Method for screening plant interaction protein by combining big data with artificial intelligence and verifying protein interaction by using yeast two-hybrid

Abstract

The invention belongs to the technical field of biology, and discloses a method for screening plant interaction protein by combining big data with artificial intelligence and verifying protein interaction by using yeast two-hybrid. According to the method, AI is combined with big data, and co-expression protein of family protein is obtained through big data screening and analysis; aI interaction prediction and molecular docking are carried out on the proteins, and interaction information is further analyzed and extracted. And further analyzing the condition of the screened interaction protein by using a yeast two-hybridization or fluorescence co-localization experiment. According to the method, aiming at the characteristics of complex genomes and abundant sequencing data of species such as wheat, public sequencing data is creatively utilized to carry out an interaction protein screening process of subgenomes. The problems of long protein screening period, high cost, large workload, false positive protein screening, no co-expression and the like are solved. The interaction relationship between important gene families can be quickly and conveniently obtained, the interaction structural domain can be accurately and clearly found, and technical support is provided for research on analysis of a wheat key gene functional mechanism.

Description

Technical Field

[0001] This invention belongs to the field of biotechnology, specifically, it relates to a method for screening plant interaction proteins using big data combined with artificial intelligence and for verifying protein interactions using yeast two-hybrid technology. Background Technology

[0002] Intermolecular protein interactions constitute the core mechanism of biological system regulation, and their functional network encompasses multidimensional biological processes such as receptor-ligand binding recognition, initiation and regulation of immune responses, transmembrane signal transduction mechanisms, and growth and differentiation regulation. In-depth research on this interaction system not only provides theoretical basis for understanding the essential laws governing life phenomena, but also demonstrates significant scientific value in applied fields such as disease mechanism elucidation, drug target screening, and synthetic biology construction.

[0003] The technology system for studying protein-protein interactions mainly encompasses two major experimental strategies: in vivo validation and in vitro analysis. In in vivo studies, the yeast two-hybrid (Y2H) system expresses target proteins by fusing them with the DNA-binding and activation domains of transcription factors, and then uses reporter gene expression to validate interactions, making it particularly suitable for large-scale screening. The bacterial two-hybrid (B2H) system, on the other hand, utilizes an adenylate cyclase complementarity mechanism to detect interactions in prokaryotic systems, making it suitable for the research needs of prokaryotic expression systems. In in vitro analysis, co-immunoprecipitation (Co-IP) combined with mass spectrometry has become the gold standard for identifying protein complexes. It captures target proteins and their interaction networks with specific antibodies, and then uses high-resolution mass spectrometry for precise identification. Pull-down assays utilize the binding properties of glutathione-S-transferase (GST) or histidine-tagged (His-tag) fusion proteins to affinity chromatography media to verify the direct interactions of specific protein pairs. Surface plasmon resonance (SPR), as an advanced technique for studying biomolecular interaction dynamics, can monitor molecular binding / dissociation processes in real time, providing key parameters such as binding constants and reaction rates. In recent years, the development of new technologies such as crosslinking-MS and fluorescence resonance energy transfer (FRET) has further expanded the spatiotemporal resolution of protein-protein interaction studies. Researchers need to select appropriate combinations of techniques for multidimensional verification based on factors such as the biological characteristics of the experimental system (prokaryotic / eukaryotic), the strength of the interaction (transient / stable), and the research objectives (qualitative screening / quantitative analysis).

[0004] Bacterial two-hybrid and yeast two-hybrid methods are widely used in identifying protein-protein interactions. However, screening libraries through experiments requires first constructing a corresponding protein library based on the research object, which is time-consuming, requires significant investment, involves complicated steps, and has a high false positive rate. Summary of the Invention

[0005] The purpose of this invention is to provide a method for screening plant interaction proteins using big data combined with artificial intelligence (AI) and for verifying protein interactions using yeast two-hybrid assays.

[0006] To achieve the objectives of this invention, the present invention provides a method for screening plant interaction proteins using big data combined with artificial intelligence and for verifying protein interactions using yeast two-hybrid assays, comprising the following steps: (1) Obtaining the amino acid sequence of the target protein P0 in the model species: Download the amino acid sequence of the target protein P0 of the model species from public biological databases; (2) Search for homologous proteins in the target species: In public biological databases, the amino acid sequence of P0 is searched for homologous proteins by Blast or by using a hidden Markov model (HMM) to obtain the homologous protein P1 in the target species. (3) Classification of homologous protein subfamilies in the research species: Analyze the protein family to which P1 belongs and find homologous proteins P2, P3, P4...P that are located on the same branch of the evolutionary tree as P1. n They form a homologous protein subfamily; (4) Acquisition of public transcriptome sequencing data and prediction of interacting proteins: Download the public transcriptome sequencing dataset corresponding to the research species from public biological databases, import it into the WGCNA package (version 1.69) of R language for analysis, find the genes co-expressed with each member of the above homologous protein subfamily, and perform functional annotation on the predicted candidate genes respectively. (5) Interaction structure prediction and molecular docking: AlphaFold3 (https: / / alphafoldserver.com) or PDB website (https: / / www.rcsb.org / ) was used to perform structural analysis on the predicted high-frequency interacting proteins, and molecular docking was performed based on the analysis results to lock the interaction domains; (6) Yeast two-hybrid verification of protein interaction and fluorescence colocalization analysis: P1 was used as the bait protein and P1′, which interacts with P1, was used as the target protein. The plasmid carrying the bait protein modified with GFP tag and the plasmid carrying the target protein modified with RFP tag were introduced into yeast for yeast two-hybrid verification and fluorescence colocalization analysis.

[0007] Further, in step (3), a phylogenetic tree is constructed using the neighbor-joining method in MEGA12 (Molecular Evolutionary Genetics Analysis Version 12, https: / / www.megasoftware.net / dload_win_beta) software to identify homologous genes.

[0008] Furthermore, the public transcriptome sequencing data obtained in step (4) requires WGCNA analysis in transcriptome samples of the target species.

[0009] Further, in step (4), based on the WGCNA analysis results, the co-expressed genes with the highest scores are functionally annotated and their protein sequences are obtained for subsequent analysis and interaction verification.

[0010] Further, in step (5), Pymol is used to perform molecular docking on the candidate proteins to predict the interaction sites.

[0011] Preferably, the plasmid carrying the bait protein modified with the GFP tag in step (6) is pGADT7 as its starting plasmid.

[0012] Preferably, the plasmid carrying the target protein modified with the RFP tag in step (6) is pGBKT7 as its starting plasmid.

[0013] Preferably, the yeast in step (6) is AH109.

[0014] Further, in step (6), the plasmids constructed for verifying the fluorescence colocalization interaction of yeast system proteins were co-transformed into the tryptophan and leucine synthesis-deficient yeast strain AH109. The transformed yeast AH109 was then subcultured in two- and three-deficient (tryptophan and leucine deficient) media, and fluorescence colocalization observations were performed. In three- and four-deficient media, namely leucine, tryptophan, histidine, and adenine media, it still grew better than the control, indicating a strong interaction.

[0015] Furthermore, to avoid the impact of self-activation on the results, a control is necessary. The experimental group's growth is only meaningful if the target protein (GFP-pGADT7) and bait protein (RFP-pGBKT7) do not produce yeast cells. The concentration gradient of the yeast culture used to detect yeast growth should be ≥3, such as an initial yeast culture dilution gradient of 0.1, 0.01, and 0.001, grown sequentially on two-, three-, and four-cell-deficient media.

[0016] Furthermore, in step (6), to observe protein colocalization and growth in tri- and quadri-deficient media, strong protein expression is required so that the fluorescence signal can be captured and imaged by a laser confocal microscope. Strong interactions are also required for growth in quadri-deficient media.

[0017] In one specific embodiment of the present invention, the model species is Arabidopsis thaliana, and the target protein PO is ATK5.

[0018] The species studied was wheat, and the homologous protein P1 was TaK14A1. P1 belongs to the Kinesin14 protein family, and the homologous proteins TaK14A1 and TaK14A2 are members of this family.

[0019] Furthermore, the wheat TaK14A1 protein interacts with the MCM4 protein at the Coiled-Coil (CC) domain on the TaK14A1 protein.

[0020] The reference sequence numbers of ATK5, TaK14A1, TaK14A2, MCM4, GFP, and RFP proteins in NCBI are: OAP00641.1, KAF7002304.1, XP_037479477.1, XP_044340218.1, P42212.1, and TID02864.1, respectively.

[0021] By employing the above technical solution, the present invention has at least the following advantages and beneficial effects: (i) This invention utilizes a large amount of public transcriptome data, which is easy to obtain and has high universality.

[0022] (ii) This invention analyzes the interaction relationships between protein families, elevating protein point-to-point prediction to face-to-face analysis, which can improve the accuracy of interaction relationship prediction and is novel in research.

[0023] (III) This invention uses AI to predict the structure of proteins analyzed from big data, which improves the speed of interaction motif research and enhances the accuracy of interaction prediction.

[0024] (iv) This invention uses an improved yeast interaction technology, adding a GFP-RFP fusion tag, and observes the fluorescent colocalization of the bait protein and the target protein, thereby further improving the accuracy of the yeast dual-hybrid interaction results. Attached Figure Description

[0025] Figure 1 The phylogenetic tree results of the family proteins identified in this invention.

[0026] Figure 2 This invention uses public transcriptome data to obtain the co-expression genes of TaK14A1 and its homologous proteins.

[0027] Figure 3 This is a molecular docking analysis of the interaction sites between Alphafold3-predicted target protein representative MCM4 and TaK14A1 in a preferred embodiment of the present invention.

[0028] Figure 4 This invention provides a preferred embodiment of the yeast interaction between the predicted CC motif by TaK14A1 and MCM4.

[0029] Figure 5 The diagram shows the backbones of two plasmids in the system constructed in a preferred embodiment of the present invention; wherein, (a) is plasmid GFP-pGADT7 and (b) is plasmid RFP-pGBKT7.

[0030] Figure 6 The fluorescence co-localization results of MCM4 and TaK14A1 in the yeast system of a preferred embodiment of the present invention are shown. Detailed Implementation

[0031] This invention provides a system for screening protein interactions using big data combined with artificial intelligence (AI), and a method for verifying interactions and observing the co-localization of interacting proteins in a yeast system. It mainly involves electronically screening a library of potential interacting proteins for key target proteins in wheat, including utilizing public sequencing data, constructing a subgenomic homologous protein co-expression network, and performing Alphafold3 interaction analysis.

[0032] The advantages of this method are: 1) The large volume of public transcriptome sequencing data improves the accuracy of interaction analysis. 2) Subgenomic homologous protein weighted gene co-expression network analysis (WGCNA) and cluster analysis can more clearly and accurately identify relationships between gene families. 3) Combining Alphafold3 and molecular docking allows for preliminary prediction of the amino acids that bind to two proteins, further improving the accuracy of the electronic screening library and providing a rapid strategy for subsequent binding domain analysis. 4) Combining the yeast two-hybrid system and laser confocal fluorescence co-localization analysis allows for the analysis of protein co-localization while observing interactions.

[0033] This method addresses the challenges of complex genomes and abundant sequencing data in species like wheat by innovatively utilizing publicly available sequencing data for subgenomic screening of interacting proteins. It solves problems such as long screening cycles, high costs, heavy workloads, false positives, and lack of co-expression. This method rapidly and conveniently reveals interactions between important gene families, accurately and clearly identifying interacting domains. Further experiments can be conducted to verify interactions in tobacco and yeast systems, as well as to verify interactions after key domain truncation, thus providing technical support for elucidating the functional mechanisms of key wheat genes.

[0034] To address the aforementioned problems, this invention provides a high-throughput screening system for protein interactions and its method of use. This high-throughput screening system is named WGCNA and AI Protein-Protein Interactions Detection (WAPD). WAPD is used to screen interacting proteins within a protein family. Taking the Arabidopsis thaliana kinesin ATK5 as an example, we identified its homolog in wheat, TaK14A1. Analysis of the Kinesin 14 protein family revealed that TaK14A1 and TaK14A2 in wheat are homologous proteins on the same branch of the phylogenetic tree. TaK14A A total of six homologous genes were available for further analysis. Using a target protein library constructed from approximately 5200 wheat transcriptome data, seven target proteins interacting with TaK14A were successfully identified. Four of these proteins belong to the MCM family, and their interactions were validated using a yeast system. The interaction site between TaK14A1 and MCM4 was predicted using Alphafold3, and the Coiled-coil (CC) motif at this site was confirmed as the key binding domain for both proteins.

[0035] The present invention adopts the following technical solution: This invention provides a system for screening protein interactions using big data combined with AI, a yeast system for verifying interactions, and a technique and method for observing the co-localization of interacting proteins. Specifically, this invention provides a method for creating an electronic library of protein interactions using AI combined with big data, comprising the following steps: (1) Acquisition of target gene amino acid sequence: Select the target gene as the object and search and download the corresponding public genome dataset from websites such as https: / / plantrnadb.com / athrdb / and http: / / wheatomics.sdau.edu.cn / . We use wheat data as an example. At the same time, download the amino acid sequence of the gene in the genome of known model species, such as the reported amino acid sequence of the protein in Arabidopsis, rice, and maize. In this invention, we take ATK5, a kinesin-14 family protein in Arabidopsis, as an example. ATK5 is a terminal tracking protein that is located in the metaphase of the mitotic spindle and in the region of vigorous growth in Arabidopsis.

[0036] (2) Homologous protein search: In the Tbtools-II software (https: / / github.com / CJ-Chen / TBtools-II / releases), the BLAST-BLAST GUI Wrapper-BLAS Zone function was used to build a wheat functional proteome library and perform BLAST operations to search for wheat homologous proteins of Arabidopsis thaliana ATK5, or a Hidden Markov model (HMM) was used for homologous protein search. The wheat ATK5 homologous protein is TaK14A1.

[0037] (3) Classification of homologous protein subfamilies: Protein subfamilies were analyzed using model plants, such as Arabidopsis thaliana. Taking wheat as an example, the TaK14A1 protein belongs to the Kinesin 14 family. TaK14A1 and TaK14A2 are two member proteins of this family.

[0038] (4) Public transcriptome acquisition: Select the corresponding research species as the object, search and download the corresponding public transcriptome dataset from websites such as https: / / plantrnadb.com / athrdb / and http: / / wheatomics.sdau.edu.cn / , and use the R package WGCNA (version 1.69) to perform weighted gene co-expression network analysis (WGCNA).

[0039] (5) Identification of homologous protein interactions: WGCNA analysis was performed on each gene in the selected target protein subfamily. For example, in wheat, the co-expression networks of TaK14A1 and TaK14A2 were selected. Co-expressed genes were annotated, and the number of genes in the same family was merged.

[0040] (6) AlphaFold3 Interaction Structure Prediction and Pymol (https: / / pymol.org / ) Molecular Docking: High-frequency interacting proteins predicted were analyzed using AlphaFold3 (https: / / alphafoldserver.com) or the Protein Data Bank (PDB) website (https: / / www.rcsb.org / ). Molecular docking was performed based on the AlphaFold3 results to pinpoint the interacting domains.

[0041] (7) Yeast interaction and fluorescence colocalization: Yeast AH109 (Clontech, CAT#YC1010) strain was simultaneously transformed with plasmids carrying bait protein-GFP (Green fluorescent protein) and target protein-RFP (Red fluorescent protein). The results of the electronic screening library were analyzed by two-hybrid growth experiment of interacting yeast and fluorescence colocalization analysis.

[0042] Furthermore, in step (3), the protein subfamily must meet the following requirements: a phylogenetic tree is constructed using the neighbor-joining method in MEGA12 (Molecular Evolutionary Genetics Analysis Version 12, https: / / www.megasoftware.net / dload_win_beta) software, and proteins on the same branch of the tree as the bait protein are selected as the protein set of that subfamily, and their sequences and structures are most similar.

[0043] Furthermore, the public transcriptome data obtained in step (4) requires all tissues to undergo WGCNA analysis, and the quantity and quality of transcriptomes determine the accuracy of the prediction results. For example, approximately 5200 wheat transcriptome datasets are available for download at https: / / plantrnadb.com / athrdb / .

[0044] Furthermore, in step (5), WGCNA selects the top 20 proteins in the WGCNA network of a total of 6 genes from wheat genomes A, B, and D that encode TaK14A1 and TaK14A2 proteins for annotation in subsequent analysis.

[0045] Further, in step (6), functional annotation is performed through step (5), the proportion of protein families in the weighted co-expression network of the 6 genes is calculated, and the protein family with the highest proportion is selected to perform interaction domain prediction and yeast system protein interaction verification using AlphaFold3.

[0046] Furthermore, in step (7), the plasmid backbone for screening the interaction between the two proteins is pGADT7 and pGBKT7, and the bait protein-GFP and the target protein-RFP are respectively linked to the backbone.

[0047] Further, in step (7), the constructed protein interaction verification system is co-transformed into yeast AH109, ​​which is deficient in tryptophan and leucine synthesis. After subculturing the transformed yeast AH109 in a two-deficient (tryptophan and leucine deficient) medium, fluorescence co-localization observation can be performed. The yeast still grows better than the control in a four-deficient medium, namely leucine, tryptophan, histidine, and adenine, indicating the presence of interaction.

[0048] Furthermore, to avoid the influence of self-activation on the results, a control is necessary. The growth of the experimental group is only meaningful if the target protein (GFP-pGADT7) and the bait protein (RFP-pGBKT7) do not produce yeast cells. The concentration gradient of the yeast culture used to detect yeast growth should be ≥3, such as an initial yeast culture dilution gradient of 0.1, 0.01, and 0.001, grown sequentially on two- and four-deficient media.

[0049] Furthermore, in step (7), to observe protein co-localization and growth in a four-deficient culture medium, strong protein expression is required so that the fluorescence signal can be captured and imaged by a laser confocal microscope. Strong interactions are also necessary for growth in the four-deficient culture medium.

[0050] This invention utilizes an AI-driven approach combined with big data analysis. First, big data analysis identifies co-expressed proteins within a protein family. AI-based interaction prediction and molecular docking are then performed on these proteins to further analyze and extract their interaction information. Finally, yeast fluorescence co-localization experiments are used to further analyze the identified interacting proteins.

[0051] The following examples are used to illustrate the present invention, but are not intended to limit the scope of the invention. Unless otherwise specified, the technical means used in the examples are conventional means well known to those skilled in the art, and the raw materials used are all commercially available products.

[0052] The plasmids pGADT7, pGBKT7, and yeast AH109 used in the following examples were all purchased from Shanghai Weidi Biotechnology Co., Ltd.

[0053] Example 1: Big Data Screening for Co-expressed Proteins of Gene Family Proteins 1. Genome Data Download: For obtaining protein sequences from the target species, select the corresponding research species and search for and download the relevant public genome datasets from websites such as https: / / plantrnadb.com / athrdb / and http: / / wheatomics.sdau.edu.cn / . We will use wheat data as an example. Simultaneously, download the protein sequences of the corresponding genes from known model organisms, such as those reported in Arabidopsis thaliana, rice, and maize. In this example, we will use Arabidopsis thaliana ATK5 as an example.

[0054] 2. Homologous protein search: In Tbtools-II software, use the BLAST-BLAST GUI Wrapper-BLAS Zone function to build a wheat functional proteome library and perform a BLAST operation to search for wheat homologous proteins of ATK5. Alternatively, use HMM for homologous protein search. The wheat ATK5 homologous protein is TaK14A1.

[0055] 3. Classification of homologous protein subfamilies: Protein subfamilies are analyzed using model plants, such as Arabidopsis thaliana. Taking wheat as an example, the TaK14A1 protein belongs to the TaK14A family. This family is divided into two proteins: TaK14A1 and TaK14A2.

[0056] 4. Obtaining Public Transcriptomes: Select the corresponding research species as the object, search for and download the corresponding public transcriptome datasets from websites such as https: / / plantrnadb.com / athrdb / and http: / / wheatomics.sdau.edu.cn / , and perform WGCNA analysis.

[0057] 5. Identification of inter-subfamily interactions: WGCNA results were analyzed for each gene in the selected subfamily. For example, in wheat, the selected... TaK14A1 and TaK14A2 Interaction networks. Interacting proteins are annotated, and the number of genes from the same family is merged.

[0058] like Figure 1 and Figure 2 As shown, the MCM family and the TaK14A family interact with each other.

[0059] Example 2: Big Data Combined with AI to Screen Co-expressed Proteins of Gene Family Proteins like Figure 3 As shown, AlphaFold3 interaction structure prediction and Pymol molecular docking: Structural analysis of predicted high-frequency interacting proteins was performed using AlphaFold3 or the PDB website. Molecular docking was performed based on the AlphaFold3 results to pinpoint the interacting domains. The Coiled-Coil (CC) domain of TaK14A1 is hydrogen-bonded to MCM4 of the MCM family.

[0060] Example 3: Yeast Interaction and Fluorescent Colocalization like Figure 4As shown, two modified vectors, GFP-pGADT7 and RFP-pGBKT7, were used. The GFP-pGADT7 plasmid carrying the bait protein and the potential target protein RFP-pGBKT7 plasmid were simultaneously transformed into strain AH109. The growth of potential positive yeast colonies on defective media was analyzed. The expression and colocalization of proteins in yeast cells were observed using a Zeiss laser confocal microscope (LSM800). The results of the electronic screening library could be analyzed simultaneously for yeast two-hybrid and fluorescence colocalization to confirm the protein interaction. Figure 5 The results show that TaK14A1 interacts with the full length of MCM2, 3, 4, and 7, and that the CC domain of TaK14A1 interacts with MCM4. Figure 6 The results showed that TaK14A1 and MCM4 proteins co-localize in the cell nucleus.

[0061] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. A method for screening plant interaction proteins using big data combined with artificial intelligence and verifying protein interactions using yeast two-hybrid assays, characterized in that, Includes the following steps: (1) Obtaining the amino acid sequence of the target protein P0 in the model species: Download the amino acid sequence of the target protein P0 of the model species from public biological databases; (2) Search for homologous proteins in the target species: In public biological databases, the amino acid sequence of P0 is searched for homologous proteins by Blast or by using a hidden Markov model to obtain the homologous protein P1 in the target species. (3) Classification of homologous protein subfamilies in the research species: Analyze the protein family to which P1 belongs and find homologous proteins P2, P3, P4...P that are located on the same branch of the evolutionary tree as P1. n They form a homologous protein subfamily; (4) Acquisition of public transcriptome sequencing data and prediction of interacting proteins: Download the public transcriptome sequencing dataset corresponding to the research species from public biological databases, import it into the WGCNA package of R language for analysis, find the genes co-expressed with each member of the above homologous protein subfamily, and perform functional annotation on the predicted candidate genes respectively. (5) Interaction structure prediction and molecular docking: AlphaFold3 or PDB website was used to perform structural analysis on the predicted high-frequency interacting proteins, and molecular docking was performed based on the analysis results to lock the interaction domains. (6) Yeast two-hybrid verification of protein interaction and fluorescence colocalization analysis: P1 was used as the bait protein and P1′, which interacts with P1, was used as the target protein. The plasmid carrying the bait protein modified with GFP tag and the plasmid carrying the target protein modified with RFP tag were introduced into yeast for yeast two-hybrid verification and fluorescence colocalization analysis.

2. The method according to claim 1, characterized in that, Step (3) Use the neighbor-joining method in MEGA12 software to construct a phylogenetic tree and identify homologous genes.

3. The method according to claim 1, characterized in that, The public transcriptome sequencing data obtained in step (4) requires WGCNA analysis in transcriptome samples of the target species.

4. The method according to claim 1, characterized in that, Step (4) Based on the WGCNA analysis results, perform functional annotation on the co-expressed genes with the highest scores and obtain their protein sequences for subsequent analysis and interaction verification.

5. The method according to claim 1, characterized in that, Step (5) Use Pymol to perform molecular docking on candidate proteins and predict interaction sites.

6. The method according to claim 1, characterized in that, The plasmid carrying the bait protein modified with the GFP tag in step (6) has pGADT7 as its starting plasmid.

7. The method according to claim 1, characterized in that, The plasmid carrying the target protein modified with the RFP tag in step (6) has pGBKT7 as its starting plasmid.

8. The method according to claim 1, characterized in that, The yeast mentioned in step (6) is AH109.

9. The method according to any one of claims 1-8, characterized in that, The model species is Arabidopsis thaliana, and the target protein P0 is ATK5. The species studied was wheat, and the homologous protein P1 was TaK14A1. P1 belongs to the Kinesin 14 family of proteins, and the homologous proteins TaK14A1 and TaK14A2 are members of this family.

10. The method according to claim 9, characterized in that, The wheat TaK14A1 protein interacts with the MCM4 protein at the Coiled-Coil domain of the TaK14A1 protein.

Images