Core Genome Multilocus Sequence Typing for Identification of Globally Distributed Clonal Groups and Differentiation of Outbreak Strains of Listeria monocytogenes

Abstract

Many listeriosis outbreaks are caused by a few globally distributed clonal groups, designated clonal complexes or epidemic clones, of Listeria monocytogenes, several of which have been defined by classic multilocus sequence typing (MLST) schemes targeting 6 to 8 housekeeping or virulence genes. We have developed and evaluated core genome MLST (cgMLST) schemes and applied them to isolates from multiple clonal groups, including those associated with 39 listeriosis outbreaks. The cgMLST clusters were congruent with MLST-defined clonal groups, which had various degrees of diversity at the whole-genome level. Notably, cgMLST could distinguish among outbreak strains and epidemiologically unrelated strains of the same clonal group, which could not be achieved using classic MLST schemes. The precise selection of cgMLST gene targets may not be critical for the general identification of clonal groups and outbreak strains. cgMLST analyses further identified outbreak strains, including those associated with recent outbreaks linked to contaminated French-style cheese, Hispanic-style cheese, stone fruit, caramel apple, ice cream, and packaged leafy green salad, as belonging to major clonal groups. We further developed lineage-specific cgMLST schemes, which can include accessory genes when core genomes do not possess sufficient diversity, and this provided additional resolution over species-specific cgMLST. Analyses of isolates from different common-source listeriosis outbreaks revealed various degrees of diversity, indicating that the numbers of allelic differences should always be combined with cgMLST clustering and epidemiological evidence to define a listeriosis outbreak.

Importance: Classic multilocus sequence typing (MLST) schemes targeting internal fragments of 6 to 8 genes that define clonal complexes or epidemic clones have been widely employed to study L. monocytogenes biodiversity and its relation to pathogenicity potential and epidemiology. We demonstrated that core genome MLST schemes can be used for the simultaneous identification of clonal groups and the differentiation of individual outbreak strains and epidemiologically unrelated strains of the same clonal group. We further developed lineage-specific cgMLST schemes that targeted more genomic regions than the species-specific cgMLST schemes. Our data revealed the genome-level diversity of clonal groups defined by classic MLST schemes. Our identification of U.S. and international outbreaks caused by major clonal groups can contribute to further understanding of the global epidemiology of L. monocytogenes.

FIG 1

Phylogenetic tree of all isolates determined by 1,827-cgMLST. Clonal complexes (CCs) or epidemic clones (ECs) were defined by MLST. CC or EC clades containing more than one isolate are collapsed into triangles. The lineage, serotype (if available), CC, singleton, or EC names are listed in the taxa for all except two isolates (BL0142 and BL0105), for which the sequence types were not observed in the Pasteur MLST database and were thus designated NO. CCs or ECs of the same serotype have the branches and taxa in the same color. Serotype IVb-v1 (4bv) is a serotype 4b variant and has a color different from the color used for serotype 4b. If a CC or EC contains more than one serotype, the color representing the serotype of the majority of the isolates is used. If a CC or EC does not have serotype information available, the color of the nearest branch is used. None of the triangles contain isolates from more than one CC or EC. The triangle length is proportional to the diversity of isolates within each clonal group. CC398 isolates and CC240 isolates had relatively low levels of diversity, and thus, their triangles appear as vertical lines.

FIG 2

(A to C, E to I, and K) Subtrees of CC1/ECI (A), CC6/ECII (B), CC2/ECIV (C), CC5/ECVI (E), CC3/ECVIII (F), ST382/ECIX (G), CC4/ECX (H), CC11/ECIII (I), CC8/ECV (K), CC7/ECVII (L), and CC101 (M) by lineage-specific cgMLST schemes without accessory genes. (D) Subtree of the 1987-1989 UK pâté outbreak and outgroup determined by 1,701-cgMLST. The subtree has the same clustering as that obtained by 1,827-cgMLST. (J) Subtree of the 1988 U.S. sporadic case linked to hot dog, 2000 U.S. turkey deli meat outbreak, and outgroups by LmLII-cgMLST with the inclusion of accessory genes. Isolate identifiers are preceded by lineage and available serotype information and followed by the MLST ST in parentheses. The outbreak names are listed on the right side of the brackets. The maximum number of pairwise allele differences (ADs) among isolates of each outbreak strain, CC, or EC is listed in parentheses after their names in the subtrees of lineage-specific cgMLST without accessory genes. STs of three isolates (CFSAN025771 and CFSAN025772 in panel C and L1023 in panel I) were not observed (NO) in the Pasteur MLST database, but they differed from existing STs by one MLST allele and thus were assigned to CCs.

FIG 2

(A to C, E to I, and K) Subtrees of CC1/ECI (A), CC6/ECII (B), CC2/ECIV (C), CC5/ECVI (E), CC3/ECVIII (F), ST382/ECIX (G), CC4/ECX (H), CC11/ECIII (I), CC8/ECV (K), CC7/ECVII (L), and CC101 (M) by lineage-specific cgMLST schemes without accessory genes. (D) Subtree of the 1987-1989 UK pâté outbreak and outgroup determined by 1,701-cgMLST. The subtree has the same clustering as that obtained by 1,827-cgMLST. (J) Subtree of the 1988 U.S. sporadic case linked to hot dog, 2000 U.S. turkey deli meat outbreak, and outgroups by LmLII-cgMLST with the inclusion of accessory genes. Isolate identifiers are preceded by lineage and available serotype information and followed by the MLST ST in parentheses. The outbreak names are listed on the right side of the brackets. The maximum number of pairwise allele differences (ADs) among isolates of each outbreak strain, CC, or EC is listed in parentheses after their names in the subtrees of lineage-specific cgMLST without accessory genes. STs of three isolates (CFSAN025771 and CFSAN025772 in panel C and L1023 in panel I) were not observed (NO) in the Pasteur MLST database, but they differed from existing STs by one MLST allele and thus were assigned to CCs.