Deciphering β-tubulin gene of carbendazim resistant Fusarium solani isolate and its comparison with other Fusarium species

Exploration of molecular structure of β-tubulin is key to understand mechanism of action of carbendazim since its activity depends on strong binding to β-tubulin. Resistance against the fungicide is often associated with mutation in β-tubulin gene. A full-length (1619 bp) β-tubulin gene has been cloned and sequenced from a carbendazim resistant and a sensitive isolates of F. solani isolated from agricultural fields of Murshidabad (24.23 °N, 88.25 °E), West Bengal, India. Phylogenetic position of the isolates was confirmed using internal transcribed spacer and β-tubulin gene sequences. In the β-tubulin based phylogenetic tree, Fusarium species with available data were clustered in nine species complexes and members of both F. solani species complex and F. fujikuroi species complex were distributed into three clades each. The β-tubulin gene of F. solani was found to be shortest due to least number of non-coding sequences indicating its primitiveness among the Fusarium species. The coding region (G + C 58.54%) was organized into five exons. The protein has 446 amino acid, 49.834 KD molecular weight and 4.64 isoelectric point. Amino acid sequence of the resistant and the sensitive isolates were identical, suggesting that the mechanism of carbendazim resistance in the F. solani isolate was not due to point mutation in β-tubulin gene. The secondary and tertiary structure of β-tubulin were similar in all the species except F. oxysporum f.sp. cubense. The identification of binding sites for GDP, carbendazim and α-tubulin would resolve how carbendazim prevents tubulin polymerization. All the data are useful to design tubulin-targeted fungicide with better performance.


Introduction
The fungal β-tubulin had attracted extensive attention since it was established that its gene had been used as molecular marker for species identification and phylogenetic studies, and the protein served as a target of many anti-fungal drugs that inhibit microtubule assembly. Due to its universal occurrence β-tubulin gene has been used to study evolutionary relationship among taxon (Einax and Voigt 2003). The gene contains 3.5-fold more phylogenetic information than the 18S rRNA gene and accumulates fewer mutations (Begerow et al. 2004). Its non-coding sequences have been used to differentiate closely related species. In addition, β-tubulin being a cytoskeletal protein is assembled with α-tubulin as heterodimer in head-to-tail manner and involved in many cellular activities, such as cell division, ciliar or flagellar motility, and intracellular transport in eukaryotic organisms (Zhao et al. 2014). Several anti-tubulin fungicides including benzimidazole compounds bind at specific site(s) of β-tubulin and inhibit cell division in target fungi. Carbendazim (methyl benzimidazole carbamate) is a systemic fungicide that is widely used for management of crop diseases caused by filamentous fungi under the genera Fusarium (with light-coloured spores), Botrytis and Penicillium (causing fruit rots), Cercospora (causing eye spot), powdery mildew fungi and selected anastomosis groups of Rhizoctonia solani.
The members of F. solani (Mart.) Sacc. are ubiquitous pathogens of many agriculturally important crops (Coleman 2016). At least 111 plant species from 87 genera are commonly infected by the pathogen (Kolattukudy and Gamble Communicated by Michael Polymenis. 1 3 1995). They represent causal agent of several plant diseases such as fruit and root rot of Cucurbita spp., root and stem rot of Pisum sativum, foot rot of Phaseolus vulgaris, dry rot of Solanum tuberosum, sudden death syndrome (SDS) of Glycine max, wilt and damping off in vegetables, and canker and dieback disease in tropical trees (Leslie and Summerell 2008). They are also responsible for post-harvest decay of many succulent fruits and vegetables. Some strains are associated with human infections (O'Donnell et al. 2020). Carbendazim was found to be effective in controlling F. solani causing wilt of mulberry (Narayanan et al. 2015).
However, exclusive and repeated use of the fungicide on crop management especially against polycyclic diseases emerged resistance among fungal populations. Benzimidazole resistance has been reported in approximately 60 genera in over 115 fungal species (https:// www. frac. info/). Among these, molecular mechanism of the resistance has been reported only in 49 species including 19 species from in vitro study. In most cases, carbendazim resistance is associated with point mutation at several target sites in the β-tubulin gene, which results in altered amino acid sequences and change in carbendazim-binding site (Ma and Michailides 2005). This causes a reduction in the affinity of the β-tubulin for the fungicide, without interfering with normal biological function of β-tubulin in the fungus. Sometimes a wide range of carbendazim-resistance could be found instead of carrying different mutations at similar amino acid residue and different substitutions at the same codon could exhibit distinct level of resistance (Albertini et al. 1999). Fusarium spp. such as F. asiaticum, F. fujikuroi, F. moniliforme, F. oxysporum f.sp. niveum, F. proliferatum, F. verticilloides, Gibberella zeae in which mechanism of benzimidazole resistance has been resolved showed point mutation at different sites in β-tubulin gene (Suga et al. 2011;Chen et al. 2014;Yan and Dickman 1996;Qiu et al. 2011;Petkar et al. 2017;Yang et al. 2018;Xu et al. 2019). Moreover, drug-efflux mechanism, detoxification of fungicides and some unknown mechanisms are responsible for benzimidazole resistance in fungi including Fusarium spp. Sequencing of β-tubulin gene from carbendazim resistant F. solani isolates will provide some idea about the resistance mechanism in this species.
Tertiary structure of any protein provides an important insight into the molecular mechanism of its function, which forms the basis for intending various approaches for structure-based drug designing or altering its structure by site-directed mutagenesis. X-ray crystallography and NMR are no doubt advance techniques to provide high definition structure of protein, but the disadvantages are that these techniques are often prolonged, costly and require large quantity of purified protein. Homology based modeling of protein is a competent computational tool that can be used for predicting the structure of unknown protein based on previously resolved three-dimensional structure of other associates of the same relative having similar folds and/ or function. To date, there are hardly any studies on structure-function relationship of β-tubulin of F. solani exhibiting carbendazim resistance and in particular there is no study on structural properties of β-tubulin of F. solani. Hence, in order to understand the functionality and role in carbendazim resistance, it is imperative to understand molecular structure of the protein.
Previously, carbendazim resistant F. solani isolates (SF0204, SF0301 and SF1303) which could grow in 100 µg/ mL carbendazim and a sensitive isolate (SF0104) which showed growth upto 10 µg/mL carbendazim have been identified from our laboratory (Tarafder et al. 2019). In this study, we have identified the phylogenetic position of the F. solani isolates SF0104 and SF0301 based on ITS sequence and reported for the first time, cloning and sequencing of the full-length of β-tubulin gene, that have been analysed to detect any mutation responsible for the carbendazim resistance and to compare their nucleotide and amino acid sequences with other Fusarium species and the β-tubulin gene was used to construct a phylogenetic tree of Fusarium species. Furthermore, we have predicted the three-dimensional structure of the β-tubulin by homology modeling and the binding sites of GDP, carbendazim and α-tubulin were identified through molecular docking.

Fungal isolates
A number of Fusarium solani isolates were isolated from agricultural fields of Murshidabad (24.23 °N,88.25 °E), West Bengal, India, identified based on morphological characterization and rRNA gene sequencing and their carbendazim sensitivity was studied (Tarafder et al. 2019). Among these, F. solani SF0104 was considered as carbendazim sensitive as it showed only 6.7% growth in 10 µg/mL carbendazim and failed to grow in 100 µg/mL carbendazim with ED 50 value 0.49 µg/mL. By contrast, F. solani SF0301 was carbendazim resistant and could grow in 100 µg/mL carbendazim with ED 50 value 0.98 µg/mL. The pure cultures of the isolates were maintained on potato dextrose agar (PDA) medium at the Mycology and Plant Pathology Research laboratory, Department of Botany, University of Kalyani, India.

Preparation of phylogenetic tree of FSSC based on ITS sequence
Genomic DNA was extracted from fresh mycelia using the DNeasy kit (Qiagen, Germany) according to the manufacturer's instruction. The ITS (internal transcribed spacer) of rDNA regions were amplified using ITS5 (F) [5′-GGA AGT AAA AGT CGT AAC AAGG-3′] and ITS4 (R) 5′-TCC  TCC GCT TAT TGA TAT GC-3′] (Dutta et al. 2018). ITS sequences of the Fusarium species of F. solani species complex (FSSC) were retrieved from GenBank (https:// www. ncbi. nlm. nih. gov/ genba nk) and Al-Hatmi et al. (2018), and used for preparation of phylogenetic tree with MEGA X software (Kumar et al. 2018). The genetic relationships were investigated by applying Neighbour-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value (Tamura and Nei 1993). The tree was drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 44 nucleotide sequences with a total of 1189 positions in the final dataset. The confidence of the branches was determined with bootstrap analysis in 1000 samplings.

Cloning, sequencing and analysis of β-tubulin gene
Forward primers βOD F (5′-AAC ATG CGT GAG ATT GTA AG-3′) was designed from partial β-tubulin gene sequence of F. solani (GenBank accession No. KF255996) and according to O'Donnell and Cigelnik (1997). The reverse primer βEx5 R (5′-TTA CTC CTC GCC CTC AGGG-3′) was designed from sequence alignment of a number of Fusarium species whose full sequences were available in database (https:// www. ncbi. nlm. nih. gov/ genba nk). The start codon and termination codon in the primers are underlined. PCR amplification was performed using Perkin-Elmer PCR system in a 50 μL reaction mixture containing 5 μL of 10X PCR buffer, 25 mM MgCl 2 , 10 mM dNTPs, 500 nM each primer, 1 U Taq polymerase (Thermo Scientific™), and 100 ng genomic DNA. The thermal programme used for PCR was as follows: an initial denaturation of 5 min at 95 °C, followed by 35 cycle of denaturation for 30 s at 94 °C, annealing for 45 s at 57 °C, and extension at 72 °C for 2 min, followed by a final extension for 10 min at 72 °C. PCR products were purified with a gel purification kit (Qiagen, Germany), directly ligated into the pGEM-T Easy vector (Promega, USA) and sequenced with the vector primers and internal primers using Automated DNA Sequencer (ABI 3500 Genetic Analyzer) (Perkin-Elmer, Applied Biosystem, Inc.) at the S. N. Bose Innovation Centre with Central Instrumentation Laboratory, University of Kalyani. The β-tubulin sequences of both the isolates were submitted to the GenBank (https:// www. ncbi. nlm. nih. gov/ genba nk) and homology search was carried out using nucleotide BLAST and Fusarium MLST (http:// fusar ium. mycob ank. org). Number, position and size of coding and non-coding regions were identified through comparison with other sequences.

Comparison of β-tubulin nucleotide sequence and phylogenetic analysis
The β-tubulin sequences of both the isolates SF104 and SF301 were compared with 53 full-length β-tubulin sequences of other Fusarium species which were available in GenBank database (https:// www. ncbi. nlm. nih. gov/ genba nk). The alignment was made by CLUSTAL W (Thompson et al. 1994) using the Blosum matrix and standard default parameters and MEGA X software was used to construct a phylogenetic tree (Kumar et al. 2018).

Alignment of β-tubulin amino acid sequence
The DNA sequences from the putative coding region were translated into amino acid sequences with the standard code using ExPASy-Translate tool (https:// web. expasy. org/ trans late). The G + C content of coding region, molecular mass, extinction coefficient and isoelectric point of the protein were calculated using web based tools (www. biolo gicsc ope. com; www. aatbio. com; www. isoel ectric. org). To detect any mutation the deduced amino acid sequences of the resistant and sensitive isolates were aligned and compared. The deduced amino acid sequences of the F. solani isolates were compared with β-tubulin sequences of other Fusarium species through alignment by CLUSTAL W (Thompson et al. 1994) using MEGA X (Kumar et al. 2018).

Homology modeling of β-tubulin
Homology model of β-tubulin of F. solani isolates and other Fusarium species were carried out using the Swiss-Model programme (https:// swiss model. expasy. org) based on the crystal structure of Bos taurus (PDB code: 4O4I of chain D). There was no experimental three-dimensional structure available in PDB for β-tubulin of F. solani. The modeling data were then visualized and analysed by BIOVIA Discovery Studio Visualizer. The secondary structure of the protein was developed in POLYWIEW-2D (http:// polyv iew. cchmc. org/) (Porollo et al. 2004). The constructed model of F. solani was compared with other Fusarium species.

Validation of the model
The backbone conformation of the modeled structure of β-tubulin of F. solani was intended by the analysis of phi (Φ) and psi (ψ) torsion angles using Ramachandran plot server (https:// zlab. umass med. edu/ bu/ rama/ and http:// vadar. wisha rtlab. com/), as resolved by Ramachandran plot statistics. The model was further scrutinized by ProSA (Wiederstein and Sippl 2007) and QMEAN (Benkert et al. 2009). ProSA was used to exhibit the Z-score and energy plots. VADAR (http:// vadar. wisha rtlab. com/) was used to calculate the volume area dihedral angle for fractional accessible surface area.

Identification of nucleotide binding site(s)
The nucleotide binding sites within the β-tubulin were identified using a topological approach. To recognize the nucleotide binding site of PDB file of the modeled β-tubulin we used software BIOVIA Discovery Studio v21.1.0 using ligand interaction menu.

Molecular docking
To recognize potential binding sites and the binding affinities of carbendazim (PubChem ID: 25429) for β-tubulin, web-based SwissDock program (www. swiss dock. ch/ docki ng) was used to perform automated molecular docking. The blind docking was performed using default parameters, with no region of interest defined. Docking results were visualized using UCSF Chimera v1.13.1 software and BIOVIA Discovery Studio v21.1.0.
To locate a consensus protein-protein binding conformation of α-tubulin and β-tubulin complex two different docking servers, viz., HADDOCK (High Ambiguity Driven protein-protein DOCKing) (Dominguez et al. 2003;de Vries et al. 2010) and ClusPro (Kozakov et al. 2017;Vajda et al. 2017) were used. The web-based docking tool ClusPro uses the rigid body docking method with the help of PIPER. In PIPER ligands were allowed to move while the conformation of the receptor was kept fixed. Top clusters were ranked with their representative centers on energy scoring functions. The best docked complex was chosen which would comply with the amino acid residues present in the binding interface. We also used HADDOCK web-based server for docking purpose. Rigid body docking server HADDOCK also followed the same steps as in ClusPro. The interfaces of all docked complexes were calculated using PISA server (www. ebi. ac. uk/ pdbe/ pisa/).

Superposition of three-dimensional protein
The constructed tertiary structure of β-tubulin of F. solani was compared with that of other Fusarium species through superimposition in UCSF Chimera (https:// www. rbvi. ucsf. edu/ chime ra) (Pettersen et al. 2004). Using web-based program SuperPose Version 1.0 (http:// super pose. wisha rtlab. com/) the three-dimensional protein of F. solani SF0301 was superimposed with that of F. oxysporum f.sp. cubense. During the superimposition, the following parameters were deposited. To look for sub-domain matches and mismatches (e.g. hinge regions) for pair-wise sequence, the identities were considered above 80%. To identify as 'similar' aligned alpha-carbon atoms with RMSD (root-mean-square deviation) of atomic positions less than 2.0 Å were used as similarity cut off value. To identify as 'dissimilar' aligned alphacarbon atoms with RMSDs greater than 3.0 Å were used as dissimilarity cut off value. To set the dissimilar sub-domain, the minimum number of contiguous alpha-carbon atoms with RMSDs above the dissimilarity cut off (above) required to be considered a 'dissimilar' sub-domain was 7 atoms.

ITS sequence based phylogeny of FSSC members
The phylogenetic analysis based on ITS sequence separated three major clades of the FSSC members and the F. solani isolates SF104 and SF301 aggregated into the members belonging to Clade 3 ( Fig. 1
By comparison with other fungi it was inferred that the full length (1619 bp) β-tubulin gene of both the F. solani isolates was organized into five exons and four introns (Fig. 3A). The intron positions were conserved and occurred after the codons for amino acids 4, 12, 53 and 316, but they varied in their lengths. All introns contained 5′ and 3′ consensus splice junctions similar to those in β-tubulin genes of other fungi that conformed to the GT-AG rule. The G + C content of the coding region was calculated as 58.61%. Both the β-tubulin genes encoded a 446 amino acid protein, with a calculated molecular mass 49.834 KD, extinction coefficient 41,640 M −1 cm −1 and an estimated isoelectric point of 4.64.
Amino acid sequence comparison between the carbendazim resistant and sensitive isolates revealed that they were all identical, indicating that the mechanism of carbendazim resistance in F. solani SF0301 was different from other

Phylogenetic analysis of Fusarium species based on β-tubulin nucleotide sequences
The phylogenetic analysis of the Fusarium species including F. solani based on β-tubulin gene sequence revealed their position in nine clusters considered as species complexes (Fig. 2). The species complexes are F. sambucinum species complex (FSaSC, nine species), F. incarnatum-equiseti species complex (FIESC, two species), F. tricinctum species complex (FTSC, two species), F. heterosporum species complex (FHSC, one species), F. fujikuroi species complex (FFSC, 25 species), F. oxysporum species complex (FOSC, four sequences), F. redolens species complex (FRSC, one species), F. decemcellulare species complex (FDSC, one species) and F. solani species complex (FSSC, 10 sequences). The phylogenetic tree showed a well-supported relationship (99% MP bootstrap) among the F. solani isolates of FSSC under the Clade 3 including our Indian isolates, SF0104 and SF0301. The formation of three clades by FSSC isolates was also confirmed by the analyses based on β-tubulin sequences. The sister groups FSaSC and FIESC, FTSC and FHSC also showed well-supported lineage with 100% bootstrap support. F. decemcellulare formed lineage from F. solani was in accordance with Aoki et al. (2014). The 25 species under the FFSC were phylogenetically placed in three different clades as American, Asian and Africanthe Asian clade was derived from the African clade with 82% bootstrap support and American clade from the Asian clade with 100% bootstrap support.

Comparison of β-tubulin gene of F. solani isolates with other Fusarium species
The β-tubulin gene of carbendazim resistant F. solani SF0301 was compared with complete β-tubulin gene of other Fusarium species available in GenBank (Fig. 3B). All Fusarium species including F. solani had β-tubulin gene with five exons except F. graminearum (AY303689), F. praegraminearum (KX260131) and F. sambucinum (AF484166) as they lack 4th intron; hence, their total β-tubulin length became shorter with 1631 bp, 1629 bp and 1643 bp, respectively. Other species also showed variation in total length due to variation of intron length and among these, 1st intron showed maximum variation. The β-tubulin gene of the F. solani isolate was found to be shortest with 1619 bp in length and their introns were shortest among the Fusarium species. Out of the total length of 1619 bp, coding region consisted of 1341 bp and non-coding region with only 278 bp was made up of 1st, 2nd, 3rd and 4th introns with 139 bp, 49 bp, 44 bp and 46 bp, respectively. All the introns in other Fusarium species were larger, thus their β-tubulin lengths were larger and in the range from 1668 bp in F. redolens (MT011043) to 1684 bp in F. andiyazi (MT011059). The nucleotide length of 2nd, 3rd and 4th introns were almost similar among the Fusarium species with an average of 58 bp, 48 bp and 50 bp, respectively. However, shorter 2nd intron was found in F. sambucinum (57 bp) and F. solani (49 bp), and larger in F. verticillioides (59 bp). The length of 3rd intron also varied in few species such as, 49 bp in F. ramigenum and F. napiforme, 47 bp in F. redolens and 44 bp in F. solani. The length of 4th intron was 49 bp in F. brevicatenulatum, F. pseudoanthophilum, F. pseudonygamai, F. verticillioides and F. redolens, whereas in F. solani it was 46 bp. The variable 1st intron was largest in F. sambucinum (197 bp) and shortest in F. solani (139 bp). Among the total 1341 bp coding region, the length of exons of all Fusarium species was highly conserved and exon 1, 2, 3, 4 and 5 contained 12 bp, 24 bp, 123 bp, 791 bp and 391 bp, respectively. Thus, exon 4 was the largest. However, exon 4 of F. graminearum (AY303689), F. praegraminearum (KX260131) and F. sambucinum (AF484166) were larger than the average length (791 bp) due to lack of 4th intron and contained 1185 bp, 1182 bp and 1182 bp, respectively. The β-tubulin of all the Fusarium species encoded 446 amino acid protein except F. graminearum which encoded 447 amino acid protein. All these data were calculated only from β 1 -tubulin nucleotide sequence from those which had two β-tubulin genes.

Comparison of β-tubulin amino acid of F. solani isolates with other Fusarium species
The most frequent difference of β-tubulin amino acid sequence of F. solani isolates with other Fusarium species was found to be at 381st position with isoleucine (I) in F. solani isolates and valine (V) in other (Fig. 4). However, F. avenaceum and F. heterosporum also contained isoleucine at 381st position similar to the F. solani isolates. But they differed from F. solani isolates at 189th position where valine in F. solani was changed as isoleucine in these two species. Another difference was identified at 335th position where asparagine (N) in F. solani isolates was changed to serine (S) in other Fusarium species. However, this change was not detected in species such as F. oxysporum, F. fujikuroi, F. verticilloides, F. decemcellulare, F. avenaceum and F. heterosporum. The amino acid proline (P) was only found in F. oxysporum f.sp. cubense at 279 th position, where as in others it was histidine (H). In addition, a few minor changes were found in different species (Fig. 4).

Homology modeling and structural analysis
To build a tertiary structural model of β-tubulin, an appropriate template has been searched using Swiss-Model (https:// swiss model. expasy. org). Of the yielded templates, the crystal structure of Bos taurus (PDB ID: 4O4I) was found to be the best template with the sequence identity of 83.15% with the query coverage of 96.64% and the QMEAN score of the built model was − 1.85. A tertiary model structure of β-tubulin was again built up by BIOVIA Discovery Studio package using the crystal structure. The three-dimensional model of the protein was found to be composed of different secondary structural elements (Fig. 5).

Model reputation
The stereo-chemical behaviour of the predicted model of β-tubulin protein of F. solani was analysed through ProSA and QMEAN server's confirmation and evaluated by the Phi/Psi Ramachandran plot inspection. ProSA was used to confirm the three-dimensional model of β-tubulin protein for potential errors. The program displayed two attributes of the input structure: its Z-score and a plot of its residue energies. The ProSA Z-score of − 10.11 indicates the overall model quality of β-tubulin protein (Fig. 6A). The deviation of total energy of the structure, measured by the Z-score, with respect to an energy distribution obtained from random conformations. The scores were well within the range of scores and indicated a highly reliable structure typically found for proteins of comparable size. The energy plot demonstrated the local model superiority by plotting the knowledge-based energies as function of the amino acid sequence position (Fig. 6B). QMEAN analysis was also employed to evaluate and validate the model. The QMEAN4 score of the model was 0.82 (Fig. 6C) and the Z-score was − 1.85 which was close to zero and this confirmed the superior quality of the model. This is because the estimated consistency of the model was projected to be in between 0 and 1 (Table 1). Assessment between regularized QMEAN score (0.82) and protein size in non-redundant set of PDB structure in the plot revealed different sets of Z value for different parameters such as C-beta interactions (0.28), interactions between all atoms (0.24), solvation (0.85) and torsion (− 2.31) ( Table 1).
The constructed homology model was also assessed for structural and stereo-chemical competency. A Ramachandran Phi/Psi plot for β-tubulin (Fig. 6D) showed 97.436% highly preferred observations, 1.865% preferred observations and 0.699% questionable observations. The analysis of the predicted structure provided strong evidence that the predicted three-dimensional structure of β-tubulin was of superior quality.

Determination of putative nucleotide binding sites
In order to explore the spatial relationships between nucleotide and β-tubulin we used a topological approach after homology modeling. The nucleotide-binding domain in β-tubulin was depicted in Fig. 7A, B. We assumed that the GDP-binding site of β-tubulin was composed of two components: a guanine-binding component and a phosphoryl-binding component. The guanine-binding component was composed of CYS12, GLN15, SER176, GLU181, ASN204, TYR222 and ASN226 and the phosphorylbinding component was composed of GLN11, SER138, GLY141, GLY142 and THR143. The binding pattern and the H-bond formation are according to the Fig. 7A, B and Table 2.

Molecular docking to identify carbendazim binding site(s)
Molecular docking was performed to map the interactions between carbendazim and β-tubulin and to find out putative binding site(s). The molecular docking of carbendazim and β-tubulin detected 48 clusters in eight different sites of the protein. The top-score cluster exhibited best "full fitness" and those parameters were calculated by averaging the 30% most favourable effective energies of a cluster's element and lower free energy than those obtained for other potential binding sites (Grosdidier et al. 2007 (Fig. 7C, D). All the putative bonds with their bond distance were in accordance with Table 3.

Dimerization of α-and β-tubulins
Hetero-dimerization of α-and β-tubulin subunits was also checked by protein-protein docking. The two monomers were interacted to each other with H-bonds and electrostatic values are represented in 0 to 10 scales; where, 0 value represents completely buried, i.e. 0-9% RSA (Black) and the value 9 represents fully exposed, i.e. 90-100% RSA (White) interactions. The H-bonds were formed through salt bridge formation, conventional H-bonds and non-conventional with carbon-hydrogen bond formation. Out of total 11 bonds/ interactions, the monomer β-tubulin contributed six amino acids as H-donor and five amino acids as H-acceptor. The amino acid representatives of α-and β-tubulin monomers with their actual positions in the peptide chain, the bond distance and bond angles are represented in Table 4. The β-tubulin of F. solani isolates had sites for the putative α-/βdomain interface at ASP128, CYS129, ARG162, LYS252, ASN256, ARG331, ASN347, GLN350 and GLY429 (Fig. 7E).

Comparison of β-tubulin tertiary structure
Homology models of β-tubulin of all the Fusarium species were constructed and to identify any discrimination with the F. solani protein, superimposition was performed. Though few regions in amino acid residues were dissimilar among the species, the tertiary structural superimpositions were nearly identical. This could be due to substitution of the   (Fig. 8A, B). In β-tubulin of F. oxysporum f.sp. cubense, PRO279 (yellow) strongly refused to form an α-helix and acted as a helix-breaker (Fig. 8D), whereas in all other Fusarium species including F. solani, HIS279 (yellow) formed an extra α-helix (Fig. 8C). In Fig. 8C, the schematic diagram of β-tubulin of F. solani SF0301 showed ARG213 (blue) as a component of an α-helix while in F. oxysporum f.sp. cubense (Fig. 8D) ARG213 was not integrated into the α-helix rather it formed loop with its C terminal amino acid residue.

Discussion
Two main objectives of the study were to identify phylogenetic position of the F. solani isolates isolated from lower Gangetic plain of Indian subcontinent and to address their carbendazim resistance mechanism. Based on phylogenetic analyses of 19 protein-coding genes, Geiser et al. (2020) (Bogale et al. 2009).
In this study, we confirmed phylogenetic position of the F. solani isolates based on ITS sequence and β-tubulin gene sequence (Figs. 1, 2). Although, β-tubulin gene sequences of both the isolates were identical, they differed slightly in ITS sequence and was placed apart in the ITS sequence based phylogenetic tree. However, both the ITS and β-tubulin gene sequence based phylogenetic trees separated the members of the FSSC into three clades and F. solani isolates, SF0104 and SF0301 were placed in Clade 3. When β-tubulin gene was used as molecular marker for construction of phylogenetic tree, the 53 Fusarium species were clustered within the nine species complexes and this clustering was almost similar with the phylogenetic analysis of O' Donnell et al. (2013), inferred from a combined gene sequences of RPB1and RPB2 datasets.
In the β-tubulin gene sequence based phylogenetic tree, 25 Fusarium species were grouped under the Fusarium fujikuroi species complex (FFSC) and placed in American (5), Asian (6) and African (14) clades (Fig. 2). The emergence of the three clades in FFSC was initially recommended by O'Donnell et al. (1998) due to the fragmentation of Gondwanaland. Afterward, the same authors reported that the complex emerged more recently (ca. 8.8 million year ago) and the apparent biogeographic clustering was probably due to long distance dispersal from South America to Africa and then to Asia in the late Miocene (O'Donnell et al. 2013). Distribution of species in three clades of FFSC was in accordance with the Maximum likelihood phylogeny of the FFSC inferred from the combined translation elongation factor 1-α and β-tubulin gene regions sequence data of Herron et al. (2015).
The members of FSSC are agriculturally important pathogens causing foot and root rot of fruits and vegetables, stem blight, vascular wilt, and sudden death syndrome of soybean. Both pre-and post-emergence stages of crop are susceptible to attack of F. solani. Control of these pathogens is very difficult since they live in soil as saprophyte by producing chlamydospores. They have broad host range with diverse formae speciales and could tolerate high temperature and salinity stress. Several in vitro studies have observed efficacy of carbendazim against F. solani (Gupta et al. 2020;Padvi et al. 2018;D'Addazio et al. 2016). Benomyl resistant F. solani causing dry rot of Amorphophallus paeoniifolious has been reported (Dorugade et al. 2021). Due to single-site action, carbendazim resistance develops rapidly in pathogenic fungi including Fusarium species Avenot et al. 2020;Liu et al. 2020). In the field, resistant isolates can arise with a very low rate of mutation in a specific gene and adapt  (Yan and Dickman 1996;Qiu et al. 2011;Suga et al. 2011;Chen et al. 2014;Petkar et al. 2017;Yang et al. 2018;Xu et al. 2019). In addition, substitutions have also been observed at the position of 6, 165, 235, 240 and 241 in other phytopathogenic fungi (Ma and Michailides 2005). Some fungi contained two β-tubulin genes and among these, only one was responsible of benzimidazole resistance. However, F. solani contained only one β-tubulin gene with four copies (Zhao et al. 2014). In this study, β-tubulin amino acid sequence of the carbendazim resistant and the sensitive isolates were identical and point mutation responsible for amino acid changes were not detected in F. solani SF0301 β-tubulin gene. This might be explained by very low (2) resistance factor (the ration of ED 50 of resistant and sensitive isolates) of F. solani SF0301 in compare to other plant pathogenic fungi (> 1000) (Chen et al. 2007). Identical amino acid sequences of the β-tubulin has been reported from several thiabendazole-resistant and -sensitive isolates of Gibberella pulicaris (syn. F. sambucinum) (Kawchuk et al. 2002). The possible mechanism(s) that confer carbendazim resistance in F. solani SF0301 include overproduction of the fungicide target, detoxification of the fungicide by fungal metabolite such as glutathione transferase, an active efflux or reduced uptake of the fungicide and some unknown mechanisms (Sevastos et al. 2017). In benomyl-resistant Colletotrichum acutatum isolates, putative leucine zipper protein CaBEN1 increased production of the target protein by enhancing β-tubulin gene (CaTUB1) expression (Nakaune and Nakano 2007). Viglas and Olejnikova (2021) reported that in filamentous fungi the ATP-binding cassette (ABC) transporters play a key role in antifungal resistance by transporting various xenobiotics, including antifungal compounds. Drug-efflux ABC transporters were involved in resistance to benzimidazole fungicides in Aspergillus nidulans (Andrade et al. 2000) and Penicillium digitatum (Nakaune et al. 1998). These reports demonstrated that the resistance to benzimidazole fungicides could be partially due to the membrane transport system for these compounds into the fungal cells. Some mutations in the α-tubulin gene of Saccharomyces cerevisiae increased benzimidazole sensitivity (Richards et al. 2000). Further investigations are warranted in understanding the existence of carbendazim resistant mechanism in F. solani SF0301.
In the present study, we characterized organization of β-tubulin gene of F. solani isolates. The gene was organized into five exons of variable length (Fig. 3A). To the best of our knowledge, this was the first report of complete β-tubulin gene sequence of F. solani. The full-length nucleotide sequences were in agreement with partial sequences of various F. solani isolates available in database (https:// www. ncbi. nlm. nih. gov/ genba nk). We also observed that β-tubulin genes of Fusarium species varied significantly in their length and the variation was due to difference of their intron lengths, especially the 1st intron, which was largest among the four introns. However, the number and length of exons among Fusarium species were almost conserved (Fig. 3B). We also compared the deduced amino acid sequence of β-tubulin of different Fusarium species (Fig. 4) and constructed the secondary and tertiary structure of β-tubulin protein of F. solani through homology modeling and compared the structure with that of other Fusarium species (Fig. 5). The constructed β-tubulin was scrutinized by ProSA and QMEAN analysis and validated by the analysis of phi (Φ) and psi (ψ) torsion angles using Ramachandran plot (Fig. 6).
During polymerization of microtubules, tubulin monomer shows GTPase activity and the GTP bound with β-tubulin is hydrolysed shortly after being incorporated. Therefore, GDP is predominantly associated in the β-tubulin subunit near the newly formed plus end (Horio et al. 2014). A highly conserved glycine-rich sequence, GGG TGA G, has been identified from residues 140 to 146 in the F. solani β-tubulin which formed part of the GDP-14 binding site as reported by Nogales et al. (1998). As similar to other fungi, the β-tubulin of F. solani isolates encoded an N terminal tetrapeptide MREI which was involved in the autoregulation of β-tubulin expression as detected in mammalian cells (Msiska and Morton 2009).
Carbendazim has been reported to bind at specific position of the β-tubulin in different fungi (Zhou et al. 2016;Vela-Corcía et al. 2018;Yang et al. 2018;Xu et al. 2019) while, to the best of our knowledge, no evidence has been raised for physical interaction between carbendazim and β-tubulin in F. solani. In this study, the interaction has been demonstrated through molecular docking. Zhou et al. (2016) showed that in F. graminearum carbendazim inhibited the dimerization of tubulin by binding with the β 2 -tubulin; but unable to affect the polymerized tubulin. In our study, we investigated the mystery behind the binding of carbendazim to the monomeric form but not to the polymerized tubulin. The carbendazim binding pocket in β-tubulin of F. solani, consisting of amino acids at the position was close to the α/β-tubulin subunit interfaces (Fig. 7). Therefore, due to the presence of carbendazim binding pocket inside the β-tubulin, we can assume that the formation of carbendazim-β-tubulin complex resulted in conformation change in α/β-tubulin subunit interfaces followed by losing the ability of these subunits to accept other tubulin molecules for further microtubule polymerization. In contrast, polymerized microtubules may have compacted organization that destroyed the accessibility of carbendazim to the β-tubulin.
Among the naturally occurring amino acids, proline is unique in that its side chain cyclically back to the backbone amide, leaving one of its dihedral angles (ϕ) fixed at − 65° (Richardson and Richardson 1989;MacArthur and Thornton 1991). Proline residue is not often found at the center of secondary structures such as the α-helix and β-sheet in globular protein due to the structural consequence of this particular arrangement Fasman 1978, 1974). When proline is located in the first turn, acting apparently as an N-capping residue, it does occur in an α-helix (Richardson and Richardson 1988). In our study, to compare the secondary structure formed in β-tubulin among different Fusarium species, homology modeling of the β-tubulin protein followed by superimposition of each β-tubulin with the β-tubulin of F. solani was carried out. Algorithms to superimpose protein's three-dimensional structures were applied to identify similarities of protein folds. The coordinates of a protein was superposed so that the backbone lies over the backbone of a reference protein. Distant homologues might not be recognized by their amino acid sequence because the sequences diverge more rapidly in evolution than the three-dimensional structure. The most significant difference was observed at the 279th position in F. oxysporum f.sp. cubense due to presence of proline instead of histidine (Fig. 5). In F. oxysporum f.sp. cubense proline was located in the last turn, acting apparently as a C-capping residue of α-helix and acted as a helix breaker whereas, in others including F. solani HIS279 was a part of α-helix (Fig. 8). Proline is unfavorable to the α-helical conformation for the several reasons. First, due to the absence of an amide proton on an x-Pro (x = any amino acid residue) bond which participates in helix stabilization during intramolecular hydrogen bonding (Williams and Deber 1991). Second, its pyrrolidine ring is too bulky to place steric constraint on the conformation of the earlier residue in the α-helix (Hurley et al. 1992). Finally, proline, being a secondary amide, is comparatively a polar residue that exhibits an enhanced tendency to form strong hydrogen bonds in non-periodic structural motifs such as prolineinduced β-turns (Smith et al. 1980) and γ-turns (Deber et al. 1990). Although, no significant differences in protein function were observed in F. oxysporum f.sp. cubense because PRO279 was not the interacting residue in both tubulin polymerization as well as carbendazim binding. The other amino acid substitution(s) in different location among the species was not significant for the changes of secondary structure probably due to the substitution by the amino acid of same chemical group.
Structural characterization of β-tubulin of Fusarium species will provide in-depth insight to design novel tubulintargeted drug whose efficacy cannot be changed through mutation in β-tubulin. Similar work might be undertaken in other phytopathogenic fungi where carbendazim resistance appears a problem. Further phylogenetic study of the F. solani isolates will be carried out by sequencing other protein coding genes such as RPB1 and RPB2, translation elongation factor 1-α and calmodulin. The exact mechanisms of carbendazim resistance in F. solani isolates will be investigated in future.

Conclusions
Carbendazim resistance of F. solani SF0301 was not associated with mutation in the β-tubulin. ITS and β-tubulin based phylogeny confirmed the placement of the Indian isolates of F. solani within Clade 3 of FSSC. In spite of variations in amino acid sequence of β-tubulin, constructed secondary and tertiary structures were in all Fusarium species except F. oxysporum f.sp. cubense. The carbendazim binding site and α-tubulin interacting site were located close to each other and binding of either of two caused conformational change in β-tubulin that resulted inaccessibility to other. The presence of proline at 279th position in β-tubulin of F. oxysporum f.sp. cubense resulted formation of 20 α-helices but did not interfere its dimerization with α-tubulin and binding with carbendazim.
Funding The work has been supported by the funds received from University of Kalyani and DST-FIST and DST-PURSE programme II, Govt. of India