1. Introduction
In Mexico, during 2020, the number of sting accidents in humans caused by scorpions was approximately 270,000 [
1]. The most recent reports indicate that 14 species of scorpion of the genus
Centruroides are possibly responsible for envenoming cases [
2]. However, apart from the 14 species described, there are at least seven other species responsible for these intoxications [
3]. Such is the case of
Centruorides huichol scorpion located in Nayarit, Mexico, where, in 2020, 14,712 intoxications were reported [
1]. This scorpion, previously classified within
C. noxius species, was reclassified through a taxonomic analysis that allowed to name
C. huichol as an independent species [
4]. Due to the great diversity of scorpion species and the novel strategies to obtain specific antibodies directed against the toxic components, in the cases of species not yet studied, systematic work is required to identify and characterize the toxic components present in the venoms. The recovery of sufficient biological material to implement the generation of antibodies and/or fragments thereof, which are capable of neutralizing toxic components, is also important.
Various groups have ventured into using single-chain antibody variable fragments (scFv) as an alternative for generating anti-venoms against venomous animals [
5,
6,
7]. This format has the advantage of being relatively small (27 kDa), which allows it to have better bioavailability and faster elimination [
8]. On the other hand, its affinity can be improved relatively quickly (even to a higher level than natural immunization). Additionally, it is possible to extend the cross-reactivity through several cycles of directed evolution using site-directed [
9] or random [
10] mutagenesis. The characterization of its kinetic constants and thermodynamic interaction parameters can be evaluated efficiently through surface plasmon resonance [
11].
From a phage-displayed repertoire of human scFvs and screened against Cn2 toxin (main toxin of
C. noxius venom), scFvs 3F and C1 [
12] were isolated, which recognize different epitopes of this toxin [
13]. From scFv 3F, scFvs 6009F and 9004G were generated by directed evolution against Cn2 and Css2 (from
C. noxius and
C. suffusus, respectively). Through a combination of mutations, scFv LR was obtained, which contains V101F (HCDR3) mutation of scFv 6009F within the context of scFv 9004G, which is an efficient antibody capable of neutralizing the main toxins (Cn2 and Css2 respectively) and venoms of
C. noxius and
C. suffussus [
10]. On the other hand, from the scFv C1, scFv 10FG2 was generated, by directed evolution, random and site-directed mutagenesis, and a combination of mutations [
12,
13,
14,
15,
16]. The individual or joint use of these scFvs allowed to neutralize various scorpion venoms and toxins [
10,
16,
17]. However, there are venoms such as
C. limpidus whose neutralization by scFv 10FG2 is partial due to the presence of the Cl13 toxin, which is not recognized by 10FG2 [
17]. Recently, the generation of scFv 11F of murine origin was published. This scFv recognized and neutralized Cl13 toxin and mixed with 10FG2, neutralized the whole venom of
C. limpidus [
18].
Identification of the toxic components for mammals of C. huichol venom, led to identify Chui5 toxin, whose sequence differs significantly from the one corresponding to toxins already recognized and neutralized by scFv 10FG2, explaining why it is not capable of efficiently neutralizing the whole venom. In this work, through the affinity maturation of 10FG2 toward Chui5 toxin by means of the construction of different libraries generated by random and/or site-directed mutagenesis, scFv HV was obtained. This scFv could neutralize Chui5 toxin and helped, in conjunction with scFv 10FG2 to optimally rescue intoxicated mice with the whole venom.
3. Discussion
In recent years, progress has been made in the study of the diversity of Mexican scorpion species of medical importance [
2,
3]. One of them recently described is the scorpion
C. huichol [
4]. The present work showed that the venom of this species is lethal to mammals. The toxicity is due to the presence of four lethal toxins in this venom: Chui2, Chui3, Chui4, and Chui5. It is noteworthy that Chui3 sequence is the same as the one of Cn3 toxin from the venom of
C. noxius [
27]. Of the three more abundant toxins, it was confirmed that they showed activity on hNav1.6 human sodium channel. In the case of Chui5 toxin, it also modified the activity of hNav1.5 channel. The presence of four components with a total abundance of 14% is significant since, typically, the abundances of the toxins correspond to less than 10% in the venom of Mexican scorpions [
16,
18,
20].
C. huichol venom was partially neutralized by scFv 10FG2, which is capable of neutralizing various venoms and toxins from scorpions of the genus
Centruroides [
16]. This partial neutralization is due to the presence of Chui5 toxin with an abundance of 5.7% and whose primary structure varies considerably from those already neutralized by this scFv. In this work, Chui5 toxin was used to perform three in vitro affinity maturation processes of scFv 10FG2. In the first process, variant 9a with L105Q mutation was obtained. The change in this position, compared to the others obtained, was the most important in terms of improvement of K
D and T
R. This variant delayed the onset of intoxication signs, as well as their intensity due to a considerable increase in T
R; however, the toxic effect of Chui5 was not neutralized. Position 105 of the parental scFv C1 was previously changed in processes of directed evolution against toxin Cll1 from
C. limpidus venom. In that work, the scFv 202F (with various mutations including M105L) was isolated, whose kinetic and neutralizing characteristics were improved when compared to its parent (scFv C1) [
14]. In addition, position 105 has been shown to be an important point of hydrogen bond-type interactions between several amino acids of various toxins (positions 5, 6, 8, 33, 55, and 56) and scFv 10FG2 [
16]. Variant 9a was used as template for the second maturation process, obtaining variant B15a that maintains L105Q mutation and incorporates K65Q mutation. This variant showed an improved K
D value compared to 9a and a higher T
R, delaying the onset of signs; however, its neutralizing capacity was lower. Previously, a model of the interaction between toxin Cn2 and scFv RJI-2 was generated which showed that position K65 of the scFv forms a hydrogen bond with N9 of the toxin. This observation prompted to identify RJI-2 as a target for mutagenesis by applying a rational design strategy. In this way, ER-5 variant was generated with K65R mutation, which neutralizes toxins Cn2, Cll1, and Css2, as well as the venom of
C. suffussus [
15]. Initially in this work, L105Q and K65Q changes were shown to be relevant for the affinity maturation toward Chui5 toxin. However, the results of the neutralization and the interaction kinetic constants indicated that the selected changes are not necessarily the optimal ones to obtain a neutralizing scFv. With this rational, positions 65 and 105 of scFv 9a were again mutagenized to saturation and random mutagenesis were performed at the same time. From this process 18 clones were isolated and sequenced, letting to analyze the frequency of occurrence of the amino acids at positions 65 and 105 (
Supplementary Figure S4b). It was observed that 50% of the sequences showed a glutamine at position 65 and 50% a glycine at position 105. Only 4 clones of the 18 evaluated carried changes K65Q and L105G, which are also present in the best variant, scFv HV. This scFv showed the highest recognition signal in ELISA assays (data not shown) and the absence of amber stop codons (
E. coli TG1 replaces amber stop codon by Q). This variant with an improved K
D despite a decreased T
R value with respect to scFv B15a, was able to neutralize Chui5 toxin. The evaluation of scFv HV with the other toxins of the
C. huichol venom showed that the mutations present decreased affinity and T
R (
Table 3). However, these K
D and T
R values are similar to those previously reported for neutralizing members of this family of scFvs [
16,
17].
Due to the affinity properties, the cross-reactivity and the ability of the scFv HV to neutralize Chui5 toxin, its neutralization capacity against
C. huichol venom was evaluated. In mix-type assays, the neutralization of 1 LD
50 with scFv HV and with the combination of scFvs 10FG2 and HV was evaluated using a 1:10:10 molar ratio (venom: scFv(s)). In both groups, the neutralization was 100%. The number of LD
50 was also increased, decreasing the venom: scFv molar ratio and it was observed that the neutralization decreased proportionally to the increase in LD
50 until a partial neutralization of 33% was obtained with 3 LD
50. These results lead to consider the use scFv 10FG2 as a neutralizing complement since it neutralizes other toxins except Chui5 (
Table 7). At molar ratios 1:10:10, neutralization occurred without signs of intoxication; thus, it was tested at lower molar ratios. The assay was made with 5 LD
50 of venom mixed with scFvs 10FG2 and HV at molar ratio 1:5:5 (venom: scFv 10FG2: scFv HV). A 100% neutralization was reached but with moderate signs of envenoming as expected due to this extreme challenge. Rescue assays were also carried out since they are closer to what happens during scorpion sting accidents. Only scFv HV was used in the rescue assays. Mice were envenomed with 3 LD
50 and 10 min later injected with the scFv; all mice survived the challenge. The disappearance of the strong signs of intoxication (sialorrhea, respiratory stress) was evident at around 2 h post injection, although discomfort in the area of the application of the venom disappeared around 48 h later. If we consider the mix-type neutralization assay of 5 LD
50 with scFvs 10FG2 and HV, the presence of 10FG2 in rescue assays can complement the neutralization of the venom and turn it even more efficient.
When analyzing the mutations of scFv HV, it was found that both the mutations in the FRs and those of the CDRs of the heavy chain variable domain contribute to the improvement of the neutralizing properties. Mutations F29Y and G154E seem to contribute to the stabilization of scFv HV at higher temperatures, which allowed the maintenance of binding to the toxin observed in the Van’t Hoff plot. It is likely that it is the result of the formation of new hydrogen bonds between Y29 and E154 with other residues of the scFv. Possibly, these mutations were favored by the use of guanidinium chloride in the screening rounds by eliminating unstable variants [
28] and/or weak interactions. The differences observed in the values of the slopes in the Van’t Hoff plot correlate with the better neutralization efficiency of scFv HV compared to that of scFv B15a that do not possess these changes. Increased stability has also been observed in scFv LR [
10] and LER [
29] scFvs with mutations in the FR region for example. Structural analysis of changes K65Q and L105G in the interaction area, suggests the conformation of a more flexible contact area, without steric hindrance for the accommodation of bulky amino acids such as Y8 of Chui5 toxin and Y59 of scFv HV. This rearrangement of the interaction forces, in terms of thermodynamic parameters, would be favoring binding. The negative ΔG of all interacting complexes indicates that the binding would be spontaneous and would form a stable complex. The ΔH of the Chui5-HV complex is slightly less negative than Chui5-B15a and Chui5-10FG2 complexes, suggesting that original interactions of L105 were possibly lost [
16], correlating with the decrease of the interaction of scFv HV with Chui2, Chui3, and Chui4 (
Figure 3 and
Table 3). This loss of interactions is reflected in the increase in ΔS in the Chui5-HV complex, which suggests a greater degree of movement of amino acid residues in the interaction zone of the toxin and scFv [
30].
Cross-reactivity with toxins from the venom of different species is a highly desirable phenomenon in those scFvs that would be selected to form part of an anti-venom. In the case of scFv HV, it was observed that it recognized a recently identified C. tecomanus toxin too (unpublished data). We have successfully expanded the cross-neutralization of scFv 10FG2, so it would be very useful to have the structure of the Chui5-HV complex to optimize this interaction and make the neutralization of these and other toxins even more efficient.