βC1, pathogenicity determinant encoded by Cotton leaf curl Multan betasatellite, interacts with calmodulin-like protein 11 (Gh-CML11) in Gossypium hirsutum

Abstract

Begomoviruses interfere with host plant machinery to evade host defense mechanism by interacting with plant proteins. In the old world, this group of viruses are usually associated with betasatellite that induces severe disease symptoms by encoding a protein, βC1, which is a pathogenicity determinant. Here, we show that βC1 encoded by Cotton leaf curl Multan betasatellite (CLCuMB) requires Gossypium hirsutum calmodulin-like protein 11 (Gh-CML11) to infect cotton. First, we used the in silico approach to predict the interaction of CLCuMB-βC1 with Gh-CML11. A number of sequence- and structure-based in-silico interaction prediction techniques suggested a strong putative binding of CLCuMB-βC1 with Gh-CML11 in a Ca+2-dependent manner. In-silico interaction prediction was then confirmed by three different experimental approaches: The Gh-CML11 interaction was confirmed using CLCuMB-βC1 in a yeast two hybrid system and pull down assay. These results were further validated using bimolecular fluorescence complementation system showing the interaction in cytoplasmic veins of Nicotiana benthamiana. Bioinformatics and molecular studies suggested that CLCuMB-βC1 induces the overexpression of Gh-CML11 protein and ultimately provides calcium as a nutrient source for virus movement and transmission. This is the first comprehensive study on the interaction between CLCuMB-βC1 and Gh-CML11 proteins which provided insights into our understating of the role of βC1 in cotton leaf curl disease.

About the Authors:

Hira Kamal

Roles Formal analysis, Investigation, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

Affiliations National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan, Pakistan Institute of Engineering and Applied Sciences (PIEAS), Nilore, Islamabad, Pakistan, Department of Plant Pathology, Washington State University, Pullman, WA, United States of America

Fayyaz-ul-Amir Afsar Minhas

Roles Conceptualization, Formal analysis, Funding acquisition, Methodology, Software, Writing – original draft, Writing – review & editing

Affiliation: Pakistan Institute of Engineering and Applied Sciences (PIEAS), Nilore, Islamabad, Pakistan

Diwaker Tripathi

Roles Investigation, Methodology

Affiliation: Department of Biology, University of Washington, Seattle, WA, United States of America

Wajid Arshad Abbasi

Roles Formal analysis, Methodology, Software

Current address: Department of Computer and Information Sciences, University of AJ&K, Muzaffarabad

Affiliation: Pakistan Institute of Engineering and Applied Sciences (PIEAS), Nilore, Islamabad, Pakistan

ORCID logo http://orcid.org/0000-0001-7691-5715

Muhammad Hamza

Roles Formal analysis, Investigation, Validation

Affiliation: National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan

Roma Mustafa

Roles Formal analysis, Investigation, Methodology

Affiliation: National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan

Muhammad Zuhaib Khan

Roles Formal analysis, Investigation, Methodology

Affiliation: National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan

Shahid Mansoor

Roles Conceptualization, Methodology, Project administration, Resources, Supervision, Writing – review & editing

Affiliation: National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan

Hanu R. Pappu

Roles Investigation, Methodology, Resources, Supervision, Writing – original draft, Writing – review & editing

Affiliation: Department of Plant Pathology, Washington State University, Pullman, WA, United States of America

Imran Amin

Roles Conceptualization, Formal analysis, Funding acquisition, Project administration, Supervision, Writing – original draft, Writing – review & editing

* E-mail: [email protected]

Affiliation: National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan

ORCID logo http://orcid.org/0000-0003-3063-4103

Introduction

Family Geminiviridae is one of the largest groups of plant viruses. Members in this family possess circular, single-stranded (ss) DNA as their genome, and infect economically important crops such as cotton, cassava, grains, maize, and vegetables [1]. On the basis of genome organization, vector transmission, genome wide pairwise sequence identity and host range, this family is classified into nine genera that include Becurto-, Begomo-, Capula-, Curto-, Eragro-, Grablo-, Mastre-, Topocu- and Turncurto-virus. Begomovirus is the largest and economically important genus of this family [2]. Begomoviruses infect dicotyledonous plant species in tropical and subtropical regions and are found in both New World (NW) and Old World (OW). The genus represents the largest number of viruses possessing more than 320 species under the family Geminiviridae [2]. It possesses either monopartite (DNA-A) or bipartite genome (DNA-A and DNA-B) of approximately 2.6–2.8 kb size that are different from each other except the region of almost 200 nucleotides (common region; CR) present within the intergenic region (IR) [3]. However, genome organization in monopartite viruses possess all proteins that are sufficient for viral replication, encapsidation and transmission across the different hosts [4]. Most of the OW monopartite begomoviruses are associated with satellite DNAs referred to as betasatellite (genus Betasatellite, family Tolecusatellitidae) and alphasatellite (family Alphasatellitidae) [5–9]. Satellites are the small molecules with the size of ~1.4 kb, and are dependent on helper virus for their replication, virus packaging (encapsidation), intracellular movement and transmission. These satellite molecules associated with begomovirus were discovered with the identification of Tomato leaf curl virus (ToLCV) [3]. Betasatellites encode βC1 protein which is a pathogenicity determinant while the role of alphasatellite is not fully known [10].

Cotton (Gossypium hirsutum) is a major cash crop of Pakistan and contributes up to 60% of foreign exchange earnings [11]. In Punjab province, cotton is grown on approximately 2.5 million hectares [12]. Cotton leaf curl disease (CLCuD) is one of the major constraints to the production of cotton in the Indian sub-continent which is caused by several begomoviruses in association with a specific betasatellite known as Cotton leaf curl Multan betasatellite (CLCuMuB). A number of distinct begomovirus species were shown to be associated with the disease, the most important of which are Cotton leaf curl Multan virus (CLCuMuV) and Cotton leaf curl Kokhran virus-Burewala strain (CLCuKoV-Bur) [13]. These viruses are poorly infectious to cotton and require CLCuMuB to cause CLCuD [11]. βC1 is the only protein encoded by Betasatellite and was shown to be involved in the induction of disease symptoms [14, 15], viral DNA movement [16], and also acts as a RNA silencing suppressor (RSS; a plant defense mechanism which target transcripts of invading viruses and generated through dsRNA) [17, 18], forms homo multimeric complexes in plants [19], interferes with host plant gene expression and was also identified to interact with a number of host factors [20]. Among various host proteins, βC1 was found to interact with the calmodulin protein family which acts as a modulator of calcium signaling in plants. βC1 encoded by begomovirus Tomato yellow leaf curl China β-satellite (TYLCCNB) upregulates calmodulin-like (CML) protein in Nicotiana benthamiana (NbrsgCaM) resulting in repression of RNA-dependent RNA polymerase 6 (RDR6), that contributes to the antiviral response in plants [21]. Transcription activator protein (TrAP; C2) encoded by begomovirus Tomato golden mosaic virus (TGMV) also binds to and upregulates CML protein rgsCaM in Arabidopsis thaliana and N. benthamiana. This C2 interaction over expresses rgsCaM in the nucleus that results into higher susceptibility to the TGMV infection [22].

Calmodulin (CaM) and CML protein belongs to the calcium binding protein family in plants that possess EF-hand (helix-loop-helix) like structure. EF-hand is more likely spread thumb and forefinger in which two alpha helices are linked with a loop region that may attain bound and unbound state of calcium ion to modulate the intracellular Ca+2 concentration. This apo (bound) and holo (free) state of Calmodulin protein is required for energy metabolism, and plays a role in host defense system [23]. The sequence and structure of CaM and CML proteins vary in A. thaliana, rice, and tobacco especially from linker (middle part connecting N- and C- lobe) region in their structures. This connecting region produces more flexibility in CML as compare to CaM using different conformations in plants, which extends strong interaction with neighboring proteins [24]. Moreover, it has been observed that CML proteins have evolved earlier than CaM [25]. In previous reports, several pathogens such as bacteria, fungi and viruses interact with different members of CML protein in different cell compartments, indicating a strong relation between Ca+2 ion channels and pathogens in plasma membrane and the nucleus [26]. For example, CML43 and its orthologs in tomato enhances the production of avirulent strain of Pseudomonas syringae to cause bacterial infection in tomato, and CML9 functions mainly in response to avirulent pathogens [27, 28]. During fungal infection by Verticillium dahliae, calcium influx is also upregulated which disturbs calcium-based responsive genes in cotton. To counteract, the Gh-CML11 in G. hirsutum can also bind with transcription factor MYB108 which forms a positive loop and enhances tolerance against this fungal attack [29]. All these studies suggest that the CML protein has different mechanisms of action with different types of pathogens and many CML proteins are still in need to be characterized with respect to their structure and function [30, 31].

Limited reports are available on the role of CaM and CML proteins in plant virus infection. CML38 in A. thaliana binds to HC-Pro to enhance Potyvirus, Turnip mosaic virus (TuMV) replication. The same viral protein (HC-Pro) that acts as RSS from another Potyvirus Tobacco etch virus (TEV) associates directly with rgsCaM from tobacco, acting as an endogenous suppressor for HC-Pro [32]. Binding of rgsCaM with HC-Pro reduces its abundance in plant cells which can facilitate viral infection to spread [32]. Therefore, plant viruses deregulates Ca+2 channels using Ca+2 dependent protein such as CaM to aid virus replication and its transcription in a host [33]. All these studies showed diverse roles of CML proteins during pathogen attack.

Bioinformatic approaches provide useful tools to investigate the role of CaM and CML members in plant-pathogen interactions. These in-silico approaches rely on several sequential steps to investigate protein-protein interaction (PPI) such as sequence-based method that extracts information from unique sequence motifs to build protein secondary structure [34], binding sites detection methods associated with the sequence, and 3D structures that predict interfacial residues in close proximity [35]. Aa a final step, the interface prediction method retrieves the information from experimentally determined interacting residues using machine learning methods and utilizes trained models to identify interacting residues of a query protein [36]. This multi-pronged approach based on sequence conservation analysis, energetics and binding site can be used to predict the interaction between virus and host proteins such as geminivirus proteins and their interacting partners in cotton and furthermore, to localize the residues in the interacting domain that are responsible for their binding affinity.

Here, we extend our analyses on CML protein in cotton that binds with one of the known begomovirus-associated satellites, CLCuMB. We report that CLCuMB-encoded βC1 protein binds to Gh-CML11 in G. hirsutum for virus pathogenicity and symptom enhancement. We first used a suite of in-silico methods to predict the putative interaction and its strength between these two proteins. Moreover, CaM binding (CaMB) motifs in CLCuMB-βC1 suggested calcium dependent interaction with Gh-CML11. This interaction prediction was further validated by three independent experimental approaches: yeast two-hybrid (Y2H), pull-down assay and bimolecular fluorescence complementation (BiFC). Taken together, our findings provide a better insight into the cotton-begomovirus interactions underlying the CLCuD.

Materials and methods

Bioinformatics analysis of Gh-CML11 and CLCuMB-βC1

Various bioinformatics tools were used for predicting the interaction between Gh-CML11 and CLCuMB-βC1 proteins, their binding affinity and binding sites. We first predicted the structures of Gh-CML11 and CLCuMB-βC1 using I-TASSER [37] and subsequently used as input to structure-based bioinformatics methods. CaMELS [38], PPA-Pred [39], PRISM [40] and PRODIGY [41] were used for interaction and binding affinity prediction. Meanwhile, CaM/CML binding motifs were identified in CLCuMB-βC1 sequence using reported data [42], CaM binding (CaMB) database including Calmodulin Target Database [43], Calmodulation database and meta-analysis predictor [44]. High binding affinity and CaMB motifs proposed a strong interaction between these two proteins. Further bioinformatics analysis to identify the binding site(s) at residue level was carried out using BSpred [45], NSPHOMPPI [46], PSIVER [47], PredictProtein [48], PPiPP [49], CPORT [50], ZDOCK [51], Docking2 at ROSETTA [52] and PAIRpred [53]. All these methods predicted different binding regions and we combined the results using an ensemble heuristic based on majority voting. Myristylation signal was identified in Gh-CML11 using Myristylator [54] and NMT server [55].

Plant lines and sample processing

For in-planta interaction study, wild type and transgenic lines of N. benthamiana seeds with the CFP-H2B marker were grown in pots containing Sunshine Mix LC1 in a greenhouse with 16 h light/8 h dark cycle. A confirmed clone for CLCuMB (AM774307), was used for amplification of βC1. CML protein was isolated from a CLCuD resistant cotton variety, UA222. RNA was obtained from cotton leaves using RNeasy Plant Mini Kit (Qiagen, Netherlands) following the manufacturer’s instructions. Purified RNA was then reverse transcribed to generate cDNA using a RevertAid first strand cDNA synthesis kit (Thermo Scientific, USA) with an oligo (dT) primer. RT-PCR was performed to amplify CLCuMB-βC1 and Gh-CML11 genes using Phusion high-fidelity DNA polymerase (Thermo Scientific, USA) and attB sequence-flanked specific primers, listed in S1 Table. For PCR amplification only 1–2 μl cDNA (50–150 ng/ μl) was used to a master-mix of total 50 μl reaction. It contained 1M MgSO4 (2 mM final conc.), F+R primers (0.5 μM each), dNTPs (0.15 Mm final conc.), 10X PCR buffer, pfx polymerase (1 U) and water to make up the volume upto 50 μl. The PCR profile consist of 1 initial denaturation cycle was set at 95 oC for 1 minute, followed by 35 cycles of denaturation step at 95 oC for 30 sec, annealing step at 58–60 oC for 30 sec and extension step at 72 oC for 45 sec (according to template size and polymerase efficiency). Then final extension was set at 72 oC for 5 minutes. All amplified products were separately cloned into pENTR-D-TOPO vector (Invitrogen, USA) and positive clones were confirmed by sequencing using M13 forward and reverse primers, and gene-specific primers.

Yeast two hybrid (Y2H) and Pull down assays

For Y2H, entry clones from Gateway® cloning system were recombined into the binary vector pEZY202 (Addgene) and pEZY45 (Addgene) [56]. Lithium acetate yeast transformation procedure as described [57, 58] was followed using yeast (Saccharomyces cerevisiae) strain EGY48 possessing pSH18-34 marker. Screening for positive interaction was performed on minimal SD Base/Gal/Raf with triple dropout medium–His/-Trp/-Ura (TDO) and Quadruple dropout medium -His/-Leu/-Trp/-Ura (QDO). Background was eliminated using 3-Amino-1,2,4-triazole (3-AT) ranging in 10-40mM serial dilution for positive interaction only between bait and pair.

The maltose binding protein (MBP) pull down assay was performed following the protocol as described previously [59]. Gateway entry clones were recombined into pull down destination vectors pMAL-c2X and pDEST15 (Invitrogen, USA). Protein samples expressed in E.coli BL21 (DE3) strain were purified by sonication and purified products were eluted after two to three washes. For Western blotting, samples were separated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and incubated with monoclonal anti-GST antibody (primary) which was subsequently probed with secondary antibody goat HRP-conjugated anti-rabbit IgG (Bio-Rad, USA). Positive signals were developed on x-ray film using ECL method based on Versa Doc Imaging System (Bio-Rad, USA) following the manufacturer’s instructions.

Bimolecular fluorescence complementation (BiFC) study

Genes were cloned from entry vector into BiFC binary vectors pSITE-cEYFP-C1+pSITE-nEYFP-C1 (ABRC; Ohio) using LR ClonaseTM enzyme mix. Agrobacterium transformation, infiltration and confocal microscopy were done by following the protocol as described [60]. Agroinfiltrated leaves containing host and viral proteins were studied after 24-48hrs incubation under a confocal microscope. Transgenic lines of N. benthamiana possessing CFP-H2B marker [61], was used to inspect under YFP and CFP fluorescent markers. A minimum of three leaves were evaluated, and images were acquired using TCS SP8 X microscope (Leica Germany) at 20X dry, 40X dry and 63X oil for fine detail images and LAS X software was used to analyze the data.

Results

In silico Gh-CML11 and CLCuMB-βC1 interaction prediction

Sequence and structure analysis were performed before identifying CLCuMB-βC1 interaction with CML11 based on molecular techniques. Previous studies have shown and rgsCaM from N. benthamiana and CML38 from A. thaliana, respectively interacts with geminiviruses (TGMV, TYLCCNV) and potyvirus (TuMV) as described [21, 22, 62]. In the current study, sequence analyses showed that Gh-CML11 protein shares only 24% identity with CaM protein in NbrsgCaM and 26% identity with AtCML38 S2 Table. However, structure alignment shows rgsCaM from N. benthamiana forms a long linker region in calcium-dependent manner and shares structure identity with Gh-CML11 under RMSD of 1.68 Å, while in calcium-independent manner shares 5.32 Å RMSD (Fig 1A). Low RMSD indicates high structure identity and vice versa. In case of AtCML38 and GhCML11, both protein complexes possess low structural identity during calcium-dependent manner showing RMSD of 6.78 Å and 1.24 Å RMSD during calcium-independent manner (Fig 1B). AtCML38 consist of 4 α-helices at N-lobe, forming 3 EF-hands connected with C-lobe through a long α-helix linker part. Gh-CML11 consists of 2 EF-hands present on each domain connected with a middle linker region (Fig 1B). This data represents CML protein in G. hirsutum possess variations with other CaM/CML protein members in N. benthamiana and A. thaliana at sequence and structure level. In geminiviruses, TYLCCNB was shown to interact with CML protein in N. benthamiana, but TYLCCNB-βC1 shares only 30% identity with CLCuMB-βC1 S2 Table. This sequence to structure analysis data suggested that CLCuMB-βC1 may or may not have the binding ability with Gh-CML11 protein which is in need to be investigated independently for CLCuMV role in the presence of calmodulin like protein.

[Figure omitted. See PDF.]

Fig 1. Structural comparison of Gh-CML11 with other CML members in Nicotiana benthamiana and Arabidopsis thaliana.

Calcium dependent and calcium independent structural variations for Gh-CML11 protein with A. NbrsgCaM and B. AtCML38.

https://doi.org/10.1371/journal.pone.0225876.g001

As Protein Data Bank (PDB) structures for both virus and host proteins were not available online, 3D structures for both of them were predicted using I-TASSER [37]. I-TASSER predicted five models for both proteins and the best model was chosen based on a higher value of C-score (confidence score). C-score for Gh-CML11 ranged from -0.27 to -2.58 and CLCuMB-βC1 had a -1.68 to 3.89 score. The model with the lowest negative value shows more stable structure and was selected for further study. Moreover, to determine the potential interaction, a number of methods were used with respect to binding affinity where ΔG (Gibbs free energy) represents binding energy or equilibrium constant. For example, when protein A binds with protein B in its neighbor forms a complex AB due to strong relationship. It will possess higher negative value (Gibbs free energy, ΔG<0) due to less conformational changes. Similarly, due to strong binding, this AB complex possess high equilibrium constant in terms of change in Gibbs free energy (ΔΔG) [ΔΔGAB = ΔGAB-(ΔGA+ΔGB)]. In simple notion, ΔΔG determines either interaction is more stable (ΔG<0) or less stable (ΔG>0). Initially interaction was determined with CaMELS [38], a sequence-based method that can be used to predict the propensity of interaction of a protein with CaM and CML proteins as well as identification of the CaM or CML binding sites on these proteins. It produces a score of 5.04 for interaction of CLCuMB-βC1 S3 Table. This high positive value corresponds to a higher binding affinity and a possible interaction between these proteins. This score was further cross checked with another sequence-based computational method PPA-pred [39] which directly predicts binding affinity between two proteins in terms of change in Gibbs free energy (ΔΔG). Based on functional relevance of a protein complex, the PPA-Pred predicted value of ΔΔG for the possible interaction between CLCuMB-βC1 and Gh-CML11 was -11.65 kcal/mol that also indicated a high probability of interaction. Though, sequence-based methods possess some limitations such as these methods are not able to assess energy for a complex during different modes of three dimensional structures of both interacting proteins. For this reason, structure-based approach PRISM [40] and PRODIGY [41] were used to predict the binding affinity of these proteins. The PRISM determined ΔΔG value was -18.32 kcal/mol and PRODIGY detected -41.2 kcal/mol between CLCuMB-βC1 and Gh-CML11 S3 Table, showing a strong binding energy between them. Using sequence and structure-based methods, a high ΔΔG value indicated a higher possibility of interaction between CLCuMB-βC1 and Gh-CML11.

Short linear amino acid motifs termed as CaM binding motifs (CaMBM) are useful in identifying all possible CaM-binding proteins in plants during biotic and abiotic stresses. The CaMB motifs in CLCuMB-βC1 were also another source of interaction prediction between these two proteins. Using databases such as Calmodulin Target database, Calmodulation database and meta-analysis predictor [42–44], almost 66 CaMB motifs either in canonical or noncanonical form were identified in CLCuMB-βC1 protein (Table 1). It has been observed that these motifs are present in overlapping form throughout the whole sequence and structure (Fig 2A). Motif 1-5-8-14 at position 66–79 in CLCuMB-βC1 is involved in α-helix formation shown in purple Fig 2B, and deletion of this motif reduces the predicted ΔΔG value indicating its role with Gh-CML11 among all reported hydrophobic CaMB motifs (Fig 2C). Due to these extensive motifs in CLCuMB-βC1 sequence, it is possible that it can bind with CML protein and a majority of these canonical CaMB forms indicated that CLCuMB-βC1 potentially modulates binding with Gh-CML11 protein in a Ca+2-dependent manner.

[Figure omitted. See PDF.]

Fig 2. Identification of CaMB motifs in CLCuMB-βC1 sequence and structure.

A. Selected CaM binding motifs are highlighted in CLCuMB-βC1 protein showing potential binding capacity with Gh-CML11. B. Deletion of CaMB motif 1-5-8-14 reduces ΔΔG value suggesting its role in binding at positions 66–79 in CLCuMB-βC1 α-helix (highlighted with purple color). C. Most known canonical CaMB motifs listed here are present in CLCuMB-βC1, indicating its potential interaction with calmodulin like protein in Ca+2-dependent manner.

https://doi.org/10.1371/journal.pone.0225876.g002

[Figure omitted. See PDF.]

Table 1. Possible canonical CaM binding motifs in CLCuMB-βC1.

https://doi.org/10.1371/journal.pone.0225876.t001

Bioinformatics study for interface prediction

We used an ensemble of structure- and sequence-based bioinformatics methods for interface and binding site prediction to identify interacting regions in CLCuMB-βC1 and Gh-CML11 proteins. The individual and ensemble residue-level binding site prediction scores from the bioinformatics analysis are given in S4 Table. Using structural information, protein docking and machine learning methods has identified possible interface site within Gh-CML11 and CLCuMB-βC1 complex (Fig 3A and 3B).

[Figure omitted. See PDF.]

Fig 3. Interaction prediction using protein docking methods and machine learning method using 3D structures of CLCuMB-βC1 and Gh-CML11 protein.

A. Top ten models predicted by docking methods ZDOCK and DOCKING2 were aligned within 5 Å, highlighted and most common residues involved in binding site were selected. Red color shows binding residues in Gh-CML11 and blue color indicates interacting residues in CLCuMB-βC1. Black arrow represents a closer view of single model among ten models. B. Machine learning method PRISM itself evaluated interacting residues within 6 Å for interaction, and interacting amino acids from PRISM are highlighted in red and blue color. C. Another machine learning method PAIRPred retrieves data from both sequence and structure. Output result produces graphical illustration of interacting residues between Gh-CML11 and CLCuMB-βC1. Color from orange to red indicates interacting region in both proteins based on B-factor.

https://doi.org/10.1371/journal.pone.0225876.g003

From the cumulative score from S4 Table, it was found that CLCuMB-βC1 possesses two main binding regions: residues 60–75 and 103–106 that form the main α-helix and a loop-turn, respectively, as shown in (Fig 4A). For host protein, linker region (60–77 amino acid) and C-lobe residues (144–150) in Gh-CML11 possess high score for binding (Fig 4A). A known CaMB motif (1-5-8-14) in CLCuMB-βC1 was also a part of this predicted domain at location 66–79, while 103–106 is part of myristylation-like motif 103‘GMDVNE’108. It has already been validated that α-helix and myristylation-like motif in CLCuMB-βC1 are mainly involved in interaction with several host proteins [63]. All predicted regions for virus and host were also localized in their respective 3D structures (Fig 4B). However, deletion of these motifs in both proteins reduces ΔΔG value suggesting the involvement of these motifs in interaction with each other. Thus, computational biology including binding affinity and interaction prediction methods for CLCuMB-βC1 and Gh-CML11 complex provided another approach for exploring the interactions between virus and host proteins.

[Figure omitted. See PDF.]

Fig 4. Cumulative score for binding site between CLCuMB-βC1 and Gh-CML11 protein based on sequence and structure prediction methods.

A. Graphical illustration shows binding site results at residue level. CLCuMB-βC1 residues from 60–75 position in α-helix and residues at 103–106 position has high binding affinity for interaction. Consensus of all computational methods shows Gh-CML11 possess higher binding affinity in the central part (residues at position 60–77) and C-lobe (residues at position 144–150) respectively. B. Predicted residues are shown in both virus and host protein structures.

https://doi.org/10.1371/journal.pone.0225876.g004

Yeast two hybrid (Y2H) and Pull down assay for interaction determination

To verify the observed in silico interactions between virus and host proteins, Y2H was performed using entry clones amplified from Gateway® cloning vector pENTR-D-TOPO. Isolated Gh-CML11 gene was submitted to GenBank after sequence confirmation (MK097275). Confirmed host and virus genes were further expressed in the binary destination vector pEZY202 and pEZY45 as bait and prey, respectively. Gh-CML11 protein was fused to the DBD of bait vector and CLCuMB-βC1 was fused to the TAD of prey vector. After successful transformation of the bait construct in EGY48/pSH18-34 strain, a number of colonies appeared for Gh-CML11 protein on selection media (SD Base/Gal/Raf) supplemented with triple dropout nutrient (-His/-Trp/-Ura) as shown +L in Fig 5A1. Further, yeast cell lines harboring bait construct were co-transformed with prey possessing CLCuMB-βC1 protein. Here, colonies appeared only for CLCuMB-βC1-Gh-CML11 complex. A number of colonies were screened on selection media SD Base/Gal/Raf supplemented with -His/-Leu/-Trp/-Ura as shown in -L in Fig 5A1. Only positive interacting colonies had appeared due to the presence of LEU reporter gene in yeast EGY48 strain.

[Figure omitted. See PDF.]

Fig 5. Interaction between Gh-CML11 and CLCuMB-βC1 using yeast two hybrid system and pull down assay.

A. Positive colonies on yeast agar media [SD-Ura-His-Trp (+L) and SD-Ura-His-Trp-Leu (-L)] indicated interaction between Gh-CML11 and CLCuMB-βC1. Interaction was further enhanced with serial dilution of 3-AT that confirmed the in-silico results for Gh-CML11 and CLCuMB-βC1. Interaction of IYSV-TSWV coded N proteins were assessed as a positive control in y2h system. Gh-CML11 fused with empty prey vector was used as a negative control during autoactivation step. B. Gh-CML11 protein cloned in pMAL-c2X/MBP was purified with CLCuMB-βC1 protein present in pDEST15/GST-tagged. Purification steps includes load samples that represents crude extracts, wash sample shows removal of unbound proteins and elution samples represents band for positive interaction for Gh-CML11 and CLCuMB-βC1. IYSV-TSWV N proteins were used as a positive control and Gh-CML11 transformation with empty prey was used as a negative control.

https://doi.org/10.1371/journal.pone.0225876.g005

To determine the strength of interaction in this CLCuMB-βC1-Gh-CML11 complex, a higher concentration of 3-Amino-1,2,4-triazole (3-AT) was added which enhance growth on selection media to confrim the observed interaction [64]. With the serial dilution of 3-AT, several colonies were found on 0-10mM. While on 20-40mM concentration of 3-AT, a fewer number of colonies eliminating false positive results showing strong interaction between CLCuMB-βC1 and Gh-CML11 (Fig 5A2). Iris yellow spotted virus (IYSV) and Tomato spotted wilt virus (TSWV) encoded nucleoproteins (N) proteins were used as a positive control [60]. IYSV-N interacted strongly with TSWV-N protein in different 3-AT serial dilutions, which confirmed the transformation protocol (Fig 5A2). For negative control, Gh-CML11 was co-transformed with empty prey. Very few or no colonies were observed on selection media as well as in serial dilution of 3-AT. This in-vivo Y2H study provided a strong experimental evidence that CLCuMB-βC1 binds with Gh-CML11 during virus infection to modulate calcium source as an energy nutrient for helper virus during CLCuD in cotton.

Results from Y2H were further verified by using Maltose binding protein (MBP) pull down assay. Gateway® entry clones were fused into destination vectors. Gh-CML11 was recombined into pMAL-c2X and CLCuMB-βC1 was fused into pDEST15 vector. In-frame cloning was done in E.coli BL21 (DE3) strain which was purified using sonication method. Following the Western blotting after protein purification, the nitrocellulose membrane was washed with specific antibodies. Bands were visualized on X-ray film using ECL method. In case of elution sample possessing CLCuMB-βC1-Gh-CML11 complex, a single band was observed with no cross reactivity (Fig 5B). This clear band in elution sample was indicative of the physical interaction of CLCuMB-βC1 and Gh-CML11. MBP-N protein from IYSV and GST-N protein from TSWV were used as a positive control (Fig 5B). A brighter band in the elution sample confirmed the test protocol, while GST alone expressed with MBP-Gh-CML11 developed no band in the elution sample (Fig 5B). These results confirmed the findings from the Y2H assay and proved a strong physical interaction between CLCuMB-βC1 and Gh-CML11 proteins.

Localization study using bimolecular fluorescence complementation (BiFC) system

To determine protein interaction and its localization between CLCuMB-βC1 and Gh-CML11, in-planta BiFC assay was performed. Entry clones from pENTR-D-TOPO vector were recombined into binary vector pSITE I series. Gh-CML11 was fused into the N terminal of pSITE-nEYFP-C1 and CLCuMB-βC1 was cloned into the C terminal of pSITE-cEYFP-C1 under 35S promoter. For agroinfiltration, 4 to 5 leaves of N. benthamiana were infiltrated. After 48 hours, leaf sections were studied under confocal microscopy. We found that CLCuMB-βC1 strongly interacted with Gh-CML11 based on the observation of predominant interaction pattern throughout cytoplasmic veins under yellow fluorescent marker (Fig 6A). To assess the cellular localization, nuclear specific histone 2B fused with CFP marker was used. A clear interaction pattern was found to be distributed throughout nucleus and cytoplasm. At higher magnification (63X), strong signals for guard cells with CLCuMB-βC1 and Gh-CML11 interaction were observed (Fig 6B). The negative control harboring only Gh-CML11 at N terminal with the empty C fragment did not produce any non-specific signals. This test construct expressed very little or zero signal under both YFP and CFP marker (Fig 6C). These results confirmed that CLCuMB-βC1 interact Gh-CML11 interact in nucleus and cytoplasm to stimulate physiological responses for virus movement and transmission.

[Figure omitted. See PDF.]

Fig 6. Subcellular localization of Calmodulin like protein Gh-CML11 interaction with CLCuMB-βC1 protein.

Gh-CML11 with CLCuMB-βC1 were cloned into GATEWAY compatible BiFC vectors for interaction analysis. For protein localization, Nicotiana benthamiana leaves was used, carrying cyan fluorescent protein fused with nuclear specific histone marker 2B (CFP-H2B). A. Interaction between Gh-CML11 with CLCuMB-βC1 was localized in plasma membrane. B. Guard cells fluorescein with cyan color indicates localization in the cytoplasm. C. Negative control showing successful transformation protocol. Images were obtained 48h post-infiltration at 20x+1.5 zoom option. Scale bar = 50 μm.

https://doi.org/10.1371/journal.pone.0225876.g006

Discussion

CML proteins have been shown to play an important role in plant’s immune system against bacterial, fungal and viral pathogens. In plants, three main families of EF-hand Ca+2 namely CaM and CML proteins, calcineurin-B like (CBL) proteins and Ca+2-dependent protein kinase (CDPK) are known in physiological pathways based on Ca+2 sensors. Among all these members, the function(s) of CML proteins are not completely understood in plants, and they do not have much identity in structure with other EF-hand proteins. Recently, data from gene expression profiling and regulatory process in downstream signaling have revealed their function in plants against biotic and abiotic stresses [65]. Here in this study, the role of CML was identified via interaction with CLCuMB-βC1 encoded by a begomovirus associated betasatellite involved in CLCuD in cotton. Based on in-silico analysis, it was found that Gh-CML11 possesses a different EF-hand formation that makes it more appropriate for binding with CLCuMB-βC1 in a calcium-dependent manner.

The OW begomoviruses are usually associated with a DNA satellite referred to as betasatellite that encodes a single gene in complementary sense orientation i.e βC1 [12]. βC1 is a multifunctional protein which interferes with metabolic and signaling pathway acts as a suppressor of transcriptional gene silencing and post-transcriptional gene silencing [66, 67] and modulates miRNA level during plant development [68]. Viral proteins such as CLCuMV-βC1 can establish mild to severe infection in the host plants using host autophagy machinery. It has been studied that autophagy related genes such as ATG5 and ATG8 in N. benthamiana helps geminiviruses (CLCuMV, TYLCV and TYLCCNV) to enhance the infectivity in plants. Whereas, silencing of glyceralaldehyde-3-phosphate dehydrogenase (GAPC) gene activates the autophagy process that delayed virus symptoms and producing resistance against CLCuMV [69].

This mode of action mediated by βC1 and other viral protein participates in the accumulation of a high DNA level of helper begomovirus to cause infectivity [70, 71]. In a previous study, it was determined that TYLCCNB-βC1 upregulates NbrgsCaM to mediate symptoms modulation that establishes the successful TYLCCNV infection in Solanaceae hosts [21]. However, low sequence identity of virus coded TYLCCNB with CLCuMB and host coded protein rgsCaM with Gh-CML11, it was essential to determine the role of CML proteins in G. hirsutum during pathogen attack and to explore the mechanism of CLCuD infection in its natural host, cotton (G. hirsutum). A higher binding affinity and several canonical CaMB motifs overlapping at different location in CLCuMB-βC1 suggested a strong bonding with Gh-CML11 in nucleus and cytoplasm. CaM binding proteins possess IQ/IQ-like motifs in canonical and non-canonical form and interacts with CAM either in Ca+2-dependent or Ca+2-independent manner [44]. However, CLCuMB-βC1 possesses canonical forms of CaMB motifs which is an indication that CLCuMB-βC1 could bind Gh-CML11 in a Ca+2 dependent manner. Further studies are warranted to validate these results.

The sequence- and structure-based interaction prediction methods showed that the myristylation-like motif of CLCuMB-βC1 ‘GMDVNE’ has a strong binding affinity with Gh-CML11. This hydrophobic region was found to interact with tomato UBC3 protein, showing its importance in the ubiquitination pathway [63]. The role of CLCuMB-βC1 has already been studied in plant ubiquitination pathway. It binds with either SKP1 or CUL1 that impairs the SCF complex during ubiquitination. This impaired binding results into a retarded leaf growth and abnormal development with severe disease symptom [72]. While the host protein, Gh-CML11, possesses the myristylation signal present at 1-MGD-3 (M-G-x), showing the basic difference between calmodulin like protein and calmodulin protein, as true calmodulin proteins lack this myristylation signal. Moreover, calmodulin and calmodulin like protein constitute N- and C-lobes that are connected via Ca+2 based EF hands. It has been shown that binding of targeted partners with CML protein at C-lobe is more selective resulting in specific conformational changes in its structure leading to a strong binding [73]. In this study, structure based analysis identified residues that are present at C-terminal lobe showing higher binding affinity for βC1 protein. This computational study suggested that structural biology could predict binding sites between two proteins that offers an efficient and fast-track approach to gain initial insights into all possible interactions of proteins associated with cotton leaf curl disease complex in cotton.

The in-silico interaction prediction was verified and confirmed by using molecular techniques that validated the interaction between CLCuMB-βC1 and Gh-CML11. Previous work has shown the localization of TGMV-encoded C2 protein and rgsCaM in cytoplasm and nucleus using YFP and H4-RFP respectively. Notably, TGMV-C2 overexpresses rgsCaM in the nucleus to enhance virus DNA transcription that promote the infection [22]. Therefore, in response to CLCuD, CLCuMB-βC1 binds with Gh-CML11 in cytoplasm and nucleus and this localization was confirmed using YFP and H2B-CFP markers. This interaction indicates that CLCuMB-βC1 is also capable to overexpress Gh-CML11 to induce pathogenicity during CLCuD. Altogether, in-silico and molecular approaches have shown the interaction between CLCuMB-βC1 and Gh-CML11 in cotton. Studying biological function of calmodulin protein like (rgsCaM), it acts as endogenous suppressor of RNA silencing during geminivirus infection. It has been determined betasatellite protein βC1 associated with TYLCCNV over expresses the rgs-CaM protein which is required to suppress PTGS pathway and symptom enhancement [21]. This interaction reduces the level of secondary siRNA by repressing RNA dependent RNA polymerase 6 (RDR6) that weakens the plant antiviral response against geminivirus. Another CaM protein in Nicotiana benthamiana does not bind with RDR6 but degrades suppressor of gene silencing 3 (SGS3) in cytoplasm that promotes TYLCCNV. In response to this viral infection, plant also blocks this SGS3 degradation via autophagy inhibitor phosphatidylinositol 3-kinase (PI3K) based on autophagy pathway [74]. Thus, it can be hypothesized that Ca+2 ion flux across the cell membrane is being affected after the onset of CLCuMB-βC1 associated with CLCuMV that modulates interaction with CML proteins. This interaction may alter the response of antiviral genes such as RDR6 in G. hirsutum to enhance the infection. Additional work is required to prove this hypothesis.

Conclusions and future prospects

Here we have shown that bioinformatics study could reveal interaction information between a geminivirus and its host proteins, and also identify discrimination between CaM and CML protein such as Gh-CML11 proteins based on sequence and structure information. The interaction of CLCuMB-βC1 with Gh-CML11 appears to utilize Ca+2 ion channels and pumps to overcome host antiviral defense mechanism, leading to the successful infection and multiplication of CLCuD. This information can be used in future to devise novel strategies to develop resistance against CLCuD. Deletion and/or site-specific mutagenesis in different EF hands in Gh-CML11 would facilitate the identification of residue-based interaction prediction that are directly involved in binding with CLCuMB-βC1. This would establish a more reliable bioinformatics approach for interaction prediction in case of CaM and CML proteins with other plant viruses.

Supporting information

[Figure omitted. See PDF.]

S1 Table. List of primers used in this study.

https://doi.org/10.1371/journal.pone.0225876.s001

(XLSX)

S2 Table. Sequence alignment data of Gh-CML11 with NbrgsCaM and AtCML38.

https://doi.org/10.1371/journal.pone.0225876.s002

(TXT)

S3 Table. Values for binding energy between Gh-CML11 and CLCuMB-βC1.

https://doi.org/10.1371/journal.pone.0225876.s003

(XLSX)

S4 Table. Consensus score for Gh-CML11 and CLCuMB-βC1 to determine interaction using several computational methods.

https://doi.org/10.1371/journal.pone.0225876.s004

(XLSX)

Acknowledgments

Authors are thankful to Dr Michael M. Goodin, University of Kentucky Lexington, KY USA for providing N. benthamiana seeds expressing cyan fluorescent protein carrying nuclear marker histone 2B (CFP-H2B). The Franceschi Microscopy and Imaging Center (FMIC) at the Washington State University, Pullman USA was gratefully acknowledged.

Citation: Kamal H, Minhas F-u-AA, Tripathi D, Abbasi WA, Hamza M, Mustafa R, et al. (2019) βC1, pathogenicity determinant encoded by Cotton leaf curl Multan betasatellite, interacts with calmodulin-like protein 11 (Gh-CML11) in Gossypium hirsutum. PLoS ONE 14(12): e0225876. https://doi.org/10.1371/journal.pone.0225876

References

1. Mansoor S, Briddon RW, Zafar Y, Stanley J. Geminivirus disease complexes: an emerging threat. Trends Plant Sci. 2003;8(3):128–34. pmid:12663223

2. Zerbini FM, Briddon RW, Idris A, Martin DP, Moriones E, Navas-Castillo J, et al. ICTV virus taxonomy profile: geminiviridae. J Gen Virol. 2017;98(2):131–3. pmid:28284245

3. Dry I, Krake LR, Rigden JE, Rezaian MA. A novel subviral agent associated with a geminivirus: the first report of a DNA satellite. Proc Natl Acad Sci U S A. 1997;94(13):7088–93. pmid:9192696

4. Navot N, Pichersky E, Zeidan M, Zamir D, Czosnek H. Tomato yellow leaf curl virus: a whitefly-transmitted geminivirus with a single genomic component. Virology. 1991;185(1):151–61. pmid:1926771

5. Willment JA, Martin DP, Rybicki EP. Analysis of the diversity of African streak mastreviruses using PCR-generated RFLPs and partial sequence data. J Virol Methods. 2001;93:75–87. pmid:11311346

6. Briddon RW, Lunness P, Chamberlin LCL, Pinner MS, Brundish H, Markham PG. The nucleotide sequence of an infectious insect-transmissible clone of the geminivirus Panicum streak virus. J Gen Virol. 1992;73(5):1041–7. pmid:1588314

7. Varsani A, Navas-Castillo J, Moriones E, Hernández-Zepeda C, Idris A, Brown J, et al. Establishment of three new genera in the family Geminiviridae: Becurtovirus, Eragrovirus and Turncurtovirus. Arch Virol. 2014;159(8):2193–203. pmid:24658781

8. Varsani A, Roumagnac P, Fuchs M, Navas-Castillo J, Moriones E, Idris A, et al. Capulavirus and Grablovirus: two new genera in the family Geminiviridae. Arch Virol. 2017;162(6):1819–31. pmid:28213872

9. Briddon RW, Martin DP, Roumagnac P, Navas-Castillo J, Fiallo-Olivé E, Moriones E, et al. Alphasatellitidae: A new family with two subfamilies for the classification of geminivirus-and nanovirus-associated alphasatellites. Arch Virol. 2018:1–14. PMID: 29740680

10. Fiallo-Olivé E, Martínez-Zubiaur Y, Moriones E, Navas-Castillo J. A novel class of DNA satellites associated with New World begomoviruses. Virology. 2012;426(1):1–6. Epub 2012/02/15. pmid:22330203

11. Mansoor S, Amin I, Iram S, Hussain M, Zafar Y, Malik K, et al. Breakdown of resistance in cotton to cotton leaf curl disease in Pakistan. Plant Pathol. 2003;52(6):784. https://doi.org/10.1111/j.1365-3059.2003.00893.x.

12. Briddon RW, Markham PG. Cotton leaf curl virus disease. Virus Res. 2000;71(1–2):151–9. pmid:11137169

13. Fauquet C, Stanley J. Geminivirus classification and nomenclature: progress and problems. Ann Appl Biol. 2003;142:165–89. https://doi.org/10.1111/j.1744-7348.2003.tb00241.x

14. Zubair M, Zaidi SS-e-A, Shakir S, Farooq M, Amin I, Scheffler JA, et al. Multiple begomoviruses found associated with cotton leaf curl disease in Pakistan in early 1990 are back in cultivated cotton. Sci Rep. 2017;7(1):680. pmid:28386113

15. Saunders K, Norman A, Gucciardo S, Stanley J. The DNA β satellite component associated with ageratum yellow vein disease encodes an essential pathogenicity protein (βC1). Virology. 2004;324(1):37–47. pmid:15183051

16. Cui X, Tao X, Xie Y, Fauquet CM, Zhou X. A DNA β associated with Tomato yellow leaf curl China virus is required for symptom induction. J Virol. 2004;78(24):13966–74. pmid:15564504

17. Akhter A, Qazi J, Saeed M, Mansoor S. A severe leaf curl disease on chilies in Pakistan is associated with multiple begomovirus components. Plant Dis. 2009;93(9):962. https://doi.org/10.1094/PDIS-93-9-0962B

18. Axtell MJ, Bartel DP. Antiquity of MicroRNAs and Their Targets in Land Plants. Plant Cell. 2005;17(6):1658–73. pmid:15849273

19. Abarshi MM, Mohammed IU, Jeremiah SC, Legg JP, Kumar PL, Hillocks RJ, et al. Multiplex RT-PCR assays for the simultaneous detection of both RNA and DNA viruses infecting cassava and the common occurrence of mixed infections by two cassava brown streak viruses in East Africa. J Virol Methods. 2011;179(1):176–84. pmid:22080852

20. Angel CA, Hsieh Y-C, Schoelz JE. Comparative analysis of the capacity of Tombusvirus P22 and P19 proteins to function as avirulence determinants in Nicotiana species. Mol Plant-Microbe Interact. 2011;24(1):91–9. pmid:20977306

21. Li F, Huang C, Li Z, Zhou X. Suppression of RNA silencing by a plant DNA virus satellite requires a host calmodulin-like protein to repress RDR6 expression. PLoS Pathog. 2014;10(2):e1003921. pmid:24516387

22. Chung HY, Lacatus G, Sunter G. Geminivirus AL2 protein induces expression of, and interacts with, a calmodulin-like gene, an endogenous regulator of gene silencing. Virology. 2014;460:108–18. pmid:25010276

23. Snedden WA, Fromm H. Calmodulin, calmodulin-related proteins and plant responses to the environment. Trends Plant Sci. 1998;3(8):299–304. https://doi.org/10.1016/S1360-1385(98)01284-9

24. Chigri F, Flosdorff S, Pilz S, Kölle E, Dolze E, Gietl C, et al. The Arabidopsis calmodulin-like proteins AtCML30 and AtCML3 are targeted to mitochondria and peroxisomes, respectively. Plant Mol Biol. 2012;78(3):211–22. pmid:22116655

25. Mohanta TK, Kumar P, Bae H. Genomics and evolutionary aspect of calcium signaling event in calmodulin and calmodulin-like proteins in plants. BMC Plant Biol. 2017;17(1):38. pmid:28158973

26. Endres MW, Gregory BD, Gao Z, Foreman AW, Mlotshwa S, Ge X, et al. Two plant viral suppressors of silencing require the ethylene-inducible host transcription factor RAV2 to block RNA silencing. PLoS Pathog. 2010;6(1):1000729. pmid:20084269

27. Chiasson D, Ekengren SK, Martin GB, Dobney SL, Snedden WA. Calmodulin-like proteins from Arabidopsis and tomato are involved in host defense against Pseudomonas syringae pv. tomato. Plant Mol Biol. 2005;58(6):887–97. pmid:16240180

28. McCormack E, Braam J. Calmodulins and related potential calcium sensors of Arabidopsis. New Phytol. 2003;159(3):585–98. PMID: 28596783

29. Cheng H-Q, Han L-B, Yang C-L, Wu X-M, Zhong N-Q, Wu J-H, et al. The cotton MYB108 forms a positive feedback regulation loop with GH-CML11 and participates in the defense response against Verticillium dahliae infection. J Exp Bot. 2016:erw016. pmid:26873979

30. McCormack E, Tsai Y-C, Braam J. Handling calcium signaling: arabidopsis CaMs and CMLs. Trends Plant Sci. 2005;10(8):383–9. pmid:16023399

31. Leba LJ, Cheval C, Ortiz‐Martín I, Ranty B, Beuzón CR, Galaud JP, et al. CML9, an Arabidopsis calmodulin‐like protein, contributes to plant innate immunity through a flagellin‐dependent signalling pathway. Plant J. 2012;71(6):976–89. pmid:22563930

32. Nakahara KS, Masuta C, Yamada S, Shimura H, Kashihara Y, Wada TS, et al. Tobacco calmodulin-like protein provides secondary defense by binding to and directing degradation of virus RNA silencing suppressors. Proc Natl Acad Sci U S A 2012;109(25):10113–8. pmid:22665793

33. Chen Y, Xue S, Zhou Y, Yang JJ. Calciomics: prediction and analysis of EF-hand calcium binding proteins by protein engineering. Sci China Chem. 2010;53(1):52–60. pmid:20802784

34. Ma D-C, Diao Y-B, Guo Y-Z, Li Y-Z, Zhang Y-Q, Wu J, et al. A novel method to predict protein-protein interactions based on the information of protein-protein interaction networks and protein sequence. Protein Pept Lett. 2011;18(9):906–11. pmid:21529343

35. Esmaielbeiki R, Krawczyk K, Knapp B, Nebel J-C, Deane CM. Progress and challenges in predicting protein interfaces. Brief Bioinform. 2015;17(1):117–31. pmid:25971595

36. Zhou H-X, Qin S. Interaction-site prediction for protein complexes: a critical assessment. Bioinformatics. 2007;23(17):2203–9. pmid:17586545

37. Zhang Y. I-TASSER server for protein 3D structure prediction. BMC Bioinformatics. 2008;9(1):40. pmid:18215316

38. Abbasi WA, Asif A, Andleeb S, Minhas FUAA. CaMELS: In silico prediction of calmodulin binding proteins and their binding sites. Proteins: Struct Funct Bioinf. 2017;85(9):1724–40. pmid:28598584

39. Yugandhar K, Gromiha MM. Protein-protein binding affinity prediction from amino acid sequence. Bioinformatics. 2014;30(24):3583–9. pmid:25172924

40. Baspinar A, Cukuroglu E, Nussinov R, Keskin O, Gursoy A. PRISM: a web server and repository for prediction of protein-protein interactions and modeling their 3D complexes. Nucleic Acids Res. 2014;42:285–9. pmid:24829450

41. Xue LC, Rodrigues JP, Kastritis PL, Bonvin AM, Vangone A. PRODIGY: a web server for predicting the binding affinity of protein–protein complexes. Bioinformatics. 2016;32(23):3676–8. pmid:27503228

42. Rhoads AR, Friedberg F. Sequence motifs for calmodulin recognition. FASEB J. 1997;11(5):331–40. pmid:9141499

43. Yap KL, Kim J, Truong K, Sherman M, Yuan T, Ikura M. Calmodulin target database. J struct Funct Genomics. 2000;1(1):8–14. pmid:12836676

44. Mruk K, Farley BM, Ritacco AW, Kobertz WR. Calmodulation meta-analysis: predicting calmodulin binding via canonical motif clustering. J Gen Physiol. 2014;144(1):105–14. pmid:24935744

45. Mukherjee S, Zhang Y. Protein-Protein Complex Structure Predictions by Multimeric Threading and Template Recombination. Structure. 2011;19(7):955–66. pmid:21742262

46. Xue LC, Dobbs D, Honavar V. HomPPI: a class of sequence homology based protein-protein interface prediction methods. BMC Bioinformatics. 2011;12(1):244. pmid:21682895

47. Murakami Y, Mizuguchi K. Applying the Naive Bayes classifier with kernel density estimation to the prediction of protein-protein interaction sites. Bioinformatics. 2010;26. pmid:20529890

48. Yachdav G, Kloppmann E, Kajan L, Hecht M, Goldberg T, Hamp T, et al. PredictProtein—an open resource for online prediction of protein structural and functional features. Nucleic Acids Res. 2014;42:337–43. pmid:24799431

49. Ahmad S, Mizuguchi K. Partner-aware prediction of interacting residues in protein-protein complexes from sequence data. PLoS ONE. 2011;6(12):29104. pmid:22194998

50. De Vries SJ, Bonvin AM. CPORT: a consensus interface predictor and its performance in prediction-driven docking with HADDOCK. PLoS ONE. 2011;6(3):17695. pmid:21464987

51. Pierce BG, Wiehe K, Hwang H, Kim BH, Vreven T, Weng Z. ZDOCK server: interactive docking prediction of protein-protein complexes and symmetric multimers. Bioinformatics. 2014;30(12):1771–3. pmid:24532726

52. Lyskov S, Chou FC, Conchuir SO, Der BS, Drew K, Kuroda D, et al. Serverification of molecular modeling applications: the Rosetta Online Server that Includes Everyone (ROSIE). PLoS ONE. 2013;8(5):e63906. Epub 2013/05/30. pmid:23717507

53. Minhas F, Geiss BJ, Ben-Hur A. PAIRpred: partner-specific prediction of interacting residues from sequence and structure. Proteins. 2014;82(7):1142–55. Epub 2013/11/19. pmid:24243399

54. Bologna G, Yvon C, Duvaud S, Veuthey AL. N‐Terminal myristoylation predictions by ensembles of neural networks. Proteomics. 2004;4(6):1626–32. pmid:15174132

55. Maurer-Stroh S, Eisenhaber B, Eisenhaber F. N-terminal N-myristoylation of proteins: prediction of substrate proteins from amino acid sequence1. J Mol Biol. 2002;317(4):541–57. pmid:11955008

56. Guo F, Chiang MY, Wang Y, Zhang YZ. An in vitro recombination method to convert restriction‐and ligation‐independent expression vectors. Biotechnol J. 2008;3(3):370–7. pmid:18064608

57. Gyuris J, Golemis E, Chertkov H, Brent R. Cdi1, a human G1 and S phase protein phosphatase that associates with Cdk2. Cell. 1993;75(4):791–803. pmid:8242750

58. Li Y, Leisner SM. Multiple domains within the Cauliflower mosaic virus gene VI product interact with the full-length protein. Mol Plant Microbe Interact. 2002;15(10):1050–7. pmid:12437303

59. Hapiak M, Li Y, Agama K, Swade S, Okenka G, Falk J, et al. Cauliflower mosaic virus gene VI product N-terminus contains regions involved in resistance-breakage, self-association and interactions with movement protein. Virus Res. 2008;138(1–2):119–29. pmid:18851998

60. Tripathi D, Raikhy G, Goodin MM, Dietzgen RG, Pappu HR. In vivo localization of iris yellow spot tospovirus (Bunyaviridae)-encoded proteins and identification of interacting regions of nucleocapsid and movement proteins. PLoS ONE. 2015;10(3):e0118973. pmid:25781476

61. Martin K, Kopperud K, Chakrabarty R, Banerjee R, Brooks R, Goodin MM. Transient expression in Nicotiana benthamiana fluorescent marker lines provides enhanced definition of protein localization, movement and interactions in planta. Plant J. 2009;59(1):150–62. pmid:19309457

62. Endres MW, Gregory BD, Gao Z, Foreman AW, Mlotshwa S, Ge X, et al. Two plant viral suppressors of silencing require the ethylene-inducible host transcription factor RAV2 to block RNA silencing. PLoS Pathog. 2010;6(1):e1000729. pmid:20084269

63. Eini O, Dogra S, Selth LA, Dry IB, Randles JW, Rezaian MA. Interaction with a host ubiquitin-conjugating enzyme is required for the pathogenicity of a geminiviral DNA b satellite. Mol Plant Microbe Interact. 2009;22(6):737–46. pmid:19445598

64. Ursic D, Chinchilla K, Finkel JS, Culbertson MR. Multiple protein/protein and protein/RNA interactions suggest roles for yeast DNA/RNA helicase Sen1p in transcription, transcription‐coupled DNA repair and RNA processing. Nucleic Acids Res. 2004;32(8):2441–52. pmid:15121901

65. DeFalco TA, Bender KW, Snedden WA. Breaking the code: Ca2+ sensors in plant signalling. Biochem J. 2010;425(1):27–40. pmid:20001960

66. Brown JK, Zerbini FM, Navas-Castillo J, Moriones E, Ramos-Sobrinho R, Silva JC, et al. Revision of Begomovirus taxonomy based on pairwise sequence comparisons. Arch Virol. 2015;160(6):1593–619. pmid:25894478

67. Zhou X. Advances in understanding begomovirus satellites. Ann Rev Phytopathol. 2013;51(1):357–81. pmid:23915133

68. Amin I, Patil BL, Briddon RW, Mansoor S, Fauquet CM. A common set of developmental miRNAs are upregulated in Nicotiana benthamiana by diverse begomoviruses. Virol J. 2011;8(1):143. pmid:21447165

69. Haxim Y, Ismayil A, Jia Q, Wang Y, Zheng X, Chen T, et al. Autophagy functions as an antiviral mechanism against geminiviruses in plants. Elife. 2017;6:e23897. pmid:28244873

70. Saeed M, Zafar Y, Randles JW, Rezaian MA. A monopartite begomovirus-associated DNA β satellite substitutes for the DNA B of a bipartite begomovirus to permit systemic infection. J Gen Virol. 2007;88(10):2881–9. pmid:17872543

71. Cui X, Li G, Wang D, Hu D, Zhou X. A begomovirus DNAβ-encoded protein binds DNA, functions as a suppressor of RNA silencing, and targets the cell nucleus. J Virol. 2005;79(16):10764–75. pmid:16051868

72. Jia Q, Liu N, Xie K, Dai Y, Han S, Zhao X, et al. CLCuMuB βC1 subverts Ubiquitination by interacting with NbSKP1s to enhance geminivirus infection in Nicotiana benthamiana. PLoS Pathog. 2016;12(6):e1005668. pmid:27315204

73. Westerlund AM, Delemotte L. Effect of Ca2+ on the promiscuous target-protein binding of calmodulin. PLoS Comput Biol. 2018;14(4):1006072. pmid:29614072

74. Li F, Zhao N, Li Z, Xu X, Wang Y, Yang X, et al. A calmodulin-like protein suppresses RNA silencing and promotes geminivirus infection by degrading SGS3 via the autophagy pathway in Nicotiana benthamiana. PLoS Pathog. 2017;13(2):e1006213. pmid:28212430