During the routine survey for the presence of erm(T) in enterocci of swine origin, a total of 159 non-duplicate enterococci isolates with erythromycin MICs of no <8 mg/L were collected in July and September 2018 from anal swabs of healthy pigs at two farms in Henan Province, China.

Antimicrobial susceptibility testing (AST) was carried out by broth microdilution according to recommendations given in document M100 (Twenty-Eighth Edition) issued by CLSI (21). S. aureus ATCC 29213 served as the quality control strain.

E. faecalis OG1RF, S. aureus RN4220, Streptococcus suis P1/7, and Streptococcus mutans UA159 served as recipients in conjugation and transformation experiments.

The aforementioned erythromycin-resistant enterococci strains were detected for the presence of macrolide resistance gene erm(T) and other resistance genes by PCR using the primers listed in Table 1 (15, 23–26). PCR products for erm(T) in E. faecalis E165 and its transconjugants and transformants were subjected to Sanger sequencing.

Whole genome DNA of E. faecalis E165 was sequenced by the PacBio RS and Illumina MiSeq platforms (Shanghai Personal Biotechnology Co., Ltd, China). The PacBio sequence reads were assembled with HGAP4 and CANU (Version 1.6), and corrected by Illumina MiSeq with pilon (Version 1.22). The prediction of ORFs and their annotation was performed using Glimmer 3.0.

To investigate whether the erm(T)-carrying plasmid could be transferred into bacteria of the same and other species, plasmid DNA extracted from E. faecalis E165 was obtained by using Qiagen Mini-prep kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol.

To determine the intraspecies transmissibility of the erm(T)-carrying plasmid pE165, conjugation experiments were performed using E. faecalis E165 as the donor and E. faecalis OG1RF as the recipient as previously described (27). Transconjugants were screened on brain heart infusion (BHI, Oxoid, British) agar supplemented with 32 mg/L rifampicin, 32 mg/L fucidin acid, and 8 mg/L erythromycin. The corresponding transconjugant was designated E. faecalis OG1RF-Tc.

Natural transformation experiment using the recipient S. suis P1/7 was performed as described to investigate the interspecies transmissibility of erm(T)-carrying plasmid pE165 (28). The peptide (GNWGTWVEE) was dissolved in Milli-Q water at a final concentration of 5 mM and was used as a pheromone for the transformation. The erythromycin-susceptible recipient strain S. suis P1/7 was grown overnight in THY broth (3 g Todd-Hewitt Broth and 2 g yeast for 100 ml, Oxoid, British) at 37°C under 5% CO₂. The overnight culture was diluted 1:40 into pre-warmed THY broth, and grown at 37°C under 5% CO₂ without shaking. Plasmid DNA (1.2 μg) and stock peptide (5 μl) were added when the recipient culture reached an OD₆₀₀ between 0.035 and 0.058, and then incubated for 2h at 37°C under 5% CO₂. The samples were diluted and plated on THA (Todd-Hewitt Broth with 1.3% agar powder, Oxoid, British) containing 8 mg/L erythromycin. The corresponding transformant was designated S. suis P1/7-Tm.

Natural transformation experiment with the recipient S. mutans UA159 was performed following procedures as previously described (29). For the screening of the transformants, THA was supplemented with 8 mg/L erythromycin. The corresponding transformant was designated S. mutans UA159-Tm.

The electrotransformation experiment with recipient strain S. aureus RN4220 was performed as described in a previous study (30). The corresponding transformant was designated S. aureus RN4220-Tm.

All colonies from the selective plates after incubation for 24–48 h at 37°C were further confirmed by erm(A), erm(B), and erm(T) gene detection (15, 23), 16S rRNA sequencing, AST, and multilocus sequence typing (MLST) following the harmonized protocols (http://www.mlst.net/). The plasmids obtained from E. faecalis E165, E. faecalis OG1RF-Tc, S. suis P1/7-Tm, S. mutans UA159-Tm, and S. aureus RN4220-Tm were digested by Sac I and Pac I (New England Biolabs, Inc., USA), and then southern bolt using erm(T) gene as the probe was performed.

Growth kinetics were determined for E. faecalis OG1RF and E. faecalis OG1RF-Tc, S. aureus RN4220 and S. aureus RN4220-Tm, S. suis P1/7 and S. suis P1/7-Tm, and S. mutans UA159 and S. mutans UA159-Tm (19, 31). Volumes of 30 ml BHI broth were inoculated independently with 10⁷ CFU of OG1RF, OG1RF-Tc, RN4220, and RN4220-Tm; cultures were grown for 12 h at 200 rpm and 37°C. Volumes of 30 ml Todd-Hewitt Broth (THB, Oxoid, British) supplemented with 5% fetal calf serum were inoculated independently with 10⁷ CFU of P1/7, P1/7-Tm, UA159, and UA159-Tm; cultures were grown for 12 h at 200 rpm and 37°C. The absorbance at 600 nm was measured every hour.

The fitness cost of pE165 was determined between E. faecalis OG1RF and E. faecalis OG1RF-Tc, S. aureus RN4220 and S. aureus RN4220-Tm, S. suis P1/7 and S. suis P1/7-Tm, and S. mutans UA159 and S. mutans UA159-Tm, as previously described (31, 32), with the following modifications. OG1RF and OG1RF-Tc, RN4220, and RN4220-Tm, were cultured in BHI broth for 24h at 37°C and 200 rpm. P1/7 and P1/7-Tm, UA159 and UA159-Tm, were cultured in THB (supplemented with 5% fetal calf serum) for 24 h at 37°C and 200 rpm. Then 1 × 10⁸ CFU of recipient strain was mixed with 1 × 10⁸ CFU of corresponding transconjugant/ transformant in 30 ml antibiotic-free BHI broth/THB (supplemented with 5% fetal calf serum). The mixtures were grown at 37°C and 200 rpm and diluted at 1:100 to fresh BHI broth/THB (supplemented with 5% fetal calf serum) every 24 h. For each sample, aliquots were plated onto non-selective and erythromycin-containing BHI agar/THA (supplemented with 5% fetal calf serum) plates. The proportion of pE165-carrying strains was calculated by the number of colonies on the selective plate divided by the number of colonies on the non-selective plate.

All erythromycin-resistant strains were investigated for the presence of the macrolide resistance genes by PCR. Of the 159 erythromycin-resistant enterococci isolates, 24(15.1%) contained solely erm(A), 33 (20.8%) contained solely erm(B), and 102 (64.2%) were positive for both erm(A) and erm(B). No strain was erm(C)/erm(F)/erm(G)/erm(X)-positive. A single E. faecalis strain E165 was positive for erm(A), erm(B), and erm(T). This is the first description of the erm(T) gene in E. faecalis.

Whole genome sequencing, assembly, and analysis for E. faecalis E165 showed a 4,244-bp small plasmid (designated pE165) harbored the erm(T) gene, with an average GC content of 33.0%. A total of three open reading frames (ORFs) encoding proteins of >100 amino acids were identified. The erm(T) gene coded for a 244-amino-acid (aa) protein identical to erm(T) of pKKS25 in S. aureus 25 (CAY48681.1) (18), and was highly similar (99.2% amino acid identity, 99.7% DNA sequence identity) to that of pGT633 from Lactobacillus reuteri 100-63 (NG_047838) (Figure 1) (32). The plasmid mobilization protein encoded by the mob gene from pE165 showed a high level of homology (identity≥90.6%) to that encoded by mob genes from Lactococcus garvieae (WP_207144600), E. faecium (HAR1670775.1), S. suis (WP_105139626), S. aureus (CCQ43999), and Escherichia coli (EFG1049274). The replication protein encoded by the rep gene from pE165 exhibited ≥99.5% identities to that encoded by rep genes from E. faecalis (HBI2052878), Listeria monocytogenes (HAB0665403), S. suis (NQK16007), S. agalactiae (WP_228308086), E. faecium (HAZ0989061), Bacillus paranthracis (AHN52261), E. coli (EFG1049261), Lactimicrobium massiliense (WP_108775117), and Amylolactobacillus amylophilus (WP_054746480), which revealed how widespread the erm(T)-carrying pE165-like plasmid is.

In addition, erm(A) gene carried by Tn6674 was located on the chromosome. Three copies of erm(B) gene were located on a 65,052 bp conjugative plasmid (designated pE165-2) and the conjugative region from pE165-2 exhibits 99% DNA identity to pL15 described in an E. faecalis isolated from swine in Brazil (CP042214). pE165-2 also includes tet(M) and tet(L) conferring resistance to tetracyclines, dfrG conferring resistance to trimethoprim, aacA-aphD conferring resistance to aminoglycosides, and cat encoding chloramphenicol acetyltransferase. A 2,836 bp small plasmid that did not carry any resistance gene was also detected in E. faecalis E165 (designated pE165-3).

Previous studies identified a complete translational attenuator immediately upstream of the erm(T) gene which consisted of two pairs of inverted-repeat sequences of 12 bp each and a reading frame for a regulatory peptide of 19 aa (15, 17, 33). Inducible erm gene expression often required an intact translational attenuator, while deletions or duplications that appeared in the regulatory region would cause constitutive erm gene expression (34). A comparison of the erm(T) regulatory region of pE165 with that of plasmids pRW35 (EU192194) revealed that the erm(T) regulatory region of pE165 had four bp point mutations and one bp deletion in the regulatory peptide ORF. This single nucleotide deletion resulted in a frame shift mutation, which extended the reading frame for the regulatory peptide from 19 aa to 22 aa (Figure 2). In addition, erm(T) regulatory region of pE165 was compared to those of pSC262 and pUR2940 with mutations in previous reports (22, 35), and the results are shown in Supplementary Figures S1, S2.

Intraspecies transmissibility of pE165 was investigated by conjugation experiments using E. faecalis E165 as the donor and E. faecalis OG1RF as the recipient. The transconjugant OG1RF-Tc was successfully obtained on the selective plates with the transfer frequencies of 1.5 × 10⁻⁵. PCR assay was used for the detection of erm(T), erm(A), and erm(B) in transconjugant, and the results revealed that only erm(T) gene was transferred into the recipient. Minimum inhibition concentrations (MICs) of E. faecalis E165, OG1RF, and OG1RF-Tc were determined and are shown in Table 2. Compared to E. faecalis OG1RF, OG1RF-Tc displayed a higher erythromycin MIC (>512 mg/L) and higher clindamycin MIC (256 mg/L).

Interspecies transmissibility of pE165 was investigated by natural transformation using S. suis P1/7 and S. mutans UA159 as the recipients and electrotransformation using S. aureus RN4220 as the recipient. Transformants P1/7-Tm, UA159-Tm, and RN4220-Tm were successfully obtained, and the transfer frequencies were 0.63 × 10² μg⁻¹,2.1 × 10² μg⁻¹ and 4.9 × 10⁴ μg⁻¹. The transformants were confirmed by AST and sequencing of 16S rRNA. PCR assay also revealed that only erm(T) could be detected in these transformants. MICs of E. faecalis E165, S. suis P1/7, S. mutans UA159, S. aureus RN4220, and their transformants are shown in Table 2. S. suis transformant P1/7-Tm displayed a higher erythromycin MIC (>512 mg/L) and a higher clindamycin MIC (32 mg/L) compared with S. suis P1/7. S. mutans transformant UA159-Tm displayed higher erythromycin MIC (>512 mg/L) and clindamycin MIC (256 mg/L) compared with S. mutans UA159. S. aureus transformant RN4220-Tm displayed higher erythromycin MIC (>512 mg/L) and higher clindamycin MIC (128 mg/L) compared with S. aureus RN4220.

The physical map of pE165 and restriction enzyme-digested plasmid profiles of the plasmids from E. faecalis E165, transconjugant E. faecalis OG1RF-Tc, transformants S. suis P1/7-Tm, S. mutans UA159-Tm, and S. aureus RN4220-Tm are shown in Figure 3. The results indicated that only pE165 can transfer into the recipient strains and pE165 can replicate in heterogenous hosts. The result of the southern bolt revealed that erm(T) gene was located on pE165 in heterogenous hosts. The results of AST indicated that pE165 can constitutively express erythromycin- and clindamycin- resistance phenotype in heterogenous hosts.

The growth curve of E. faecalis E165, S. suis P1/7, S. mutans UA159, S. aureus RN4220, and their transconjugants and transformants in the absence of erythromycin are shown in Figure 4. The results showed that no significant fitness burden for pE165-carrying transconjugants and transformants was observed compared with the recipient strains in the absence of selective pressure.

Competition experiments can offer a more discriminative and precise measurement of fitness, and the competitive disadvantage of the fitness burden caused by pE165 can be reflected during all the phases of the growth cycle and in successive cycles. During the competition experiment between E. faecalis OG1RF and OG1RF-Tc, from day 1 on, a successive decrease in the proportion of E. faecalis OG1RF-Tc was observed, and all the colonies tested were pE165 free on day 14 (Figure 5A). From day 1 on, a fast and constant decrease in the proportion of S. suis P1/7-Tm was observed and all the strains were tested pE165 free on day 6 (Figure 5B). In the process of competition experiment between S. mutans UA159 and UA159-Tm, an obvious decrease in the proportion of UA159-Tm was observed. On day 9, all the colonies tested were pE165 free (Figure 5B). For the result of the competition experiment between S. aureus RN4220 and RN4220-Tm, it had an obvious decrease from day 3 on, and RN4220-Tm could not be detected on day 15 (Figure 5A). The above results suggested that all the pE165-carrying transconjugants and transformants had a fitness cost compared to the recipient strains without pE165, but the fitness cost among the different transconjugants and transformants differed.