Bacterial isolates

Sixteen pure cultures of heavy growth of K. pneumoniae isolated from stool of different patients from two clinical laboratory facilities in Nigeria (diagnosed with various diseases) were identified using cultural morphology, Gram reaction (Bartholomew and Mittwer, 1952), standard biochemical tests; indole, citrate, motility, methyl red, voges proskauer and sugars (Ewing, 1986; Barrows and Feltham, 1993), and API 20E strips (BioMérieux, Basingstoke, UK). Ethical approval was obtained for the study from Federal Capital Territory, Health Research Ethics Committee with approval number FHREC/2018/01/95/14-08-18, including informed written consent from the participants.

Antibiotic susceptibility testing

Susceptibility of all isolates to a panel of antibiotics classes in common clinical use in these hospitals were determined by the agar dilution method on Mueller–Hinton agar according to the recommendations of CLSI breakpoints (Wayne, 2018). All runs included the control organism E. coli (ATCC 25922).

Whole genome sequencing and bioinformatics

DNA was extracted using a QIAamp1 DNA Mini Kit (QIAGEN, Crawley, UK) according to the manufacturer's instructions. Paired-end Illumina whole genome sequencing was completed using a NextSeq instrument at the Quadram Institute Bioscience. Bioinformatics used an in-house pipeline hosted on an IRIDA instance; sequences were assembled with shovill and annotated with prokka and core snps identified with snippy. Furthermore, assemblies were used to search for plasmid content using the ‘PlasmidFinder’ tool hosted at the Centre for Genomic Epidemiology(https://cge.cbs.dtu.dk/services/PlasmidFinder), and for isolates likely to carry significant resistance genes in plasmids, ‘plasmidSPAdes’ was used to assemble likely plasmid contigs from the trimmed reads.

Comparative genomics

Strains belonging to sequence types implicated in globally disseminated disease were compared against completed genomes were available to determine relationships between sources of strains, Nigerian strains and others in global circulation. Reads were also mapped against reference strain HS11286, SNPs identified and phylogenetic relationships determined to determine whether clones in Nigeria are divergent or highly similar from those seen globally.  

In the 16 isolates of K. pneumoniae, 5 encoded virulence genes namely, astA, senB, and gad. All of these 5 isolates had the gad gene while only 2 had all the 3 genes (Figure 1). astA encodes enteroaggregative E. coli heat-stable enterotoxin (EAST 1), senB encodes TieB enterotoxin, while gad is a glutamate decarboxylase protein. The contribution to pathogenicity of gad has been debatable; this enzyme has been reported to be essential in the survival of enteric pathogens in the acidic conditions of the mammalian stomach (Lin et al., 1996). Two isoforms of glutamate decarboxylase (GAD) encoded by gadA and gadB are the most effective E. coli acid resistance system (De Biase et al., 1999). The major role of this system might be to facilitate the colonization of the intestines by commensal strains of E. coli. This is a housekeeping gene in E. coli, but not part of the K. pneumoniae core genome and here it was 31.3%.

EAST 1 has been identified as a plasmid-mediated enterotoxin of low molecular weight and associated with enteroaggregative E. coli. It shares about 50% protein identity with heat-stable enterotoxin (STa), and the gene has also been found in many enterotoxigenic E. coli strains (Contreras et al., 2011) and other members of Enterobacteriaceae such as Salmonella (Paiva de Sousa et al., 2001). The authors demonstrated the presence of the gene in different strains of K. pneumoniae (ST914 and ST1962) from two different geographic regions of Nigeria, and these patients presented with diarrhoea episodes. There are reports debating whether astA is sufficient to cause diarrhoea without other virulence factors in E. coli, but Soto et al. (2009) and Mirzarazi et al. (2015) reported that E. coli acquired this toxin to become a diarrhoea-causing agent, which may be the situation in these K. pneumoniae strains.

The two K. pneumoniae that encoded astA also encoded senB, a plasmid-mediated gene associated with enterotoxicity of enteroinvasive E. coli (EIEC) that codes the TieB protein. It has also been described to have some role in uropathogenic E. coli. Multiple plasmid replicons were present in all strains; however IncF and Col plasmids play a major role in the dissemination of these genes (Figure 1).

Susceptibility testing identified all 5 isolates with these virulence genes were sensitive to meropenem (MIC values of ≤ 0.03 µg/ml) while isolates with both astA and senB had low level resistance to ciprofloxacin and cefotaxime (MIC value of 8 µg/ml each), whereas other strains had high level resistance to ciprofloxacin (MIC value of >64 µg/ml). Resistance to colistin varied with MICs between 1 and >64 µg/ml. The isolates carried a range of important resistance genes to cephalosporins, fluoroquinolones, aminoglycosides, with the presence of various PMQR genes, aph(3'')-Ib, aac(3)-IId, variants of blaCTX-M including blaCTX-M-15 found in one of the isolates.

There is presence of diarrhoeagenic K. pneumoniae linked to carriage of plasmid mediated virulence genes such as astA, senB, and gad in diversity strains in Nigeria. This may represent a greater burden of diarrhoea caused by K. pneumoniae in future.

DOO was the recipient of an overseas scholarship from the Royal Society Newton International Fellowship (NF110504) and their Follow-on funding which supported this work, including support from Quadram Institute of Bioscience.

The authors have not declared any conflicts of interests.

The authors thank the staff within the laboratory of the study site hospitals for their technical support during this study.