In the current study, 37 carbapenem-resistant K. pneumoniae samples were collected in Shandong Provincial Hospital, China, between April 2011 and October 2013. Of these, 31 isolates were from the neonatal unit, 5 were from the pediatric ICU and 1 was from the cardiac ICU. All isolates were identified using a VITEK-2 compact 60 system (bioMérieux, Marcy l'Etoile, France). It is of note that 21 of these K. pneumoniae isolates carried with NDM-1 have been described in a previous article (11). The novel isolates were Kpn1, 3–8, 12, 14–18, 21, 24 and 39. Antibiotic susceptibility testing was performed by the agar dilution method (12). All antibiotics, except tigecycline and colisin, were administered according to the approved standard of the Clinical and Laboratory Standards Institute 2014 guidelines (13). The minimum inhibitory concentrations (MICs) of meropenem, imipenem, ertapenem, tigecycline, colisin and fosfomycin were determined by an Etest (bioMérieux). The 2014 European Committee on Antimicrobial Susceptibility Testing breakpoint (www.eucast.org/clinical_breakpoint) was used to determine resistance in colistin and tigectcline. All isolates were screened for carbapenemase production using the modified Hodge test (MHT) and imipenem-ethylenediaminetetraacetic acid (EDTA) double-disc synergistic test (14). E. coli ATCC25922 (American Type Culture Collection Center, Manassas, VA, USA) acted as the control.

Samples were screened for the presence of carbapenem resistance genes (blaKPC, blaSME, blaIMI/blaNMC, blaGES, blaIMP, blaVIM, blaGIM, blaSIM-1, blaSPM, blaNDM-1 and blaOXA), common extended-spectrum β-lactamase (ESBL) genes (blaCTX-M, blaTEM and blaSHV) and AmpC β-lactamase (AMPC) genes (blaMOX, blaFOX, blaDHA, blaCIT and blaEBC) in all 37 strains using PCR and previously described primers (15–18). The primers used are listed in Table I. The genome used as a template was extracted using a Rapid Bacterial Genomic DNA Isolation Kit (Sangon Biotech Co., Ltd., Shanghai, Chin). PCR was performed according to the manufacturer's instruction of the PCR (Taq) kit (B532501-0100; Sangon Biotech Co., Ltd.). The reaction was performed using the following thermal cycling conditions: 1 Cycle at 94°C for 5 min; 30 cycles at 94°C for 1 min, 55–58°C for 45 sec and 72°C for 1 min; and 1 cycle at 72°C for 10 min. The results of PCR were screened by electrophoresis on a 1% agarose gel and the results were sequenced by Sangon Biotech Co., Ltd.

Transfer experiments were performed using azide-resistant E. coli J53 (donated by Peking Union Medical College Hospital, Beijing, China) as the recipient strain. Overnight single colonies of donor and acceptor strains on MacConkey agar plates were inoculated into 5 ml nutrient broth containing imipenem (0.5 µg/ml) or sodium azide (200 µg/ml). These were incubated at 37°C for 8 h. Cultures of the donor strain (10 µl) and recipient strain (20 µl) were mixed with 10 ml fresh Mueller-Hinton broth and incubated for 24 h at 35°C. The mixture was then inoculated onto MacConkey agar plates containing sodium azide (200 µg/ml) and imipenem (0.5 µg/ml) for 24 h at 35°C. Conjugation was confirmed with a VITEK-2 compact system for drug sensitivity testing. The presence of carbapenemase was confirmed by MHT, imipenem-EDTA double-disc synergistic test and PCR analysis.

An overnight bacterial culture was suspended in cell suspension buffer [100 mM EDTA, 100 mM Tris-HCl (pH 8.0)] and adjusted to an optical density of 4.0 at a wavelength of 600 nm. The suspension was mixed with equal volumes of low-melting agarose in Tris-EDTA (TE) buffer in a 2% solution of low-melting-temperature agarose in TE buffer [1 mM EDTA, 10 mM Tris-HCl (pH 8.0)]. Following cooling, the agarose sections were incubated for 4 h at 54°C in cell lysis buffer [50 mM Tris-HCl, 50 mM EDTA (pH 8.0), 0.01 g/ml N-lauroyl-sarcosine, sodium salt, 0.1 mg/ml proteinase K]. These were then washed thoroughly with TE buffer and digested overnight with XbaI restriction endonuclease (Takara Bio, Inc., Otsu, Japan). DNA separation was performed in 0.5X Tris/borate/EDTA (TBE) buffer in a PFGE system (CHEF Mapper; Bio-Rad Laboratories, Inc., Hercules, CA, USA) at 14°C, using a voltage of 6 V/cm, a switch angle of 120° and a switch ramp of 4–40 sec for 21 h. The Salmonella enterica serotype Braenderup H9812 (Peking Union Medical College Hospital) was used as a marker for PFGE. The restriction patterns were analyzed and interpreted in accordance with the protocol implemented by Tenover et al (19).

The plasmid number and size of the 37 isolates, and their transconjugants, were analyzed by S1-PFGE. In summary, the gels embedded with bacterial DNA were digested with proteinase K and Sl nuclease. DNA separation was performed in 0.5X TBE buffer in the above PFGE system at 14°C with a voltage of 6 V/cm, a switch angle of 120° and a switch ramp of 2.16–63.8 sec for 20 h. S. enterica H9812 fragment, digested by XbaI enzyme, was used as a molecular weight standard. Subsequent to PFGE, the DNA products of the positive transconjugants were recovered and used as templates to amplify the carbapenem-resistance genes. The PCR reaction conditions and primers used are the same as described above. The PCR products were sequenced and searched on GenBank (www.ncbi.nlm.nih.gov/genbank).

MLST was performed on K. pneumoniae using 7 conserved housekeeping genes (gapA, infB, mdh, pgi, phoE, rpoB and tonB) according to protocols available on the MLST Pasteur website (bigsdb.pasteur.fr/klebsiella/primers_used.html).

The OMP genes were screened in all clinical isolates with previously described primers, using PCR (20). The primers used are listed in Table I. The reaction used the following thermal cycling conditions: 1 Cycle at 94°C for 5 min; 30 cycles at 94°C for 45 sec, 55°C for 45 sec and 72°C for 1 min; and 1 cycle at 72°C for 10 min. The results of PCR were screened by electrophoresis on a 1% agarose gel. The PCR negative isolates were investigated for alterations in the OMPs by SDS-PAGE, as previously described (21).

Between April 2011 and October 2013, 37 carbapenem-resistant K. pneumoniae samples were isolated from neonatal inpatients in Shandong Provincial Hospital in Jinan, China. Of these, 32 isolates were from sputum specimens, 3 were from umbilical secretions and 2 were from blood (Table II). As indicated, 21 of these K. pneumoniae isolates carried with NDM-1 have been described in a previous article (11). Of these samples, 31 were isolated from patients in the neonatal department, 5 were from the pediatric ICU and 1 was from a patient in the cardiac ICU. These patients included 23 males and 14 females. The 31 patients from the neonatal ward were all premature, with a low birth weight and neonatal pneumonia. A total of 5 infant patients had an intrauterine infection at birth. All patients were initially treated with various antibiotics, including cloxacillin, cefuroxime, and cefotetan. A total of 8 patients were treated with imipenem or meropenem. The cardiac ICU patient had a respiratory tract infection and fever 5 days after their cardiac operation. The prognosis of this cohort was poor; 2 patients succumbed to bloodstream infection.

Drug sensitivity was tested with the standard agar dilution method (Table III). These features in 21 previously described K. pneumoniae isolates are included in this analysis (11). All 37 isolates were resistant to penicillin, piperacillin-tazobactam, cephalosporins and their composite agents including the enzyme inhibitors cefoxitin and carbapenem. The percentage of the number of bacteria senstive to antibiotics divided by total strains, the sensitivity rate, to tigecycline, levofloxacin, amikacin and polymyxin were all 100%. The sensitivity rates of aztreonam, gentamicin, trimethoprim-sulfamethoxazole and fosfomycin were 13.51, 48.64, 78.38 and 86.49%, respectively.

The 37 K. pneumoniae samples displayed different degrees of resistance to carbapenems. Kpn4 was sensitive to ertapenem, with intermediate sensitivity to imipenem and meropenem. Of the 7 strains with mild sensitivity to ertapenem, 4 strains (Kpn16, 21, 3 and 8) were sensitive to imipenem and meropenem, 2 strains (Kpn5 and 6) were markedly resistant to imipenem and meropenem and Kpn39 was resistant to imipenem, with intermediate resistance to meropenem. The 19 isolates all had high resistance to the three carbapenem antibiotics; the MICs of the three carbapenems were ≥32 µg/ml.

The 37 carbapenem-resistant K. pneumoniae, including those from our previous study, all showed positive results in the EDTA synergistic test, and 7 attained positive or weakly positive results in the MHT, at a rate of 18.92% (11). Metallo-β-lactamases (MBL) were confirmed to be present in all isolates through sequencing of the PCR products. Of these, 23 isolates expressed only bla_NDM-1, and 11 isolates expressed only bla_IMP-4. Kpn4, isolated from blood, expressed only bla_IMP-8, and 2 isolates simultaneously expressed bla_NDM-1 and bla_IMP-4 (11). In addition to MBL, other types of β-lactamase were examined, including ESBLs and AMPCs. The distribution of the resistance genes in these strains is reported in Table II. The ESBL genes were determined to be bla_CTX-M-15, bla_CTX-M-14 and bla_TEM-1 in 24 (64.87%), 13 (35.14%) and 34 (91.89%) isolates, respectively. Genotyping of the AMPC genes confirmed the presence of bla_DHA-1 in 28 (75.68%) isolates.

In our previous study, the outer membrane protein genes OmpK35 and OmpK36 were detected in 21 K. pneumoniae isolates expressing NDM-1 (Kpn9, Kpn10, Kpn11, Kpn13, Kpn19, Kpn20, Kpn23 and Kpn25-38) (11). The remaining 16 strains revealed no loss of OmpK35 and OmpK36 at the gene or the protein level.

A total of 28 transconjugants were obtained from 37 isolates by transfer experiments, with a success rate of 75.68%. Of these, 20 transconjugants were obtained from 23 isolates encoding bla_NDM-1 and 8 transconjugants were obtained from 11 isolates encoding bla_IMP-4 (11). The transfer experiment failed in 2 isolates simultaneously encoding bla_NDM-1 and bla_IMP-4. All transconjugants were confirmed to have the same biochemical spectrum as E. coli strain J53, using the VITEK-2 compact system. All transconjugants demonstrated similar carbapenem susceptibility to donor strains Table IV. Compared with E. coli strain J53, the MIC of imipenem, meropenem and etarpenem increased 8- to 64-fold, 32- to 512-fold and 256- to 2,048-fold, respectively. The MIC of piperacillin tazobactam, ceftazidime, Cefatriaxone, cefoxitin and cefepime increased 8- to 2,048-fold.

MLST revealed that the 37 isolates had 5 different sequence types: ST17, ST20, ST54, ST705 and ST290 (Table III) (11). The ST20 and ST54 sequences represented the majority of the clones. PFGE revealed 7 distinct clusters amongst the 37 isolates (Fig. 1). The 3 strains from the umbilical cord all had ST54 sequences, while two strains (Kpn4 and Kpn9) isolated from blood had ST54 and ST290 sequences. Cluster B included 17 isolates, all of which had ST20 sequences. The 12 isolates with ST54 sequences were divided into clusters D and E. A cluster was defined as strains with homology >80%. There were 5 strains isolated from the pediatric ICU, 3 of which had ST705 sequences and were allocated into cluster F. The Kpn35, Kpn37 and Kpn38 samples belonged to ST17 and the same cluster E in PFGE (11).

The plasmid number and size of all isolates were analyzed by S1-PFGE, which revealed that these isolates contained 2–4 plasmids. As the primary sequences, ST20- and ST54-encoding isolates contained 2 or 3 plasmids, but their transconjugants only had 1 plasmid. The sizes of the plasmid from ST20- and ST54-encoding bacteria were 336 kb and 55 kb, respectively. The bla_NDM-1 and bla_IMP-4 were located in the 336 kb plasmid and 55 kb plasmid, respectively, as determined by PCR and sequencing.