Molecular Basis of Macrolide, Triamilide, and Lincosamide Resistance in Pasteurella multocida from Bovine Respiratory Disease

Abstract

The mechanism of macrolide-triamilide resistance in Pasteurella multocida has been unknown. During whole-genome sequencing of a multiresistant bovine P. multocida isolate, three new resistance genes, the rRNA methylase gene erm(42), the macrolide transporter gene msr(E), and the macrolide phosphotransferase gene mph(E), were detected. The three genes were PCR amplified, cloned into suitable plasmid vectors, and shown to confer either macrolide-lincosamide resistance [erm(42)] or macrolide-triamilide resistance [msr(E)-mph(E)] in macrolide-susceptible Escherichia coli and P. multocida hosts.

INTRODUCTION

Pasteurella multocida is one of the economically most relevant bacterial pathogens involved in bovine respiratory disease (9, 14). Antimicrobial agents are commonly applied to control respiratory tract infections in cattle (9, 14). Based on the available susceptibility data, most bovine P. multocida strains are still susceptible to newer antimicrobial agents, including the 15-membered ring macrolide tulathromycin (4, 7, 16). Since tulathromycin has a unique chemical structure characterized by the presence of three amine groups, it has been referred to in the literature as the first representative of the triamilide subclass among the macrolides (3). Tulathromycin was approved in 2005 for use in the treatment and control of bovine respiratory disease and the treatment of swine respiratory disease in the European Union and the United States (3). Since its approval, continuous surveillance of the susceptibility status of the target pathogens has been conducted by Pfizer Animal Health, and from 2005 on, single tulathromycin-resistant strains have been detected. Previous analyses, which included bioassays for drug inactivation, efflux pump inhibitor assays, PCR assays for the known macrolide resistance genes, and sequencing of the genes for ribosomal proteins L4 and L22 as well as the separate amplification of part of the six rRNA operons with subsequent sequencing of the domain V of the 23S rRNA, did not elucidate the mechanism of tulathromycin resistance in these P. multocida strains (15).

To identify the mechanism(s) of macrolide resistance in P. multocida, the plasmid-free, tilmicosin- and tulathromycin-resistant P. multocida strain 36950, obtained from a case of bovine respiratory tract infection in a Nebraska feedlot in 2005, was subjected to whole-genome sequencing. For this, genomic DNA was prepared as previously described (10). A shot-gun library and a long-tag paired-end library were produced using the 454 Life Sciences (Roche) GS-FLX system (Eurofins MWG, Ebersberg, Germany). A matched assembly against the sequence of the largely susceptible P. multocida strain Pm70 (11) was performed to identify contigs that differed from the Pm70 sequence. These contigs, which were the most likely ones to contain resistance genes, were analyzed with the BLAST programs available at http://blast.ncbi.nlm.nih.gov/Blast.cgi. The matched assembly identified ca. 220 small contigs of 1.5 to 25.7 kb that differed from the Pm70 sequence. One of these contigs contained a gene for an Erm-like rRNA methylase, while another contig harbored two genes organized in an operon structure: a gene for an Msr-like ABC transporter and a gene for a Mph-like macrolide phosphotransferase. These three genes received the novel designations erm(42), msr(E), and mph(E) from the macrolide-lincosamide-streptogramin (MLS) nomenclature center (http://faculty.washington.edu/marilynr/).

The erm(42) gene codes for a protein of 301 amino acids (aa) which is only distantly related (28% identity) to the 284-aa Erm(X) proteins of Bifidobacterium thermophilum (19) and Corynebacterium diphtheriae (18). However, during the course of this study, part of a sequence of a resistance plasmid from a fish-pathogenic Photobacterium damselae subsp. piscicida isolate has been released (DDBJ accession number BAJ34818). This sequence contained a gene for an erythromycin resistance protein of 303 aa which showed 99.3% identity to the Erm(42) protein from P. multocida. The msr(E) gene codes for a protein of 491 aa which showed identities of 60.6% to Msr(D) (GenBank accession number AF227521), 36.8% to Msr(C) (GenBank accession number AY004350), and 22.3% to Msr(A) (EMBL accession number X52085). The mph(E) gene coded for a protein of 294 aa, which proved to be 37.8 and 38.1% identical to Mph(A) variants (GenBank accession number U36578 and DDBJ accession number D16251, respectively), 27.2% identical to Mph(B) (DDBJ accession number D85892) and 26.7% identical to Mph(C) (GenBank accession number AB013298). The msr(E) and mph(E) genes are separated by a spacer sequence of 55 bp that contains the mph(E)-associated ribosome binding site, but no other regulatory elements. As a consequence, both genes are likely to be transcribed from a promoter upstream of msr(E). A putative promoter structure with a −35 sequence (TTGGAC), a −10 sequence (TTTAAT), and an A at position +1 is available in this msr(E) upstream region. A BlastN search revealed that these two genes (including the spacer) have previously been found on plasmids in Klebsiella pneumoniae (17), Acinetobacter baumannii (12, 20), Citrobacter freundii (5), Escherichia coli (6), and an uncultured bacterium from a sewage treatment plant (13). It should be noted that the msr(E) and mph(E) genes have been named mel and mef(E) or mph or mph2, respectively, in the aforementioned publications.

To confirm their role in macrolide resistance, the erm(42) gene as well as the msr(E)-mph(E) genes and their flanking regions were amplified by PCR (Table 1) and resequenced to confirm identity with the sequence obtained by whole-genome sequencing. Different approaches were conducted for cloning these amplicons in plasmid vectors that replicate well in E. coli and in P. multocida. For expression in the macrolide-susceptible E. coli AS19, vector pCR2.1-TOPO and the TA cloning kit (Invitrogen, Groningen, Netherlands) were used. For expression in the pan-susceptible P. multocida B130, the blunt-ended erm(42) amplicon was cloned into the EcoRV site of plasmid pLS88 (2), while the msr(E)-mph(E) amplicon was inserted into the EcoRV site within the tetracycline resistance gene tet(L) of plasmid pCCK3259 (8). The selection of E. coli AS19 and P. multocida B130 transformants occurred overnight on Luria-Bertani agar supplemented with 10 μg/ml erythromycin and 5% (vol/vol) sheep blood (for P. multocida B130). The transformants were confirmed by plasmid profiling. Antimicrobial susceptibility testing of the original P. multocida strain 36950 and of the transformants in both recipient strains was performed at least three times on independent occasions by broth microdilution according to document M31-A3 of the Clinical and Laboratory Standards Institute (1) using custom-made microtiter plates (MCS Diagnostics, Swalmen, Netherlands). Staphylococcus aureus ATCC 29213 served as the quality control strain. For erythromycin, concentrations from 0.015 to 32 μg/ml were tested, whereas for tilmicosin, tulathromycin, and clindamycin, the test concentrations ranged from 0.03 to 64 μg/ml.

Table 1.

PCR primers and amplicon sizes^a

Gene region	Primer directionb	Primer sequence (5′ → 3′)	Amplicon size (bp)
erm(42)	fw	GGGTGAAAAGGGCGTTTATT	1,254
	rv	ACGTTGCACTTGGTTTGACA
msr(E)-mph(E)	fw	TACCGGAACAACGTGATTGA	3,277
	rv	GAAGGGTTACGCCAGTACCA

^aThe annealing temperature for both gene regions was 55°C.

^bfw, forward; rv, reverse.

Comparative analysis of the MIC values of the recipient strains with and without the cloned erm(42) gene (Table 2) showed up to 128-fold increases in the MICs of erythromycin, tilmicosin, and clindamycin when the erm(42) gene was present in either one of the two recipient strains. In contrast, expression of erm(42) resulted in only moderate 8- or 4-fold increases in the MICs of tulathromycin in the P. multocida and E. coli recipients, respectively. As expected, erm(42) conferred a macrolide-lincosamide resistance phenotype. Clones carrying the msr(E)-mph(E) amplicons exhibited up to 128-fold increases in the MICs of erythromycin, tilmicosin, and tulathromycin, whereas the clindamycin MIC remained unchanged in both hosts (Table 2). This observation strongly suggested that the substrate spectrum of msr(E)-mph(E) includes macrolides and triamilides, but not lincosamides.

Table 2.

MICs of the original P. multocida 36950 and P. multocida B130 and E. coli AS19 with and without a cloned erm(42) or msr(E)-mph(E) amplicon

Antimicrobial agent	MIC (μg/ml) in the P. multocida host						MIC (μg/ml) in the E. coli host
	P. multocida 36950	P. multocida B130	P. multocida B130 + pLS88	P. multocida B130 + pLS88 + cloned erm(42)	P. multocida B130 + pCCK3259	P. multocida B130 + pCCK3259 + cloned msr(E)-mph(E)	E. coli AS19	E. coli AS19 + pCR2.1	E. coli AS19 + pCR2.1 + cloned erm(42)	E. coli AS19 + pCR2.1 + cloned msr(E)-mph(E)
Erythromycin	≥64	4	4	≥64	4	≥64	0.5	0.5	≥64	≥64
Tilmicosin	≥128	4	4	≥128	4	≥128	1	1	32	8
Tulathromycin	≥128	2	2	16	1	≥128	8	8	32	≥128
Clindamycin	≥128	16	16	≥128	16	16	8	8	≥128	8

In conclusion, to the best of our knowledge, this is the first report on the genetics of macrolide, triamilide, and lincosamide resistance in P. multocida. The observation that the three genes identified have—even only rarely—also been seen in other bacteria supports the assumption that P. multocida is able to acquire mostly plasmid-borne resistance genes from other Gram-negative bacteria. Further work is warranted to elucidate the exact mechanisms by which these resistance genes have become integrated into the chromosomal DNA of P. multocida strain 36950.

Nucleotide sequence accession numbers.

The sequence of the erm(42) gene has been deposited in the EBI database under accession number FR734406, and the sequence of the msr(E)-mph(E) genes has been deposited in the EBI database under accession number FR751518.