Exploring the genetic background of the botulism neurotoxin BoNT/B2 in Spain

ABSTRACT To determine whether the neurotoxin BoNT/B2 causing botulism in Spain is clonal, the genetic diversity and phylogenetic relationships of Clostridium botulinum from food-borne episodes and infant cases of the condition were explored. The botulinum toxin gene (bont) subtype, the variable region of the flagellin gene (flaVR), and a seven-gene multi-locus sequence type were examined by sequencing 37 BoNT-positive cultures obtained over the period 2010 to 2022. Out of 37 botulism events, 16 food-borne episodes and 16 infant cases were associated with bont/b2. Eight bont/b2 alleles were detected [nucleotide distance range 0.0259–0.415%, Hunter and Gaston discrimination index (HGDI) 0.71]. The most common bont/b2 allele corresponded to that of strain Prevot 25 NCASE and its single and double locus variations (87.5%). Four known flaVR types were identified (HGDI 0.79), along with one previously unknown (flaVR-15). Sixteen sequence types (STs) (HGDI 0.89) were recorded including seven new STs (ST164–ST170; 10 new alleles) and five new STs (ST171–ST175; with new allele combinations) were also noted. Correlations among some STs and flaVR types were seen. Overall, the present results show that the combined analysis of bont/b2-flaVR-ST at the nucleotide level could be used to track botulism events in Spain. The neurotoxin BoNT/B2 has largely been responsible for human botulism in Spain. The polymorphism analysis of bont/b2, flaVR typing, and sequence type determinations, revealed a wide variety of clones to be responsible for human botulism, ruling out a common source of acquisition. IMPORTANCE Botulism, a potentially fatal disease, is classically characterized by a symmetrical descending flaccid paralysis, which if left untreated can lead to respiratory failure and death. Botulinum neurotoxin (BoNT), produced by certain species of Clostridium, is the most potent biological toxin known, and the direct cause of botulism. This study characterizes the acquisition in Spain of two forms of botulism, i.e., food-borne and infant botulism, which are largely caused by the main neurotoxin BoNT/B2. Polymorphism analysis of the bont/b2 gene, typing of the flagellin variable region sequence (flaVR), and multilocus sequence typing, were used to explore the genetic background of Clostridium botulinum group I. To our knowledge, this is the first phylogenetic and typing study of botulism undertaken in Spain.

the ingestion of just 70 µg of BoNT/A toxin is lethal.Death would also ensue if some 0.7-0.9µg were inhaled or 0.09-0.15µg were intravenously administered (4).
The BoNT types actually produced are strain-independent.Most strains produce only one type, but others are bivalent [Ab, Ba, Af, Bf (the lowercase letter denoting lesser production)] or trivalent (e.g., strain Af84, which produces BoNT/A2, F4, and F5), and yet others possess silent neurotoxin genes (5,8,9).All bont genes are located within a neurotoxin cluster, either on the chromosome, on plasmids, or on phages, depending on the species and the bacterial strain.This supports the idea that horizontal bont transfer occurs between Clostridium and non-Clostridium strains.
Human botulism, mainly associated with groups I and II, occurs in three main forms: food-borne botulism, botulism by intestinal colonization (including infant botulism), and wound botulism.Food-borne botulism occurs after the ingestion of pre-formed BoNT in food, while infant botulism is caused by spores germinating in the colon, leading to BoNTs being produced in situ.The ability to form endospores is critical to clostridial pathogenicity.
The active neurotoxin consists of a light chain (L, about 50 kDa) and a heavy chain (H, about 100 kDa), which remain linked by a disulfide bridge.Three distinct domains: L-chain containing α-helices, and β-strands and including the catalytic zinc-binding protease motif (His-Glu-XX-His); the N-terminal part of the H-chain (HN) forming two unusually long and twisted α-helices involved in the translocation of the L-chain into the cytosol; and the C-terminal part of the H-chain (HC) consisting of two distinct subdomains (H CN and H CC ).Mainly through the Hcc subdomain (the most divergent), the different BoNTs types/subtypes interact with distinct membrane receptors on demyelina ted terminal nerve endings.The L-chain cleaves different intracellular SNARE proteins (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) and form a complex that is essential for synaptic vesicle exocytosis in neurons: SNAP-25 (synapto somal-associated protein 25 , the target for BoNT/A, C1, and E), syntaxin 1 (the target for BoNT/C1), and VAMP1/2/3 (vesicle-associated membrane protein, the target for BoNT/B, D, F, and G).The cleavage of the SNARE proteins blocks neurotransmission and causes flaccid muscle paralysis.After the disease develops, its treatment is problematic and mortality is high.Treatment includes trivalent or heptavalent equine botulinum antitoxin plus intensive care support, but even available between 5% and 10% of patients with food-borne botulism die (1,3,12,13).

Patients and samples
From 2010 to 2022, samples from 268 patients with suspected botulism or related neurological disorders (range 11-41 cases/year; mean 19.5) were studied at the National Centre of Microbiology of the Institute of Health Carlos III (Madrid, Spain).Seventy patients had laboratory-confirmed botulism (patients with at least one positive diagnostic test, range 1-13 cases/year; mean 4.4 per year), 34 food-borne and 36 infant botulism, as revealed by positive standard mouse bioassay (SMB, authorized by the corresponding institutional animal care) involving serum, stools, or gastric fluid, and/or by multiplex PCR of bont genes in stool cultures (14)(15)(16)(17).Forty-eight patients were SMB-positive, while PCR confirmed botulism in 22 patients (31.43%) with SMB-negative results (19 food-borne and 3 infant cases).Thirty-seven positive cultures out of 70 patients [15 SMB(+)/PCR(+) and 22SMB(−)/PCR(+)] could be BoNT subtyped (10).

Molecular surveillance by bont/b2, flaVR, and multi-locus sequence type analysis
To determine the alleles (variant forms of a gene) of the main subtype involved in Spain, bont/b2 (3,876 nucleotides, 1,292 amino acids) was analyzed (10), and blastn/ blastp comparisons made with available sequences in the GenBank database.Maximum likelihood (ML) phylogenies of the bont/b2 gene, with the analysis of 62 sequences including those from other locations and representative of BoNT/B serotypes, were conducted using MEGA7 software.
To assess genetic relatedness, cultures and strains were typed by sequencing the variable region of flagellin A (flaVR) (18), and by multi-locus sequence typing with a seven-loci scheme (aroE, mdh, aceK, opp, rpoB, recA, and hsp60) for the elucidation of evolutionary lineages (19).Allelic numbers and sequence types (STs) were identified by querying the C. botulinum multi-locus sequence type (MLST) PubMLST database (https:// pubmlst.org/organisms/clostridium-botulinum)(20).New alleles and new STs were submitted to this database.The genetic relationships among the STs of C. botulinum BoNT/B2 within Spain, and with respect to those BoNT/B found in other countries were conducted using the goe-BURST routine (provided with PHYLOVIZ 2.0 software: https:// online.phyloviz.net/index).

Predominance of botulism subtype BoNT/B2
BoNT subtyping was performed with 37 positive cultures (10); Fig. 1 shows the geo graphic distribution of these subtypes.Twenty-one food-borne episodes (involving one to five people) were caused by C. botulinum A1 (2 episodes), A2 (1 episode), B2 (16 episodes), F1 (1 episode), or Clostridium baratti F7 (1 episode).Sixteen recorded cases of infant botulism all involved BoNT/B2.All cases (food-borne and infant) involving this neurotoxin were caused by proteolytic C. botulinum group I (positive for the phenyl lactate dehydratase subunit B gene, fldB, associated with phenylalanine metabolism in proteolytic clostridia) (21).Table 1 shows the epidemiological characteristics of the laboratory-confirmed botulism cases caused by BoNT/B2.

Distribution of bont/b2 alleles
Eight bont/b2 alleles were detected.An allele identical to bont/B2 of C. botulinum strain 111 (sequence accession BAC22064), the representative strain of BoNT/B2, was seen in four infant cases only.Another allele was identical to that of the strain Prevot 25 NCASE (sequence accession EF033129.1)(10); this was associated with nine food-borne episodes and eight infant cases.The bont/b2 of the latter strain shows a difference of just 0.15% (two amino acids and six nucleotides) with respect to strain 111.Table 1 shows the distribution of the different bont/b2 alleles.

DISCUSSION
In the European Union, botulism case numbers have remained stable at ≈100 cases/ year with an overall notification rate of 0.02 cases/100,000 inhabitants (as occurs in Spain) (1,24), although the botulism form (food-borne or infant) and incidence vary according to the reporting countries.The highest incidences have been recorded in Italy and Romania.Food-borne botulism outbreaks are, however, reported worldwide, and depend notably on dietary habits and culinary traditions (1,25).An unusually high prevalence of infant botulism is reported in the United States (100 cases per year), especially northern California, eastern Pennsylvania, southern New Jersey, and northern Delaware.The same has been described for Argentina (Buenos Aires and Mendoza) and Spain (23,26).A higher incidence of infant cases in Andalusia (10/16), especially Seville, with no clear association with any particular food or environmental source or risk factor or sex (nine males and seven females).This form of the disease only occurs in infants (typically when they are between 1 and 6 months old, but sometimes up to 12 months of age) given their immature gut physiology and/or underdeveloped gut microbiota.The source of the spores in affected infants remains unknown.They may come from contami nated soil and dust from nearby construction sites, via the consumption of powdered milk, natural sweeteners, corn syrup, or medicinal herbs.The real impact of infant botulism is likely underestimated worldwide given diagnostic difficulties and the condition's non-specific presentation (12,13).At the time of the present study, iatrogenic botulism caused by the therapeutic use of BoNT (for the treatment of spasticity, focal dystonia, hemifacial spasm, hyperhydrosis, strabismus, chronic migraine, and cosmetic) was unreported.
In Spain, the subtype BoNT/B2 was by far the most frequently encountered neuro toxin.The predominance of specific BoNT types/subtypes in different geographical areas had already been noted, with BoNT/B involved in ≈90% of cases from the European Union, BoNT/A1 and BoNT/B1 predominant in the United States, and BoNT/B2 predom inant in Argentina (1,23).Moreover, the present findings show strong predominance (87.5%) of the Prevot 25 NCASE bont/b2 allele and its variations (SLV and DLV) in Spain.Analysis of the nucleotide differences of the bont/b2 gene [Hunter and Gaston discrimination index (HGDI) 0.71] may be of help as a first step in botulism surveillance.Strong bont/b2 identity with the Prevot 25 NCASE strain has been also reported for Italian strains (8).It should be noted that BoNT/B has the greatest degree of inter-subtype and intra-subtype variability of all BoNTs (0.8-1.9%), perhaps as the consequence of more frequent recombination events allowing the subtype's proliferation (2).
Two studies, based on the sequence of the variable region of flaA revealed the genetic diversity of the flagellin genes of the C. botulinum group I and II strains, affording a simple genotyping method (18,22).Different strains with the same or no BoNT can share the same flaVR type.Seventy-five percent of the European strains carried four types (flaVR-1, flaVR-3, flaVR-4, and flaVR-5), and their distribution may vary with geographical origin (22).In Spain, two flaVR types out of four, flaVR-4 and flaVR-15, were the most frequent (each one accounting for nearly one third of all cases).No correlation with geographical origin could be confirmed.
MLST has been used for genetic diversity and phylogenetic analyses of C. botulinum, but with varying success.It has been reported an efficient method for strain discrimi nation and for phylogenetic analyses of C. botulinum BoNT/A2 (19) and BoNT/B2 (27).However, in other studies involving C. botulinum BoNT/A2, BoNT/A3 and BoNT/B5, and BoNT/A(B) it has not given such good results, probably due to the limited genetic diversity of these latter strains (28)(29)(30).Sixteen STs were detected, 12 of which were new (with 10 new alleles).ST58, the most common ST, was detected mainly in cases of infant botulism (six out of eight cases).No other known ST for bont/b2 described in other countries [i.e., ST19, ST29, ST31-33, ST37, ST38, ST51, ST54-57, ST59, and ST92 (20)] was identified.In the present study, the seven-loci scheme returned a sufficiently good diversity coefficient (HGDI 0.89), and is therefore useful for the molecular typing and phylogenetic characterization of Spanish C. botulinum subtype BoNT/B2.Discrepancies between bont/b genes and chromosome clonal phylogeny have been described, probably because of these genes' considerable lateral spread by recombina tion among different lineages leading to wide heterogeneity in C. botulinum group I populations (2,5,8,10,23,27).Pangenomic analysis of the C. botulinum group I strains, which have relatively stable genomic components, is an almost perfect way of studying genetic population structures and botulism events (8,9,23,(27)(28)(29)(30).In its absence, however, the combined analysis of bont/b2-flaVR-ST at the nucleotide level appears to offer a suitable way of tracking botulism events in Spain.Despite the predominance of the Prevot 25 NCASE strain bont/b2 allele and its SLVs and DLVs, the four flaVR-types and the 16 STs revealed different evolutionary lineages to be responsible for human botulism in our country.In France, C. botulinum BoNT/A (the main neurotoxin in this country), and BoNT/B also showed a wide genetic diversity, probably due to multiple and independent genetic rearrangements (27), but not to a single evolutionary lineage as could have happened in other countries (28)(29)(30).High heterogeneity of the circulating strains is also observed via multiple-locus variable number of tandem repeat analysis during botulism surveillance activities in Italy, other Mediterranean countries, without detection of any regional or temporal clustering (31).
In this scenario of diversity, the behavior of the flaVR and aceK genes in typing (HGDI ≥ 0.8, Table 1) indicated their potential use as first-line markers in the epide miological surveillance of botulism via BoNT/B2 in Spain.Despite the temporal and geographical relationships between the examined disease events, no clustering or any further epidemiological link was detected.
In conclusion, the BoNT/B2 subtype is mainly responsible for food-borne episodes of botulism in Spain, and, for now, is the only one involved in infant cases.Polymorphism analysis of bont/b2, flaVR typing, and sequence type determination, showed, for first time in Spain, that different clones are responsible for human botulism.The genetic background of the neurotoxin BoNT/B2 has been explored with success to track botulism events in Spain.

FIG 1 4 TABLE 1 5 TABLE 1 6 FIG 2
FIG 1 Distribution of BoNT subtypes for 21 food-borne episodes (on the left) and 16 infant cases (on the right) of laboratory-confirmed botulism in Spain over the period 2010-2022.Number of episodes and cases are included between parentheses.The maps were obtained from https://d-maps.com/carte.php?num_car=2208&lang=es.

FIG 3
FIG 3 Genetic relationships among the STs of C. botulinum BoNT/B within Spain, and with respect to those found in other countries.The goe-BURST routine (provided with PHYLOVIZ 2.0 software: https://online.phyloviz.net/index),and the criterion of seven shared alleles, were used to distinguish clonal complexes among 83 isolates.Forty-seven STs for BoNT/B were available in PubMLST (https://pubmlst.org/organisms/clostridium-botulinum).Black and red numbers indicate STs and genetic distances.The flaVR types for the received Spanish samples are included.