Construction and validation of safe Clostridium botulinum Group II surrogate strain producing inactive botulinum neurotoxin type E toxoid

Botulinum neurotoxins (BoNTs), produced by the spore-forming bacterium Clostridium botulinum, cause botulism, a rare but fatal illness affecting humans and animals. Despite causing a life-threatening disease, BoNT is a multipurpose therapeutic. Nevertheless, as the most potent natural toxin, BoNT is classified as a Select Agent in the US, placing C. botulinum research under stringent governmental regulations. The extreme toxicity of BoNT, its impact on public safety, and its diverse therapeutic applications urge to devise safe solutions to expand C. botulinum research. Accordingly, we exploited CRISPR/Cas9-mediated genome editing to introduce inactivating point mutations into chromosomal bont/e gene of C. botulinum Beluga E. The resulting Beluga Ei strain displays unchanged physiology and produces inactive BoNT (BoNT/Ei) recognized in serological assays, but lacking biological activity detectable ex- and in vivo. Neither native single-chain, nor trypsinized di-chain form of BoNT/Ei show in vivo toxicity, even if isolated from Beluga Ei sub-cultured for 25 generations. Beluga Ei strain constitutes a safe alternative for the BoNT research necessary for public health risk management, the development of food preservation strategies, understanding toxinogenesis, and for structural BoNT studies. The example of Beluga Ei generation serves as template for future development of C. botulinum producing different inactive BoNT serotypes.

www.nature.com/scientificreports/ botulinum Group II physiology. This study also constitutes a solid base for future modifications of BoNT genes in other C. botulinum strains.

Results and discussion
Construction of the C. botulinum Beluga Ei toxoid strain using CRISPR/Cas9. The constructed C. botulinum Beluga Ei strain contains a chromosomally located bont/ei encoding enzymatically inactive toxin BoNT/Ei with three amino acid substitutions in the LC. We designed these modifications based on previous crystallography and functional studies of non-toxic recombinant variants of BoNT 19,20,22,27 . Because all BoNT serotypes demonstrate strict conservation of the catalytic motifs 18,28 , we were able to localize the previously determined residues within the BoNT/E polypeptide by sequence alignment and crystal structure analysis. Utilizing the CRISPR/Cas9 tool 41 , we introduced point mutations into the genome of the Beluga Ei strain, altering the amino acid residues of the zinc-binding motif ( 638 A > C resulting in E213A within HExxH + E motif) and the transition state stabilization motif ( 1042 A > G, 1043 G > C, 1044 G > A, 1052 A > T resulting in R348A and Y351F within RxxY motif). In addition to these BoNT-inactivating alterations, we introduced silent modifications into the Cas9-targeted sequence of bont/e. All modified DNA regions were designed to introduce unique restriction sites to allow rapid discrimination between Beluga WT and Beluga Ei by PCR amplification of the modified region followed by restriction enzyme digestion. The resulting mutant clones were screened for the presence of the desired genome changes. The restriction digestion of the modification-harboring amplicon as well as Sanger and Illumina sequencing confirmed the successful modification of the bont/ei sequence. Altogether, we show that CRISPR/Cas9 is a convenient tool for introducing chromosomal single-nucleotide alterations into C. botulinum. Our approach can be used to generate a library of C. botulinum strains producing other inactive BoNT serotypes. The Beluga Ei strain can be further genetically modified using CRISPR/ Cas9 or any other gene manipulation tool and may serve as a base for safe research of almost every aspect of C. botulinum Group II physiology, including BoNT gene expression and its regulation, as well as mechanistic and structural research.
General characterization of the C. botulinum Beluga Ei strain. The C. botulinum Beluga Ei strain can serve as a reliable surrogate strain for future studies on the overall physiology and toxinogenesis of Group II strains, which are of major concern for the modern food industry. To demonstrate the applicability of the C. botulinum Beluga Ei strain for bacterial physiology studies, we characterized its ability to grow, display regular cell morphology, form heat-resistant spores, and produce and release detectable BoNT/Ei. These steps aimed at verifying that the introduced genomic modifications do not impact key physiological processes of the strain and, most importantly, do not affect BoNT/Ei production and its recognition in immune-based detection assays. For this purpose, we performed an in-depth characterization of growing cultures of Beluga Ei and parental Beluga WT strains in cooked meat medium-tryptone/peptone/glucose/yeast extract (CMM-TPGY) sporulation Group II medium 42 for 96 h. The growth curves of Beluga Ei and WT showed comparable shapes and reached similar optical density (OD 600 ) values, confirming that the introduced mutations in Beluga Ei did not affect its growth (Fig. 1a). Phase-contrast micrographs of late-logarithmic (9 h) and late-stationary phases (72 h post inoculation) did not reveal any morphological differences between the mutant and parental strain (Fig. 1b). Also concentrations of viable cells and heat-resistant spores were similar at most sampling time points (Figs. 1c, d). Only 6 h post inoculation, Beluga Ei demonstrated nearly 2-log higher viable cell counts than Beluga WT (Fig. 1c). However, cell division in the mid-logarithmic phase is rapid and the difference likely results from a minor delay in the sampling procedure.
Next, we verified the strains for BoNT production. We determined the concentrations of BoNT/Ei and BoNT/E in intracellular and extracellular culture fractions, respectively, using BoNT/E-directed sandwich ELISA based on the polyclonal capture antibody KE97 and the monoclonal BoNT/E-specific antibody E136 directed against the H N -domain of the toxin, unaltered in BoNT/Ei 43 . We additionally confirmed the ELISA-measured concentrations in densitometric analysis (Fig. S1). The antibodies applied in ELISA recognized BoNT/Ei at comparable levels to BoNT/E produced by Beluga WT (Fig. 2). The distribution of BoNT between the cell pellet and supernatant (S/N) collected at different time points was similar between the strains, confirming that the modifications within BoNT/Ei did not alter its release from the cells to the S/N. In conclusion, the introduced genome modifications did not detectably affect central cellular processes or toxin production of Beluga Ei. This successful validation provides strong evidence for the safe applicability of C. botulinum Beluga Ei in comprehensive studies, such as food challenge testing.

Verification of biological activity of BoNT/Ei ex vivo and in vivo and genetic stability of C. botulinum Beluga Ei strain.
To verify that the BoNT/Ei produced by the Beluga Ei strain is biologically inactive, we challenged the single chain BoNT/Ei (scBoNT/Ei) ex vivo using the MPN assay representing the physiological BoNT target and allowing detailed studies on BoNT pharmacodynamics 44 . Firstly, the scBoNT/E wild-type produced by Beluga WT strain was tested in the MPN assay and yielded a mean paralytic half-time of 74.5 ± 8.9 min (mean ± SD; n = 4) corresponding to 151 ± 13% neurotoxicity of recombinantly expressed, singlechain BoNTE wild-type (scBoNTE). Hence, recombinantly expressed scBoNTE wild-type and native scBoNT/E wild-type from Beluga WT display virtually identical neurotoxicity. In contrast, the samples containing scBoNT/ Ei produced by the Beluga Ei strain showed no sign of paralysis of the hemidiaphragm that could be observed within the lifetime of the tissue (> 300 min). Any theoretical residual biological activity of scBoNT/Ei is well below the limit of detection of the MPN assay (LoD = 0.5 pM scBoNT/E wild-type).
The CRISPR/Cas9 approach allowed us to construct a markerless C. botulinum Beluga Ei strain free from external coding sequences which might display unpredictable genetic recombination events. Nevertheless, the www.nature.com/scientificreports/  www.nature.com/scientificreports/ nature of C. botulinum genetics has to be considered when assessing the safety of genetically modified non-toxic strains encoding specific point mutations in their genome. C. botulinum genomes harbor numerous mobile genetic elements and prophages that could unpredictably rearrange the BoNT cluster sequence 3,8 . Additionally, the existence of numerous different BoNT sero-and subtypes is indicating that spontaneous alterations within bont can occur 45 . Considering the extreme toxicity of BoNT, it is therefore crucial to verify the stability of the introduced mutations in any novel non-toxic C. botulinum strain to ensure its safe use.
To investigate whether the introduced mutations are stable in a long-term evolutionary scenario and therefore to confirm that the constructed strain will remain safe to handle, we performed 25 single-colony serial passages of C. botulinum Beluga Ei strain. We subjected the resulting 25 th generation (G25), together with the 1 st generation Beluga Ei (G1), to in vivo toxicity testing and verified its genetic stability. Using a restriction endonuclease digestion approach (Fig. 3a) and Sanger sequencing, we screened a randomly picked colony of G25 for the presence of the previously introduced genomic modifications in bont/ei. All PCR amplicons of the targeted bont region of Beluga WT, Beluga Ei G1 and G25 showed the expected restriction patterns after separate digestions with SphI, ApaLI and MspI (Fig. 3b). All introduced modifications within the catalytic motifs in bont/ei were confirmed in Sanger (Fig. S2) and Illumina sequencing (Table S1).
To verify that the BoNT/Ei produced by G1 and G25 was biologically inactive, we tested the culture S/Ns in vivo using the mouse bioassay 46 . We determined the mouse intraperitoneal (i.p.) 50% lethal dose (LD 50 ) of Beluga WT culture S/Ns (Table 1). After this we examined Beluga Ei G1 and G25 S/Ns to detect any residual toxicity of BoNT/Ei. We tested the BoNT toxicity of both native (scBoNT/E) and trypsinized (di-chain BoNT/E) forms, as trypsin treatment of scBoNT/E increases its i.p. toxicity more than 50-fold 47 by cleaving the full-length Mice treated with i.p. injections of S/N preparations were monitored for any symptoms of botulism, which include "wasp-waist", moving difficulties, respiratory problems, and death, for four days 48 . Initially, we performed an i.p. titration of trypsinized and native BoNT/E S/Ns of Beluga WT. The i.p. LD 50 of trypsinized BoNT/E was 2.25 ng/kg of body weight (kgbw) and of native scBoNT/E was 225 ng/kgbw, as determined by the Weil's method 49 . To assess the toxicity of BoNT/Ei G1 and G25, we administered four different S/Ns at the highest BoNT/Ei concentration that could be obtained into four groups of five mice: group i) native scBoNT/Ei G1 in amount equivalent to 480 i.p. LD 50 of native WT scBoNT/E, ii) trypsinized BoNT/Ei G1 in amount equivalent to 23,979 i.p. LD 50 of trypsinized WT BoNT/E, iii) native scBoNT/Ei G25 amount equivalent to 464 i.p. LD 50 of native WT scBoNT/E, iv) and trypsinized BoNT/Ei G25 amount equivalent to 23,196 i.p. LD 50 of WT trypsinized BoNT/E (Table 1).
During four days of observation, all the BoNT/Ei treated animals remained asymptomatic. These results confirm that the Beluga Ei strain does not produce biologically active BoNT detectable in the standard mouse bioassay and BoNT/Ei is more than 20,000-fold less toxic than WT BoNT/E thereby confirming MPN data. Moreover, BoNT/Ei produced by the 25 th generation of serially passaged C. botulinum Beluga Ei still displays no measurable biological activity. In conclusion, we demonstrated that the constructed C. botulinum Beluga Ei strain produces biologically inactive BoNT/Ei and that the introduced mutations within the bont/ei are stable for at least 25 generations, thus the likelihood of reversion to the wild-type genotype is negligible.

Conclusions
The results demonstrate that the C. botulinum Beluga Ei strain, constructed in this study using the CRISPR/ Cas9 tool, produces a non-toxic, biologically inactive botulinum neurotoxin structurally similar to WT BoNT/E, termed BoNT/Ei, in comparable amounts as the wild-type strain. BoNT/Ei carries three amino acid modifications in catalytically active sites: E213A in the zinc-binding HExxH + E motif 19 , and R348A and Y351F within the RxxY motif responsible for the transition state stabilization 22 , causing a loss of catalytic activity. The introduced genomic modifications did not notably affect important physiological features of the Beluga Ei strain, like growth and vegetative cell morphology, sporulation rate and the ability to form heat-resistant endospores. Most importantly, the total amount of produced BoNT/Ei and its distribution over time were similar to BoNT/E produced by the parental Beluga WT strain. Polyclonal and monoclonal antibodies directed against native BoNT/E recognized BoNT/Ei comparable to wild-type BoNT/E, and trypsin correctly cleaved the 150-kDa full-length BoNT/Ei into the expected 100-kDa HC and 50-kDa LC, indicating an unaltered BoNT/Ei tertiary structure. Mouse bioassays demonstrated absence of biological activity of trypsinized and native BoNT/Ei administered i.p. to mice in amounts equivalent to 23,979 i.p. LD 50 of trypsinized WT BoNT/E, and to 480 i.p. LD 50 of native WT scBoNT/E. To determine the stability of the introduced point mutations, we sub-cultured the C. botulinum Beluga Ei strain in 25 subsequent passages. Sanger sequencing and the restriction digestion of PCR-amplified bont/ei fragments of the 25 th mutant generation confirmed the presence of the intended mutations. Trypsinized and native culture S/N from the 25 th generation, containing BoNT/Ei in an amount equivalent to 23,196 i.p. LD 50 of active trypsinized WT BoNT/E and 464 i.p. LD 50 of native WT BoNT/E, respectively, showed no biological activity in the mouse lethality assay in groups of five mice. The results confirm that the introduced mutations within bont/ei are stable and reversion to the wild-type genotype is unlikely. Additionally, we confirmed the Table 1. Detection of the activity of wild-type and inactive botulinum neurotoxin type E (BoNT/E and BoNT/ Ei, respectively) using the mouse bioassay. n.a. not applicable, n number of animals used; n* number of deaths, **%, percent of deaths. www.nature.com/scientificreports/ absence of detectable BoNT/Ei neurotoxicity in an ex vivo MPN assay. Here, scBoNT/Ei did not cause any sign of paralysis of the hemidiaphragm tissue at maximal achievable protein concentrations. The constructed surrogate strain C. botulinum Beluga Ei carries unique restriction digestion sites as molecular markers to confirm the presence of the inactivating mutations and to facilitate identification. Being a genetically modified organism, C. botulinum Beluga Ei is intended exclusively for use in contained research facilities. Even in the very unlikely event of accidental release, potentially leading to Beluga Ei being erroneously detected as pathogenic C. botulinum strain when applying standard serological and DNA-based detection methods 50 , restriction digestion and sequencing would identify the strain as non-pathogenic. Therefore, it is important to highlight that new surrogate strains for C. botulinum should carry toxoid-specific molecular markers and that they must be handled only in the specialized research laboratories. In conclusion, the constructed C. botulinum Beluga Ei strain producing biologically inactive BoNT/Ei provides a safe and stable surrogate for botulinum neurotoxin research and for food safety risk assessment.
The C. botulinum Beluga Ei strain has been officially excluded from the US select agent regulations of the Division of Select Agents and Toxins (DSAT) at Centers for Disease Control and Prevention (Atlanta, GA), fulfilling all the requirements of an efficiently attenuated strain and thus not posing a threat to public health and safety (decision effective 12th December 2019 and available online 51 ).  Supplementary Figures S4 and S5 Construction of the mutant strain. C. botulinum Beluga Ei was constructed using the CRISPR/Cas9 method 41,42 . Primers are listed in Table S3. The construction of the CRISPR/Cas9 vector pMTL431511-bont/ ei included: (i) construction of 2,900-bp modification cassette (MC) where four DNA segments carrying the described single nucleotide mutations were PCR-synthetized using Beluga WT gDNA as template and fused by splicing by overhang extension-polymerase chain reaction (SOE-PCR); (ii) and generation of a fragment encoding the single guide RNA (sgRNA) designed using CRISPR/Cas9 tool (benchling.com/) and constructed in primer dimer-PCR reaction 41,42 . (iii) Ligation was performed as previously described 42 . E. coli NEB5-alpha were transformed with the ligation mixture and screened in colony-PCR. Positive plasmids were Sanger-sequenced. The conjugation donor E. coli CA434 56 was transformed with pMTL431511-bont/ei. Conjugation was performed as previously described 42 . Antibiotic-resistant C. botulinum colonies were screened by colony-PCR and amplicons were Sanger-sequenced. Positive clones were cured of pMTL431511-bont/ei as described 42 . To rapidly confirm the presence of the mutations, amplicons from colony-PCR were digested separately with SphI, ApaLI and MspI enzymes and separated in agarose gel. Optical density measurement, spore and viable cell count assays. Spores of C. botulinum Beluga WT and Beluga Ei were used to inoculate CMM-TPGY in three biological replicates. Culture OD 600 was measured at a wavelength of 600 nm using a Thermo Spectronic Genesys 10 UV/Vis Spectrophotometer. Heat-resistant spore and viable cell concentrations were determined as described 42 applying the most probable number technique 57 . One-way ANOVA was used to compare the total viable cell and spore counts.
Phase-contrast microscopy imaging. C. botulinum culture pellets were re-suspended in phosphatebuffered saline (PBS) and immobilized in 1.7% agarose coated on a microscopy slide. Micrographs were captured with a DMi8 Leica microscope, equipped with an HC PL APO 100x/1.40 OIL PH3 objective and a Hamamatsu Orca Flash V2 LT camera, and analyzed using the MetaMorph software (Molecular Devices, San Jose, CA, USA).
BoNT/E-directed sandwich enzyme-linked immunosorbent assay. Culture aliquots were centrifuged (15 min; 6800 × g; + 4 °C) and cell pellets re-suspended in PBS to extract the intracellular toxin. BoNT/Edirected sandwich ELISA was carried out as described 43  In vivo determination of wild-type BoNT/E intraperitoneal LD 50 and analysis of BoNT/Ei toxicity. CMM-TPGY media were inoculated with C. botulinum Beluga WT and G1 in three biological replicates and incubated for 96 h. S/Ns were collected as before (4.7) and BoNT was quantified applying ELISA (Table S5). When applicable, the samples were trypsinized 59 . To determine the toxicity of wild-type S/N, the samples were initially ten-fold serially diluted in a standard phosphate-gelatin buffer of pH = 6.2 and a volume of 0.5 ml of each sample was injected i.p. into two randomly grouped four-week-old female outbread HsdWin:NMRI mice (Envigo, Horst, The Netherlands) of 20-22 g bodyweight. The animals were monitored for symptoms of botulism for four days and euthanized by an overdose of 200 mg/kg pentobarbital (Mebunat Vet, Orion Pharma, Espoo, Finland) injected i.p. or by cervical dislocation in case of reaching the humane end-point. The highest dilution leading to death of both animals was further serially two-fold diluted and administered to four mice each, following the same protocol, to enable calculation of the i.p. LD 50 using the Weil method 49 . To detect any residual toxicity of BoNT/Ei, each sample was injected i.p. as described above into five randomly assigned mice which were then monitored for symptoms of botulism for four days. BoNT/E-directed Western blot. Aliquots of C. botulinum Beluga WT and Beluga Ei culture S/Ns before and after trypsinization were separated in 4-12% SDS-PAGE. Purified 15 ng BoNT/E standard was used as a Western blot reference. Proteins were transferred onto a PVDF 0.2 µM membrane, subsequently blocked and incubated in BoNT/E-directed KE97 rabbit polyclonal antibodies (1:1,250 in PBS with 0.05% Tween 20 (PBST) with 5% bovine serum albumin) at 4 °C overnight. The membrane was washed, incubated in goat anti-rabbit IgG:HRP antibody (Bio-Rad, Hercules, CA, USA) (1:50,000 in PBST) for 2 h at RT and washed again. Subsequently, the membrane was incubated in Clarity Western solution (Bio-Rad) and chemiluminescence was detected using Fujifilm LAS-3000 Imager (Fuji Photo Film, Tokyo, Japan) applying different exposure times.

Data availability
Whole genome sequencing data are available in the National Center for Biotechnology Information database under the BioProject PRJNA751216.