Background

Integrons are genetic elements containing a site-specific recombination system capable of integrating, exchanging, and expressing gene cassettes (GCs). Each GC is composed of an exogenous and often promoterless gene together with a recombination site, attC [1,2,3,4,5]. The recombination system possesses an integrase gene, intI, needed for site-specific recombination, an adjacent recombination site, attI, recognized by the IntI integrase, and a promoter, Pc, located upstream of attI and necessary for efficient transcription and expression of GCs [1,2,3,4,5]. The attC site is also recognized by IntI, and recombination between attC and attI leads to the addition and exchange of GCs and further generation of a multi-GC array within the integron structure [1,2,3,4,5].

Based on the amino acid sequences of IntI integrase, integrons can be divided into different classes, with those carrying intI1 defined as class 1, intI2 as class 2, intI3 as class 3, and so on. Class 1 integrons are the most common type present among Enterobacteriaceae isolates [6,7,8]. Class 1 integrons constitute a substantial reservoir of resistance genes that confer a selective advantage upon strong selection pressure imposed by human use of antimicrobial compounds, leading to the horizontal transfer of integron-carrying resistance markers from the community to hospitals and the development of multidrug resistance (MDR) among Enterobacteriaceae, independent of species or isolate origin [6,7,8].

The ancestors of class 1 integrons are not considered to be mobile elements, and the connection of class 1 integrons with Tn402 with a complete tniABQC transposition module generates a hybrid structure flanked by the 25 bp terminal inverted repeat initial (IRi) and inverted repeat terminal (IRt), making class 1 integrons capable of self-mobility [1,2,3]. The capture and formation of a quaternary ammonium compound resistance (qacEΔ1)-sulphonamide resistance (sul1)-orf5 region occurs immediately downstream of the GC array [1,2,3]. Eventually, class 1integrons manifest as a prototype structure organized in order of IRi, a 5′-conserved segment (5′-CS: intI-attI), a central variable region (the GC array), a 3′-conserved segment (3′-CS: qacEΔ1-sul1-orf5-tniABQC), and IRt [1,2,3]. Most class 1 integrons from clinical contexts carry modifications at their 5′ and 3′ ends, especially partial or complete deletions of the tniABQC module of Tn402, which impairs their mobility [4, 5]. These integrons are often inserted within mobile DNA elements such as plasmids and transposons, facilitating their rapid spread in the community and within hospitals [4, 5].

This work presents the sequences of six novel class 1 integrons, In0, In1069 and In1287 to In1290, together with the previously characterized In893. These integrons were obtained from clinical Escherichia coli, Klebsiella pneumoniae, and Enterobacter cloacae isolates. The bla KPC-2-carrying unusual In0 and the bla IMP-30-carrying In1069 coexist in a single Escherichia coli isolate. The detailed genetic characterization of genetically related integrons In1069, In893 and In1287 to In1290 denotes a scheme of stepwise integration or excision of individual GCs to generate these integron variations.

Methods

Bacterial isolates and identification

A total of 139 non-redundant Enterobacteriaceae strains, including Escherichia coli EC6335, Escherichia coli EC4212, Klebsiella pneumoniae KP1262, Enterobacter cloacae ECL2236, Escherichia coli EC7328 and Klebsiella pneumoniae KP5325, were recovered from hospitalized patients with nosocomial infections in a teaching hospital of Taizhou, China, from January 2014 to September 2015. Bacterial species were identified by 16S rRNA gene sequencing [9]. All DNA markers listed in Table 1 were screened by PCR amplification (ThermoFisher scientific, USA) using the listed primers, followed by amplicon sequencing on an ABI 3730 Sequencer (ThermoFisher scientific, USA). PCR was run for 3 min at 94 °C followed by 30 cycles of 1 min of denaturing at 94 °C and annealing at 50 to 59 °C according to various primers, with a final elongation of 10 min at 72 °C on Life Veriti® PCR machine (Invitrogen, USA). The total reaction volume was 20 μL containing 4 μL 5 × PCR buffer, 0.4 μL of 10 mM dNTPs, 1 μL each of 10 μM primers and 0.2 μL Polymerase, with nuclease-free water filled up to 20 μL. PCR amplification and amplicon sequencing were run in according with operation manual.

Table 1 Oligonucleotide primers used in this study
Full size table

Integron cloning and sequencing

For cloning of In0 and In1069, the bla KPC- and bla IMP- positive strain EC6335 was identified by PCR. Plasmid DNA was then isolated from this strain using the AxyPrep Plasmid Miniprep kit (Axygen, USA), digested with BamHI, and ligated into the cloning vector pMD19-T. This was transformed into host bacteria Escherichia coli TOP10, which were screened for bla KPC- or bla IMP- positivity. Inserts within recombinant pMD19-T vectors from the above transformants were sequenced using the primer walking method. For cloning of In893 and In1287 to In1290, the strains EC4212, KP1262, ECL2236, EC7328, and KP5325, each co-harboring the three genes aacA4, intI1, and tniR, were identified by PCR, then DNA fragments were amplified from these strains using the primer pair intI1-F/tniR-R2 and subsequently sequenced as above.

Sequence annotation and comparison

Open reading frames were predicted using RAST 2.0 [10] combined with BLASTP/BLASTN [11] searches against the UniProtKB/Swiss-Prot database [12] and the RefSeq database [13]. Annotation of resistance genes, mobile elements, and other features was carried out using the online databases including CARD [14], ResFinder [15], ISfinder [16] and INTEGRALL [17]. Multiple and pairwise sequence comparisons were performed using MUSCLE 3.8.31 [18] and BLASTN, respectively. Gene organization diagrams were drawn in Inkscape 0.48.1 (https://inkscape.org).

Detection of carbapenemase activity

The activity of class A/B/D carbapenemases in bacterial cell extracts was determined via a modified Carba NP test [19]. Overnight bacterial cell cultures in MH broth were diluted 1: 100 into 3 mL of fresh MH broth, and bacteria were allowed to grow at 37 °C with shaking at 200 rpm to reach an OD600 of 1.0 to 1.4. If required, ampicillin was used at 200 μg/mL. Bacterial cells were harvested from 2 mL of the above culture, and washed twice with 20 mM Tris-HCl (pH 7.8). Cell pellets were resuspended in 500 μL of 20 mM Tris-HCl (pH 7.8), and lysed by sonication, followed by centrifugation at 10,000×g at 4 °C for 5 min. A total of 50 μL of the supernatant (the enzymatic bacterial suspension) was separately mixed with 50 μL each of substrates I to V: 50 μL supernatant was added to 50 μL substrate I, then separately 50 μL supernatant was added to 50 μL substrate II, and so on; followed by incubation at 37 °C for a maximum of 2 h. Substrate I: 0.054% phenol red plus 0.1 mM ZnSO4 (pH 7.8); substrate II: 0.054% phenol red plus 0.1 mM ZnSO4 (pH 7.8), and 0.6 mg/μL imipenem; substrate III: 0.054% phenol red plus 0.1 mM ZnSO4 (pH 7.8), 0.6 mg/μL imipenem, and 0.8 mg/μL tazobactam; substrate IV: 0.054% phenol red plus 0.1 mM ZnSO4 (pH 7.8), 0.6 mg/μL imipenem, and 3 mM EDTA (pH 7.8); substrate V: 0.054% phenol red plus 0.1 mM ZnSO4 (pH 7.8), 0.6 mg/μL imipenem, 0.8 mg/μL tazobactam, and 3 mM EDTA (pH 7.8).

Bacterial antimicrobial susceptibility testing

Bacterial antimicrobial susceptibility was tested by the MicroScan broth dilution method (MicroScan, USA) and interpreted according to Clinical and Laboratory Standards Institute (CLSI) guidelines [20].

Nucleotide sequence accession numbers

The sequences of In0, In1069, In893, and In1287 to In1290 were deposited in GenBank under accession numbers KP870110, KM589497, KX434463, KX371912, and KX387648 to KX387650, respectively.

Results

Clinical bacterial isolates containing integrons

Escherichia coli EC6335, Escherichia coli EC4212, Klebsiella pneumoniae KP1262, Enterobacter cloacae ECL2236, Escherichia coli EC7328, and Klebsiella pneumoniae KP5325 were determind to harbor class 1 intergrons In0 and In1069, In893, In1287, In1288, In1289, and In1290 (see Additional file 1), respetively. Modular stuctures and sequence comprasion of these integrons and related reference sequences were showed in Figs. 1 and 2. These strains were isolated from diffrent patients from the single hospital.

Fig. 1
figure 1

Genetic structure of In0EC6335 and comparison with related regions. Genes are denoted by arrows and colored according to gene function classification. Shaded areas denote regions of homology (>95% nucleotide identity). (a) Comparison of In0EC6335 to InpVS1; (b) Comparison of In0EC6335 to Tn6296 and Tn3

Full size image
Fig. 2
figure 2

Genetic structures and proposed evolutionary history of In1069 and related integrons. Genes are denoted by arrows and colored according to gene function classification. Shaded areas denote regions of homology (>95% nucleotide identity). Double-headed arrows indicate the addition or excision of relevant gene cassttes. Asterisks denote the partially sequenced coding regions of relevant genes

Full size image

A 53-year-old male with vomiting, high fever, and septic shock was admitted in the hospital in August 2015. The patient had underlying diabetes, and complained of an infected leg wound. Large doses of vasoactive drugs were needed to maintain his blood pressure, but empirical intravenous antimicrobial administration of teicoplanin plus meropenem was ineffective. Carbapenem-nonsusceptible Escherichia coli EC6335 was subsequently isolated from the wound secretions. The patient was switched to intravenous administration of levofloxacin plus fluconazole based on antimicrobial susceptibility test results. His symptoms associated with infection progressively improved and he was discharged after 10 days of antimicrobial treatment.

Klebsiella pneumoniae KP1262 and Escherichia coli EC4212 were isolated from the blood specimens in the infection unit in July 2015. Both patients complained of suffering high fevers for two to three days, and were diagnosed to have bacteriaemia. Enterobacter cloacae ECL2236 was isolated from the sputum specimens in the respiration medicine unit in January 2014. Escherichia coli EC7328 was recovered from the urine specimens in the urinary surgery ward in September 2015. Klebsiella pneumoniae KP5325 was a cultivated from the sputum in the neurosurgery unit in May 2014. All of these three patients were treated by intravenous administration of ceftriaxone according to the antimicrobial susceptibility test results, and their symptoms associated with infections were gradually recovered after several days of treatments.

Coexistence of In0 and In1069 in Escherichia coli EC6335

PCR screening indicated the presence of blaKPC and blaIMP, but none of the other tested carbapenemase genes blaNDM, blaVIM, blaOXA-23, blaOXA-48, or blaOXA-58, in strain EC6335. Further cloning and sequencing disclosed that EC6335 harbored two novel class 1 integrons: an unusual blaKPC-2-carrying In0EC6335 and a blaIMP-30-carrying In1069. In0EC6335 and In1069 (see Additional file 2) were present in two different BamHI-digested EC6335 DNA fragments, which were independently cloned into pMD19-T and transferred into Escherichia coli TOP10, generating the Escherichia coli transformants In0-TOP10 and In1069-TOP10, respectively.

Strains In0-TOP10 and In1069-TOP10 have class A and B carbapenemase activities, respectively, while EC6335 appears to have class A + B activity (data not shown). All the above strains were resistant to the cephalosporin, carbapenem, and aminoglycoside drugs tested but remained susceptible to the fluoroquinolone drugs tested (Table 2). EC6335 displayed much higher levels of resistance to cephalosporins/carbapenems than In0-TOP10 and In1069-TOP10 (Table 2), which was consistent with the fact that EC6335 harbors two carbapenemase genes, blaKPC-2 and blaIMP-30, while In0-TOP10 and In1069-TOP10 carry only one gene, blaKPC-2 or blaIMP-30, respectively.

Table 2 Antimicrobial drug susceptibility profiles
Full size table

Discussion

Genetic futures of novel In0EC6335

In0 EC6335 differs dramatically from the prototype In0 pVS1 (accession number U49101) from the Pseudomonas aeruginosa plasmid pVS1, and they only overlap each other by a majority region of 5′-CS (Fig. 1). In0 pVS1 has a weak PcW promoter and an unoccupied attI site with no gene cassettes, thereby representing an ancestor of more complex integrons [21]. In0 pVS1 possesses an intact 5′-CS, whereas its 3′-CS is disrupted by the insertion of IS1326, leading to the truncation of tniABQC of Tn402 into tniA-tniBD1 [21].

In0 EC6335 contains a strong PcH1 promoter and a truncated GCaacA4’-3 composed of ∆aacA4’-3 and attC aacA4; a bla KPC-2-containing region of Tn6296 has been integrated between the PcH1 promoter and the truncated GCaacA4’-3, and located downstream of the GCaacA4’-3 remnant are a bla TEM-1-containing region of Tn6296 and an ISKpn27- containing region from Tn6296 (Fig. 1). AacA4’-3 (555 bp in length; aminoglycoside resistance) is a derivative of the reference aacA4 gene (accession number AF034958), and encodes the variations Asn5Thr and Leu102Ser compared with aacA4. The ∆aacA4’-3 gene of In0 EC6335 has a 36 bp deletion at its 5′ end. It is most likely that In0 EC6335 originates from a GCaacA4’-3-carrying integron that has recombined with Tn6296, which is one of the major mobile platforms of bla KPC genes in China [22,23,24], and Tn3 as a major mobile platform of bla TEM-1 [22,23,24].

Inferred evolution of In1069 and related integrons in Enterobacteriae

In1069 carries two GCs, GCbla IMP-30 (carbapenem resistance) and GCaacA4’-3, in this order, and its intI1 gene is truncated because of the connection of IS26 at its 3′ end. We also determined the sequences of four genetically related integrons, In893 and In1287 to In1290, from three different Enterobacteriaceae species: Escherichia coli, Klebsiella pneumoniae, and Enterobacter cloacae (Fig. 2). In1069 and the other four integrons carry PcS and PcH1, which are both strong promoters known to drive GC expression [21].

These integrons appear to have conserved 5′-CS and 3′-CS (tni of Tn402), and possess a core GCaacA4’-3 in which the attC site has an unusual 9 bp deletion (CCCTTCCAT) (Fig. 2). Five more GCs, including GCbla IMP-30, GCaadA16 (aminoglycoside resistance), GCcatB3 (phenicol resistance), GCarr3, and GCdfrA27 (rifampin resistance), are also present in these integrons. The integration/excision of one or more of these GCs upstream/downstream of the GCaacA4’-3 core generates various organizations of GC arrays, which are mediated by the IntI-based, attC-recognizing site-specific recombination system [25]. In893 seems to represent the most primitive form of these integrons. In893, In1069, and In1287 carry one or two different resistance markers, while In1288 to In1290 have evolved to capture the determinants for at least three different classes of antibiotics, mostly likely conferring MDR. These findings provide the insight into stepwise and parallel evolution of In1069-associated integron variations (Fig. 2).

Conclusions

Excessive use of antibiotics causes the spread of MDR Enterobacteriaceae strains in clinical settings, most of which harbor class 1 integrons. The characterization of novel class 1 integrons In0, In1069 and In1287 to In1290 denotes a step-by-step and parallel evolution scheme involving massive genetic changes in integron GC arrays under high levels of antibiotic selection pressure in clinical settings.