Sequencing the Botulinum Neurotoxin Gene and Related Genes in Clostridium botulinum Type E Strains Reveals orfx3 and a Novel Type E Neurotoxin Subtype

Bacterial strains, media, and growth conditions.A total of 14 C. botulinum type E strains were analyzed in this study. Strains K3, K8, K14, K15, K35, K36, K44, K51, K117, and K119 were Finnish fish isolates, while strain S16 was isolated from sea sediment in Finland. Strains K81 and K101 were German fish isolates. Strain 31-2570 of an unknown source was isolated in the United States. Strains K35, K81, and 31-2570 were sequenced for the entire type E neurotoxin gene cluster. The other 11 strains were sequenced for the bont/E gene. All strains, apart from 31-2570, have been isolated in our laboratory and were selected for this study based on the wide geographical and genetic diversity revealed by pulsed-field gel electrophoresis (10), randomly amplified polymorphic DNA (13), and amplified fragment length polymorphism (16) analyses. Egg yolk agar plates were used to subculture these strains and confirm their purity. A single typical colony was picked and inoculated into 10 ml of anaerobic tryptone-peptone-glucose-yeast extract (Difco, Detroit, MI) broth. All cultures were incubated overnight at 30°C under anaerobic conditions.

PCR amplification.Genomic DNA was extracted from C. botulinum cultures as described by Keto-Timonen et al. (16). All primers for amplification and sequencing are listed in Table 1.

When the genes corresponding to previously published ones were being sequenced, PCR was applied to amplify suitable DNA fragments using the primers targeted to the published genes (9, 19, 30). To consecutively sequence the whole gene cluster, all adjacent fragments were amplified in an overlapping fashion.

For sequencing the region upstream of the known genes, inverse PCR was used to amplify the unknown fragment with primers orfx1E-r1 and orfx2E-f2 in the K35 strain (24). The genomic DNA (∼27 μg) was cut with restriction enzyme HindIII (New England Biolabs, Inc., Ipswich, MA) at 37°C overnight and then ligated with T4 DNA ligase (New England Biolabs, Inc.) to form circular DNA at a DNA concentration of ≤3 ng/μl. The circular DNA was used as a template at a concentration of 2.5 ng/μl (26). After the upstream region of the orfx2 gene in K35 was sequenced, the primer orfx2E-r4 was designed and used in combination with orfx2E-f2 to amplify the upstream region of the neurotoxin gene cluster in strains K81 and 31-2570.

In sequencing the previously unidentified upstream region of the gene cluster, primers targeted to the consensus sequence of orfx3 from NCTC 2916 and Kyoto-F (4) were successfully used. This method was considered justified since high similarities were observed in the nucleotide sequences and arrangements of the orfx2, orfx1, and p47 genes between the type E strains sequenced here and the previously published type A strains NCTC 2916 and Kyoto-F (4).

The PCRs were conducted using standard protocols (24). Amplification conditions consisted of 30 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 30 s, and polymerization at 72°C for 80 s (for 4 min if the fragment was 2 kb or more), followed by a final extension at 72°C for 3 min (for 10 min if the fragment was 2 kb or more). Titanium Taq DNA polymerase (Clontech Laboratories, Inc., Mountain View, CA) was used in all PCRs. Amplification was performed with a PTC-200 thermal cycler (MJ Research, Watertown, MA).

DNA sequencing.Template DNA was purified from PCR products using a QIAquick PCR purification kit (Qiagen GmbH, Hilden, Germany) or a Montage PCR96 cleanup kit (Millipore Corp., Billerica, MA). Sequencing PCR was conducted using an ABI BigDye Terminator 3.1 cycle sequencing kit (Applied Biosystems, Inc., Foster City, CA) according to the manufacturer's instructions. After precipitation with ethanol or purification with an Agencourt CleanSEQ kit (Agencourt Bioscience Corp., Beverly, MA), the sequencing samples were loaded onto an ABI 310 genetic analyzer or an ABI 3730 genetic analyzer, respectively. The former method was used to sequence the entire neurotoxin gene clusters, and the latter method was used to sequence the 11 neurotoxin genes for more efficient throughput.

Sequence analysis.The sequences of the entire neurotoxin gene clusters and bont/E were assembled and aligned by Sequencher software (Gene Codes, Ann Arbor, MI). Bidirectional sequence information from two independent PCRs was used, and the sequences were considered final when threefold coverage for each gene was obtained. Sequences were edited by BioEdit software (v. 7.0.5) (http://www.mbio.ncsu.edu/BioEdit/bioedit.html ), and open reading frames were predicted manually and via ORF Finder software (http://www.ncbi.nlm.nih.gov/gorf/gorf.html ). Alignment and comparison of two or more sequences were performed using Align (http://www.ebi.ac.uk/emboss/align/index.html ) or ClustalW software (3).

Type E neurotoxin genes from 33 C. botulinum type E strains, including the 14 genes sequenced in this study and 19 previously published ones, were compared using ClustalW software (3). The GenBank accession numbers of the previously published sequences are shown in Table 2.

Nucleotide sequence accession numbers.The sequence data reported in this paper have been submitted to the EMBL Nucleotide Sequence Database under accession numbers AM695752 through AM695765 and are listed in Table 2.

Structure and organization of BoNT/E gene cluster in three type E strains.The neurotoxin gene clusters of C. botulinum K35, K81, and 31-2570 consisted of six genes and were each 12,908 bp long (Table 3). For the first time, the entire nucleotide sequences of orfx2 (2,247 bp) and a partial orfx3 gene (1,355 bp), located immediately upstream of orfx2 and having the same orientation as orfx2, were determined for the type E BoNT gene cluster (Fig. 1). The structures of the neurotoxin gene clusters in the three C. botulinum strains were identical to each other and, apart from the novel elements, to published ones.

• Open in new tab
• Download powerpoint

FIG. 1.

Schematic representation of the neurotoxin gene cluster in Clostridium botulinum type E, A1, and A2 strains. The gene orientation is shown by arrows. Only orfx3 in the type E gene cluster is a partial gene. Type A1 and A2 neurotoxin gene clusters are derived from C. botulinum A1 strain NCTC 2916 (GenBank accession numbers AY497357, Y14238, and X52066) and Kyoto-F (accession numbers AY497358, AB004778, X96493, X87974, and X73423).

K81 and 31-2570 had identical sequences for p47, orfx1, orfx2, and orfx3. These two strains were also highly similar in their bont/E and ntnh sequences. K35 differed from the other two strains in bont/E and orfx2, with 98% and 99.1% identities, respectively; however, K35 was highly identical to K81 and 31-2570 in the p47, orfx1, and orfx3 gene sequences (Table 4). Marked differences were observed in the GC content of the six genes in the bont/E cluster, with orfx1 showing the lowest and the partial orfx3 gene the highest proportions of GC content (Table 3).

Diversity of type E neurotoxin genes.The bont/E genes from a total of 33 strains of BoNT/E-producing clostridia were analyzed (Table 2). These strains included the 14 C. botulinum type E strains of mainly Finnish origin described in this study and 19 C. botulinum and C. butyricum strains with bont/E sequences available in GenBank. The 33 strains could be classified into six subtypes based on the nucleotide sequences of bont/E (Fig. 2). Five Finnish and four other C. botulinum strains of diverse origin formed subtype E1. All of these strains were identical in their bont/E gene sequences, with the exception of C. botulinum Beluga (GenBank accession number X62089), which showed a 1% difference. Subtype E2 included one C. botulinum strain E544 (EF028404) alone, while subtype E3 included six C. botulinum strains from Finland, Germany, and the United States. Except for strain 31-2570, the strains in subtype E3 had 100% identical bont/E genes. All C. butyricum strains formed subtypes E4 and E5, with E4 including ATCC 43755 (X62088), BL5262 (AB088207), and BL6340 (AB039264). Subtype E5 was a uniform group containing 11 C. butyricum strains. A previously unidentified subtype designated E6 included three C. botulinum strains of Finnish origin. Of these, K36 and S16 had 100% sequence identity for bont/E.

• Open in new tab
• Download powerpoint

FIG. 2.

The bont/E genes from 14 Clostridium botulinum type E strains sequenced here and the 19 previously published ones were compared using ClustalW software. Six distinct neurotoxin subtypes were described and designated E1 to E6. C. botulinum strains fell into subtypes E1, E2, E3, and E6, while subtypes E4 and E5 consisted of C. butyricum strains alone. Three Finnish isolates formed a previously unidentified neurotoxin subtype, E6.

The homology analysis of the nucleotide and amino acid sequences of BoNT/E within and between the six subtypes is shown in Table 5. Strains within a subtype had only 1% differences in their nucleotide and amino acid sequences, while 1 to 3% and 1 to 5% differences in nucleotide and amino acid sequences, respectively, were observed between the four subtypes of C. botulinum. Between the C. botulinum and C. butyricum subtypes, 2 to 3% and 3 to 6% differences in the nucleotide and amino acid sequences, respectively, were observed. Moreover, the two C. butyricum subtypes had corresponding differences of 3% and 5% in their nucleotide and amino acid sequences.

Homology of ntnh and p47.Among C. botulinum type E strains K35, K81, 31-2570, and Mashike (GenBank accession number D12697), the ntnh genes were in general 99% homologous in nucleotide and corresponding amino acid sequences; however, strains K35 and Mashike shared 98% amino acid sequence homology for ntnh. Furthermore, the differences between the four C. botulinum strains and C. butyricum BL6340 were 1 to 2% in nucleotide sequences and 2% in amino acid sequences.

All the BoNT/E-producing clostridia shared a high homology in the nucleotide and amino acid sequences for p47. C. botulinum K35 shared 99% homology in the nucleotide and amino acid sequences for p47 with K81, 31-2570, Iwanai (accession number D88418), and Alaska (U70780, previously named p48), the four strains being identical in p47 sequence. With C. butyricum BL6340 (D88419), all C. botulinum type E strains had 98 to 99% and 98% homologies in nucleotide and amino acid sequences, respectively (Table 6).

The p47 genes in the BoNT/E-producing clostridia were markedly different from those in C. botulinum type A strains NCTC 2916 (AY497357), Kyoto-F (X96493), and Mascarpone (DQ310546). The differences in nucleotide and amino acid sequences were 15 to 21% and 19 to 29%, respectively (Table 6).

Homology of orfx1, orfx2, and orfx3.No difference was found in orfx1 among C. botulinum type E strains K35, K81, 31-2570, and Iwanai (D88418). However, differences of 21 to 22% and 28% in nucleotide and amino acid sequences, respectively, for orfx1 were observed among type E, type A1, and type A2 strains (Table 6).

For orfx2, there were only differences of 1% and 2% in the nucleotide and amino acid sequences, respectively, among the C. botulinum type E strains (Table 7), while again considerable differences were observed among strain types E, A1, and A2. Moreover, for orfx3, high homologies (99 to 100%) of nucleotide and amino acid sequences were present among C. botulinum type E strains, while homologies of only 81% and 75 to 76% were observed among strain types E, A1, and A2 (Table 7).

Analysis of amino acid sequence of BoNT/E.A total of 149 variable sites were found in the alignment of the BoNT/E amino acid sequences of seven strains representing the different subtypes (Table 8). The variable sites were distributed mainly in the light chain (37.6%) and the C-terminal half of the heavy chain (43.0%), while less variation was observed in the N-terminal half of the heavy chain (19.4%). No variation was seen in the conserved motifs of the functional domains. Compared with the consensus sequence, the sequences of the C. butyricum strains forming subtype E5 had maximum variability, with all their variable residues being within the heavy chain. By contrast, the C. botulinum K8 sequence had no variation at all. Sequences of strains within the new subtype E6 had the second highest variability distributed along the functional domains of the light chain and the N- and C-terminal halves of the heavy chain. The number and distribution of variable amino acid residues for subtype E6 differed markedly from those of the other subtypes.