Introduction

Leclercia adecarboxylata, which is ubiquitously distributed in nature, is a motile, aerobic member of Enterobacteriaceae and it shows high degree of phenotypic similarity to Escherichia coli. L. adecarboxylata infections are rarely reported in humans, emphasizing the nature of this bacterium as an opportunistic pathogen1,2,3. In most cases, L. adecarboxylata is isolated as a pure culture from immunocompromised persons or patients with underlying medical conditions and its can be occasionally found as a part of polymicrobial cultures in immunocompetent patients suggesting the dependence of this microorganism on co-flora to cause a disease1,2,3. In addition, only two cases of L. adecarboxylata-induced monomicrobial infections–without other coinciding pathogens–have also been reported in immunocompetent patients, indicating the relevant isolates may possess unique virulence factors not found in the other reported clinical isolates4,5. It has been postulated that L. adecarboxylata infections are underestimated and have been under-reported for a long time due to the fact of misidentification of this microorganism as Escherichia coli6.

L. adecarboxylata strains are naturally susceptible to all but two β-lactams (see below), azithromycin, tetracyclines, aminoglycosides, quinolones and amphenicols, but resistant to penicillin G and oxacillin, erythromycin, roxithromycin and clarithromycin, fosfomycin, ketolides, lincosamides, glycopeptides and rifampicin6. There are few reports describing the antimicrobial resistance of L. adecarboxylata due to acquisition of foreign resistance gene(s). The SHV-12-mediated cephalosporin resistance7, or the carbapenem resistance due to production of carbapenemase KPC-28 or VIM-19 has been observed in clinical L. adecarboxylata. In addition, reported are two clinical isolates of multidrug-resistant (MDR) L. adecarboxylata, one harboring blaTEM-1, blaCTX-M-3 and a class 1 integron cassette array dfrA12-orfF-aadA210 and the other possessing blaSHV-12, blaDHA-1-ampR and a class 1 integron cassette array aacA4cr-blaOXA-1-catB3-arr311.

We recently reported a fully sequenced resistance plasmid pP10164-NDM, harboring a total of two resistance genes blaNDM (conferring carbapenem resistance) and bleMBL (conferring bleomycin resistance), from the clinical L. adecarboxylata isolate P1016412. Strain P10164 is resistant to β-lactams including carbapenems, quinolones, aminoglycosides, macrolides, fosfomycin, tetracyclines, amphenicols and trimethoprim/sulfamethoxazole but remained susceptible to tigecycline and polymyxin E. This follow-up study provides the evidence for the presence of two additional resistance plasmids pP10164-2 and pP10164-3 in L. adecarboxylata P10164. These two multidrug-resistant (MDR) plasmids were fully sequenced and shown to carry a large amount of antibiotic and heavy metal resistance genes.

Results and Discussion

Overview of plasmids pP10164-2 and pP10164-3

The complete sequences of pP10164-2 and pP10164-3 were determined from the genomic DNA of strain P10164 by high-throughput shotgun sequencing (the mean sequencing coverages are 79 × and 93 × respectively) and PCR-based gap closing. These two plasmids have circularly closed DNA sequences, 313,395 bp and 80,460 bp in length with mean G + C contents of 47.3% and 54.1%, respectively and they contain 356 and 91 predicted open reading frames (ORFs) in total, respectively (Fig. 1). The modular structure of each plasmid is discriminated as the backbone with insertion of multiple separate accessory modules.

Figure 1
figure 1

Schematic maps of pP10164-2 (a) and pP10164-3 (b). Genes are denoted by arrows and colored based on gene function classification. The innermost circle presents GC-Skew [(G − C)/(G + C)] with a window size of 500 bp and a step size of 20 bp. The blue circle presents GC content. Shown also are backbone and accessory module regions.

The pP10164-2 backbone, 205 kb in length, is closely related (97% query coverage and maximum 99% nucleotide identity) to the prototype IncHI2 plasmid R478 from Serratia marcescens13 and almost identical (100% coverage and 99% identity) to another IncHI2 plasmid pKST313 from Salmonella enterica serotype Typhimurium14. Located in the pP10164-2 backbone are genes or gene clusters that encode the core IncHI2 plasmid determinants such as repHIA and repHI2 (replication initiation), the tra1 and tra2 regions (conjugal transfer), parAB and parM-parR (partition) within tra2, ter (tellurium resistance), klaABC (plasmid maintenance) and ars (arsenic resistance). It has been proposed that the repHIA replicon, the essential trh (conjugal transfer), tra and oriT (origin of transfer) sequences within tra1 and tra2 and the parAB partitioning module might represent the minimal IncHI2 determinants13.

The pP10164-2 accessory regions, which are dramatically different from R478 and pKST313, are composed of the group IIB1 intron Kl.pn.I2, ISKpn26, two IS903D elements, a novel insertion sequence (IS) of IS3-family designated ISLad1, a novel IS element of IS1202 group named ISLad2 and two novel MDR regions designated MDR-1 and MDR-2. The MDR-1 and MDR-2 regions, 61.3 kb and 40.3 kb in length respectively, are adjacent and isolated by a 1.9 kb backbone region composing of two ORFs orf381 and ∆orf666.

The pP10164-3 backbone encodes the plasmid replication (repA) and maintenance (parFG and umuCD) functions as well as the residual conjugal transfer determinants (traA, mutated nikAB and mobC) and overall it exhibits no significant sequence similarity to any known DNA sequences. The deduced replication initiator protein RepA belongs to the Rep_3 superfamily and cannot be assigned into any known incompatibility groups and it matches various plasmid RepA proteins of unknown incompatibility groups from Leclercia, Cronobacter and Enterobacter with above 93% amino acid identity.

pP10164-3 is quite unusual because it has a relatively small (19 kb in length) backbone but carries much larger accessory contents including the 2.3 kb Kl.pn.I2 intron and a 37.8 kb region composed of a MDR region and a carbohydrate utilization region. The carbohydrate utilization region is sequentially organized as a mutated sequence of a novel IS1-family member (MIS1), a novel 14-gene locus probably accounting for galactan utilization, IS1F, a mutated sequence of a novel IS3-family member (MIS3) and a Tn2555 remnant. Both MIS1 and MIS3 cannot be discriminated as intact IS elements because their transposase genes insB and tnpA, respectively, becomes pseudogenes due to frameshift. The sucrose transposon Tn2555 from E. coli is an IS26-based composite transposon that carries the sucrose utilization gene cluster scrKYABR, two direct IS26 copies on its flanks and, sometimes, a third inverted IS26 copy inside the transposon15, while the Tn2555 remnant from pP10164-3 containing only scrK and ∆scrY.

The MDR-1 region of pP10164-2

The pP10164-2 MDR-1 region (Fig. 2a) is organized sequentially as a novel Tn3-family unit transposon designated Tn6317, the Tn3-IS26-blaSFO-1 unit, a Tn1548-associated region, IS26, In27pP10164-2, the IS26-tetA(C)-tetR(C) unit, IS26 and a Tn5396-like transposon remnant.

Figure 2
figure 2

The pP10164-2 MDR-1 region and comparison to related regions.

Genes are denoted by arrows and colored based on gene function classification. Shading regions denote regions of homology (>95% nucleotide similarity).

Tn6317 is generated from the insertion of Tn5058b into a backbone remnant of Tn6256, a Tn3-family TnPa38-related transposon from clinical Citrobacter freundii from Italy16. Each of the two 39 bp terminal inverted repeats (IRL: inverted repeat left; IRR: inverted repeat right) of Tn6317 is disrupted by IS4321R into two separate parts (IR-5′ plus IR-3′), which is also observed in Tn6256. It seems that the Tn5058b insertion is accompanied by not only the truncation of IS4321R but also the loss of downstream IRL-3′ and the core transposition module tnpA (transposase) at the 5′ region of Tn6317 relative to Tn6256 (Fig. 2b). Tn5058b is composed of a Tn5053-family core transposition module tniA (transposase)-tniB (ATP-binding protein)-tniQ (transposition auxiliary protein)-res (resolution site)-tniR (serine resolvase) and two mercury resistance gene clusters named mer1 and mer2, which is delimited by terminal 25 bp IRL and IRR.Tn5058b differs from the prototype Tn5058 (accession number Y17897) from Pseudomonas sp. ED23-33 by the insertion of IS5075 into each of the two internal inverted repeats IIRmerT and IIRmerR2. The IS1111-family IS4321 and its close derivative IS5075 are known to target the terminal inverted repeats of the Tn21-subgroup transposons of Tn3 family17.

The Tn3-IS26-blaSFO-1 unit is likely derived from a precursor Tn3 [IRL-tnpA-res-tnpR (resolvase)-blaTEM-1-IRR], which has undergone at least two major evolutionary events (Fig. 2c): i) The disruption of the 38 bp IRL of Tn3 into IRL-5′ and IRL-3′ by IS4321R; and then ii) the insertion of the IS26-blaSFO-1-IS26 unit (which is known to be transposable among plasmids18) upstream of IS4321R, leading to the truncation of IS4321R, the loss of IRL-3′-tnpA-res of Tn3 and the truncation of tnpR of Tn3. The connection of Tn3-IS26-blaSFO-1 with Tn6317 orientated in opposite directions likely results in the loss of the IRR of Tn3, making Tn3-IS26-blaSFO-1 cannot to be discriminated as a transposon due to the absence of one of the paired IRL/IRR routinely bracketing at both ends. Both blaTEM-1 and blaSFO-1 encode class A β-lactamases, whose activity can be inhibited by clavulanic acid. TEM-1 is able to hydrolyze penicillins but not extended-spectrum cephalosporins; by contrast, SFO-1 exhibits significant hydrolytic activity against both penicillins and extended-spectrum cephalosporins, but it has no detectable activity against carbapenems and cephamycins19. The blaSFO-1 expression is inducible, which is regulated by the transcriptional activator encoded by ampR that is inversely orientated upstream of blaSFO-120.

Tn1548 is an IS26-based composite transposon from the C. freundii plasmid pCTX-M3 and displays a modular structure IS26-In27-ISCR1-∆ISEc28-armA-ISEc29- msr(E)-mph(E)-orf543-IS2621,22. Notably, Tn1548 lacks the paired short direct repeats (DRs), which represent the target site duplication signals routinely bracketing at both ends of a composite transposon. Tn1548 and various Tn1548-asscoiated elements (with insertion of different class 1 integrons or integron-like sequences between IS26 and ISCR1) are thought to promote the dissemination of the aminoglycoside resistance gene armA, the macrolide resistance operon msr(E)-mph(E) and other classes of antibiotic resistance genes within the inserted integrons23. The Tn1548-asscoiated region from pP10164-2 differs from Tn1548 by the replacement of In27 by a novel class 1 integron named In1262 and the deletion of orf543-IS26 originally at the 3′ region of Tn1548 (Fig. 2d). The connection of immediately upstream IS26 and immediately downstream ISCR1 with In1262 leads to the loss of two terminal 25 bp inverted repeats (IRi: inverted repeat initial; IRt: inverted repeat terminal) and the truncation of intI1 (integrase) occurred for this integron. In1262 carries two gene cassettes gcu167 and aacA3 (aminoglycoside resistance):attCaacA3. The novel gene cassette gcu167 of unknown function contains two consecutive ORFs gcu167a (putative nudix hydrolase) and gcu167b (putative nucleotidase), followed by a single attCgcu167 site.

In27pP10164-2 resembles a complex class 1 integron, whose modular structure can be generally divided sequentially into 5′-conserved segment [5′CS: intl1-attI], variable region 1 (VR1), the first copy of 3′-conserved segment [3′CS1: qacE∆1 (quaternary ammonium compound resistance)-sul1 (sulfonamide resistance)], common region ISCR1, VR2 and the second copy of 3′CS (3′CS2), bordered by terminal 25 bp IRi and IRt24. In27pP10164-2 comprises ∆5′CS (∆intl1-attI), VR1 [three sequentially arranged gene cassettes: dfrA12 (trimethoprim resistance):attCdfrA12, gcuF (unknown function):attCgcuF and aadA2 (aminoglycoside resistance):attCaadA2], 3′CS1, VR2 [orf639 (putative β-lactamase)-IS1 × 4], an ISCR-like element, 3′CS2, orf5 and ∆orf6; ∆5′CS and ∆orf6 are in truncated formats and IRi and IRt are absent, which is likely resulted from the connection of IS26 at both ends of In27pP10164-2. The common region ISCR1, which is commonly located between 3′CS1 and VR2 of a typical complex class 1 integron, is not found in In27pP10164-2, but a 1.4 kb element (which encodes a putative protein with 80% amino acid sequence similarity to the ISCR20 transposase but lacks the oriIS element characteristic of ISCRs) is found between VR2 and 3′CS2 of In27pP10164-2.

IS26-tetA(C)-tetR(C) represents a putative mobile unit carrying a tetracycline resistance module tetA(C) (tetracycline efflux protein)-tetR(C) (transcriptional repressor of tetA); moreover, similar genetic elements are found in various plasmids such as the IncN1 plasmid N325, the IncHI2 plasmids pMRVIM0813 (accession number KP975077), pSTm-A54650 (LK056646) and pKST31314 and the partially sequenced plasmid pQKp274H26. Located at the 3′ terminus of the MDR-1 region is a 2.9 kb transposon remnant, which contains the 38 bp IRL and a pseudogene of tnpA (with truncation and frameshift) and shows 96% nucleotide sequence identity to the Tn3-family transposon Tn539627.

At least 6 copies of IS26 are found in the MDR-1 region and can be arbitrarily assigned into the four structures IS26-In1262-ISCR1-∆ISEc28-armA-ISEc29-msr(E)-mph(E)-IS26, IS26- blaSFO-1-IS26, IS26-In27pP10164-2-IS26 and IS26-ydiB-tetA(C)-orf378-tetR(C)-orf477 -IS26. Each of them contains two terminally flanking IS26 elements but cannot be annotated as a composite transposon, because the paired DR sequences are not identified. The common component IS26 would act as an adaptor to mediate massive recombination and transposition events28,29, facilitating the assembly of the MDR-1 region with a very complex mosaic structure.

The MDR-2 region of pP10164-2

The pP10164-2 MDR-2 region (Fig. 3) is mainly composed of IS26, ∆In705, ∆Tn2670, a novel Tn3-family unit transposon designated Tn6322, ISKpn19, ∆ISPpu12 lacking IRL, sil, IS1R, IS903D, ∆cop and rcn in order of their priority. ∆In705 contains ∆5′CS (∆intI1-attI, truncated by connection of IS26 upstream of ∆In705) and a single gene cassette aadA1ai (aminoglycoside resistance):attCaadA1ai. The aadA1ai gene is a derivative of the prototype aadA1 gene (accession number X12870), displaying the Val5Met amino acid substitution.

Figure 3
figure 3

The pP10164-2 MDR-2 region and comparison to related region.

Genes are denoted by arrows and colored based on gene function classification. Shading regions denote regions of homology (>95% nucleotide similarity).

Tn2670 is an IS1-based composite transposon, which is composed of a backbone region with Tn21 inserted within it30 and originally found in the MDR plasmid R100 (accession number AP000342) from Shigella flexneri. The In2670 backbone consists of two IS1 elements flanking a 1.5 kb central region that harbors the amphenicol resistance gene catA1 and the ybjA gene encoding putative acetyl transferase30. ∆Tn2670 from the pP10164-2 MDR-2 region resembles the In2670 backbone but lacks the right terminal IS1 and, notably, similar structures are found in other IncHI2 plasmids such as pRH-R2731 and in the chromosomally located resistance island AbaR1 and its derivatives from Acinetobacter baumannii32.

Tn6322 is composed of the Tn21 core transposition module tnpAR-res33 together with a novel mercury resistance gene cluster designated mer3 and the mer3 region differs dramatically (92% coverage and maximum 86% nucleotide identity) from the mer locus from Tn21, indicating the capture of mer3 by the Tn21 core transposition module during the genesis of Tn6322. The mer3 region is mostly similar (100% coverage and maximum 96% nucleotide identity) to the counterpart of the Enterobacter cloacae transposon Tn6005 belonging to the Tn5036/Tn3926 subgroup of Tn3 family34. Tn6322 is flanked of 38 bp IRL/IRR resembling those of Tn21: the IRR is intact, while the IRL shows the insertion of IS5075.

Silver and copper compounds are used as antimicrobial agents in hospitals and the relevant resistance determinants could serve as hygienic fitness factors and thus improve bacterial survival in hospital environments. In R478, the silver and copper resistance gene clusters, called sil and cop respectively, are located adjacently and associated with an upstream Tn7-like core transposition module tnsABCD. Similar tnsABCD-sil-cop structures are widely found in IncHI2 plasmids such as pMRVIM0813 (accession number KP975077), pSTm-A54650 (LK056646), pKST31314 and pRH-R2731, although considerable variations in both genetic content and nucleotide sequence are observed among different plasmids. Similarly, a multi-heavy metal resistance region ISKpn19-∆ISPpu12-sil-IS1R-orf1623-∆cop-rcn is found in the pP10164-2 MDR-2 region: compared with the prototype tnsABCD-sil-cop structure, ISKpn19-∆ISPpu12 replaces tnsABCD, the insertion of IS1R-orf1623 (putative metal-dependent hydrolase)-IS903D between sil and cop marked the truncation of cop into ∆copS-copE and a rcn locus (encoding the RcnA efflux pump responsible for nickel/cobalt detoxification and the rcnA repressor RcnR) is added immediately downstream of ∆copS-copE. Notably, the IncHI2 plasmid pRH-R2731 carries a very similar structure from ∆ISPpu12 to rcn with further insertion of a fragment composed of three hypothetical ORFs between IS1R and orf162331. The MDR-2 region of pP10164-2 and the corresponding MDR region of pRH-R2731 are genetically related and might share a much more recent ancestor, although they contains dramatically different sets of resistance genes upstream of the ∆ISPpu12 to rcn region.

The MDR region of pP10164-3

The pP10164-3 MDR region (Fig. 4) is 35.5 kb in length and can be divided into two components, namely a 4.9 kb qnrS1 (quinolone resistance) region and a novel Tn3-family unit transposon designated Tn6308. The qnrS1 genetic platform ∆ISEcl2-qnrS1- ∆tnpR (truncated Tn3-family resolvase)-ISKpn19 is widely found in resistance plasmids from Enterobacteriaceae species35. Replacement of the 5′ terminal ∆ISEcl2 by ∆ISKpn19 generates a novel qnrS1 region ∆ISKpn19-qnrS1-∆tnpR-ISKpn19 as observed in the pP10164-3 MDR region.

Figure 4
figure 4

The pP10164-3 MDR region.

Genes are denoted by arrows and colored based on gene function classification.

The Tn6308 backbone is a hybrid of the core transposition module tnpAR-res of Tn1696 and the mer region of Tn21 and it is bordered by the intact 39 bp IRL and the IS5075-disrupted IRR at both ends in the absence of DRs. Tn1696 and Tn21 are both the members of the Tn21 subgroup of Tn3 family, but they have independent histories and origins with limited nucleotide sequence similarity (79 to 96%) between corresponding backbone genes36. The res site, originally 120 bp in length, is truncated into an 83 bp remnant in Tn6308 due to the insertion of a class 1 integron In37b. Notably, all the three novel Tn3-family transposons Tn6317, Tn6322 and Tn6308 identified in this work have undergone at least two evolutionary events after their initial transposition into pP10164-2 or pP10164-3: i) the disruption of one or both terminal IR sequences by insertion of IS5075 or IS4321R, making them deficient in further mobilization; and ii) the removal of target site duplication signals, making them lack of terminal DR sequences.

In37b from pP10164-3, In37c from the C. freundii plasmid p112298-KPC37 and In37d from the Aeromonas plasmid pP2G138 are all derivatives of the typical complex class 1 integron In37 from Escherichia coli39 (Fig. 5). In37 is sequentially organized as 5′CS, VR1 [aacA4cr (quinolone and aminoglycoside resistance):attCaacA4cr], blaOXA-1 (β-lactam resistance):attCblaOXA-1, catB3 (amphenicol resistance):attCcatB3 and arr3 (rifampicin resistance):attCarr3], 3′CS1, ISCR1, VR2 [qnrA1 (quinolone resistance) and ampR (LysR-family regulator)], 3′CS2 and orf5-orf6-IS6100, which is bracketed by 25 bp IRi/IRt and associated 5 bp DRs39.

Figure 5
figure 5

Comparison of In37b with its derivatives.

Genes are denoted by arrows and colored based on gene function classification. Shading regions denote regions of homology (>95% nucleotide similarity).

Compared to In37, In37b has undergone the insertion of Tn6309 into intI1, the loss of ISCR1-VR2-3′CS2, the truncation of orf5-orf6 into ∆orf5 due to the insertion of the chromate resistance unit IRLchrA-chrA-orf9840 (the 38 bp IRLchrA is further disrupted by IS5075) and the replacement of IS6100 by the macrolide resistance unit IS26-mph(A)-mrx-mphR(A)-IS610040 followed by ∆tniATn21 (Fig. 5). Tn6309 is an IS26-based composite transposon containing the tetracycline resistance module tetA(C)-tetR(C); although just being named in this work, Tn6309 has been found in the genomic island Sm1-MDRGI from Stenotrophomonas maltophilia41 and also in the three sequenced plasmids pB3 from Pseudomonas sp. GFP142, pKAZ3 from an uncultured bacterium43 and pNDM-116-14 (accession number LN831184) from Vibrio cholerae. The absence of IRi and ∆tniATn21 and the truncation of intI1 are observed but there is no insertion of Tn6309 in In37c compared to In37b (Fig. 5). Insertion of Tn6309 and IRLchrA-chrA-orf98 are not found in In37d, leaving intI1 and orf5-orf6 intact, but ∆tniATn21-IRt is absent from In37d compared to In37b (Fig. 5). The above observations indicate that extensive recombination and transposition events have occurred during derivation of In37, In37b, In37c and In37d from an In37-like precursor, making them to integrate different sets of additional resistance genes, but the core resistance gene cassette array aacA4cr-blaOXA-1-catB3-arr3 is shared by these integrons.

Concluding remarks

This is the first report of detection of MDR plasmids and determination of their complete sequences in L. adecarboxylata. Coexistence of three resistance plasmids pP10164-NDM, pP10164-2 and pP10164-3 makes L. adecarboxylata P10164 tend to become extensively drug-resistant. This bacterial species may serve as a potential reservoir of antimicrobial resistance genes in clinical settings. Data presented here would promote us to gain deeper understanding of plasmid-mediated mechanisms of drug resistance in L. adecarboxylata. Prevalence of the resistance plasmids pP10164-NDM, pP10164-2 and pP10164-3 in L. adecarboxylata and other bacterial species from the clinical settings cultures especially those from immunocompromised patients needs to be elucidated.

Methods

Bacterial genomic DNA were isolated by classical phenol/chloroform method followed by diethyl ether removal of polysaccharides that contaminate genomic DNA44 and then sequenced with a paired-end library with an average insert size of 500 bp and a mate-pair library with average insert size of 5,000 bp, using HiSeq 2500 sequencer (Illumina, CA, USA). In order to get complete plasmid sequences, the contigs were assembled with Velvet and the gaps were filled through combinatorial PCR and Sanger sequencing on ABI 3730 Sequencer (LifeTechnologies, CA, USA).

The open reading frames and pseudogenes were predicted with GeneMarkS™ (http://topaz.gatech.edu/GeneMark), RAST (http://rast.nmpdr.org/) and Prodigal (http://compbio.ornl.gov/prodigal) and further annotated by BLASTP and BLASTN against UniProtKB/Swiss-Prot (http://web.expasy.org/docs/swiss-prot_guideline.html) and NCBI NR databases.

Annotation of resistance genes, mobile elements and other gene futures was based on the relevant databases including CARD (http://arpcard.mcmaster.ca), BacMet (http://bacmet.biomedicine.gu.se/), β-lactamases Database (http://www.ncbi.nlm.nih.gov/pathogens/submit_beta_lactamase), ISfinder (https://www-is.biotoul.fr/), ISCR Elements Databases (http://medicine.cf.ac.uk/infect-immun/research/infection/antibacterial-agents/iscr-elements), INTEGRALL (http://integrall.bio.ua.pt/), Tn Number Registry (http://www.ucl.ac.uk/eastman/research/departments/microbial-diseases/tn) and Group II Introns Databases (http://webapps2.ucalgary.ca/~groupii/blast.html). Sequence comparison was performed with BLASTN and CLUSTALW and gene organization diagrams were drawn with Inkscape (https://inkscape.org). The complete sequences of pP10164-2 and pP10164-3 were submitted to GenBank under accession numbers KX710093 and KX710094, respectively.

Additional Information

How to cite this article: Sun, F. et al. Genetic characterization of two fully sequenced multi-drug resistant plasmids pP10164-2 and pP10164-3 from Leclercia adecarboxylata. Sci. Rep. 6, 33982; doi: 10.1038/srep33982 (2016).