Plasmid-Borne Type E Neurotoxin Gene Clusters in Clostridium botulinum Strains

Abstract

A collection of 36 Clostridium botulinum type E strains was examined by pulsed-field gel electrophoresis (PFGE) and Southern hybridization with probes targeted to botE and orfX1 in the neurotoxin gene cluster. Three strains were found to contain neurotoxin subtype E1 gene clusters in large plasmids of about 146 kb in size.

TEXT

Botulinum neurotoxin (BoNT) is the most potent natural toxin, produced by Clostridium botulinum and some strains of Clostridium butyricum and Clostridium baratii. The toxin blocks neurotransmitter release in cholinergic nerve ends, causing flaccid paralysis known as botulism (1). Based on their antigenic properties, BoNTs are classified into types A to G, of which toxins A, B, E, and F cause botulism in humans. Moreover, sequence analysis and binding studies with monoclonal antibodies have revealed several subtypes within most neurotoxin serotypes (2–6).

The genomic location of the neurotoxin gene cluster is variable and partly serotype dependent. In the early 1970s, the botC and botD clusters were observed to be encoded by bacteriophages (7). Recent genomic studies revealed that their prophages exist as large plasmids (8, 9). The botG cluster is carried by a plasmid (10). The botA, botB, and botF clusters were initially found in the bacterial chromosome; however, plasmid-borne toxin genes have recently been shown for subtypes A3 (11), B1, B2, the nonproteolytic type B (12–14), and some bivalent subtypes, such as Ab, Ba, and Bf (15). The two available C. botulinum type E genome sequences (NCBI) show chromosomal botE clusters, as is the case with neurotoxigenic C. butyricum (16). Type E neurotoxin genes carried by extrachromosomal elements have not been reported.

Here we evaluated the genomic location of the botE gene cluster in 36 C. botulinum type E strains isolated from aquatic environments, fish, and fish products (Table 1). All strains were incubated anaerobically in 10 ml of tryptone-peptone-glucose-yeast extract (TPGY) medium at 30°C. Agarose plugs containing genomic DNA of C. botulinum type E strains were prepared as described previously (1) and analyzed by pulsed-field gel electrophoresis (PFGE) in a 1% agarose gel. All samples including chromosomal DNA showed migration to a short distance. Smeared backgrounds were observed for some strains in the PFGE gel, indicating that the genomic DNA had degraded to some degree. This is probably due to the high levels of endogenous DNases produced by C. botulinum type E strains (1). In the gel lanes, DNA bands ranging in size from 9 to 146 kb were shown for 11 strains, including 250, CB11/1-1, K51, K22, K15, K101, K119, K23, K25, K28, and K8 (Fig. 1A), suggesting the presence of extrachromosomal elements. To measure the size of the potential plasmids accurately, the DNA embedded in PFGE plugs were incubated with S1 nuclease (Promega, Madison, WI) at 37°C for 45 min prior to electrophoresis to linearize the supercoiled plasmid DNA (17). However, for an unknown reason, S1 nuclease caused degradation of the genomic DNA of C. botulinum type E. Therefore, the sizes of the extrachromosomal elements were estimated based on non-S1-treated DNA and may thus represent supercoiled plasmids, making the size estimate inaccurate.

Table 1.

Clostridium strains used in this study

Strain	Lane no. in PFGE gel	Origin	Country	Yr	Sourcea	botE subtype
250	1	Canned salmon	UK	1978	Crowther/Lindrothb	NDc
CB11/1-1	2	Whitefish roe	Finland	1999	DFHEH	E1
K51	3	Rainbow trout surface	Finland	1996	DFHEH	E1
Beluga	4	Fermented white whale flippers	USA	1951	Dolman/Lindrothb	E1
K22	5	Burbot surface	Finland	1996	DFHEH	E1
341	6	European river lamprey	Finland	2005	DFHEH	E6
350	7	European river lamprey	Finland	2005	DFHEH	ND
4062	8	Muktuk	USA	1981	Hatheway/Lindrothb	ND
92	9	Marine environment	USA	1960s	Eklund/Lindrothb	ND
BL81/31	10	Smoked salmon	Canada	1934	IFR	E1
BL86/21	11	Fish	ND	ND	IFR	ND
BL87/1	12	Fish	USA	ND	IFR	ND
BL93/7	13	Fish	USA	ND	IFR	E1
BL93/8	14	Human	USA	ND	IFR	E2
C60	15	Dried mutton	Denmark	1989	SSI	ND
K15	16	Rainbow trout intestines	Finland	1995	DFHEH	E1
K101	17	Chub surface	Germany	1997	DFHEH	E1
K117	18	Smoked Canadian whitefish	Germany	1997	DFHEH	E3
K119	9	Rainbow trout intestines	Finland	1996	DFHEH	E1
K126	20	Smoked Alaskan salmon	Finland	1996	DFHEH	E3
K23	21	Burbot intestines	Finland	1996	DFHEH	ND
K25	22	Burbot surface	Finland	1996	DFHEH	E1
K28	23	Burbot surface	Finland	1996	DFHEH	ND
K3	24	Rainbow trout surface	Finland	1995	DFHEH	E3
K35	25	Vendace	Finland	1996	DFHEH	E6
K36	26	Smoked rainbow trout	Finland	1996	DFHEH	E6
K37	27	Smoked whitefish	Finland	1996	DFHEH	E1
K44	28	Rainbow trout intestines	Finland	1996	DFHEH	E3
K47	29	Rainbow trout surface	Finland	1996	DFHEH	ND
K7	30	Rainbow trout intestines	Finland	1995	DFHEH	E6
K76	31	Hot-smoked vendace	Finland	1997	DFHEH	E1
K8	32	Rainbow trout intestines	Finland	1995	DFHEH	E1
KA2	33	Seola Creek strain	USA	ND	Riemann/Lindrothb	ND
M3	34	Vegetarian sausage	Finland	2001	DFHEH	E3
RS1	35	Pacific red snapper	USA	1983	Lindrothb	ND
S16	36	Fish pond sediment	Finland	1997	DFHEH	E6

^aDFHEH, Department of Food Hygiene and Environmental Health, University of Helsinki, Helsinki, Finland; IFR, Institute of Food Research, Norwich, United Kingdom; SSI, Statens Serum Institut, Copenhagen, Denmark.

^bCollected from John Crowther, Claude Dolman, Charles Hatheway, Melvin Eklund, and Hans Riemann by Seppo Lindroth (University of California, Davis).

^cND, no data available.

Fig 1.

PFGE of nondigested genomic DNA from C. botulinum type E (A) and Southern hybridization with probes specific for 16S rrn (B), botE (C), and orfX1 (D). Strain numbers correspond to those listed in Table 1. Each strain was loaded with genomic DNA (left) and S1 nuclease-treated genomic DNA (right) in two adjacent gel wells. For the strains shown in lanes 14, 23, 27, and 32, the marker (M) was loaded in the lane between the genomic DNA and S1 nuclease-treated genomic DNA. The white arrows indicate plasmid bands (A), and the black arrows indicate plasmid bands hybridized with probes specific for botE (C) and orfX1 (D).

To further define the extrachromosomal elements, DNA of the PFGE gels was transferred to charged nylon membranes (Roche Applied Science, Mannheim, Germany) and hybridized with digoxigenin (DIG)-labeled 16S rrn probe. DIG-labeled hybridization probes were prepared by PCR labeling using primers listed in Table 2. Strong signals were observed for the large genomic DNA bands, whereas no hybridization signal with 16S rrn probe was observed for the smaller bands (Fig. 1B), suggesting that these elements represented plasmids. These results suggest that 30% of the C. botulinum type E strains studied harbor one or more plasmids ranging in size from 9 to 146 kb. The proportion is consistent with our previous study of 54 C. botulinum type E isolates (18). Of the 11 plasmid-containing strains, 8 represented the botE subtype E1 (Table 1).

Table 2.

Oligonucleotide primers used in this study

Primer	Sequence (5′→3′)
botE-F	TTTTTGTGGCTTCCGAGAAT
botE-R	TATTTTCACCTTCGGGCACT
orfX1-F	AAAATTACATTTATTGATGGTTACCTG
orfX1-R	AGAATTCCATTTTTAGTTATCCTTTTT
16S rrn-F	GAAGGCGACTTTCTGGACTG
16S rrn-R	TAAGGTTCTTCGCGTTGCTT

To test whether any of the detected plasmids included the neurotoxin genes, the transfer membranes were stripped and subsequently rehybridized with probes specific for botE and finally with those specific for orfX1. Specific and distinct hybridization signals with the botE probe, consistent with the hybridization signals obtained with the 16S rrn probe, were observed for the chromosomal DNA from all but three strains, suggesting botE is in a chromosomal locus in these strains (Fig. 1C). In contrast, for strains CB11/1-1, K51, and K22, the botE probe hybridization signals were observed in extrachromosomal bands of approximately 146 kb in size, indicating their botE gene is located in a large plasmid (Fig. 1C). This finding was further confirmed by stripping and rehybridization with the orfX1 probe (Fig. 1D). The strains with plasmid-borne botE were analyzed at least in triplicate (see Fig. S1 in the supplemental material). These results suggest that C. botulinum type E strains CB11/1-1, K51, and K22 carry their neurotoxin gene clusters in large plasmids.

All three strains carrying a plasmid-borne botE gene were isolated in the 1990s from fish in Finland (Table 1). CB11/1-1 was involved in a food botulism outbreak in Finland in 1999 (19) and was epidemiologically unrelated to K51 and K22. Previous genotyping studies showed CB11/1-1 and K51 cluster together and are genetically distant from K22 (20–22), suggesting that the botE-containing plasmids are carried by C. botulinum strains representing at least two distinct genetic backgrounds. Our in-house amplified fragment length polymorphism (AFLP) typing collection shows that CB11/1-1 is identical to the nontoxigenic C. botulinum strain H61 (96% similarity) in AFLP analysis with an 89% cutoff value (data not shown), suggesting lateral transfer of the botE plasmid between toxigenic and nontoxigenic strains. Strain K51 showed a close genetic relatedness (higher than 85% similarity) to the chromosomal botE-carrying strains K23 and K28 (22). It is not known if the plasmids detected in these two strains, being smaller than the botE plasmid of K51, show similarity to the botE plasmid or if these two strains obtained their chromosomal neurotoxin genes as a result of plasmid transfer and further integration into their chromosome. Further comparative genome study will provide more evidence to understand the role of botE-containing plasmids in the evolution of C. botulinum.

Previous studies suggested that the plasmids carrying neurotoxin subtype A3 or A4 genes were approximately 270 kb in size, while those carrying neurotoxin type B genes varied considerably in size, ranging from 55 to 245 kb (11, 12). However, all three botE-containing plasmids have a similar size of approximately 146 kb. Although a previous study on 13 C. butyricum isolates showed all strains to carry their botE in the chromosome, the PFGE image indicated that four isolates, including those from two infant botulism cases, might harbor plasmids ranging in size from 120 to 170 kb (16). Whether such large plasmids in C. butyricum show regions of homology to the botE-containing plasmids detected here requires further investigation. Mobility of plasmids is often mediated by conjugation mechanism (23). A recent study revealed that bot-containing plasmids were able to conjugatively transfer from strains CDC-A3, 657 Ba4, and Eklund 17Bnp to a nontoxigenic mutant of strain 62A1 (24). Further studies are warranted to evaluate if the botE-containing plasmids can be conjugatively transferred to other C. botulinum or C. butyricum strains.

In conclusion, we provide the first evidence of plasmid-borne type E neurotoxin gene clusters. All three plasmid-borne neurotoxin genes represented subtype E1 and were located within large plasmids of approximately 146 kb in size. These results may have important implications for understanding the evolution and toxigenicity of C. botulinum type E.