Detection of poxtA2, a Presumptive poxtA Ancestor, in a Plasmid from a Linezolid-Resistant Enterococcus gallinarum Isolate

LETTER

The poxtA gene encodes a protein belonging to the F lineage of the ATP-binding cassette superfamily, which can protect the bacterial ribosome from some antiribosomal antibiotics, including oxazolidinones (1). After the first identification in a methicillin-resistant Staphylococcus aureus of clinical origin (1), several reports of poxtA-positive isolates of Enterococcus spp. of animal and human origin from different countries have documented the broad and intersectoral dissemination of this transferable resistance gene, mainly among enterococci (2–8).

During a survey on antibiotic-resistant bacteria carried out in 2018 among the rural population living in the Bolivian Chaco region, an Enterococcus gallinarum (Eg-IV02) resistant to linezolid (MIC of 8 μg/ml by reference broth microdilution [9]) was isolated from a fecal swab collected from a healthy child after plating the swab on CNA-cv Sh medium (bioMérieux, Marcy l'Etoile, France) supplemented with 16 μg/ml of florfenicol.

Multiplex real-time PCR was positive for the poxtA gene but not for the optrA, cfr, or cfr(B) genes, which are the other known transferable linezolid resistance determinants encountered in Gram-positive cocci.

Whole-genome sequencing (WGS) analysis of E. gallinarum Eg-IV02, performed using both Illumina MiSeq (Illumina, San Diego, CA) and MinION (Oxford Nanopore Technologies, Oxford, UK) platforms to yield a de novo hybrid assembly generated by Unicycler v0.4.6 (10) revealed the presence of a poxtA-like gene, named poxtA2, carried on a 13,746-bp plasmid, named pIB-BOL (GenBank accession number MZ171245) (Fig. 1a). The plasmid carried a repB gene closely related (98% identity) to that of an 11-kb plasmid from a cfr-harboring Enterococcus faecalis of animal origin from China (GenBank accession number CP028840.1) and also a fexA phenicol resistance determinant, located upstream of poxtA2. Both genes were flanked by IS1216-like insertion sequences, likely involved with their mobilization (Fig. 1a).

FIG 1.

(a) Map of pIB-BOL plasmid from E. gallinarum Eg-IV02. (b) Comparison of the genetic context of poxtA in the chromosome of S. aureus AOUC-09/15 (1) and in plasmid pIB-BOL; regions with >99% nucleotide identity are connected by gray zones, transposase-encoding genes are indicated by striped arrows, and IRs are indicated by black triangles. (c) Nucleotide sequences showing the insertion points of IS1216_D at the 3′-end of poxtA and of IS1216_E downstream poxtA2. The sequence of IS1216 elements is capitalized. Translation of the poxtA and poxtA2 coding sequences is also shown, and the different residues at the C terminus of PoxtA2 are boldfaced. Alignment of the C terminus of PoxtA and PoxtA2 with the homologous region of OptrA is also shown. (d) Comparison of the genetic contexts of poxtA2 and of the identical gene recently reported from an E. faecalis EFS0019 of animal origin from Korea (GenBank accession number NZ_QUSQ01000020.1) (12). The gray shaded zones indicate 100% nucleotide sequence homology. The sequences of the junctions are also shown (uppercase letters for IS1216 elements), revealing the identity between the two genetic contexts.

Unlike poxtA, poxtA2 was not truncated by an IS1216 insertion at the 3′ end (Fig. 1b). Consequently, PoxtA2 differed from PoxtA by a few amino acids at the C terminus, which showed some detectable homology with the closest relative of the same protein family, namely, OptrA (Fig. 1c). Altogether, these findings suggest that poxtA2 likely represents the ancestor of poxtA, mobilized by a recombination event that had not truncated the gene at the 3′-end. In fact, the genetic context of poxtA2 was different from that of poxtA (Fig. 1b), supporting the notion that the two genes had been mobilized by independent recombination events.

Plasmid pIB-BOL was transferred to E. faecalis JH2-2 by electrotransformation (11). Transformants were selected on tryptic soy agar containing 8 μg/ml florfenicol. Acquisition of pIB-BOL by E. faecalis JH2-2 was associated with a significant increase of linezolid, florfenicol, and chloramphenicol MICs and with a lower increase of doxycycline MIC, while susceptibility to tedizolid, tetracycline, and tigecycline was apparently not affected (see Table S1 in the supplemental material). Since poxtA2 and fexA were the only resistance determinants carried by the plasmid (Fig. 1a), these results confirmed that poxtA2 was functional in conferring protection from linezolid, while the effect on phenicols could be ascribed to fexA.

A search of the NCBI sequence databases (carried out on 25 May 2021) revealed only another poxtA-like gene identical to poxtA2 from an E. faecalis isolate of animal origin from Korea (GenBank accession number NZ_QUSQ01000020.1) (12). Even that gene was flanked by two IS1216-like elements and exhibited a genetic context identical to that of poxtA2 from pIB-BOL (Fig. 1d), suggesting a common origin from a unique mobilization event. Interestingly, the same database search also revealed two poxtA-like genes encoding proteins identical to PoxtA but for a single amino acid change (G33H or R256H, respectively) from different Enterococcus faecium isolates in China (NCBI:protein accession numbers WP_212481470.1 and WP_159373727.1, respectively).

Overall, these findings identified a poxtA-related gene, which (i) appears to be the presumptive ancestor of poxtA and (ii) has been mobilized to enterococci circulating in human and animal settings. The intersectoral spreading of these resistance determinants, mediating transferable linezolid resistance, is concerning and mandates for surveillance. In this perspective, it should be noted that the prevalence of poxtA2 might be underestimated in case of screening using primers targeting external regions of the poxtA gene.

Data availability.

Sequence data have been deposited in GenBank under the accession number MZ171245.