Whole genome sequencing and genotyping Klebsiella pneumoniae multi-drug resistant hospital isolates from Western Kenya

  1. Victor Dinda
  2. Andrew Nyerere Kimang’a
  3. Daniel Kariuki
  4. Anthony Wawire Sifuna
  5. Thomas James O’Brien
  6. Martin Welch
  7. Oleg N. Reva

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

Objectives. Klebsiella pneumoniae are a frequent cause of nosocomial infections worldwide. Sequence type 147 (ST147) has been reported as a major circulating high-risk lineage in many countries, and appears to be a formidable platform for the dissemination of antimicrobial resistance (AMR) determinants. However, the distribution of this pathogen in Western African hospitals has been scarcely studied. The main objective of this work was to perform whole genome sequencing of K. pneumoniae isolates from a referral hospital in Kakamega (Kenya) for genotyping and identification of AMR and virulence determinants.

Methods. In total, 15 K . pneumoniae isolates showing a broad spectrum antimicrobial resistance were selected for whole genome sequencing by Illumina HiSeq 2500 platform.

Results. ST147 was the dominant lineage among the highly-resistant K. pneumoniae isolates that we sequenced. ST147 was associated with both community- and the hospital-acquired infections, and with different infection sites, whereas other STs were predominantly uropathogens. Multiple antibiotic resistance and virulence determinants were detected in the genomes including extended-spectrum β-lactamases (ESBL) and carbapenemases. Many of these genes were plasmid-borne.

Conclusions. Our data suggest that the evolutionary success of ST147 may be linked with the acquisition of broad host-range plasmids, and their propensity to accrue AMR and virulence determinants. Although ST147 is a dominant lineage in many countries worldwide, it has not been previously reported as prevalent in Africa. Our data suggest an influx of new nosocomial pathogens with new virulence genes into African hospitals from other continents.

Article activity feed

  2. Version published to 10.1099/acmi.0.000667.v4 on Access Microbiology
  3. Version published to 10.1099/acmi.0.000667.v3 on Access Microbiology
  4. Access Microbiology

    Thank you for your revisions, reviewers were happy with your responses, there are only now minor changes required (for example some software versions). No additional data processing is required, however where possible please include the information requested.

  5. Access Microbiology

    Comments to Author

    1. Methodological rigour, reproducibility and availability of underlying data The Methods section is now written clear and the reanalysis using Kleborate suggested by reviewer one adds valuable information to the manuscript. * The information of how many isolates were tested for AMR in total and what is the percentage of multi-drug resistant isolates is missing in the method section. This is only stated in the Results section. The information is on Ln122. Given that the % resistant isolates is a result, rather than a method, we would prefer to leave this where it currently is (Ln200). It would be beneficial to know how many K. pneumonia strains were isolated during the sampling time and how many of those are multi drug resistant. The authors stated that 5-10 patients a day present with K. pneumonia currently. Is that 2023? Is there a change in the presents of multidrug resistance K. pneumonia? If there is an increase from the sampling time in early 2016 to the present, it would underline the influx of multidrug resistant K. pneumonia supporting the presented data. 2. Presentation of results The amendments to the result section are satisfactory. *Figure 1 is a good representation of the results. However, the genes in the black box are confusing. They are described as: "Sequence variants of marker genes traditionally used for MLST are depicted by framed boxes of different colour." The genes seem to be MLST marker genes and virulence genes; aerobactin is already included in the virulence score. It is confusing that the ST and capsule types have overlapping colours which makes the Figure difficult to read. *Ln 188 Hospital- and community-acquired Klebsiella pathogens were isolated 3. How the style and organization of the paper communicates and represents key findings The amendments to the discussion section are satisfactory. *Ln 327 currently change to at the time of publication How much was it during sampling time? See comment above.

    Please rate the manuscript for methodological rigour

    Good

    Please rate the quality of the presentation and structure of the manuscript

    Satisfactory

    To what extent are the conclusions supported by the data?

    Partially support

    Do you have any concerns of possible image manipulation, plagiarism or any other unethical practices?

    No

    Is there a potential financial or other conflict of interest between yourself and the author(s)?

    No

    If this manuscript involves human and/or animal work, have the subjects been treated in an ethical manner and the authors complied with the appropriate guidelines?

    Yes

  6. Access Microbiology

    Comments to Author

    I am happy that the authors have sufficiently addressed all my comments from the previous round of review. Their work contributes to the surveillance of an important pathogen in an under-reported region. Software version numbers for CheckM2, Kleborate and Kaptive are missing and should be included.

    Please rate the manuscript for methodological rigour

    Very good

    Please rate the quality of the presentation and structure of the manuscript

    Good

    To what extent are the conclusions supported by the data?

    Strongly support

    Do you have any concerns of possible image manipulation, plagiarism or any other unethical practices?

    No

    Is there a potential financial or other conflict of interest between yourself and the author(s)?

    No

    If this manuscript involves human and/or animal work, have the subjects been treated in an ethical manner and the authors complied with the appropriate guidelines?

    Yes

  7. Version published to 10.1099/acmi.0.000667.v2 on Access Microbiology
  8. Access Microbiology

    This study would be a valuable contribution to the existing literature. This is a study that would be of interest to the field and community. The reviewers have highlighted major concerns with the work presented. Please ensure that you address their comments.

  9. Access Microbiology

    Comments to Author

    1. Methodological rigour, reproducibility and availability of underlying data * The information of how many isolates were tested for AMR in total and what is the percentage of multi-drug resistant isolates is missing in the method section. This is only stated in the Results section. * The isolates were obtained by standard techniques and ethics approval was appropriate. * Genome sequencing and annotation was performed using standard methods. However, the authors state in the data summary that the DNA reads were quality controlled but do not provide the method. * The genome data is deposited on the NCBI server. The authors state in the method section that the genomes were annotated using Prokka v1.14.5. Contrary, the annotation details on the NCBI website state that the genome was annotated using NCBI Prokaryotic Genome Annotation Pipeline (PGAP). Additionally, in the results section, line 183, RAST genome annotation robot was used as basis for the phylogenetic tree. Hence, it is not clear which annotation methods was used to analyse the data. * The authors performed a search for virulence genes stated in the methodology line 148 but do not show any results regarding virulence genes. * The table with NCBI accession number is more appropriate in the supplementary data. 2. Presentation of results * Figure 1 shows the phylogenetic tree of sequenced isolates. It is a good representation of the relationship of the isolates. The authors write that five isolates were sensitive to one respective tested antibiotic. It would be interesting to add this data to the phylogenetic tree of the isolates and further which drug related genes (DRG) relating to the sensitive antibiotic is missing in these isolates either in the genome or plasmid. * Figure 2 was created using Mauve alignment. The figure shows a root, but it is not clear which sequence that is. * The authors make the intuitive assumption that the larger plasmids in K10, K12, K13 and K15 evolved from the smaller ancestral plasmids found in CK2 and CK4. Are there any transposable elements on the plasmids to support this hypothesis? It is also possible that the ancestral plasmid was larger in size and lost DRG in CK2 and CK4. * Similar conclusions were made for the phylogenetic group containing plasmids CK9, CK7, CK6 and CK1. The larger plasmids and smaller plasmids share a common ancestor. It is difficult to say if the larger plasmids evolved from the smaller ones with the small sample size. The common ancestor plasmid could have lost some DRG. * Figure 2 could be improved by adding the presents or absence of DGR in a heatmap. This heatmap could also include virulence genes. 3. How the style and organization of the paper communicates and represents key findings * The authors use standard techniques to demonstrate the importance of WGS as a surveillance tool for AMR in under surveyed regions. They clearly demonstrate the presence of multi-drug resistant K. pneumonia strains. However, the evolutionary pathway of the acquisition of DGR needs to be investigated further. This could be achieved by including additional genome and/or plasmid sequences in the phylogenetic trees. * The authors state that the evolutionary success of the investigated isolates is because of the acquisition of DGR genes and virulence genes. They present data of DGR confirming this but miss to show how virulence genes support this hypothesis. 4. Literature analysis or discussion * Line 261: Do the authors mean references [23] and [24] instead of [25]? * The distribution of ST needs to be discussed more in a geographical and temporal context. Reference [25] was published in 2022 but sampling occurred from May 2015 till March 2020. * ST147 and ST14 have been reported to spread DRG. This could be included to discuss the evolutionary success of the sequenced isolates.

    Please rate the manuscript for methodological rigour

    Good

    Please rate the quality of the presentation and structure of the manuscript

    Satisfactory

    To what extent are the conclusions supported by the data?

    Partially support

    Do you have any concerns of possible image manipulation, plagiarism or any other unethical practices?

    No

    Is there a potential financial or other conflict of interest between yourself and the author(s)?

    No

    If this manuscript involves human and/or animal work, have the subjects been treated in an ethical manner and the authors complied with the appropriate guidelines?

    Yes

  10. Access Microbiology

    Comments to Author

    Summary: The article by Dinda et al. presents data on a collection of K. pneumoniae isolates gathered from a Kenyan hospital over a 6-month period circa. 2016. They identify isolates that display alarming levels of antibiotic resistance and proceed to whole genome sequence and genomically characterise them. They find a number of resistance genes present in the isolates and locate them on genomic islands and plasmids. Their primary finding is that the ST147 lineage has become a dominant strain in Kenyan hospitals. This article provides much needed phenotypic and genotypic observations on a major MDR pathogen from an under-surveyed region. 1. Methodological rigour, reproducibility and availability of underlying data Several analyses which are considered standard by the field for this pathogen are missing from this article. All these analyses can be performed using Kleborate (https://github.com/klebgenomics/Kleborate) which is the standard analysis pipeline for this pathogen. This tool is even implemented in Pathogen Watch (https://pathogen.watch/). Analyses that are specifically missing: No genomic confirmation of species. The isolates in this study were identified as K. pneumoniae by culture and biochemical assay only. It has been previously published that biochemical and even MALDI-TOF analysis is not capable of accurately discriminating between Klebsiella species (see: 10.1128/mSphereDirect.00290-17). I'd like to see the genomic data used to confirm species with Kleborate, or even Kraken2 + GTDB (https://github.com/DerrickWood/kraken2 & https://gtdb.ecogenomic.org/). No data were presented on virulence factors. The authors state that isolates were characterised using the Virulence Factor Database yet no results for virulence factors are presented. VFDB is a multi-species virulence factor database and not optimised for Klebsiella, again Kleborate would be more appropriate. The isolates have not been capsule typed. The capsule is important for both virulence and AMR in Klebsiella. Moreover, it is particularly relevant for ST147 as previous studies have identified specific K loci associated with this lineage. Again, Kleborate can perform capsule typing. There is a mismatch between results and methods, were the genomes annotated with Prokka as per the methods section or with RAST as per the results section? If RAST was used it is not cited and the version number is absent. It is not clear to the reader how the authors distinguish between community and hospital acquired infections. I commend the authors for making their genome assemblies publicly available with the publication however I would also like the raw FASTQ files to be made available as well. The genome assembly records can be linked to SRA records through the Biosample ID's. The FASTQ files are considered the raw data and their availability contributes to open access, reproducible science. The genome assemblies are not considered as raw data. NCBI Biosamples also allow submission of antibiogram data such as zones of inhibition from disk diffusion assays, making this data available (if possible) would be highly beneficial to the research community. There was no mention of quality control on the genomic data. Illumina reads were trimmed with Trimmomatic but there was no assessment of levels of contamination though Kraken2 or quantification of sequencing coverage; were the isolates sequenced at 30x coverage or 100x coverage? There was also no assessment of resulting genome assembly metrics (number of contigs, genome size, N50 etc.), these metrics could be calculated with CheckM (https://github.com/Ecogenomics/CheckM) or Quast (https://github.com/ablab/quast). The authors use the tool SeqWord Genomic Island Sniffer to identify Genomic Islands, I am unfamiliar with this tool. The cited article is not readily available (no results from a Google or Pubmed search), nor could I locate a code repository for this tool (available hyperlinks returned errors). The paper would benefit from a brief overview of the tool's methodology in predicting horizontally transferred islands, as well as providing a link to the tool. Were there any additional analyses performed to confirm predictions such as identifying transposons at genomic island borders? Or prediction of phage? Version numbers and parameters for software and databases should be listed, as well as cited. Some examples: the Mauve version number is missing and is not cited. RAST (if used, see comment above) is also missing a version number and citation. Resistance Gene Identifier (RGI) version number. CARD database version or date of access. MUSCLE parameters used and version number. Gblocks version, etc. The phylogenetic method used is unusual for such a small set of closely related isolates. The authors opt to produce a peptide alignment and build a Neighbour Joining tree from this, while this approach is valid a more typical approach used in the field would be a Maximum Likelihood tree from a nucleotide core genome alignment. This could be achieved by read mapping to a reference genome to produce a core genome alignment with Snippy (https://github.com/tseemann/snippy). A phylogenetic tree could then be built using IQTree or RAxML (IQTree: http://www.iqtree.org/ RAxML: https://github.com/amkozlov/raxml-ng). Another option is the tool Parsnp which performs both alignment and tree construction (https://github.com/marbl/parsnp). The peptide alignment of orthologs is perhaps better suited to a more diverse dataset spanning multiple species / genera. 2. Presentation of results The isolates assayed display resistance to the carbapenem antibiotics Imipenem and Meropenem yet a carbapenemase was not found? Could the authors comment on this observation? Was carbapenem resistance borderline? What mechanism do they think is driving resistance? See: 10.1128/JCM.00651-08. In Supplementary Table 1 the carbapenemases NDM-6 and OXA-232 are listed. Why were these results not discussed in the main body? The discovery of carbapenemase genes on plasmids is an important observation. The software used to produce the figures is not clearly stated. Was iTOL (https://itol.embl.de/) used to create the trees? Or were they produced in ggtree (https://github.com/YuLab-SMU/ggtree)? Or other software? It is also not clear to the reader what data the dendrogram in Figure 2 is based on. What distance metric was used? Blast identity? Coverage? Average nucleotide identity? What does the scale bar represent? It would also be beneficial to add virulence gene results to this figure. Supplementary Figure 1 is difficult to interpret. What do the colour blocks represent? What are the lines connecting the plasmid contigs? I'd suggest the plasmid contigs identified are grouped based on some characteristic such as size or gene content (virulence / AMR) and then compared as separate panels. 3. How the style and organization of the paper communicates and represents key findings Overall, the paper is well written and has a clear flow of reasoning. However, a key discovery of carbapenemase genes located on plasmids is not discussed in the main body of the paper. Some results (strain details, gene detection) would be better suited to presentation in tables and not listed extensively in the paper (Section 3.4). The conclusion that ST147 is the dominant lineage from such a small sample collection over a limited time period from a single hospital is premature. Especially given that the authors only sequenced 15 out of the 27 isolates collected, focussing on the most resistant ones. I'd recommend the authors amend their conclusion to reflect this limitation, perhaps clarifying that ST147 is the dominant lineage in highly resistant K. pneumoniae. 4. Literature analysis or discussion The authors aptly describe the threat posed by multi-drug resistant K. pneumoniae and the lack of surveillance data from certain regions. The authors discuss their findings well, particularly with results from recent studies of the same region. They postulate that human migration or pathogenesis may drive the success of ST147. 5. Any other relevant comments Typos and lack of clarity: I would also recommend using the acronym Antibiotic / Antimicrobial Resistance Genes (ARG's) instead of Drug Resistance Genes (DRG) to be more consistent with the literature. Line 148 and 198: correct GRI-CARD to RGI-CARD. Clearer to specify that Resistance Gene Identifier (RGI) tool was used with the CARD database. Tool and database versions are also missing. Line 275: "the infection is either patchy and differs from clinic to clinic" - unclear what this statement means. Do the authors mean that there are geographic differences in the presence of ST147? Line 40, 154 and 262: Correct pneumonia to pneumoniae Line 55: Why is there a question mark after SPAdes version number? Line 264: correct multigrug to multidrug Line 274: correct discrepensy to discrepancy Line 276: correct resent to recent Line 284: correct plismids to plasmids Line 288: correct turkey to Turkey / Türkiye Line 427: correct resustance to resistance The comments below should be considered as optional. I believe addressing them could elevate the article. The authors mention an increase in MDR K. pneumoniae infections triggered a targeted investigation, more information on typical case numbers seen at the hospital would be beneficial. The authors state that they performed AST on 27 isolates, is 27 K. pneumoniae in a 6-month period typical? If not, what are the typical numbers? A more systematic comparison of phenotypic and genotypic resistance profiles would be beneficial. Specifically, can all the phenotypic resistances be explained by genomic factors? Do all the pan-resistant isolates contain genes conferring resistance to each antibiotic tested? Are there any isolates with unexplained phenotypic resistance? Further characterisation of the plasmid contigs identified. Specifically, incompatibility group typing, replicon identification and conjugation machinery. The tool MOB-typer would be useful for this (https://github.com/phac-nml/mob-suite). Moreover, K. pneumoniae virulence and AMR determinants are known to circulate on plasmids, the convergence of these genes onto the same plasmid is of great concern and subject to surveillance. More detailed plasmid characterisation with respect to virulence and AMR gene content would be beneficial; do the plasmids harbouring carbapenemase genes also possess virulence factors or conjugation machinery? Did the plasmid contigs identified show any similarity to previously published plasmids? A blast database search would identify similar sequences.

    Please rate the manuscript for methodological rigour

    Poor

    Please rate the quality of the presentation and structure of the manuscript

    Good

    To what extent are the conclusions supported by the data?

    Partially support

    Do you have any concerns of possible image manipulation, plagiarism or any other unethical practices?

    No

    Is there a potential financial or other conflict of interest between yourself and the author(s)?

    No

    If this manuscript involves human and/or animal work, have the subjects been treated in an ethical manner and the authors complied with the appropriate guidelines?

    Yes

  11. Version published to 10.1099/acmi.0.000667.v1 on Access Microbiology