Clonal diversity and genetic profiling of antibiotic resistance among multidrug/carbapenem-resistant Klebsiella pneumoniae isolates from a tertiary care hospital in Saudi Arabia

Bacterial isolates and testing of antibiotic resistance

Seventy-one non-redundant multidrug-resistant CRKP isolates were obtained from patients attending the Microbiology Section of the King Abdulaziz Medical City, Riyadh, Saudi Arabia, between January 2011 and December 2012. The isolates were identified to the species level using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (bioMérieux, Marcy-l’Étoile, France). The frozen isolates were sub-cultured on blood agar at 37 °C overnight. Single colonies were picked up and grown in liquid tryptic soy broth(TSB) medium 37 °C in a shaking incubator at 250 rpm overnight. Minimum inhibitory concentrations (MICs) were determined using a Micro VITEK® 2 microbial identification instrument (bioMérieux). The MIC breakpoints for carbapenem were defined according to the modified 2010 Clinical and Laboratory Standards Institute guidelines [7]. Isolates found to have elevated MICs for carbapenem by Micro VITEK® 2 analysis were confirmed by the manual ETEST® (bioMérieux) to have reduced susceptibility to either imipenem or/and meropenem.

Characterization of resistance genes

Bacterial DNA was isolated using a MagNA Pure kit (Roche Diagnostics, Risch-Rotkreuz, Switzerland). Uniplex polymerase chain reaction (PCR) was performed for bla-SHV, bla-TEM-1, bla-CTX-M, bla-KPC, bla-OXA-48, bla-VIM, and bla-IMP using primers reported previously [8]. Appropriate positive and negative controls were run simultaneously. Each PCR reaction was conducted at least twice. The primers 5’-GGTTTGGCGATCTGGTTTTC-3′ (F) and 5’-CGGAATGGCTCATCACGATC-3′ (R) were used for the amplification of bla-NDM-1, and the primers 5‘-CACCTCATGTTTGAATTCGCC-3′ (F) and 5‘-CTCTCTCACATCGAAATCGC-3′ (R) were used to determine its genetic environment [9]. The PCR products were sequenced on an ABI 3100 DNA analyzer (Life Technologies, Carlsbad, CA, USA). The DNA sequences were analyzed using SeqMan (DNASTAR, Inc., Madison, WI, USA) and MEGA4[10] and blasted against GenBank (National Center for Biotechnology Information, Bethesda, MD, USA). PCR-based replicon typing (PBRT) was performed on these isolates to identify the plasmids as per the protocol described in the PubMLST database (https://pubmlst.org/plasmid/).

Conjugation experiments

Conjugation experiments were performed on selected isolates (Table 2) using J53Az^r Escherichia coli, a strain resistant to sodium azide, as a recipient (a kind gift from Dr. Timothy Walsh, University of Cardiff, Wales, UK). Cultures from both donor and recipient bacteria were grown in lysogeny broth (LB) to the logarithmic phase, mixed at a donor:recipient ratio of 1:3, and incubated at 37 °C overnight without shaking. Transconjugants were selected on LB agar plates supplemented with 100 μg/ml of sodium azide and 100 μg/ml of ampicillin. Positive clones were screened by colony PCR for the target resistance genes.

Detection of hypermucoviscocity phenotyping and virulence associated genes

Bacterial isolates grown on blood agar plates were checked for hypermucoviscocity. Hypermucoviscocity was considered present when an inoculation loop formed a viscous string of 5 mm or more in length from a bacterial colony [11]. Capsular polysaccharide synthesis (CPS) and virulence-associated genes were detected by multiplex PCR as described by Compain et al. [12]. The presence of CPS genes was reconfirmed by uniplex PCR.

Multilocus sequence typing, allelic diversity, and population structure analysis

Multilocus sequence typing (MLST) was conducted according to the method of Diancourt et al. [13] and sequence types (STs) were assigned through the Institut Pasteur (Paris, France) database (http://bigsdb.pasteur.fr/klebsiella/klebsiella.html). Individual locus sequences, as well as the concatenated sequences of all seven MLST loci for each isolate, were analyzed for their diversity using DIVEIN [14]. Clonal groups were defined based on STs sharing six loci (single-locus variants) using eBURST [15]. A population snapshot was also drawn for the clonal relationship of these STs with those in the database of the Institut Pasteur using eBURST. Using the alignments of the concatenated sequences from the seven MLST loci for all isolates, a phylogenetic tree was constructed by the neighbor-net method using SplitsTree4 [16].

Isolates and their minimum inhibitory concentrations

Results on resistance determinants, virulence factors, plasmid types, and clonality are presented in Table 1. The clinical sources of the specimen were as follows; respiratory (n = 24); surgical wound (n = 14); rectal swabs (n = 14); urine (n = 13); blood (n = 9); and abdominal drainage (n = 2). Isolates RD-121 and RD-122, which were retrieved from different sites (rectal swab and urine sample) of a single patient 6 months apart, belonged to different STs (STs-15 and -11, respectively) and were considered independent isolates. All isolates were resistant to more than three types of antibiotic. Isolates with ST-199 were all resistant to amikacin (15/15), ‘intermediate’ to tobramycin (9/11), and susceptible to gentamicin (14/15). In contrast, isolates with ST-16 tended to be more susceptible to amikacin and gentamicin than tobramycin. ST-340 isolates were susceptible to amikacin alone. MIC levels are shown in Additional file 1.

Genes of resistance and their clonal distributions

The OXA-48 gene was the most common carbapenemase gene, found in 54/71 isolates, followed by the metallo-β-lactamase gene NDM-1. Of the 54 isolates positive for OXA-48, 48 exhibited OXA-48 alone and six exhibited OXA-48 in combination with NDM-1. Sequence analysis of the ~ 740-base pair PCR product revealed an OXA-48 gene allele sensu stricto in all the positive isolates. OXA-48 was found in the majority of the clones in this study; however, it was less prevalent in STs-152, − 340, and − 15. Among them, four of 12 isolates of ST-152, none of five isolates of ST-340, and one of three isolates of ST-15 were positive for this OXA-48 gene. The gene NDM-1 was less common, existing as the lone carbapenemase in nine isolates and in combination with OXA-48 in six isolates. The main clone carrying this gene was ST-152, with all of its 12 isolates positive for NDM-1. No other allelic form of this gene was detected in any isolate. The ~ 1-k base amplicon generated to study the genetic environment surrounding NDM-1 revealed the presence of ISAba125 and ble_MBL flanking this gene. The co-existence of two carbapenemase genes (NDM-1 + OXA-48) was seen in six of 71 (8.5%) isolates. Eight isolates were negative for both the OXA-48 and NDM-1 genes. This included three isolates from ST-340, two from ST-15, and one each from STs-37, − 199, and − 694. None of the isolates, including those eight that were negative for the OXA-48 and NDM-1 genes, were positive for other carbapenemase genes such as KPC, IMP, and VIM. The CTX-M-15 gene allele was the major extended-spectrum β-lactamase (ESBL), with gene positivity in 47 isolates (69%), whereas CTX-M-14 was found in 16 isolates (21.1%). CTX-M-15 was carried by isolates of all STs except STs-199 and -709, whose main CTX-M phenotype was CTX-M-14. Eleven of 15 isolates of ST-199 and two of three isolates of ST-709 were carriers of the CTX-M-14 gene. The other β-lactamase genes identified in this collection of isolates were TEM-1 in 60/71 (84.5%) and SHV (SHV-1 and SHV-11 together) in 71/71 (100%) isolates. Isolates belonging to STs-340 and -709 were all negative for the TEM-1 gene. Two of three isolates of ST-15 were also negative for the TEM-1 gene. Isolates belonging to STs-37, − 340, − 709, and − 974 carried the SHV-11 phenotype, whereas SHV-1 was the main allele in the remaining STs (Table 1).

Plasmid profile

PBRT detected seven plasmid replicon types in the whole population of 71 isolates. The replicons identified were: L/M; IncFII; IncFIIK; IncFIA; IncFIB; IncI1; and IncN. Sixty-seven isolates carried one or more plasmid replicons, whereas four isolates were negative for any plasmid type studied. IncFIIK was the predominant plasmid, found in 49/71 (69%) isolates, followed by L/M, found in 47/71 (66.2%). Similarly, the incidence of IncN was 10/71 (14.1%), that of FIA was eight of 71 (11.3%), that of FIB was eight of 71 (11.3%), that of FII was five of 71 (7%), and that of IncI1 was one of 71 (1.4%). The four isolates negative for any plasmid type belonged to ST-152. RD-122 was the only isolate carrying the IncI1 plasmid replicon (Table 1).

Virulence genes and hypermucoviscocity phenotyping

The status of different virulence genes is shown in Table 1. All isolates were positive for entB and 68/71 (95.7%) isolates were positive for mrkD. Isolates belonging to ST-709 alone showed the plasmid borne iutA virulence factor. All were negative for rmpA and the allS, magA (K1), and K2 CPS genes. Similarly, none of the isolates showed the hypermucoviscocity phenotype (Table 1).

Conjugation experiments

Multilocus sequence typing, clonal distribution, and genetic diversity

The 71 isolates sequenced for the seven loci of the MLST scheme of Diancourt et al. [13] exhibited a total of 37 alleles and were distributed into 18 haplotypes/STs. The clonal distribution of the isolates was as follows: ST-199, n = 15; ST-152, n = 12; ST-37, n = 14; ST-48, n = 5; ST-340, n = 5; ST-29, n = 4; ST-16, n = 4; ST-709, n = 3; ST-15, n = 4; ST-348, n = 2; ST-974, n = 2; and seven other STs with single isolates each. Thirty-four isolates (48%) were distributed in four clonal complexes/groups (CGs) and remained as singletons. A SplitsTree4 diagram was drawn to show their phylogenetic relationship based on the concatenated sequences of all seven loci (Fig. 1). A population snapshot showed their relationship with the STs in the database of the Institut Pasteur (Fig. 2). The 37 alleles ranged from two (rpoB)–10 (tonB). There were no indels or tri-allelic single-nucleotide polymorphisms (SNPs). Of the 436 SNPs observed, 326 were transitional changes, whereas the remaining 110 were transvertional. The diversity index (π) ranged from 0.00055 (pgi)–0.0089 (tonB). Of the 39 mutations observed, 31 were synonymous, whereas eight were non-synonymous. Of the eight non-synonymous mutations, six were found at the tonB locus. The non-synonymous:synonymous mutations ratio (Ka:Ks) was 0.11 for the concatenated sequences of the seven loci (Table 3). The diversities of the concatenated sequences of the seven loci and individual locus sequences are given in Additional files 2 and 3.

Locus	Length (bp)	No. alleles present	No. Polymorphic sites (%)	No. Synonymous mutations	No. Non-synonymous mutations	# SNPs	Ts	Tv	Ks	Ka	Ka/Ks	Diversity Index (π)
gapA	450	4	4 (0.89)	4	0	32	26	6	0.0060	NANa	–	0.00161126
InfB	318	5	4 (1.25)	4	0	69	69	0	0.0160	NAN	–	0.00393538
mdh	477	5	7 (1.47)	6	1	35	33	2	0.0050	0.0000	0.0000	0.0014328
pgi	432	3	2 (0.46)	2	0	9	0	9	0.0020	NAN	–	0.00055167
phoE	420	8	8 (1.90)	8	0	77	73	4	0.0230	NAN	–	0.00533477
rpoB	501	2	1 (0.20)	0	1	22	22	0	NAN	0.0010	–	0.00083368
tonB	414	10	13 (3.14)	7	6	192	103	89	0.0160	0.0060	0.3750	0.0089006
Concatenated	3012	0	39 (1.29)	31	8	436	326	110	0.0090	0.0010	0.1100	0.00304244

^aNot a number; SNP single nucleotide polymorphism, Ts Transitions, Tv Transversions, Ks Average number of nucleotide substitution per aynonymous site, Ka Average number of nucleotide substitution per non-synonymous site

Clonality and resistance profile

Multidrug resistance in members of the Enterobacteriaceae family has become a global phenomenon and for several reasons Saudi Arabia is viewed as a growing pool of these pathogens [4, 18–20]. The presence of large numbers of resistance determinants such as CTX-M, OXA-48, NDM-1, SHV, and TEM-1 in our KP isolates is reasonable, because all of them are resistant to several classes of antibiotic (Table 1, Additional file 1). A polyclonal OXA-48 gene was the most common carbapenemase found in our isolates and was distributed among all the STs. The OXA-48 gene has been reported to be endemic in Saudi Arabia [4, 8]. Although variants of OXA-48 such as OXA-204, − 232, − 181, − 162, and − 163 have also been reported in this region [4, 17, 21], none of our isolates showed any other OXA-48 variant. Similarly, the incidence of infections involving the NDM-1 gene is also being regularly reported in Asian countries, but limited data are available on the characteristics of this newly described gene in this region [4]. More recently, Mantilla-Calderon et al. [22] found the NDM-1 gene on an IncF plasmid in E. coli present in a non-nosocomial sample (sewage water) collected in Jeddah, Saudi Arabia, in 2013. The gene appeared to be on a conjugative plasmid because all of their E. coli transconjugants were positive for NDM-1. However, the genetic environment of NDM-1 on plasmids carrying the gene in our isolates appeared similar to that reported previously in KP and most other NDM-1 producers of the Enterobacteriaceae family: surrounded upstream by an ISAba125 insertion element and downstream by the bleomycin resistance gene ble_MBL [23, 24]. Regarding the eight isolates that did not present any of the carbapenemase genes and yet were resistant to carbapenem, an alternative mechanism such as a defective or non-functional outer membrane porin, frequently reported in CRKP, may be present [25–27]. The CTX-M gene is associated with antibiotic resistance worldwide and its positivity rate of 93% (66/71) in our isolates indicates its endemicity in this region, in agreement with previous reports [8, 28–30]. The presence of CTX-M-15 (77%) and CTX-M-14 (23%) variants is also in agreement with previous reports from Saudi Arabia [31]. The conjugation experiment showed that CTX-M-15, OXA-48, and NDM-1 genes can be transferred to recipient E. coli, whereas CTX-M-14 appears to be on a non-conjugative plasmid, also evident in its limited dissemination.

Plasmid typing

Limited data are available on plasmid replicon typing in multidrug-resistant KP in Saudi Arabia. The incidence of seven types of plasmid replicon, ranging from one to four in a single isolate, is reasonable in the context of the large number of antibiotic-resistance determinants found in these isolates. The high incidence of conjugative FIIK (69%) and L/M plasmids (66%) found in our isolates is in agreement with reports from European and Asian countries [29, 32–34]. IncI1 is carried only by an isolate with ST-11, a member of clonal group 258, which is famous for carrying the KPC gene worldwide. The reason that four isolates lack any plasmid type may be due to differences in their geographic distribution or limitations in the PBRT protocol for their detection [35]. Except for IncI1, FIA, and FIB, all other plasmids could be transferred to recipient E. coli during conjugation, indicating that these plasmids contribute to the multidrug-resistant phenotype of these isolates. Because the conjugation experiments were carried out on a limited number of isolates in this study (Table 2), it is not possible to determine the relationship between the replicons identified and resistance determinants. However, these data may provide the international community with information on local plasmid profiles and their evolutionary origins [35].

Virulence factors

Hypervirulence is usually linked to the presence of the rmpA, magA (K1), and K2 genes and the hypermucoviscous phenotype is regarded as critical for an isolate to be virulent, whereas other virulence factors appear to be only contributory to this property [36]. Because all our isolates were negative for the rmpA, magA (K1), and K2 genes and did not present a hypermucoviscous phenotype (Table 1), they may have not been hypervirulent-KP. These results also support the notion that virulence and antibiotic resistance are independent and not overlapping traits of KP [37]. Although none of our isolates showed a hypervirulent phenotype, the presence of siderophores (such as ybtS, entB, and iutA) may contribute to the severity of infections in patients, as reported previously [38]. The virulence factor iutA and hypermucoviscous phenotype gene rmpA are usually carried together and co-transferred simultaneously by the plasmid pLVPK and its homologs; thus, the presence of iutA alone in ST-709 isolates warrants further investigation.

Clonality and genetic diversity

MLST is a valuable tool for drawing inferences on genetic diversity and population structure in an epidemiologic setting (https://pubmlst.org/references.shtml) [39]. A significant observation of this study is the presence of a polyclonal pool of several genetically unrelated and non-popular antibiotic-resistant KP STs at our hospital. Using the Breurec et al.[40]. Approach of grouping of single-and double-locus variants into a clonal group and applying e-BURST analysis, 35/71 (49%) isolates (belonging to 10 STs) were grouped in three clonal groups (CG-37, CG258, and CG17), whereas 51% remained singletons belonging to eight diverse and unrelated non-epidemic genetic lineage STs (Fig. 1). A population snapshot drawn against the KP MLST database of the Institut Pasteur also showed that only four of 18 (22%) STs aligned with any of known clonal group (Fig. 2). These findings are consistent with those of previous reports that described the spread of ESBL-producing KP largely as multi-clonal, in contrast to the spread of KPC-producing KP, which is limited to specific clones (CG258) [41]. None of our isolates, including those from CG258, carried a KPC gene. Several clones in this study, such as STs-29, − 37, − 709, and − 111, have been previously reported at this hospital as well as elsewhere [8, 42, 43]. These results also support previous observations that this region is becoming a pool for new antibiotic-resistant KP, but is still mostly free of KPC [4]. The most interesting finding is the presence in large numbers of the non-popular, singleton clone ST-199, which requires further investigation.

Nucleotide diversity in the seven loci sequences remained much lower and the Ka:Ks ratio of the concatenated sequences of these loci was 0.11, indicating a purifying selection against deleterious mutations in these isolates (Table 3). The diversity analysis of individual locus sequences showed a significant diversity in three loci: viz., infB, rpoB, and tonB. Of these, the tonB gene appears to be under more selection pressure because six of its 13 polymorphic sites showed non-synonymous mutations (Table 3). A clone-wise analysis of these sequences showed significant diversity between the tonB and infB genes of the ST-199 and -48 isolates (Fig. 1, Additional file 2). The maximum number of non-synonymous mutations in tonB in a single ST (three of six) belonged to ST-199, indicating strong positive selection of this clone.

A significant observation of our study is the association of certain resistance genes with certain clonal types/genotypes of KP. For example, the majority of isolates belonging to ST-152 were positive for the NDM-1 gene and negative for the OXA-48 gene, whereas isolates of STs-199 and -37 were positive for OXA-48 and negative for NDM-1. Although the NDM-1 gene is regularly reported [4, 17], OXA-48 remains the main carbapenemase in this country. We also found this gene in only 21% of the isolates. The interesting feature of this gene was that ST-152 was its main carrier and that all 12 isolates belonging to ST-152 were positive for it. Both this finding and a previous observation from the United Arab Emirates [18] indicate that clone ST-152 of KP is a “frequent carrier” of the NDM-1 gene exclusive to Saudi Arabia. What makes this more interesting is that all of our isolates belonging to clonal group ST-152 were of the KP subspecies ozaenae, a species not represented in large numbers among KP members [44]. This is the first report to describe the involvement of the ozaenae subspecies of KP in the carriage of the NDM-1 gene. The production of the OXA-48 and NDM-1 carbapenemases alone or in combination with other resistance genes, in isolates of genetic lineages that have not previously been described as high-risk clones, is alarming. It suggests that these clones in an outbreak-like situation may limit the spread of other high-risk clones and/or add to their increasing numbers. The CTX-M-15 phenotype has previously been associated with STs-11, − 15, − 23, and − 37 [45, 46]. In our study population, it was uniformly distributed among all STs except for STs-199 and -709. Isolates from these two clonal types exclusively exhibited CTX-M-14. To the best of our knowledge, this association of the CTX-M-14 phenotype with any ST has not been previously reported. Similarly, the majority of isolates of STs-199 and -709 were CTX-M-14 carriers. Another example of such an association is that of the SHV gene: most of the isolates belonging to STs-199 and -48 were SHV-11 carriers, whereas SHV-1 was the primary SHV phenotype in the other clonal groups. Similar selective resistance was observed in STs-199, − 340, and − 16 for different types of aminoglycoside, such as amikacin, gentamycin, and tobramycin.