A total of twenty (n = 20) multidrug resistant Klebsiella-like clinical isolates were collected from the culture collection bank of the Microbiology laboratory of Kenyatta National Hospital (KNH). These isolates were sourced from various clinical specimens including tracheal aspirate, blood, urine and pus swab. Bacterial clinical isolates were subcultured on MacConkey media. Then, pure colonies were picked from the overnight culture plate based on the morphological characteristics of Klebsiella species on MacConkey followed by oxidase reaction. Another pure colony was picked from the same culture plate and inoculated into tryptic soy broth (TSB) in order to obtain only pure bacterial colonies and re-identify the Klebsiella clinical isolates using the analytical profile index (API) 20E Kit (BioMerieux, Inc., Hazelwood, MO). API is a biochemical panel from BioMerieux, Inc. for the identification and differentiation of Gram-negative bacteria of the family Enterobacteriaceae. Briefly, each well of the API strip was filled up with bacterial suspension to its neck, except the wells labelled as citrate (CIT), voges-porskauer (VP), and charcoal gelatin (GEL). The wells labelled as lysine decarboxylase (LDC), ornithine decarboxylase (ODC), arginine decarboxylase (ADH), Na-thiosulfate (H2S), and urea (URE) were filled up to the top with bacterial suspension and sterile mineral oil, respectively. After 18 - 24-hour incubation period,

Kovac’s reagent, Barritt’s reagents and ferric chloride (FeCl₃) were added the wells labelled as indole (IND), VP and tryptophan deaminase (TDA), respectively prior to the interpretation of the results. The reactions of three tests were added together at a time to obtain a 7-digit number, which was then looked up in the API WEB software for identification of the bacterial species and/or subspecies. The NCTC 13438 K. pneumoniae and sterile saline solution served as positive and negative control samples, respectively.

The multidrug-resistant profile of the bacterial clinical isolates was reconfirmed by checking the susceptibility profile of the isolates to Ceftazidime (CAZ, 30 μg), Cefotaxime (CTX, 30 μg), Cefoxitin (FOX, 30 μg), Cefepime (CPM, 30 μg), Ampicillin (AMP, 2 μg), Imipenem (IPM, 10 μg), Meropenem (MER, 10 μg), Colistin (CT, 10 μg) (HiMedia Laboratories Pvt. Limited, Mumbai, India). The isolates were further processed for detection of extended spectrum beta-lactamases (ESBL) and carbapenemase production. The antibiotic susceptibility testing was performed by following and Kirby-Bauer disk diffusion susceptibility test protocol [14]. Each bacterial suspension was adjusted to a 0.5 McFarland standard and used within fifteen minutes. A disk of augmentin (20 μg of amoxicillin plus 10 μg of clavulanic acid) was placed at the center of the culture plate, and a disk of Ceftriaxone (CTX) and Ceftazidime (CAZ) 30 mm apart (center to center) in order to analyze the production of Extended β-lactamases and carbapenemases (ESBL). Once all disks were placed, the lid was replaced and the plate was inverted, and incubated at 37˚C in Wised Cube for 18 - 22 hours. The production of carbapenemase was also analysed in bacterial clinical isolates, which showed resistant to imipenem and/or meropenem by following the guidelines of clinical and laboratory standards institute (CLSI)s [15]. Briefly, a disk of imipenem (10 μg) or meropenem (10 μg) was applied on the culture plate of E. coli ATCC® 25922™ at its center. Using a 10-μl sterile inoculating loop, 2 to 3 colonies of each clinical isolate was inoculated in a straight line at least 20 to 25 mm in length out from the edge of the disk by streaking. The plate was closed, inverted, and incubated at 37˚C in Wised Cube for 16 to 18 hours. For quality control test, Escherichia coli ATCC® 25922™ was used as negative control sample whereas Escherichia coli ATCC® 35218™ and K. pneumoniae NCTC 13438 served as positive control samples to validate the production of extended spectrum beta-lactamases and carbapenemase, respectively. All the results were interpreted in accordance with the Performance Standards for Antimicrobial Disk Susceptibility tests of CLSI and European Committee on Antimicrobial Susceptibility Testing [15] [16].

Bacterial genomic DNA and plasmid DNA were extracted from all clinical isolates and ATCC strains by using Quick-DNA™ Fungal/Bacterial Miniprep Kit (Zymo Research, California, USA) and the Zyppy™ Plasmid Miniprep Kit (Zymo Research, California, USA), respectively. Approximately 60 mg of wet colonies was resuspended in 200 μl sterile Phosphate-buffered saline in a ZR BashingBead™ Lysis Tube for extraction of genomic DNA. Regarding the Plasmid DNA extraction, pellet approach was considered over the broth medium approach. All the bacterial samples including the NCTC 13438 K. pneumoniae were processed according to the manufacturers’ instructions. The genomic DNA and plasmid DNA extracts were then separated by electrophoresis on a 1% agarose gel (AmpliSize; Bio-Rad Laboratories) for 40 minutes at 70 volts using 1× TAE Buffer containing GelRed Nucleic Acid and visualized by UV transillumination. O’GeneRuler 1 kb DNA Ladder, 0.1 μg/μL (Thermo Fisher Scientific, California, USA) was used as a marker for the amplicons’ product sizes. The mixture of the master mix and sterile saline solution was used as negative control. The quantity and quality of genomic plasmid DNA extracts were also determined using Nanodrop. The genomic and plasmid DNA extracts were stored at -20˚C for use in PCR assay.

Bacterial genomic DNA and plasmid DNA extracts were used as templates during the PCR reactions. All the clinical isolates (21) including the positive control were screened for the presence of blaNDM-1_, blaKPC_, blaOXA-48 and blaIMP genes using PCR 5x Master Mix (New England BioLabs, Massachusetts, USA) and the published primers shown in Table 1 [17] [18].

The 16S rRNA gene amplification was also performed. The mix for the detection of each gene contained 11.5 μl of nuclease free buffer, 3.75 μl of each primer, 5 μl of Multiplex PCR 5× Master Mix (New England BioLabs, Massachusetts, USA) 1 μl of template (genomic DNA). The total volume was 25 μl for the reaction. Amplification was carried out with the following thermal cycling conditions: 1 min at 95˚C for initial denaturation and 35 cycles of amplification consisting of 20 s at 95˚C, 1 min s at 51˚C for annealing, 1 minute at 68˚C, and 5 min at 72˚C for the final extension. The PCR products were separated by electrophoresis at 70 volts for 50 minutes on a 1% agarose gel (AmpliSize; Bio-Rad Laboratories) using 1× TAE Buffer containing GelRed Nucleic Acid and visualized under UV transillumination. O’GeneRuler 1 kb DNA Ladder, 0.1 μg/μL (Thermo Fisher Scientific, California, USA) was used as a marker for the amplicons’ product sizes. The clinical, which carried carbapenemase genes were also screened for the presence of ESBL genes such as blaTEM and blaOXA. DNA extracts of K. pneumoniae NCTC 13438 and Escherichia coli ATCC® 35218™ represented the control positive samples whereas the master mix plus blank control sample without DNA served as positive and negative control samples, respectively. The quantity and quality of PCR products were determined using the Nanodrop. The PCR reaction was performed in duplicate to ensure reproducibility.

The sequence analysis was performed on 16s rRNA sequences from the clinical isolate forward and reverse sequences of each sample were assembled using ChormasPro software version 2.3.1 (Technelysium Pty. Ltd., http://www.technelysium.com.au). A copy of revised sequences can be viewed on GenBank (https://submit.ncbi.nlm.nih.gov/subs/?search=SUB10035180). Briefly, the forward and the reverse sequences of each sample were imported in the programme by opening new sequencing project the “file” tab, and then adding the files in FASTA format. The wrong bases were automatically trimmed out from the raw sequences. The reference sequence, Klebsiella pneumoniae subsp. pneumoniae NTUH-K2044 (NC_012731.1) was uploaded in FASTA format and the corrected sequences were assembled and a contiguous sequence was generated. The contiguous sequence was edited by making base call while checking the quality of each base and the right peak of the chromatogram. Each corrected contiguous sequence was blasted in order to sort out the closed relative species. The best hits were selected based on the species names, the E-values and the percentage of identity. The sequences of the closed relative species to Klebsiella species and the most distant ones were considered for pairwise and global alignment using ClustalW algorithm in Molecular Evolutionary Genetics Analysis software (MEGA MX, Version 10.2.3). Yersinia enterocolitica an E. coli species were taken as outgroups. The evolutionary history was inferred by using the Maximum Likelihood method and Tamura-Nei model. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Tamura-Nei model, and then selecting the topology with superior log likelihood value. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. This analysis involved 17 nucleotide sequences. Codon positions included were 1^st + 2^nd + 3^rd + Noncoding. There were a total of 1091 positions in the final dataset. Evolutionary analyses were conducted in MEGA X.

All the twenty-one (21) isolates including the K. pneumoniae NCTC 13438 were pink (lactose fermenter), mucoid on MacConkey and oxidase-negative. The API 20 E analysis showed that all the clinical isolates did not consist of only Klebsiella pneumoniae species. The taxon of two isolates was Klebsiella oxytoca and the isolates were positive for indole and negative for Voges-Proskauer (VP). The isolates classified as Kluyver spp. were both positive for indole and VP. The other isolates (17) including the positive control were negative for indole and positive for VP, and were identified as Klebsiella pneumoniae subsp. pneumoniae (Table 2).

ONPG = Ortho-nitrophenyl-beta-D-galactopyranoside, ADH = Arginine dihydolase, LDC = Lysine decarboxylase, ODC = Ornithine decarboxylase, CIT = Citrate, H₂S = Hydrogen sulfide, URE = Urea; TDA = Tryptophan deaminase, IND = Indole, VP = Voges-Porskauer (Na pyruvate), GEL = Charcoal gelatin, GLU = Glucose, MAN = Manitol, INO = Inositol, SOR = Sorbitol, RHA = Rhamnose, SAC = Sucrose, MEL = Melibiose, AMY = Amygdalin, ARA = Arabinose, OX = Oxidase.

The susceptibility profiles of all the isolates to the antibiotics used in our study revealed that two (2) out of the 20 clinical isolates were resistant both to imipenem and meropenem (Table 3).

CAZ = Ceftazidime, CTX = Cefotaxime, FOX = Cefoxitin, CPM = Cefepime, AMP = Ampicillin, IMP = Imipenem, MER = Meropenem, CT = Colistin AMC = Amoxicillin-clavulanic acid, ESBL = Extended β-lactamases, CP = Carbapenemase, R = Resistant, S = Susceptible, I = Intermediate.

These isolates were carbapenemase producers (Figure 1) but not extended β-lactamases.

The other clinical isolates (17) were ESBL producers (Figure 2), except for one isolate.

The genomic and plasmid DNA extracts differed from each other based on the thickness of their bands (Figure 3(a)).

Out of the twenty Klebsiella clinical isolates screened for presence of four carbapenemase genes including blaNDM-1, blaKPC, blaOXA-48 and blaIMP, it appeared that three (3) isolates harboured blaNDM-1 and blaIMP as shown in Figure 4 (Figure 4(b) and Figure 4(c)) below. The positive control used separately yielded the expected band sizes for blaOXA-48 (438 bp) and BlaKPC (798). However, different band size was observed for blaIMP (275 bp). BlaOXA-48 was found in one clinical isolate on the plasmid DNA. Furthermore, all these three Klebsiella clinical isolates carried extended spectrum beta-lactamases: blaTEM and bla-OXA (Figure 4(d) and Figure 4(e)). The resistant genes were both found on the genomic and plasmid DNA as described in Table 4. The band sizes of amplicons for blaTEM and BlaOXA were 1000 and 750 bp, respectively.

N = Total number of clinical isolates carrying each resistance gene; n = Number of clinical isolate resistant to each antibiotic in association with resistance gene(s) carriage.

The band sizes of the 16S rRNA amplicons were about 1500 bp (Figure 5). The similarity analysis of carbapenem-resistant bacterial clinical isolates 16S rRNA gene sequences and other Klebsiella isolates showed that the percentage of identities between the K13, K16, K20 and Klebsiella pneumoniae strain DSM 30104 were 99.58%, 99.50% and 97.29%, respectively. Their E-values were all zero and the length of contigs ranged from 1422 to 1450 base pairs. Based on the ruling concept, strains having 16S rRNA similarity less than 97.5% most likely belong to different species whereas 16S rRNA similarity values upper than this threshold require more investigation to determine the taxonomic status [19].