Highly drug resistant clone of *Salmonella* Kentucky ST198 in clinical infections and poultry in Zimbabwe

The study investigates the genetic diversity, antimicrobial resistance determinants, and phylogenetic relationships of Salmonella enterica serotype Kentucky in Zimbabwe, with the aim of identifying potential sources of human infection. The context is the global rise of antimicrobial resistance (AMR), especially in low-income settings, driven by antibiotic overuse in human medicine and agriculture. Non-typhoidal Salmonella (NTS) is a significant public health burden; multidrug-resistant lineages complicate treatment, particularly for invasive disease where fluoroquinolones and extended-spectrum cephalosporins are commonly used in sub-Saharan Africa. S. Kentucky ST198, an MDR clone that emerged in Egypt around 1989, has spread internationally and is associated with SGI1-mediated resistance and quinolone resistance via gyrA/parC mutations. The prevalence and molecular epidemiology of ESBL-producing S. Kentucky in Zimbabwe were previously unknown. This study uses whole-genome sequencing (WGS) to characterize isolates from human clinical cases and poultry-associated sources to understand transmission dynamics and resistance mechanisms.

Prior work documents the emergence and global spread of ciprofloxacin-resistant S. Kentucky ST198, initially linked to North Africa and travel-related cases in Europe, with high rates of fluoroquinolone resistance reported worldwide. SGI1 and its variants (e.g., SGI1-K, SGI1-KIV) encode MDR phenotypes in Salmonella, and additional resistance to extended-spectrum cephalosporins has been reported via ESBLs (e.g., blaCTX-M variants), sometimes chromosomally integrated (e.g., blaCTX-M-146 near hcp1) or plasmid-borne (e.g., blaCMY, blaCTX-M-15). Studies from Europe and Asia have described ESBL-producing ST198 and their genomic contexts. However, data from sub-Saharan Africa on ESBL-producing S. Kentucky were lacking. The literature also highlights the role of poultry as a common source of S. Kentucky and the implications of antibiotic use in animal production for AMR dissemination.

- Study design and isolates: 245 non-typhoidal Salmonella isolates from Zimbabwe (2016–2020) were analyzed, including human clinical cases (n=162), chicken farms (n=82), crocodile meat (n=1), and a dining table swab (n=1). Among these, 37 were S. Kentucky (human n=11; poultry and related sources n=26). Ethical approvals: University of Pretoria (779/2018) and Medical Research Council of Zimbabwe (MRCZ/A/2369). - Isolation, serotyping, and AST: Serotyping followed ISO 6579-1:2017 using agglutination with Mast antisera; interpretation via Kauffmann-White-Le Minor scheme. Antimicrobial susceptibility testing used Kirby-Bauer disk diffusion on MH agar with a panel including ciprofloxacin, ceftriaxone, chloramphenicol, tetracycline, azithromycin, ertapenem, ampicillin, ceftazidime (± clavulanic acid), cefotaxime (± clavulanic acid). E. coli ATCC 25922 served as QC; CLSI M100 (2020) criteria via WHONET 5.6. - DNA extraction and sequencing: Genomic DNA from overnight cultures was extracted using Maxwell RSC 48. Concentrations measured by Qubit. Short-read libraries prepared with Nextera Flex; sequenced on Illumina NextSeq. Long-read libraries prepared with ONT native barcoding (SQK-LSK109); sequenced on MinION (R9.4.1). - Data availability: SRA accession PRJNA762287 for raw reads. - QC and typing: Read quality assessed with fastp and summarized with MultiQC. Samples with depth <20x or <80% Salmonella (Bracken) were excluded. Serotypes predicted with SeqSero2. MLST, AMR genes, and plasmid replicons identified using ARIBA with ResFinder and PlasmidFinder databases. Assemblies generated with SPAdes v3.13.0. Chromosomal mutations in gyrA, gyrB, parC, parE identified with RGI. - Phylogenomics (Zimbabwe set and global context): Reads mapped to reference S. Kentucky 201001922 (CP028357) using snippy (mapqual 60, basequal 13, mincov 4, minfrac 0.75). Recombination masked with Gubbins v2.2.0. ML trees inferred with RAxML v8.2.8 (GTR, auto-mre bootstraps); visualized with ggtree. Population structure via HierBAPS (max depth 3, n.pops 10). Conservation of SGI1-K assessed by mapping reads to SGI1-K reference (AY463797.8) using minimap2 and coverage via bedtools; SGI1 genes assessed with ARIBA. - Long-read assembly and plasmid analysis: ONT reads filtered with filtlong; Trycycler used to generate consensus assemblies from Flye, Raven, and Miniasm assemblies; polished with Pilon using Illumina reads; assembly quality via QUAST; chromosome orientation checked with Socru; chromosome start set with Circlator. BLAST and genoPlotR used for comparative genomics of SGI1 and plasmids. Plasmid taxonomic units identified with COPLA; closest relatives searched against PLSDB.

- Serotype and sequence types: Among 37 S. Kentucky isolates, 36 (97.3%) were ST198 and 1 (2.7%) was ST152. - AMR gene burden: All ST198 isolates harbored 6–15 AMR genes; 92% carried at least 10 AMR genes. - SGI1-KIV: All ST198 isolates carried SGI1-KIV integrated into the chromosome, typically encoding aac(3)-Id, aac(6)-Iaa, aadA7, blaTEM-1, sul1, and tetA, with sporadic loss of one or more genes in five isolates. SGI1-KIV was split across the rbsk and trmE/ydiY loci and lacked aph(6′) and aph(3′) found in canonical SGI1-K. - Fluoroquinolone resistance: All ST198 had quinolone resistance-determining region mutations in gyrA and parC. One strain (ZM1151) carried qnrB and was the only one lacking gyrA mutations. - ESBL and additional resistance: 92% of ST198 isolates carried blaCTX-M-14.1 and fosA3 on an IncHI2 plasmid along with aadA2b, aadA1, aph(3′)-Ib, aph(6′)-Id, cmlA1, and sul3. ColpVC and IncHI2/IncHI2A replicons were detected in 100% (36/36) and 92% (33/36) of ST198 isolates, respectively. - Plasmid characterization: Long-read assemblies identified IncHI2 plasmid pGTZIM1 (~157 kb) carrying blaCTX-M-14.1 and fosA3 on a composite transposon; closest relative was E. coli plasmid pF218CHI2 lacking blaCTX-M-14.1 and fosA3. An IncI plasmid pGTZIM2 (~92.5 kb) carrying blaCMY was found in strain NM17-56, related to E. coli plasmid p92; a small ColpVC plasmid (pGTZIM3) without known AMR/virulence genes was also detected. - Phylogenetics within Zimbabwe: ST198 isolates formed three first-order clades with low SNP distances (mean root-to-tip ~12 SNPs), indicating a recent common ancestor. Several human and poultry/environmental isolates differed by ≤5 SNPs, consistent with recent transmission and inter-farm spread; identical strains were found across different towns and years. - Global context: Zimbabwe ST198 isolates formed a distinct subclade within a globally dispersed epidemic lineage, closely related to strains from the UK, India, Denmark, Pakistan, the Netherlands, the USA, and Belgium. Travel data implicated South Asia in related cases. Zimbabwe isolates carried more AMR genes, partly due to acquisition of pGTZIM1; IncHI2-associated genes (aph(6′), blaCTX-M-14.1, cml, fosA3, sul3) were generally absent outside Zimbabwe. - Epidemiological implication: The close genetic relationship between human and poultry isolates supports poultry as a significant source for human infections in Zimbabwe.

The study addresses the gap in knowledge about ESBL-producing S. Kentucky in Zimbabwe by integrating WGS-based phylogenomics with resistome and plasmid analyses across human clinical and poultry-associated isolates. The tight clustering and minimal SNP distances among ST198 isolates from different sources, locations, and years indicate rapid recent spread of a single epidemic clone within Zimbabwe, with inter-farm transmission and potential contamination of the food chain (e.g., feed). The presence of near-identical isolates across human and poultry sources supports poultry as an important source of human infection, even though direct transmission events cannot be conclusively demonstrated from the available sampling. The detection of SGI1-KIV in all ST198 isolates, combined with widespread fluoroquinolone resistance (gyrA/parC mutations) and high prevalence of an IncHI2 plasmid carrying blaCTX-M-14.1 and fosA3, underscores extensive and clinically significant resistance, narrowing therapeutic options. In the global phylogenetic context, the Zimbabwe clade is nested within the well-known international ST198 epidemic lineage, but exhibits a distinctive resistome likely due to local acquisition of an IncHI2 plasmid (pGTZIM1). These findings emphasize the role of poultry production and antimicrobial usage in driving and maintaining MDR in S. Kentucky and highlight the need for One Health surveillance and stewardship.

This work provides the first WGS-based characterization of ESBL-producing S. Kentucky ST198 from human and poultry sources in sub-Saharan Africa, revealing a highly resistant clone circulating in Zimbabwe. Key contributions include: identification of a closely related ST198 clade across human and poultry sectors; universal presence of SGI1-KIV; high prevalence of an IncHI2 plasmid (pGTZIM1) carrying blaCTX-M-14.1 and fosA3; and widespread quinolone resistance mutations. The data implicate poultry as a major reservoir and transmission source for human infections. Future work should expand genomic surveillance across the food chain and environment, quantify transmission pathways (including feed and inter-farm spread), monitor plasmid dynamics, and implement antimicrobial stewardship and biosecurity interventions under a One Health framework.

The principal limitation is the relatively small number of S. Kentucky isolates (n=37) and uneven sampling across sources, geography, and time, which constrains inference of direct transmission events and quantification of transmission routes. Overrepresentation of poultry-associated samples may bias serotype frequencies. Some conclusions about spread and source attribution are therefore inferential and would benefit from broader, longitudinal, and systematic sampling.