|Home | About | Journals | Submit | Contact Us | Français|
Listeria monocytogenes epidemic clone II (ECII) strains have been responsible for two major multistate outbreaks of food-borne listeriosis in the United States, but their prevalence and ecology remain poorly understood. In this study, we describe DNA probes that unambiguously identify this clonal group. These probes were able to differentiate ECII strains of outbreak, sporadic, or environmental origin from other L. monocytogenes strains of the same serotype (4b).
Most outbreaks of food-borne listeriosis have involved a small number of well-defined clonal groups. Population genetic and epidemiologic studies have revealed three serotype 4b clonal groups, designated epidemic clones (ECs). ECI, ECIa, and ECII have been responsible for repeated outbreaks of food-borne listeriosis (9, 19, 28). ECI and ECIa were recognized in the earliest epidemiologically characterized outbreaks of listeriosis (e.g., coleslaw outbreak in the Maritime Provinces in 1979 and Mexican-style cheese outbreak in California in 1985, both involving ECI, as well as an outbreak in Massachusetts in 1983, involving ECIa) (2, 9, 19, 24). In contrast, ECII was not recognized until the hot dog-associated multistate outbreak in the United States in 1998-1999 (4, 5, 16, 28). A subsequent multistate outbreak in 2002 was attributed to contaminated turkey deli meats and also involved ECII strains (6, 20). In addition, ECII strains have contributed to sporadic human listeriosis and were isolated from environmental samples of food-processing plants (13, 18, 25, 26).
Our understanding of the epidemiology and ecology of ECII remains limited, partially due to the unavailability of tools for unambiguous detection of these strains by using relatively simple formats (e.g., hybridizations with specific DNA probes or PCR with specific primers). ECII strains were found to have diverged in a genomic region (“region 18”) that was otherwise specific to serotype 4b and conserved among other serotype 4b strains (14). The genome of ECII strain H7858 (1998-1999 outbreak) has been sequenced (23), and DNA probes derived from region 18 of this strain hybridized with genomic DNA of all tested ECII strains (20). However, hybridization was also observed with two serotype 4b strains with pulsed-field gel electrophoreses (PFGE) profiles clearly different from ECII, suggesting that region 18 sequences of H7858 were also harbored by certain unrelated serotype 4b strains (20). Similar findings were reported by Chen and Knabel, who employed PCR with primers derived from region 18 of H7858 (8). Thus, a continuing need for DNA-based tools for unambiguous detection of ECII strains exists. This would supplement other subtyping tools (e.g., PFGE) to facilitate rapid detection and monitoring of clones of L. monocytogenes for epidemiological and risk assessment purposes.
In this study, we assessed the ECII specificity of selected DNA probes derived from four different genomic regions present in the chromosome of ECII strain H7858 but not in the fully sequenced genome of epidemic clone I strain F2365, also of serotype 4b, or the genomes of the serotype 1/2a strains EGD-e and F6854 (15, 23; www.broadinstitute.org). These regions included (i) the previously identified region 18 (14, 20) and fragments from wap, which encodes a putative wall-associated protein and flanks region 18 on one side (the other side is flanked by inlA and inlB, encoding internalins A and B, respectively), (ii) region 1168, immediately upstream of the gene cluster that includes the gene for listeriolysin S and is harbored by certain lineage I (serotypes 1/2b, 3b, and 4b) strains (10), (iii) region 0088 (corresponding to LMOh7858_0088), downstream of LMOh7858_0075 (encoding a putative FitsK/SpoIII), and (iv) region 2753, flanked by LMOh7858_2752 and rpsL and recently implicated in temperature-dependent phage resistance in ECII strains (21; J.-W. Kim and S. Kathariou, unpublished data).
The DNA probes were obtained by PCR using relevant primers (Table (Table1)1) and tested with a panel of 60 serotype 4b strains from our laboratory's Listeria strain collection (Fig. (Fig.1).1). The strain panel was chosen so as to include 12 ECII strains from the 1998-1999 and 2002 outbreaks, as well as 19 clinical and environmental isolates with AscI PFGE profiles identical to those of the ECII outbreak strains but without known involvement in outbreaks. PFGE profiles of representative ECII and non-ECII strains from the strain panel are shown in Fig. Fig.2;2; the ECII profiles represent those previously encountered among ECII strains (16, 20). The non-ECII panel included 29 serotype 4b strains with PFGE profiles clearly distinct from those of ECII (Fig. (Fig.22 and data not shown). Non-ECII strains included eight that were assigned to another major epidemic clone, ECI, by previously described procedures (13, 17, 29) and 21 strains that (based on PFGE and other assays) were not members of either ECI or ECII. The latter group (non-ECI, non-ECII strains) included the North Carolina outbreak strain WS1 (22). Strain designations, sources (epidemic, sporadic, or environmental), and years of isolation are indicated in Fig. Fig.11.
Genomic DNAs of the isolates were prepared as described previously (29) and arrayed in multiple (three to six) replicates (each 0.4 μl, ca. 20 ng DNA/spot) on nylon membranes (Osmonics Inc., Minnetonka, MN) by using a VP408 Multi-Blot replicator (V&P Scientific, Inc., San Diego, CA) as described previously (17). Each membrane was hybridized with a different digoxigenin-labeled DNA probe; hybridizations were done as described previously (29) in at least two independent experiments, each of which included hybridizations with a probe internal to the listeriolysin O gene hly as a positive control and in order to assess DNA quantities on the membrane. The serotype 1/2a strain F6854 (23) was used as a negative control for the serotype 4b-derived probes.
The six probes derived from this region (Fig. (Fig.3)3) included three that were employed previously (probes 2, 3, and 7). Probes 2 (H18RP11/12) and 3 (H18RP9/10) were derived from sequences within region 18 present in the genome of H7858 but absent from the sequenced genomes of other strains, including the serotype 4b strain F2365 (ECI) (20, 23). In contrast, probe 7 (4bSF18) was derived from region 18 of F2365 and earlier was shown to hybridize with DNA of F2365 and other serotype 4b strains but not with outbreak-derived ECII strains (14).
Probes 2 and 3 hybridized with all ECII strains. However, they also hybridized with the genomes of eight other strains for which the PFGE profiles were distinct from ECII (Fig. (Fig.1).1). The reactivities of two of these strains (J2255 and J2621) with probes 2 and 3 were reported before (20); the additional strains included J3139, J3215, J3195, 2003-196R, and 2004-471, from sporadic cases of human listeriosis, as well as WS1, implicated in the North Carolina outbreak in 2000 and (based on PFGE and other subtyping data) clearly distinct from ECII (see also Fig. Fig.2)2) (7, 12, 22, 27). Probe 7 hybridized with non-ECII strains only; strains hybridizing with this probe were negative with probes 2 or 3 (and vice versa), with two exceptions: J2621 (also shown to have such dual reactivity earlier ) and J3195 (Fig. (Fig.1).1). These unusual strains may harbor two different types of region 18 cassettes (ECII-like and non-ECII-like) and would be worthy of further genomic characterization. Since several non-ECII strains (including the North Carolina outbreak strain) hybridized with probes 2 and 3, derived from ECII strain H7858, our data confirm and extend previous evidence (8, 20) that region 18 content provides only tentative evidence for ECII or non-ECII status.
Region 18 is flanked on one side by wap (Fig. (Fig.3),3), a serotype 4b-specific gene of ca. 6 kb (with homologs also harbored by L. innocua and L. welshimeri) (14, 17, 23). Results from our laboratory, to be described in detail elsewhere, have shown that the terminal fragment of wap has undergone pronounced sequence divergence between ECII and other serotype 4b strains. We assessed three probes derived from wap sequences; probes 5 (ECIIC-WAP) and 6 (non-ECIIC-WAP) corresponded to 3′ wap terminal fragments of the ECII strain H7858 and of F2365, respectively, whereas probe 8 (Inter-WAP) corresponded to an internal fragment of wap, earlier found to be conserved among various serotype 4b strains (14, 17, 23).
Probe 8 hybridized with all serotype 4b strains in the panel except two (2001-75R and 2003-151R) of sporadic clinical origins, and these were not assigned to any clonal group (Fig. (Fig.1).1). Non-ECII wap terminal probe 6 (non-ECIIC-WAP) hybridized only with a subset of non-ECII strains (including all ECI strains). On the other hand, ECII wap terminal probe 5 (ECIIC-WAP) hybridized with DNA from all ECII strains as well as from three of the eight strains that had also hybridized with probes 2 and 3 (J3215, 2003-196R, and North Carolina outbreak strain WS1) (Fig. (Fig.1).1). Thus, even though more limited to ECII than probes 2 and 3, wap-derived probes did not yield unambiguous results regarding ECII or non-ECII status.
Probes 1 and 4 yielded promising results in an earlier study, hybridizing with DNA from outbreak-derived ECII strains but not from J2255 or J2621 (20), and these were therefore tested further with the current strain panel. Probe 1 was derived from open reading frame (ORF) LMOh7858_1168, which encodes a putative ATPase and was identified through comparative genomic analysis as a member of a DNA gene cassette present in the chromosome of the ECII strain H7858 but not in the other sequenced genomes of L. monocytogenes (10, 23; www.broadinstitute.org). The cassette is here designated as “region 1168.” In F2365, the counterparts of the two genes flanking region 1168 (LMOf2365_1110 [guaA] and LMOf2365_1111, homologous to LMOh7858_1160 [guaA] and LMOh7858_1175, respectively), were adjacent to each other without any intervening genes (Fig. (Fig.4).4). This H7858-specific genomic region also harbors phage-related genes (LMOh7858_1170 and LMOh7858_1163, which encode a putative immunity repressor protein and a putative phage integrase, respectively), and its G+C content (27 to 35%) is lower than the average for the genome (38%). Interestingly, the cassette in H7858 is immediately followed by the listeriolysin S gene cluster (Listeria pathogenicity island 3 [LIPI-3]), also present in the genome of F2365 (Fig. (Fig.4).4). LIPI-3 has been identified only among certain serotype 4b and 1/2b strains (lineage I) and was implicated in virulence (10). It is also of interest that in the serotype 1/2a strain EGD-e, a gene cluster in this genomic location contains a cadmium resistance cassette (cadA cadC) in association with a Tn916-like conjugative integrative chromosomal element (Fig. (Fig.4)4) (15). Recently, cadC was found to be highly expressed in the spleen following intravenous infection of mice with strain EGD-e and was also implicated in virulence (3).
Probe 4 was derived from ORF LMOh7858_0088, encoding a hypothetical protein. The genomic region harboring LMOh7858_0088 (indicated here as region 0088) appears to be diversified across the genomes of different L. monocytogenes strains (Fig. (Fig.5).5). Interestingly, within this genomic region, F2365 shares more sequence similarity with EGD-e than with other strains, while H7858 harbors several ORFs conserved in the serotype 1/2a strain F6854 (Fig. (Fig.5).5). It is noteworthy that both H7858 and F6854 harbor in this genomic region ORF LMOh7858_0080 (encoding putative toxin b), which is absent from the genomes of F2365 and EGD-e. H7858 harbors in addition a related ORF (LMOh7858_0086, with 87% nucleotide sequence identity with LMOh7858_0080), also annotated as encoding a putative toxin (Fig. (Fig.55).
Probe 1 hybridized with all tested ECII strains, regardless of their origin. Furthermore, no hybridization was observed with any of the tested non-ECII serotype 4b strains, including the eight strains with ECII-like region 18 hybridization patterns (Fig. (Fig.1).1). The data suggest that LMOh7858_1168 is unique for, and conserved among, ECII strains. On the other hand, probe 4 hybridized with all ECII strains but also hybridized with nine non-ECII, non-ECI strains (Fig. (Fig.1),1), suggesting that ORF LMOh7858_0088 is a conserved, but not unique, genomic feature of ECII. Analysis of sequenced genomes of L. monocytogenes indeed revealed that sequences corresponding to probe 1 and the 1168 cassette were harbored only by ECII strain H7858. However, sequences homologous to probe 4 and the 0088 cassette were harbored by H7858 as well as serotype 4b strain HPB2262 (23; www.broadinstitute.org); the latter was implicated in a febrile gastroenteritis outbreak of listeriosis (1) and appears to be a member of ECIa (11).
Probes 10 to 13 were derived from LMOh7858_2753, LMOh7858_2754, LMOh7858_2759, and LMOh7858_2764, respectively, members of a gene cassette (“region 2753”) identified through the genome sequencing of H7858 (23). The cassette has a lower-than-average G+C content (ca. 33%) and was recently implicated in growth temperature-dependent phage resistance of ECII strains (21; Kim and Kathariou, unpublished). The gene cassette is flanked by a putative lipoprotein gene (LMOh7858_2752) and a ribosomal protein gene (rpsI); these genes were adjacent to each other in EGD-e, whereas a cluster of genes encoding clustered regularly interspaced short palindromic repeat (CRISPR)-associated proteins was harbored in this location in the genome of F6854 (Fig. (Fig.66).
Testing of a subset of the strain panel with probe 10 revealed that the probe had the same ECII specificity as probe 1, hybridizing with ECII strains only (Fig. (Fig.7);7); identical results were obtained with probes 11 to 13 (data not shown). BLAST searches of sequenced genomes indicated that, as with probe 1 and region 1168, these probes and region 2753 were harbored only by the genome of ECII strain H7858 (23; www.broadinstitute.org).
One may speculate that regions 1168 and 2753 were acquired through horizontal gene transfer (as suggested by their atypical G+C contents) by an ECII ancestral strain and have been maintained in this clonal group, possibly due to fitness advantages they may confer (e.g., phage resistance conferred by genes in region 2753). On the other hand, regions 18 and 0088 may have been acquired by horizontal gene transfer (again suggested by their atypical G+C contents) by a serotype 4b ancestral strain and were maintained only in ECII and certain other strains or, alternatively, may have transferred from ECII to other serotype 4b strains. Further genome content analysis of these strains will be needed to assess their phylogenetic relationships among themselves and with ECII.
In addition to identifying a panel of probes suitable for unambiguous detection of ECII strains, the hybridization data also suggested other clusters of serotype 4b strains with distinct and shared hybridization profiles. One of these strain clusters (strains 4b1, 266, and 82-2a) appears to correspond to ECIa. These strains had DNA that hybridized with probe 4 and with the non-ECII probes for region 18 and wap but not with probe 1 or 10 (Fig. (Fig.11 and and7).7). The strains had closely related PFGE profiles (Fig. (Fig.2),2), and their genotype based on multilocus variable-number tandem-repeat analysis (27) was identical to that of the ECIa strain implicated in the Massachusetts outbreak of 1983 (K. Sperry and S. Kathariou, unpublished data). Genome sequence analysis of strain HPB2262, which appears to be a member of ECIa (11), also revealed that the genome harbored region 0088, as well as the non-ECII region 18 and non-ECII wap terminal region, whereas sequences homologous to region 1168 or 2753 were not detected (Fig. (Fig.1).1). Other apparent clusters included one with strains 2003-196R, WS1 (North Carolina outbreak), and J3215, which could be differentiated from all other tested serotype 4b strains by their positive results with ECII-derived region 18 and wap probes but lack of hybridization with probe 1, 4, or 10 (Fig. (Fig.11 and and7).7). Strains J2255, 2004-471, and J3139 could also be differentiated from all others by positivity with region 18 ECII probes 2 and 3 and positivity with probe 4 but negativity with probes 1 and 5 (Fig. (Fig.11).
In conclusion, the hybridization data from this study clearly indicated that probes derived from two ECII-specific genomic regions (regions 1168 and 2753) could unambiguously differentiate ECII strains from other strains of serotype 4b. These probes (probes 1 and 10 to 13) hybridized with all tested ECII strains regardless of their source but not with any of the other serotype 4b strains that were tested. On the other hand, sequences corresponding to probes from region 18, the terminal portion of wap, and region 0088 were indeed harbored by all tested ECII strains regardless of their source but were also harbored by certain non-ECII strains. We expect that including the probes described here (or other DNA-based tools from genomic regions 18, 1168, 0088, and 2753) along with ECI-specific probes such as 17B (7, 17, 29) in strain subtyping or detection schemes (i.e., conventional or real-time PCR, DNA microarrays, and Luminex xMAP system, etc.) will facilitate accurate monitoring of serotype 4b strains of L. monocytogenes for those strains representing major epidemic clones, such as ECI and ECII, and other clonal groups that may not yet have been recognized.
Funding for this research was partially provided by USDA grant 2006-35201-17377 and by a grant from the American Meat Institute Foundation.
We thank L. M. Graves and K. Sperry for strains and are grateful to all members of our laboratory for encouragement and support.
Published ahead of print on 19 March 2010.