Carbapenem and amikacin resistance on a large conjugative Acinetobacter baumannii plasmid

Sir, 
 
Worldwide, most multiply antibiotic-resistant (MAR), and particularly carbapenem-resistant, Acinetobacter baumannii causing infections belong to global clone 1 (GC1) or global clone 2 (GC2). Though there have been incidences where a different clonal group was predominantly responsible for an epidemic outbreak, MAR isolates that are not members of these clones have been observed to cause infections in hospitals but are not as common.1 
 
Isolate D46, collected at Royal North Shore Hospital, Sydney in 2010, was shown to be resistant to ampicillin, imipenem, meropenem, ticarcillin/clavulanate, ceftazidime, cefotaxime, streptomycin, spectinomycin, sulfamethoxazole, tetracycline, trimethoprim, chloramphenicol, florfenicol, kanamycin, neomycin, gentamicin, amikacin, tobramycin, nalidixic acid and ciprofloxacin, making it extensively antibiotic resistant. D46 was previously shown to be ST110 (Oxford MLST) and to harbour the aphA6 amikacin resistance gene within TnaphA6 and an aadB gene cassette (gentamicin, kanamycin and tobramycin resistance) in the small plasmid pRAY.2 An ISAba1 upstream of the chromosomal ampC confers resistance to third-generation cephalosporins, ceftazidime and cefotaxime.3 
 
Here, we have further examined the causes of resistance. Using PCR as described previously,4 D46 was shown to contain the strA-strB, sul2 and tetA(B) genes, responsible for streptomycin, sulfamethoxazole and tetracycline resistance, respectively. These genes were in the same configuration as in AbGRI1-2 (Tn6167),4 but D46 does not have an island in comM. Carbapenem and ticarcillin/clavulanate resistance was due to the blaOXA-23 carbapenemase gene, which was within Tn2006. 
 
To better understand how Tn2006 was acquired, conjugation was performed as described previously,5,6 using D46 as a donor and a rifampicin-resistant mutant of A. baumannii ATCC 17978 as a recipient. Transconjugants resistant to meropenem, imipenem and ticarcillin/clavulanate, indicative of blaOXA-23, and resistant to kanamycin, neomycin and amikacin, indicative of aphA6, were recovered. Hence, both Tn2006 and TnaphA6 were located on a conjugative plasmid. 
 
The whole genome sequence of D46 was determined using Illumina HiSeq and assembled as described previously.5 The draft genome comprised 115 contigs with an average read depth of 88.7× coverage and D46 was determined to be ST25 according to the Pasteur MLST scheme. Three plasmids were detected. The sequence of the smallest plasmid, pD46-1 (6078 bp), contained aadB and was almost identical to pRAY* (a single base difference).2 pD46-2 is an 8731 bp cryptic plasmid that was identical to p1ABTCDC0715 (GenBank accession number {"type":"entrez-nucleotide","attrs":{"text":"CP002523","term_id":"323519903","term_text":"CP002523"}}CP002523) from a GC2 isolate. 
 
To determine whether D46 harboured Tn2006 together with TnaphA6 on a plasmid related to pAb-G7-25 and pD72-2,7 contigs that matched them were retrieved from the draft genome. Seven contigs were recovered and assembled using PCR.5,7 The amplicons were sequenced, finalizing the assembly. Plasmid pD46-3 has a 67 027 bp backbone sharing 99.99% identity (five single-base changes) with that of pD72-2 (Figure 1a). pD46-3 also harboured TnaphA6 in the same position as pD72-2 (Figure 1a). One end of two backbone contigs had inversely oriented fragments of ISAba1 and they were linked with blaOXA-23 in Tn2006 using PCR (Figure 1a). The complete 74 916 bp sequence of pD46-3 was deposited in GenBank under accession number {"type":"entrez-nucleotide","attrs":{"text":"KM977710","term_id":"730910803","term_text":"KM977710"}}KM977710. As the backbone of pD46-3 was almost identical to that of pD72-2 and pAb-G7-2, pD46-3 also contained the complete set of transfer genes (Figure 1). This is the first report of a completely sequenced A. baumannii plasmid that has been demonstrated to simultaneously transfer resistance to carbapenems and aminoglycosides. 
 
 
 
Figure 1. 
 
Comparison of repAci6 plasmids pD72-2, pD46-3 (a), pABUH1 and pAb-G7-2 (b). The plasmid backbones, linearized and opened at repA, are represented by horizontal lines. The extent of ORFs and genes are shown as arrows beneath the lines, with names given ... 
 
 
 
A recent genomic study of a GC2 A. baumannii collection from a US hospital reported the sequence of pABUH1 (GenBank accession number {"type":"entrez-nucleotide","attrs":{"text":"AYOH01000010","term_id":"573041124","term_text":"AYOH01000010"}}AYOH01000010), which the authors suggested could be transferable.8 It was described as a plasmid related to pACICU2, harbouring ISAba1 upstream of aphA6 and having a copy of blaOXA-23 flanked by two copies of ISAba125 that is in the same position as ISAba125 in pACICU2.8 However, it was recently shown that pACICU2 is likely identical to pAb-G7-2 and in fact harbours TnaphA6 at this location.9 Our analysis of the sequence of pABUH1 revealed that it in fact carries TnaphA6 in this position (Figure 1b). Furthermore, ISAba1 is actually upstream of blaOXA-23, in a structure similar to Tn2008 (GenBank accession number {"type":"entrez-nucleotide","attrs":{"text":"GQ861438","term_id":"281486593","term_text":"GQ861438"}}GQ861438) that is flanked by a 9 bp direct repeat. However, in pABUH1 the sequence adjacent to the ISAba1 was 1613 bp whereas it is only 1351 bp in Tn2008. 
 
The backbones of the repAci6 plasmids, pAb-G7-2, pD72-2, pD46-3, pABUH1 and pACICU2, are all very closely related, but they can be separated into two lineages based on the position of TnaphA6 (Figure 1). Representatives of the two lineages, pD46-3 and pABUH1, have each acquired the blaOXA-23 gene on separate occasions and a sixth repAci6 plasmid, pA85-3, has also gained blaOXA-23 within AbaR4.6 The acquisition of blaOXA-23 in different structures and in separate events indicates that this group of related plasmids plays a vital role in the dissemination of this carbapenemase gene and could be one of the factors responsible for making it a worldwide problem. Hence, surveillance of this group of A. baumannii plasmids will be vital in curtailing the spread of carbapenem resistance in A. baumannii.

Sir, Worldwide, most multiply antibiotic-resistant (MAR), and particularly carbapenem-resistant, Acinetobacter baumannii causing infections belong to global clone 1 (GC1) or global clone 2 (GC2). Though there have been incidences where a different clonal group was predominantly responsible for an epidemic outbreak, MAR isolates that are not members of these clones have been observed to cause infections in hospitals but are not as common. 1 Isolate D46, collected at Royal North Shore Hospital, Sydney in 2010, was shown to be resistant to ampicillin, imipenem, meropenem, ticarcillin/clavulanate, ceftazidime, cefotaxime, streptomycin, spectinomycin, sulfamethoxazole, tetracycline, trimethoprim, chloramphenicol, florfenicol, kanamycin, neomycin, gentamicin, amikacin, tobramycin, nalidixic acid and ciprofloxacin, making it extensively antibiotic resistant. D46 was previously shown to be ST110 (Oxford MLST) and to harbour the aphA6 amikacin resistance gene within TnaphA6 and an aadB gene cassette (gentamicin, kanamycin and tobramycin resistance) in the small plasmid pRAY. 2 An ISAba1 upstream of the chromosomal ampC confers resistance to third-generation cephalosporins, ceftazidime and cefotaxime. 3 Here, we have further examined the causes of resistance. Using PCR as described previously, 4 D46 was shown to contain the strA-strB, sul2 and tetA(B) genes, responsible for streptomycin, sulfamethoxazole and tetracycline resistance, respectively. These genes were in the same configuration as in AbGRI1-2 (Tn6167), 4 but D46 does not have an island in comM. Carbapenem and ticarcillin/clavulanate resistance was due to the bla OXA-23 carbapenemase gene, which was within Tn2006.
To better understand how Tn2006 was acquired, conjugation was performed as described previously, 5,6 using D46 as a donor and a rifampicin-resistant mutant of A. baumannii ATCC 17978 as a recipient. Transconjugants resistant to meropenem, imipenem and ticarcillin/clavulanate, indicative of bla OXA-23 , and resistant to kanamycin, neomycin and amikacin, indicative of aphA6, were recovered. Hence, both Tn2006 and TnaphA6 were located on a conjugative plasmid.
The whole genome sequence of D46 was determined using Illumina HiSeq and assembled as described previously. 5 The draft genome comprised 115 contigs with an average read depth of 88.7× coverage and D46 was determined to be ST25 according to the Pasteur MLST scheme. Three plasmids were detected. The sequence of the smallest plasmid, pD46-1 (6078 bp), contained aadB and was almost identical to pRAY* (a single base difference). 2 pD46-2 is an 8731 bp cryptic plasmid that was identical to p1ABTCDC0715 (GenBank accession number CP002523) from a GC2 isolate.
To determine whether D46 harboured Tn2006 together with TnaphA6 on a plasmid related to pAb-G7-2 5 and pD72-2, 7 contigs that matched them were retrieved from the draft genome. Seven contigs were recovered and assembled using PCR. 5,7 The amplicons were sequenced, finalizing the assembly. Plasmid pD46-3 has a 67 027 bp backbone sharing 99.99% identity (five singlebase changes) with that of pD72-2 ( Figure 1a). pD46-3 also harboured TnaphA6 in the same position as pD72-2 ( Figure 1a). One end of two backbone contigs had inversely oriented fragments of ISAba1 and they were linked with bla OXA-23 in Tn2006 using PCR (Figure 1a). The complete 74 916 bp sequence of # The Author 2014. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/ 4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
pD46-3 was deposited in GenBank under accession number KM977710. As the backbone of pD46-3 was almost identical to that of pD72-2 and pAb-G7-2, pD46-3 also contained the complete set of transfer genes ( Figure 1). This is the first report of a completely sequenced A. baumannii plasmid that has been demonstrated to simultaneously transfer resistance to carbapenems and aminoglycosides.
A recent genomic study of a GC2 A. baumannii collection from a US hospital reported the sequence of pABUH1 (GenBank accession number AYOH01000010), which the authors suggested could be transferable. 8 It was described as a plasmid related to pACICU2, harbouring ISAba1 upstream of aphA6 and having a copy of bla OXA-23 flanked by two copies of ISAba125 that is in the same position as ISAba125 in pACICU2. 8 However, it was recently shown that pACICU2 is likely identical to pAb-G7-2 and in fact harbours TnaphA6 at this location. 9 Our analysis of the sequence of pABUH1 revealed that it in fact carries TnaphA6 in this position (Figure 1b). Furthermore, ISAba1 is actually upstream of bla OXA-23 , in a structure similar to Tn2008 (GenBank accession number GQ861438) that is flanked by a 9 bp direct repeat. However, in pABUH1 the sequence adjacent to the ISAba1 was 1613 bp whereas it is only 1351 bp in Tn2008.
The backbones of the repAci6 plasmids, pAb-G7-2, pD72-2, pD46-3, pABUH1 and pACICU2, are all very closely related, but they can be separated into two lineages based on the position of TnaphA6 (Figure 1). Representatives of the two lineages, pD46-3 and pABUH1, have each acquired the bla OXA-23 gene on separate occasions and a sixth repAci6 plasmid, pA85-3, has also gained bla OXA-23 within AbaR4. 6 The acquisition of bla OXA-23 in different structures and in separate events indicates that this group of related plasmids plays a vital role in the dissemination of this carbapenemase gene and could be one of the factors responsible for making it a worldwide problem. Hence, surveillance of this group of A. baumannii plasmids will be vital in curtailing the spread of carbapenem resistance in A. baumannii.

Transparency declarations
None to declare.
Sir, Fosfomycin [(2R,3S)-3-methyloxiran-2-yl]phosphonic acid is a broad-spectrum antibiotic with bactericidal activity against both Gram-positive and Gram-negative bacteria. 1 Fosfomycin inhibits bacterial cell wall synthesis by inactivating the enzyme UDP-N-acetylglucosamine-3-enolpyruvyl transferase (MurA), which is essential for the catalysis of bacterial peptidoglycan biosynthesis. Thus, modification or overexpression of MurA can contribute to the development of fosfomycin resistance in bacteria. 2 Other common mechanisms of resistance to fosfomycin include decreased drug uptake and inactivation of the antibiotic by addition of various groups. 3 To date, several fosfomycin resistance genes have been functionally characterized in bacteria. FomA and FomB enzymes can inactivate fosfomycin by phosphorylation. 4 FosA, FosB and FosX can inactivate fosfomycin by catalysing the addition of glutathione, L-cysteine and H 2 O, respectively, to its epoxide ring. 2 Campylobacter species are the leading foodborne pathogens worldwide, accounting for 400 -500 million cases of diarrhoea each year. 5 Over the past few decades, Campylobacter species have become increasingly resistant to several clinically important antimicrobial agents, compromising the effectiveness of clinical therapy. 6 Various mechanisms of resistance to macrolides, fluoroquinolones, tetracyclines, aminoglycosides and b-lactams in Campylobacter have been extensively described. 7 In addition, clinical trials have indicated that fosfomycin can effectively treat Campylobacter enteritis in humans. 8 To date, no mechanism of fosfomycin resistance has been characterized in Campylobacter species. Here, we identified a novel fosX-like gene from Campylobacter coli (designated fosX CC ), located in the multidrug resistance genomic island (MDRGI), which confers fosfomycin resistance to Campylobacter species.
C. coli DZB4 was isolated from swine faeces in China in 2012. Our recent study indicated that DZB4 contains a type II MDRGI consisting of 10 ORFs that is inserted between cadF and CCO1582 on the chromosome (GenBank accession number KC876749) and this MDRGI can mediate antibiotic resistance, at least to macrolides, aminoglycosides and tetracycline. 9 Further analysis revealed that the fifth ORF, immediately upstream of erm(B), in type II MDRGI encodes a 136 amino acid protein that exhibits 26.9%, 34.2% and 63.9% identity to fosfomycin resistance determinants FosA, FosB and FosX found in Serratia marcescens (M85195), 10 Staphylococcus haemolyticus (X89875) 11 and Listeria monocytogenes (NP_465227.1), 2 respectively. Protein structure prediction indicated that this putative protein has identical quaternary structure to FosX and contains the conserved active site residue E44, which plays a key role in catalysis in FosX but not in FosA or FosB. 2,12,13 Considering that the fosfomycin MIC for isolate DZB4 was .512 mg/L, which is much higher than that for reference strains