Interaction estimation of pathogenicity determinant protein βC1 encoded by Cotton leaf curl Multan Betasatellite with Nicotiana benthamiana Nuclear Transport Factor 2

Sequence retrieval for βC1 and construction of expression cassette

The full-length βC1 gene was amplified from the isolate H181 of CLCuMB using the primers having restriction sites HindIII and BamHI (F: CCAAGCTTATGACAACGAGCG and R: CCGGATCCTTAAACGGTGAACT). The polymerase chain reaction (PCR) profile was adjusted as an initial denaturation at 95 °C for 5 min, denaturation at 95 °C for 30 s, annealing at 50 °C for 30 s, extension at 72 °C for 30 s and then final extension at 72 °C for 10 min followed by 35 cycles.

The amplified product of βC1 (357 bps) was phenol-chloroform purified and digested to clone under CaMV (Cauliflower Mosaic Virus) 35S promoter of expression vector pJIT-163. Positive clones were confirmed by restriction analysis and further confirmed by standard PCR. Gene expression cassette (1,888 bps) was excised from positive clone samples in pJIT-163 vector using SacI and XhoI. This expression cassette was cloned in binary vector pGreen-0029 at the SacI and XhoI sites.

Agroinfiltration of expression cassette of βC1 gene

For in-planta interaction evaluation, N. benthamiana plants were grown under controlled conditions including 16 h of light, 8 h of dark photoperiod at 28 °C and 25 °C day and night temperature and approximately 65% humidity. Seeds of plants were grown in pots containing peat moss for three weeks.

The expression cassette of βC1 gene and empty vector pGreen-0029 was electroporated into Agrobacterium tumefaciens strain GV3101 within electroporation cuvette of (two mm space) according to the previously reported procedure (Shen & Forde, 1989). Agrobacterium culture nurtured this construct of expression cassette and mock inoculation (empty vector pGreen-0029) was grown within 100 µl of kanamycin and 25 µl of ampicillin per 100 ml media for 48 h (at 28 °C). Bacterial cells were pelleted (5,000×g for 15 min at 4 °C) on an OD at 600 nm of 1 and then resuspended in 10 mM MgCl₂ and 100 µM acetosyringone. Cells were left for 2 to 3 h before infiltration at room temperature. The Agrobacterium suspension was then infiltrated using 1 ml needle-free disposable syringe on the lower side of three weeks old healthy, young, and fully expanded leaves directly. After infiltration, plants were placed in control conditions of 16 h light photoperiod at 25 °C to 28 °C in an insect-free chamber.

Generating a bait clone

The bait vector was constructed for yeast-two-hybrid analysis, primers designed having restriction sites NcoI and BamHI (F: 5′AA CC ATG G AA ATG ACA ACG AGC GGA ACA and R: 5′CC G GAT CC T TAA ACG GTG AAC TTT TTA). The primers were designed to avoid any codon change and frame-shift mutation. βC1 was amplified using these primers. The conditions for PCR using Phusion High-Fidelity DNA polymerase (New England Biolabs, Ipswich, MA, USA) with the following parameters: 98 °C pre-denaturation temperature for 30 s, denaturation at 98 °C for 10 s: annealing at 55 °C for 30 s: extension at 72 °C for 60 s following 35 cycles. A final extension was 72 °C for 7 min and hold temperature 25 °C. The amplified DNA fragment of βC1 was cloned into the vector pGBKT7 (Clontech Laboratories Inc., Mountain View, CA, USA), It has a GAL4 DNA-binding domain, Kan^r for Escherichia coli selection, and TRP1 for yeast selection. As a result, the gene encoding for βC1 was included in the pGBKT7 vector, which was created as bait for a GAL4-based two-hybrid system and given the name as pGBKT7-βC1.Y2H Gold cells were transformed with 100ng of pGBKT7-βC1 construct along with positive and negative control, whereas, pGBKT7-53 was a positive control and pGBKT7-lam was used as the negative control. Preparation of master mix by using denatured salmon sperm DNA (ssDNA) at 100 °C for 5 min, 50% PEG and 1M Lithium Acetate (LiAC) was done. The mixture was heated at 30 °C for 30 min, 42 °C for 20 min, and cool the tubes on ice for 10 min. 100 µL of transformants was spread on a single SD/-Trp plate while spreading of positive and negative was done on SD/-Leu/-Trp plates. Bait autoactivation and toxicity test were done by transforming pGBKT7-βC1 construct into Y2H Gold cells along with positive and negative controls.

Screening of βC1 interacting partner

The N. benthamiana cDNA library was generated with an activation domain (AD) having GAL4-AD (Clontech Laboratories Inc., Mountain View, CA, USA). It was generated at Plant Virology Lab at The University of Arizona, USA. Concentrated bait culture of pGBKT7-βC1 working as binding domains (BD) was prepared and combined with library vial AD, after mating the culture was spread on an array of selective agar plates. Initially on 150 mm double dropout SD/–Leu/–Trp (DDO), quarter dropout SD/-Ade/His/Leu/-Trp QDO/X-α-Gal agar plates following incubation at 30 °C for 3–5 days.

Mating efficiency was determined by the following formula: ${\displaystyle \text{Mating efficiency}\left(\text{Diploids \%}\right)=\frac{\text{Number of cfu}/\text{ml of diploids}}{\text{Number of cfu}/\text{ml of limiting partner}}\times 100}$ whereas, the number of colony forming units per mL (cfu/mL) of diploid indicated the diploid viability that was equal to number of cfu/mL on SD/–Leu/–Trp, number of cfu/mL of limiting partner was estimated as viability of the prey library that was equal to the number of cfu/mL on SD/–Leu, while the viability of bait was equal to the number of cfu/mL on SD/–Trp.

Confirmation of positive interaction and rescuing prey plasmid

The true positive prey plasmids were rescued from a single blue colony selected from Y2H experiment and re-streaked on DDO/X- α-Gal agar plates (Table 1). The cultures were inoculated overnight in DDO medium at 30 °C. The plasmids were isolated from yeast cells and used for transformation into E. coli DH5α cells. Plasmids showed positive interaction was further evaluated and confirmed by transformation and co-transformation.

Prey plasmid PCR for sequencing

Ten samples showing genuine positive interaction were used for sequence analysis. Plasmids isolated from yeast cells were used for PCR (Table 2). The PCR amplicons were analyzed on 1% Agarose gel. The positive clones were confirmed by amplification with SMART III primer and CDS III primer from both directions (Table 2).

Sequence homology, alignment, and evolution analyses

NTF2 gene sequence and protein sequences were subjected to the Basic Local alignment search tool for nucleotides (BLASTN) for the identification of their close relatives. Multiple sequence alignments were performed to construct the neighbor-joining phylogenetic tree using MEGA7 (Kumar, Stecher & Tamura, 2016) and sequence pairwise identity was analyzed by the Clustal W algorithm in MegAlign application Lasergene package of sequence analysis software (DNA Star Inc., Madison, WI, USA).

Protein structure prediction of viral proteins

The protein structure of NTF2 was retrieved from the National Center of Biotechnology Information (NCBI) database (XP_019236968). Evaluation of primary protein structure including the amino acid composition, atomic composition, molecular weight, theoretical isoelectric point (pI), estimated half-life, extinction coefficient, instability index, aliphatic index, and grand average of hydropathicity (GRAVY) was determined by using ProtParam (https://web.expasy.org/protparam/) (Gasteiger et al., 2005). PSIPRED tools were used for secondary protein structure analysis (http://bioinf.cs.ucl.ac.uk/psipred/). As Protein Data Bank (PDB) did not have a three-dimensional (3D) structure of the NTF2 protein, therefore, I-TASSER online server (https://zhanglab.ccmb.med.umich.edu/I-TASSER/) (Zhang, 2008) was used for 3D structure prediction. Then, the model was further refined by using ModRefiner (https://bio.tools/modrefiner). After 3D model prediction and refinement of the NTF2, further structural evaluation, validation, and stereo-chemical analysis were performed. Evaluation of backbone confirmation was done by examining the ψ/φ Ramachandran plot acquired from RAMPAGE (https://www.ccp4.ac.uk/html/rampage.html) (Lovell et al., 2003). The 3D model was assessed by Verify-3D (https://www.doe-mbi.ucla.edu/verify3d/#: :text=Determines%20the%20compatibility%20of%20an,the%20results%20to%20good%20structures) to determine three-dimensional protein profiles and compatibility of the atomic model with its own amino acid sequence in comparison to the results of good structure (Eisenberg, Lüthy & Bowie, 1997).

Protein-protein interaction (PPI) evaluation

A web-based server ClusPro 2.0 (https://cluspro.bu.edu), was used to perform Protein-Protein interaction analysis. It performs rigid-body docking of two proteins by giving a basic interface and gives the lowest energy of an interacting complex. ClusPro 2.0 required only two pdb files, receptor, and ligand (Comeau et al., 2004). βC1 model was used as a receptor, while predicted NTF2 was used as a ligand. Furthermore, PyMOL v1.7.4 software and Chimera (DeLano, 2002) were used to visualize and analyze the interactions between receptor and ligand complexes, thus obtained.

Transient expression of 35S-βC1 from CLCuMB strain in model host plant Nicotiana benthamiana

Non-transgenic N. benthamiana plants were inoculated with Agrobacterium cultures harboring expression cassette of βC1 gene from CLCuMB strain and analyzed the infectivity and symptoms induction ability of 35S-βC1 gene expression cassette. Inoculated leaves sample developed severe phenotypic symptoms of virus infection at 14 days post-inoculation (dpi). The health plants (Fig. 1A) and the plants inoculated with empty vector (Fig. 1B) were compared to the agroinfiltrated plants (Fig. 1C). These symptoms were downward leaf curling and yellowing of inoculated leaves area were observed on the infiltrated leaves (Fig. 1C). With the passage of time, the symptoms became prominent and deeply intense. In all of the four inoculated N. benthamiana plants, the leaf showed these typical symptoms at 21–25 dpi when compared to control leaves infiltrated with empty vector pGreen-0029 and healthy plants (Figs. 1A and 1B). The appearance of infectious symptoms in all the nurture leaves confirmed the working efficiency of the 35S-βC1 construct in the model plants and justified the major role of the βC1 gene in diseased symptoms induction in infected plants.

Yeast two hybrid assay using βC1 as bait and N. benthamiana proteins as prey

Based on the available virus genome sequence information (LN886545.1), primers for the βC1 gene were designed. βC1 was cloned into the pGBKT7 vector and confirmed by sequencing showing no point mutation. The results showed 100% identity with the βC1 gene encoded by CLCuMB. Transformation for confirmed clones was done into Y2H Gold competent cells.

A preliminarily study for the autoactivation of βC1 was conducted to investigate the interaction among bait and prey fusion proteins. Yeast cells expressing the reporter genes HIS3 and His3 auxotrophic on histidine deficient medium were employed for verification. No growth was observed in yeast cells when pGBKT7-βC1 construct was co-expressed with pGADT7-empty in the histidine-deficient medium. In contrast, positive control showed blue colonies indicating that there was a positive interaction between 53+t, similarly, negative control lam+t showed no interaction at all (Fig. 2).

For mating, the screening of more than 2 × 10⁷ cells per mL was confirmed by library tittering. Mating efficiency was 5.6% which was sufficient to screen the library successfully (Table 3). Positively interacted proteins showing blue colonies from 150 mm DDO/ X-α-Gal plates were further screened and purified by patching on DDO/ X-α-Gal plates and QDO/ X-α-Gal agar plates (Fig. 3). Positive interacted protein’s plasmids were rescued by transformation into Dh5α of E. coli.

Validation of positive interaction through transformation and co-transformation

After being propagated in E. coli, the extracted plasmids were again transformed into yeast and co-expressed with the pGBKT7-βC1 construct to examine the reproducibility of the candidate proteins in retransformed yeast cells. On SD/Leu-/Trp-/His-plates, yeast transformants were dispersed, and the survival rate was used to assess the interaction’s efficacy. Positive clones were further verified to have strong positive interaction by spreading onto SD/Leu–/Trp–/His–/Ade–plates. As a final point, ten positive clones with strong βC1 interacting proteins were identified.

Identification of in vivoβC1 interacting proteins using Y2H

Ten candidate protein plasmids showing genuine positive interaction were sequenced for further analyses (Table S1). The sequencing data was validated for true interaction by homology study, sequence alignment, and open reading frames (ORFs) detection. The sequence homolog was searched in NCBI by the BLASTN tool. All the identified sequences were translated, a significant chunk of sequences was revealed, and ORFs were extracted and aligned. Finally, the similarity index was estimated by the BLAST tool of NCBI. Among the identified sequences, the LOC109217199 mRNA of Nicotiana tabacum nuclear transport factor 2-like occurred repeatedly. Therefore, the sequences encoding for Nuclear Transport Factor 2 occurring more than once indicated the best interacting partner of β C1 protein.

Phylogenetic analysis of NTF2 encoded genes and encoded proteins

Gene sequences of NTF2 from N. benthamiana and other plant species were aligned to find its close relatives. Twenty closely related sequences were retrieved from databases with the help of BLASTN. In phylogenetic analysis of N. benthamiana NTF2 sequence showed the closest relationship with the members of its own genus Nicotiana tabacum and Nicotiana attenuata. Next closest clade of NTF2 gene sequences consisted of Solanum tuberosum (potato), Solanum lycopersicum (tomato) from the genus Solanum of the Solanaceae family (Fig. 4, Table S2).

In phylogenetic analysis, the studied NTF2 protein sequence also showed the closest relationship with same members of its own genus Nicotiana as revealed for gene sequence. Further, the NTF2 protein sequence was rooted to the Brassica rapa (turnip) and Arabidopsis family suggesting the wide pedigree of NTF2 from these species (Fig. 5, Table S3).

In silico prediction of viral protein structure

The NTF2 data retrieved from the online server revealed 123 amino acid residues. Its primary structure showed a molecular weight of 13747.67 Dalton, and a 5.87 pI-value (Fig. 6A). The pI value was below 7, which indicated the negatively charged protein. Significant stability of protein was revealed by computed instability index of 39.37. The availability of methionine (M) at the N-terminus of the protein, and the negative large average of hydropathicity (−0.089) indicated that the protein was hydrophobic in nature. The glutamine (Q), leucine (L), glycine (G), phenylalanine (F), serine (S), and valine (V) residues were observed in excess (Fig. 6A).

Secondary structure analysis exposed that NTF2 had 20.32% helices and 29.26% extended strands (β -sheets) (Fig. 6B). The predicted 3D model of NTF2 (Fig. 6C) was also evaluated through the Ramachandran plot. The 96.7% of the residues were found in the favored region, 2.5% of the residues were in the allowed region and only 0.8% residues were found in the outlier region (Fig. 6D). I-TASSER was used to predict the 3D structure of NTF2 and the predicted structure was refined by ModRefiner. Having more than 99% residues of the predicted structure in the favored and allowed regions, the structure was considered to be validated (Fig. 6E).

Estimation of protein-protein interaction (PPI)

To study the PPI of βC1 and NTF2, βC1 protein was used as a receptor in cluspro while NTF2 was used as the ligand (Fig. 6F). Interaction studies of βC1 with NTF2 showed a strong binding affinity with releasing energy value of −730.6 kJ/mol (Table 4). Analysis of interacting residues showed that 10 amino acids from the middle portion towards the C-terminus and five amino acid residues (TYR48, TYR50, ASP52, ASP58, ASN60) from the middle portion of βC1 interacted with six amino acids of NTF2 (Table 4).