Cloning, Sequencing, and Characterization of Genomic Subtracted Sequences from Listeria monocytogenes

Abstract

Individual sequences of a genomic subtracted, PCR-amplified, mixed-sequence probe (GS probe) were cloned and sequenced. The GS probe differentiated restriction fragment length polymorphism patterns for Listeria monocytogenes but did not hybridize with members of other bacterial genera. Sequence analysis identified several L. monocytogenes sequences already present in the GenBank database; the putative identities of other sequences were inferred from homology data, and still other sequences did not exhibit significant levels of homology with any GenBank sequences.

Listeria monocytogenes is a gram-positive bacterial food-borne pathogen which may cause listeriosis, a potentially life-threatening disease in members of susceptible at-risk groups, such as pregnant women, infants, the elderly, and immunocompromised individuals (10, 11). High fatality rates in large outbreaks, the ability of L. monocytogenes to grow at refrigeration temperatures, and other recalcitrant physical attributes of this organism have prompted the U.S. Department of Agriculture and the Food and Drug Administration to require zero-tolerance levels in ready-to-eat foods (3, 16). Studies to determine the involvement of L. monocytogenes in food-borne illness have included efforts to develop efficient isolation procedures, efforts to establish the presence and distribution of this organism in foods, efforts to develop rapid identification methods, and efforts to increase our understanding of virulence factors and pathogenesis. Various rapid detection methods have been developed for L. monocytogenes; these methods include immunoassays and nucleic acid-based techniques (9, 10, 20). The nucleic acid-based techniques used to study Listeria spp. have included analyzing cloned genes and virulence factors, developing nucleic acid probes (15), using known sequences to develop PCR detection methods (4), and using molecular probes and/or electrophoretic methods to differentiate strains for epidemiological analysis (2, 5). Generally, analyzing unidentified virulence factors in new and emerging pathogens is a slow and tedious process; it has taken more than a decade to identify the various genes involved in virulence in L. monocytogenes (17). However, subtractive hybridization is a convenient method for isolating multiple unique DNA sequences that may be present in two nucleic acid pools.

MATERIALS AND METHODS

The bacterial strains used in this study are listed in Table 1. Listeria strains were cultivated in brain heart infusion broth (Accumedia, Baltimore, Md.) overnight at 37°C before DNA isolation. The L. monocytogenes strains used for restriction fragment length polymorphism (RFLP) analysis were isolated previously from commercial frankfurters (21). Bacterial strains other than the Listeria strains were grown in Trypticase soy broth (Difco, Detroit, Mich.) overnight at 30 or 37°C before genomic DNA isolation. Escherichia coli DH5α, which was used as a host strain for DNA cloning, was propagated in Luria-Bertani medium at 37°C (19).

TABLE 1.

Bacterial strains and RFLP results

Strain	PMM culture collection no.	EcoRI RFLP group	Source
L. monocytogenes CW 2	PMM 631	2	Hot dog
L. monocytogenes CW 7	PMM 635	4	Hot dog
L. monocytogenes CW 14	PMM 641	4	Hot dog
L. monocytogenes CW 28	PMM 643	5	Hot dog
L. monocytogenes CW 32	PMM 645	8	Hot dog
L. monocytogenes CW 33	PMM 646	6	Hot dog
L. monocytogenes CW 34	PMM 647	7	Hot dog
L. monocytogenes CW 35	PMM 648	9	Hot dog
L. monocytogenes CW 39	PMM 652	3	Hot dog
L. monocytogenes CW 43	PMM 654	10	Hot dog
L. monocytogenes CW 44	PMM 655	11	Hot dog
L. monocytogenes CW 45	PMM 656	8	Hot dog
L. monocytogenes CW 50	PMM 660	5	Hot dog
L. monocytogenes CW 52	PMM 662	9	Hot dog
L. monocytogenes CW 70	PMM 667	10	Hot dog
L. monocytogenes CW 72	PMM 669	9	Hot dog
L. monocytogenes Scott A	PMM 376	1	C. Donelly
L. monocytogenes V23 (MJ105)	PMM 139		M. Johnson
L. monocytogenes 4233 (MJ23)	PMM 141		M. Johnson
L. monocytogenes ATCC 15313	PMM 142		M. Johnson
L. monocytogenes V7 (MJ1)	PMM 144		M. Johnson
L. monocytogenes ATCC 19112	PMM 362		M. Loessner
L. monocytogenes ATCC 19114	PMM 366		M. Loessner
L. monocytogenes ATCC 19116	PMM 368		M. Loessner
L. monocytogenes ATCC 19118	PMM 370		M. Loessner
L. innocua ATCC 33090	PMM 345		J. Radosevic
L. welshimeri (fish isolate)	PMM 355		J. B. Luchansky
Escherichia coli DH5α	PMM 152		Gibco BRL
Escherichia coli M177 (O157:H7)	PMM 159		R. Linton
Aeromonas hydrophila ATCC 49140	PMM 158		ATCCa
Campylobacter jejuni ATCC 29428	PMM 157		ATCC
Salmonella enteritidis ATCC 13076	PMM 111		ATCC
Shigella sonnei ATCC 11060	PMM 156		ATCC
Yersinia enterocolitica ATCC 23715	PMM 155		ATCC
Bacillus cereus ATCC 11778	PMM 512		ATCC
Clostridium perfringens P.I.	PMM 514		Presque Isle cultures
Enterococcus faecalis ATCC 19433	PMM 229		ATCC
Lactobacillus acidophilus N2	PMM 71		P. Muriana
Micrococcus luteus ATCC 8166	PMM 150		ATCC
Staphylococcus aureus ATCC 12600	PMM 510		ATCC

^aATCC, American Type Culture Collection. 

The genomic DNA of Listeria and other bacterial species used for Southern hybridization were isolated by using a modified alkaline lysis method (19, 22). Previously, we described a genomic subtraction process for L. monocytogenes (22). Briefly, we utilized partially digested (AluI) L. monocytogenes target DNA to which linkers were ligated and biotinylated subtracter DNA from Listeria innocua or Listeria ivanovii, which was subsequently sheared (22). L. monocytogenes-specific probe DNA sequences were then obtained by liquid hybridization of the subtracter DNA with L. monocytogenes DNA. After hybridization, the mixture was treated with streptavidin and extracted with phenol, which removed excess subtracter DNA and L. monocytogenes-subtracter DNA hybrids. We expected that the remaining sequences would be enriched for L. monocytogenes DNA that was not complementary to the subtracting sequences. These sequences were increased by PCR amplification of primer-specific sequences with the previously added linkers, and the resulting mixture was referred to as the genomic subtracted mixed probe (GS probe) (22). L. monocytogenes-specific probe DNA sequences were cloned by using two sources, genomic DNA and the GS probe. DNA was digested with either EcoRI (for cloning of genomic DNA) or SalI (for cloning of the GS probe sequences). The digested DNA was ligated into the equivalent restriction site on a plasmid vector, either pBluescript KS+ or pSA3 (8), and transformed into E. coli DH5α by using electroporation (Gene Pulser; Bio-Rad, Richmond, Calif.) at 2,500 V, 25 μF, and 100 Ω. Transformants were selected on Luria-Bertani medium plates supplemented with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) and ampicillin and were screened as described by Sambrook et al. (19). The identities of putative clones were confirmed by colony hybridization with the mixed probe by using a colony lift procedure described by Sambrook et al. (19). The GS probe used in the hybridization experiments was labeled with biotin by incorporating biotin-11-dUTP (Boehringer Mannheim, Indianapolis, Ind.) or biotin-14-dCTP (Tropix, Bedford, Mass.) during PCR amplification, as described previously (22).

Southern hybridization was carried out by using an alkaline phosphatase-based chemiluminescent detection system as previously described (22). In order to demonstrate the specificity of the mixed probe for Listeria spp., Southern blot hybridization experiments were done with representatives of food-borne gram-positive and gram-negative pathogenic bacteria and several other genera that may be found in food. The specificity of individually cloned GS probe sequences was also determined by Southern blot hybridization with genomic digests of L. monocytogenes, L. innocua, and Listeria welshimeri (the only other Listeria species exhibiting a cross-hybridization reaction with the GS probe). Southern hybridization experiments were also done with strains of L. monocytogenes previously isolated from frankfurters (21). DNA sequencing was carried out with alkaline-denatured plasmid DNA by using a DNA sequencing kit (Sequenase 2.0; U.S. Biochemicals, Cleveland, Ohio) as recommended by the supplier. Sequence information was analyzed by using the BLAST algorithm obtained through the National Center for Biotechnology Information network service (1). Nucleotide and peptide sequences were aligned by using the GCG sequence computer analysis system (Genetics Computer Group, Madison, Wis.).

RESULTS AND DISCUSSION

A mixed-sequence, genomic subtracted probe for L. monocytogenes was isolated previously; this probe did not hybridize to L. innocua, L. ivanovii, Listeria seeligeri, Listeria grayi, or Listeria murrayi, although a cross-hybridization reaction occurred with L. welshimeri (22). In this study, the specificity of the GS probe was tested by using representatives of 12 bacterial genera comprised of gram-positive food-borne pathogens (Bacillus cereus, Clostridium perfringens, L. monocytogenes, Staphylococcus aureus), gram-negative enteric food-borne pathogens (Aeromonas hydrophila, Campylobacter jejuni, E. coli, Salmonella enteritidis, Shigella sonnei, Yersinia enterocolitica), and other food-related bacteria (Enterococcus faecalis, Lactobacillus acidophilus, Micrococcus luteus). A cross-hybridization reaction occurred with L. welshimeri (22), but the multiple-sequence GS probe did not hybridize to members of five other Listeria species (22) or 12 other bacterial genera (Fig. 1A), indicating that it contained predominantly L. monocytogenes-specific sequences, as intended by the subtraction process.

FIG. 1.

(A) Southern blot hybridization of the GS probe with EcoRI-digested genomic DNA from L. monocytogenes ScottA and representative strains of other bacterial genera. Lanes 1 and 15, 1-kb DNA ladder size standards; lane 2, A. hydrophila ATCC 49140; lane 3, C. jejuni ATCC 29428; lane 4, E. coli M177 (serotype O157:H7); lane 5, S. enteritidis ATCC 13076; lane 6, S. sonnei; lane 7, Y. enterocolitica ATCC 23715; lane 8, L. monocytogenes; lane 9, B. cereus ATCC 11778; lane 10, C. perfringens P. I.; lane 11, E. faecalis ATCC 19433; lane 12, L. acidophilus N2; lane 13, M. luteus; lane 14, S. aureus ATCC 12600. (B) RFLP analysis of EcoRI-digested genomic DNA from 16 strains of L. monocytogenes isolated from six brands of hot dogs. Lanes 1 and 20, 1-kb DNA ladder size standard; lane 2, L. monocytogenes Scott A; lanes 3 to 18, L. monocytogenes hot dog isolates CW 2, CW 39, CW 7, CW 28, CW 33, CW 14, CW 34, CW 32, CW 35, CW 43, CW 44, CW 45, CW 50, CW 52, CW 70, and CW 72, respectively; lane 19, L. innocua PMM 355.

Since the L. monocytogenes-specific GS probe could identify multiple DNA fragments in a Southern blot analysis of the L. monocytogenes genome (Fig. 1A) (22), we examined whether it could be used in an RFLP analysis to differentiate L. monocytogenes strains isolated from foods. The GS probe identified 11 RFLP patterns among 16 strains of L. monocytogenes isolated from six brands of frankfurters obtained from local supermarkets, which demonstrated the resolving power of this probe for epidemiological analysis and confirmed the genetic diversity in L. monocytogenes (Fig. 1B) (21). As determined by the RFLP patterns, the same strains were isolated from different brands of frankfurters (Fig. 1B, lanes 5 and 8 and lanes 6 and 15), indicating that these foods may have had the same contamination source (i.e., the same supplier of a raw product ingredient) or, in the case of copackaged products, the same manufacturer (i.e., the processing equipment or environment was contaminated). Likewise, the same strains were isolated from samples of different lots of the same brand over the course of 1.5 years (Fig. 1B, lanes 10 and 14, lanes 12 and 17, and lanes 11, 16, and 18). The variety of strains isolated over time from this brand suggests that product contamination was a recurring problem and may well have been due to postprocess contamination during packaging, as previously suggested (21). The data confirmed that a mixed genomic subtracted probe is an excellent diagnostic probe for epidemiological investigations of food-borne pathogens. Some nucleic acid-based differentiation methods may also recognize multiple DNA fragments in target microorganisms; these methods include ribotyping (2, 6), RFLP analysis (12, 18) and randomly amplified polymorphic DNA analysis (7, 13, 14). However, many of the sequences may be harbored by many nontarget species, whereas a genomic subtracted probe that identifies numerous sequences unique to the target microorganism (e.g., L. monocytogenes) may be very useful for distinguishing related strains and may provide a source of DNA for direct cloning and analysis of the sequences that comprise the DNA.

We obtained five genomic clones whose sizes ranged from 1.9 to 5.9 kb (data not shown) and 94 GS probe fragments whose sizes ranged from 44 to 520 bp. A comparison of the 94 sequences resulted in identification of 21 unique sets of fragments (Table 2). Nucleotide sequences (data not shown) and deduced amino acid sequences were aligned in order to determine the putative identities of the PCR clones (Fig. 2). On the basis of the sequence analysis we identified two different partial clones of the listeriolysin O hemolysin (hlyA) and putatively identified sequences encoding ATPase (two clones), a fructose transport enzyme (PTS fru II BC; three clones), a repressor protein for the SOS DNA repair system (lexA; one clone), and a bacteriophage-related clone (φX174; one clone). The identities of the remaining 12 unique GS probe clones could not be determined (Table 2).

TABLE 2.

PCR clones of genomic subtracted sequences and putative identities

PCR clone	GenBank accession no.	Size (bp)	Gene function or putative identity	Reaction with L. monocytogenes	Reaction with other bacteria
1-27	AF180026	264	Hemolysin	13/13a
1-43	AF180027	260	Hemolysin	13/13
1-14	AF180028	162	ATPase	11/13
3-15	AF180029	257	ATPase	11/13
4-31	AF180030	208	PTS fru II BC	11/13
2-14	AF180031	211	PTS fru II BC	10/13
2-28	AF180032	144	PTS fru II BC	2/13
4-16	AF180033	216	LexA	11/13
3-50	AF180034	300	Bacteriophage	3/13
1-22	AF180035	196	Unknown	11/13
3-36	AF180036	284	Unknown	11/13
2-31	AF180037	262	Unknown	11/13
3-18	AF180038	170	Unknown	10/13
3-4	AF180039	396	Unknown	8/13
4-22	AF180040	123	Unknown	8/13
3-7	—b	44	Unknown	5/13
1-16	AF180041	216	Unknown	3/13
1-35	AF180042	164	Unknown	11/13	L. welshimeri
2-18	AF180043	196	Unknown	10/13	L. welshimeri
2-22	AF180044	131	Unknown	5/13	L. welshimeri
2-40	AF180045	219	Unknown	4/13	L. welshimeri

^aNumber of positive reactions/number of experiments. All of the PCR clones reacted with L. monocytogenes Scott A (i.e., the source of the genomic DNA). 

^b—, the sequence was too small to obtain an accession number. 

FIG. 2.

Alignment of partial deduced amino acid sequences for several putative listerial genes determined in this study with amino acid sequences of other microorganisms. (A) Alignment of the putative listerial PTS fructose-permease IIBC component (PCR clone 4-31) with the fructose-specific IIBC components from E. coli (accession no. P20966), Xanthomonas campestris (P23355), Rhodobacter capsulatus (P23387), and Bacillus amyloliquefaciens (P41029). (B) Alignment of the putative listerial ATPase (PCR clone 3-15) with the ATPases from Propionigenium modestum (accession no. S25827), Thiobacillus ferrooxidans (P41167), Haemophilus influenzae (P43714), and Vibrio alginolyticus (P12985). (C) Alignment of the putative listerial LexA repressor protein (PCR clone 4-16) with the LexA repressor proteins from Pseudomonas aeruginosa (accession no. P37452), Pseudomonas putida (P37453), and Bacillus subtilis (P31080).

The specificity of each GS probe clone was determined by performing Southern hybridization with 13 strains of L. monocytogenes selected from our culture collection (Table 2) (data not shown). The two hlyA fragments hybridized to all 13 strains of L. monocytogenes but not to control DNA from L. innocua and L. welshimeri. Although we expected more known virulence-related genes for L. monocytogenes to be among our GS probe clones, it is possible there were few such genes because (i) all of the available GS sequences may not have been recovered from the GS mixture or (ii) the GS sequences may have been removed by subtraction by contiguous regions that were homologous to subtracter sequences.

The putatively identified sequences for ATPase (clones 1-14 and 3-15), PTS fru II BC (clone 4-31), and LexA (clone 4-16) and three clones whose identities were not determined (clones 1-22, 3-36, and 2-31) hybridized to 11 of the 13 L. monocytogenes strains, although not necessarily the same 11 strains in each case (Table 2). The other GS probe clones did not react with as many of the 13 test strains. We also identified three GS probe clones which encoded part of the fructose transport enzyme, PTS fru II BC (clones 4-31, 2-14, and 2-28) (Table 2), and four GS probe clones that also hybridized to L. welshimeri (Table 2) (clones 1-35, 2-18, 2-22, and 2-40) (Table 2), as well as a 10-kb genomic fragment (data not shown).

Our data confirm that the genomic subtraction process can be used for identification and isolation of unique sequences in prokaryotes. Genomic subtraction may be a quick and facile mechanism to obtain important, perhaps subtle, genetic information for strains for which little or no information is available and could be important in investigations of virulent and avirulent food-borne pathogens.