ISEcp1-Mediated Transposition Leads to Fosfomycin and Broad-Spectrum Cephalosporin Resistance in Klebsiella pneumoniae

A fosfomycin-resistant and carbapenemase (OXA-48)-producing Klebsiella pneumoniae isolate was recovered, and whole-genome sequencing revealed ISEcp1-bla_CTX-M-14b tandemly inserted upstream of the chromosomally encoded lysR-fosA locus. Quantitative evaluation of the expression of lysR and fosA genes showed that this insertion brought a strong hybrid promoter leading to overexpression of the fosA gene, resulting in fosfomycin resistance.

ABSTRACT

A fosfomycin-resistant and carbapenemase (OXA-48)-producing Klebsiella pneumoniae isolate was recovered, and whole-genome sequencing revealed ISEcp1-bla_CTX-M-14b tandemly inserted upstream of the chromosomally encoded lysR-fosA locus. Quantitative evaluation of the expression of lysR and fosA genes showed that this insertion brought a strong hybrid promoter leading to overexpression of the fosA gene, resulting in fosfomycin resistance. This work showed the concomitant acquisition of resistance to broad-spectrum cephalosporins and fosfomycin due to a single genetic event.

TEXT

Acquired resistance to fosfomycin in Enterobacterales isolates is often related to the production of Fos-type glutathione S-transferases encoded by fos genes, among which fosA-like genes are the most common and fosL1 gene was the latest identified (1, 2). It was recently shown that the natural reservoir of fosA5 and fosA6 genes is Klebsiella pneumoniae (3), whereas that of the latest fosA8 gene is Leclercia adecarboxylata (4). In this species, the fosA gene is chromosomally located and is overlapping the end of a putative lysR regulator gene (Fig. 1). Scarce information about the role of this regulator in the expression of fosA is available. Despite the presence of the fosA gene in their chromosome, wild-type K. pneumoniae isolates remain susceptible to fosfomycin (<16 μg/ml), suggesting that the chromosomal fosA gene is poorly (or not) expressed. In this study, we recovered a carbapenemase-producing K. pneumoniae isolate presenting high-level resistance to fosfomycin and submitted it to analysis of its resistance pattern.

FIG 1.

Schematic representation of insertion of the ISEcp1-bla_CTX-M-14 tandem into the K. pneumoniae S54 chromosome leading to the formation of a hybrid promoter. (A) Original sequence found in a wild-type K. pneumoniae genome. (B) Sequence identified in isolate 54 once the ISEcp1-bla_CTX-M-14 tandem has been acquired by transposition. Arrows show promoters that are described in detail at the bottom. The lysR gene encodes a putative LysR regulator; the fosA gene encodes the FosA glutathione S-transferase; the mdaB gene encodes an NAD(P)H-dependent oxidoreductase. Promoter −35 and −10 sequences are boxed. IRL, inverted long repeat; DR, direct repeat; RBS, ribosome binding site; +1, start codon of the lysR gene.

Isolate S54 was recovered from a urine sample of a patient from Switzerland in 2015. It was resistant to all β-lactams, including ceftazidime (MIC, 8 μg/ml), cefepime (MIC, 32 μg/ml), carbapenems (MICs, 128, 8, and 32 μg/ml for ertapenem, imipenem, and meropenem, respectively), chloramphenicol (MIC, >256 μg/ml), ciprofloxacin (MIC, >256 μg/ml), kanamycin (MIC, >256 μg/ml), tobramycin (MIC, 32 μg/ml), and fosfomycin. Isolate S54 also presented reduced susceptibility to amikacin (MIC, 16 μg/ml) and gentamicin (MIC, 1 μg/ml). The isolate remained susceptible only to colistin (MIC, 0.25 μg/ml) and tigecycline (MIC, 0.25 μg/ml).

Fosfomycin resistance of isolate S54 was confirmed by the agar dilution method following CLSI recommendations (5), with an MIC of >1,024 μg/ml. Addition of phosphonoformate in the medium (5 mM) showed a decreased MIC (128 μg/ml), suggesting that the mechanism leading to fosfomycin resistance in that isolate was likely related to the production of a glutathione transferase (6). PCR amplifications targeting the common plasmid-mediated fosfomycin resistance genes (fosA1 through fosA9 and fosC2) remained negative except for the fosA5- and fos6-like genes, being intrinsic to K. pneumoniae isolates as mentioned above (3).

Whole-genome sequencing was therefore performed with Illumina technology (Illumina, Inc., San Diego, CA). The genomic DNA was extracted using the Sigma-Aldrich GenElute bacterial genomic DNA kit. Genomic libraries were assessed using the NexteraXT library preparation kit (Illumina, Inc.), and sequencing was performed using the MiniSeq system with 150-bp paired-end reads and 50× coverage. Generated FastQ files were compiled and analyzed using the CLC genomic workbench 7.5.1 (CLC bio, Aarthus, Denmark). The resulting contigs were uploaded to the Center for Genomic Epidemiology server (http://www.genomicepidemiology.org/). Multilocus sequence type and resistome were determined using MLST 2.0 and ResFinder 3.2, respectively (7, 8).

Using those data, the sequence type (ST) of isolate S54 was determined, corresponding to ST101. It carried the bla_OXA-48 carbapenemase gene, along with the bla_CTX-M-14b extended-spectrum β-lactamase (ESBL) gene, bla_OXA-9 class D narrow-spectrum β-lactamase gene, and bla_TEM-1D broad-spectrum β-lactamase genes. Additionally, it harbored the dfrA14, sul2, aadA1, and aac(6′)-Ib-cr genes, conferring resistance to trimethoprim, sulfonamides, and aminoglycosides, respectively, together with reduced susceptibility to ciprofloxacin. No putative chloramphenicol resistance gene was identified, despite the observed resistance phenotype, suggesting that upregulation of efflux systems might be involved (9).

Detailed in silico analysis showed that the bla_CTX-M-14 gene was associated with insertion sequence ISEcp1 located upstream of it. ISEcp1 is a member of the IS1380 family known to mobilize adjacent DNA sequences by a so-called one-ended transposition mechanism (10). Furthermore, ISEcp1 has been shown to provide a promoter sequence to different antibiotic resistance genes located at its right-hand extremity, such as bla_CTX-M genes, which is the case here with bla_CTX-M-14 (11). Note that the bla_CTX-M-14 ESBL gene was previously reported in animal, environmental, and human isolates from Switzerland (12). In addition, a similar association between ISEcp1 and a bla_CTX-M-14 gene has already been observed in another K. pneumoniae isolate from Switzerland, but it belonged to another ST (ST23) for which the fosfomycin susceptibility status remains unknown (13). The tandem formed by ISEcp1 and bla_CTX-M-14b was located onto the chromosome of isolate S54, where it had inserted upstream of the lysR-fosA tandem, at position −2 from the start codon of the proximate mdaB gene (Fig. 1).

Given the long distance separating ISEcp1 from the lysR-fosA tandem, it was unlikely that the promoter sequence provided by ISEcp1 could play a significant role in fosA expression. Therefore, in silico promoter identification was performed using BPROM software (http://www.softberry.com/berry.phtml?topic=bprom) (14), identifying two putative promoters named P1 and P2. Promoter P1 was located upstream from the lysR-fosA tandem and was therefore hypothesized to be the intrinsic promoter of these genes. Promoter P2 was identified as a putative hybrid promoter constituted by a −35 box located at the end of the ISEcp1-bla_CTX-M-14b tandem combined with a −10 box intrinsic to the K. pneumoniae sequence. This −10 box actually corresponded to the entire 5-bp direct repeat flanking the ISEcp1-bla_CTX-M-14b tandem, together with an additional nucleotide (Fig. 1). The presence of this putative hybrid promoter P2 was therefore hypothesized to be responsible for overexpression of the fosA gene and the source of acquired resistance to fosfomycin in isolate S54.

To confirm the impact of both promoters, two plasmid constructs were generated in the pBAD_b vector carrying an l-arabinose-inducible promoter and expressed in Escherichia coli (15). These two recombinant plasmids, namely, pBAD-P1 and pBAD-P2, respectively, encompassed (i) the lysR-fosA tandem together with promoter P1 and (ii) the same tandem with both P1 and P2 promoters. Both plasmids were transformed into E. coli strain TOP10 (ThermoFisher Scientific), selection being performed with ampicillin 50 μg/ml (resistance marker of plasmid pBAD). Of note, E. coli is a bacterial species that does not possess any naturally occurring fosA gene. The recombinant E. coli strains carrying the pBAD-P1 and pBAD-P2 plasmids showed fosfomycin MICs of 128 and 2,048 μg/ml, respectively, whereas that of E. coli TOP10 was <4 μg/ml. This confirmed the significant role of P2 as an efficient promoter of fosA expression.

Experiments were performed to evaluate the expression of lysR and fosA genes in the two recombinant strains. Briefly, total RNAs of the two recombinant strains were extracted using the Quick-RNA MiniPrep kit (Zymo Research, CA). Then a total of 200 ng of each RNA sample was reverse transcribed using the qScript cDNA SuperMix (Quantabio, Beverly, MA) according to the recommendation of the manufacturer. Quantitative real-time PCR was performed using RotorGene Q (Qiagen, Hilden, Germany). Reactions were set up in a total volume of 25 μl with PerfeCTa SYBR green FastMix (Quantabio). The selected genes in this assay were 16S rRNA (internal control), the putative regulator gene lysR, and fosA. Three independent replicates were performed, with three different RNA extractions, cDNA amplifications, and measurements.

Results showed that the recombinant strain carrying both P1 and P2 promoters had an overexpression of lysR and fosA genes of 380- and 340-fold, respectively, in E. coli isolates compared with those measured from the recombinant strain harboring promoter P1 only.

Altogether, these results showed that P2 constitutes an additional promoter responsible for overexpression of the fosA gene leading to the observed fosfomycin resistance phenotype in isolate S54. It is well known that IS elements may provide promoter sequences to resistance genes, such as for bla_TEM-6 or bla_SHV-2a genes (16, 17). However, the identified structure here was singular. First, the promoter was chimeric with a −35 box belonging to the sequence bracketing the ISEcp1-bla_CTX-M-14b tandem (with this sequence actually corresponding to the one recognized by ISEcp1 as an imperfect right inverted repeat) and a −10 box sequence being the target-site duplication generated on transposition by ISEcp1. Second, insertion of the ISEcp1-bla_CTX-M-14b tandem led to overexpression of another resistance gene, namely, fosA, providing a coresistance phenotype. Hence, coacquisition of fosfomycin and broad-spectrum cephalosporin occurred through a single genetic event. A similar phenomenon has been reported with the insertion of an ISEcp1-bla_CTX-M-15 tandem into the chromosomal mgrB gene of a K. pneumoniae isolate, leading to coresistance to broad-spectrum cephalosporins and colistin. However, in the latter case, coresistance related to ISEcp1 insertion was associated with expression of a resistance gene (bla_CTX-M-15) and truncation of a regulatory gene (mgrB) involved in the lipopolysaccharide biosynthesis pathway (18).

Finally, our finding further highlights how the genetic plasticity involving an IS element may play a pivotal role in acquired resistance to antibiotics by creation of mosaic structures (19). In the present report, a single IS-mediated event compromised the efficacy of two important classes of antibiotics, i.e., β-lactams and fosfomycin, in a significant nosocomial pathogen, i.e., K. pneumoniae.

Data availability.

The whole genome of isolate S54 has been deposited in GenBank under accession number JAAIKF000000000.