet(M) and 30 tet(W/N/W) gene sequences identified in this paper together with the same genotype sequences retrieved from the NCBI database were performed by ClustalW, and the phylogenetic tree was constructed using the Neighbor-Joining method by MEGA X (
Figures S1 and S2). Most of the same species were in the same branch, suggesting that the genes had adaptive mutations when transferred to different species, and the tet(M) and tet(W/N/W) genes of LAB of the same species may have come from the same host.
5. Conclusions
Based on the MIC distribution data of 478 strains of LAB, LAB showed moderate resistance to tetracycline. Further, formulating the breakpoint value at the species level was found to be necessary. Therefore, the species-specific microbiological cutoff values for
L. (para)gasseri,
L. crispatus, and
L. johnsonii against tetracycline were formulated, and new susceptibility-resistance cutoff values for
L. plantarum were defined. The genes tet(M), tet(W/N/W), tet(L), tet(S), and tet(45) were the key resistance genes for the tetracycline resistance phenotype and were found to widely exist in LAB. The determination of antibiotic resistance in probiotic strains is related to food safety issues. The findings of this study provide certain guiding significance and reference values at the phenotype and genotype levels for the safe application of LAB in the food industry and the formulation of probiotic resistance evaluation standards.