An Omics-Guided Investigation of a Hospital Outbreak Caused by blaNDM-1-Producing Pseudocitrobacter faecalis

Abstract

Carbapenemase-producing Enterobacterales (CPE) pose a major healthcare challenge. We report the first hospital outbreak of Pseudocitrobacter faecalis carrying blaNDM-1 using an omics-based approach. Short- and long-read sequencing enabled genomic epidemiological investigation to track its spread, characterize its resistome, and analyze the genomic context of blaNDM-1. Additionally, we developed and implemented a matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS)-based method for rapid outbreak isolate typing using protein biomarkers. Our investigation identified 2 independent blaNDM-1-producing clonal clusters of multidrug- and carbapenem-resistant P faecalis circulating for over 3 years, carrying blaNDM-1 either chromosomally or on a plasmid. MALDI-TOF MS spectra analysis revealed candidate protein markers corresponding to genomic clusters, with 1 predicted biomarker applicable for rapid typing. Pseudocitrobacter faecalis is an emerging CPE taxon requiring hospital surveillance. Early whole genome sequencing unexpectedly revealed 2 intertwined clones with independent carbapenemase acquisition routes. Cluster-specific markers enabled rapid typing, serving as proof of concept for validating proteomics in future surveillance.

Keywords: Pseudocitrobacter faecalis; bla NDM-1; carbapenemases; genomics; outbreak; proteomics.

Figure 1.

Epidemic curve of isolations of 37 bla_NDM-positive Pseudocitrobacter faecalis isolates from inpatients at the Hadassah Hebrew University Medical Center, spanning March 2019 until July 2022, and an environmental isolate. Break in the x-axis (May–September 2019) refers to lack of bla_NDM-positive isolations. The peak comprising of 11 sentinel isolates leading to outbreak recognition is notable in December 2020. The pie chart shows the distribution of bla_NDM-positive isolates by source.

Figure 2.

Core genome multilocus sequence typing ad hoc scheme of Pseudocitrobacter faecalis. Color coding denotes the epidemiological context of isolate detection. Node size is proportional to the number of isolates assigned to clone types. Numbers denote the allelic distances between nodes. The minimum spanning tree shows 2 main clusters. Cluster 1 (to the left) consisted of retrospective and prospective isolates, whereas cluster 2 (to the right) mainly consisted of the sentinel cluster (see Supplementary Table 4 for more details). The environmental J0005 isolate belonged to cluster 1. Two singletons are noted in the tree, including 1 carbapenemase-producing Enterobacterales (CPE) isolate (3447) and 1 non-CPE isolate (9929) shown as outliers. Three reference genomes for comparison are Pseudocitrobacter vendiensis ERR3255970 and P faecalis ERR3278183 and ERR3278184.

Figure 3.

Presence or absence of a selection of 47 resistance genes from the AMRFinderPlus database in short-read assemblies of sequenced isolates. The figure includes core genome multilocus sequence typing (cgMLST) cluster 1 isolates (blue, lines 14–26 ) and cgMLST cluster 2 isolates (outbreak, red, lines 1–10), 2 singletons (3447 and 9929, beige, lines 27–28), and 3 reference Pseudocitrobacter spp (P vendiensis ERR3255970 and P faecalis ERR3278183 and ERR3278184, lines 29–31). Abbreviation: ARG, antibiotic resistance gene.

Figure 4.

Genomic context of bla_NDM-1 among the 2 investigated clusters. A, Alignment of the 3 hybrid-assembled isolates from cluster 1 harboring a chromosomal carbapenemase. B, Alignment of isolate 1621 (representing cluster 2) and 2 selected homolog plasmids from an Enterobacter cloacae isolate (accession number KR559890) and a Klebsiella pneumoniae isolate (accession number JN157804). Gray scale reflects direct alignment, while red reflects inverted orientation. Yellow arrows show annotated genes, except for the bla_NDM-1 gene position (red arrows).

Figure 5.

Identification and validation based on cluster-specific peaks. Principal component analysis plot of 17 training (circle) and 3 validation (square) Pseudocitrobacter faecalis sequenced isolates based on cluster-specific peaks from matrix-assisted laser desorption/ionization–time of flight mass spectra. The analysis included 7 isolates from cluster 1 (including 4 retrospective, 2 prospective, and 1 environmental), 12 isolates from cluster 2 (all from the sentinel outbreak, including 1 from January 2021), and 1 retrospective singleton (3447). Abbreviations: cgMLST, core genome multilocus sequence typing; PC, principal component.

Figure 6.

Implementation of the cluster-defining marker. Analysis of 41 matrix-assisted laser desorption/ionization–time of flight spectra from Pseudocitrobacter faecalis isolates based on a single cluster-defining marker (YjbJ) and the related cluster-specific peaks. Isolates included were 17 from the identification, 3 from the validation, and 21 from implementation steps (ie, nonsequenced isolates, designated cluster not available [NA]). Core genome multilocus sequence typing (cgMLST) cluster 1, blue; cgMLST cluster 2, red; cgMLST singleton 3447, yellow; and cluster NA, green.