Genomic characterization of Hafnia paralvei isolated from milk and dairy products

he nutrient richness of milk and dairy products makes these food items highly susceptible to bacterial contamination during production, processing, storage, and distribution. Within the genus Hafnia, the species H. alvei and H. paralvei group together strains used for technological purposes, strains related to food spoilage, and pathogenic strains associated with infectious diseases of the respiratory and gastrointestinal tracts. This study aims to characterize Hafnia paralvei at the genomic level, isolated from raw milk and dairy products, in order to assess its potential hazard to human health. Methods. Following species confirmation by biotyping, the genomes of 20 Hafnia alvei isolates collected from raw milk, cheese, and the cheese processing environment were sequenced on an Illumina NovaSeq platform. The reads were assembled using Unicycler V0.5.1, and the assembly metrics were evaluated using QUAST. Taxonomic assignment was performed using RefSseeker v1.8.0. The resistome and virulome were analyzed using the ABRIcate v1.0.1 tool with the ResFinder and Ecoli_VF databases, respectively. The phylogenetic tree was constructed based on SNPs using kSNP4.1 and visualized with iTOL v6. The SNP distance matrix was calculated using snp-dists v0.6. Three The heatmaps were created using ggplot in R Studio. The search for genetic mutations associated with the antibiotic resistance phenotype was conducted using the ResFinder v4.7.1 tool available on the website of the Center for Genomic Epidemiology (http://genepi.food.dtu.dk/resfinder). Results. Following biotyping, all 20 isolates were found to belong to the species Hafnia alvei. Using WGS, 18 isolates (90%) were reassigned to the species Hafnia paralvei, 1 to Escherichia coli, and 1 to Lelliottia adecarboxylata. A total of 12 antimicrobial resistance genes (ARGs) were found, with blaACC-1 present in all isolates. The genome of an isolate (HAF_8010_norm_3 isolated from raw milk) was positive for 7 ARGs (mdf(A), blaACC, aph(3'')-Ib, aph(3')-Ia, aph(6)Id, sul2, and tet(Y), predicting phenotypic resistance to class C beta-lactams, aminoglycosides, sulfonamides, and tetracycline. Two genomes (HAF_5SWP4 and HAF_6SWM4 from the cheese processing environment) contained the mcr-9 gene located on the chromosome, predicting resistance to colistin. A total of 42 virulence genes were detected, with the following genes present in all isolates: cheBDRWYZ, clp, fepA, flgBCFGI, flhABC, fliAGIMPQS, galU, htpB, kdsA, lpxAC, mgtB, motA, msbA, ompA, and rfaDF. The remaining genes such as chuA, flgH, glf, gtrB, hcp-2, rfaE, rffG, shuA, and vip/mglB were present in a subset of isolates. Conclusions. In conclusion, the application of WGS technology has revealed its great potential in identifying H. paralvei, a genetically similar species often confused with H. alvei. The presence of ARG like mcr-9, which confers resistance to a critically important antibiotic like colistin, and the co-occurrence of virulence genes previously identified in extraintestinal Escherichia coli, suggest the potential of some H. paralvei strains isolated from milk and dairy products as potential hazards to human health.