3D binding pocket protein structure similarity virtual screening

Introduction

Algorithm

Examples

Run 3Dclick

References

Contacts

Help

Examples

1. Query structure of A.fulgidus Rio2 Kinase with ligand ATP and possible targets of ATP

  • In this example, the binding pocket/site of the ligand ATP (Adenosine triphosphate) on the structure of A.fulgidus Rio2 Kinase (PDB code: 1zaoA) is used for structure similarity virtual screening by 3Dclick on the database of different SCOP families of 2868 PDB chains

  • Please click here to look at the output of this query example.

  • The virtual screening on this database will help us identify possible target proteins that are unrelated by protein fold/ family with the query structure. On the database of different SCOP families, PurT-encoded glycinamide ribonucleotide transformylase (PDB code: 1kj9A, Thoden et al., J Biol Chem , 2002) is identified as a target protein of ATP using 3Dclick that has a significant Z-score=2.98 with the ATP binding site of A.fulgidus Rio2 Kinase. The PurT-encoded glycinamide ribonucleotide transformylase has the 9th rank in the output table after the energy minimization step on our 3Dclick server. Moreover, A.fulgidus Rio2 Kinase and PurT-encoded glycinamide ribonucleotide transformylase are unrelated by protein fold and sequence. The Structural Classification of Proteins (SCOP) of A.fulgidus Rio2 Kinase are a.4.5.56 and d.144.1.9 while SCOP of PurT-encoded glycinamide ribonucleotide transformylase are c.30.1.1, d.142.1.2, and b.84.2.1. Figure 1 shows the possible complex structure of PurT-encoded glycinamide ribonucleotide transformylase and ATP after the energy minimization step on our 3Dclick server using the 3D visualization of NGL (https://nglviewer.org/).

  • As displayed in Figure 2 using Chimera (https://www.cgl.ucsf.edu/chimera/), the ATP in X-ray structure of PurT-encoded glycinamide ribonucleotide transformylase (white and orange sticks) has the same location as ATP predicted by 3Dclick (plum and orange sticks). This result indicates that 3Dclick has capability of predicting potential target proteins of ligands/small molecules even for unrelated protein folds with query structures.

  • For this example, both ProBiS-ligands and GrAfSS web servers could not find PurT-encoded glycinamide ribonucleotide transformylase (PDB code: 1kj9A) as a target protein for ATP from the binding site of A.fulgidus Rio2 Kinase (PDB code: 1zaoA) while idTarget web server is currently unreachable.

  • Figure 1: 3Dclick displays the visualization of 3D structures for the complex of PurT-encoded glycinamide ribonucleotide transformylase (PDB code: 1kj9A) and ATP using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target 1kj9A (rank 9th) in the output table after the energy minimization step on our 3Dclick server.

    Figure 2: The ATP in X-ray structure of PurT-encoded glycinamide ribonucleotide transformylase (white and orange sticks) has the same location in 3D structure as ATP predicted by 3Dclick (plum and orange sticks).

    [back to top]

    2. Possible target proteins of anti-cancer drug Lapatinib

  • Lapatinib developed by GlaxoSmithKline (GSK) is an anti-cancer drug approved by the U.S. Food and Drug Administration (FDA) in 2007 for the treatment of for solid tumours such as breast, lung cancer and other solid tumours (https://go.drugbank.com/drugs/DB01259).

  • In this study, the binding pocket/site of the drug Lapatinib (ligand FMM) on the structure of epidermal growth factor receptor (PDB code: 1xkkA, Wood et al., Cancer Res, 2004) is used for structure similarity virtual screening by 3Dclick on the database of 2942 structures from kinases and AlphaFold structures as well as proteins of different SCOP families (as shown in the below query).



  • On this database, the virtual screening step of 3Dclick identifies the list of 87 possible target structures from PDB and AlphaFold with significant Z-scores (Z-score ≥2) and the number of ugly contacts < 2.2*(the number non-hydrogen atoms of Lapatinib) in their complex model with Lapatinib. The detailed visualization of each complex is shown when users click on "3D view" of each possible target in the output table before the energy minimization step on our 3Dclick server.

  • After the energy minimization step to attempt removing any possible ugly and bad contacts of the complexes, 3Dclick identifies the list of 76 possible target structures from PDB and AlphaFold with the number of ugly contacts = 0 in their complex model with Lapatinib. The detailed visualization of each complex is shown when users click on "3D view" of each possible target in the output table after the energy minimization step.

  • The top hits after the energy minimization step (rank-ordered by Z-score) are from EGFR, ERBB-4, ERBB-2, ERBB-3, RIPK2, MEK5, LOK, SLK, MKK7, PIK4CB. Interestingly, the DFG motif of the activation loop on the binding pocket of Lipatinib is not only conserved in the sequences but also has the same location in the 3D structures for these top hits (Figures 3 and 4).

  • Current PDB has only one other structure co-crystalized with Lipatinib that is ERBB-4 kinase (PDB code: 3bbtB, Qiu et al., Structure, 2008 ). 3Dclick also identifies this structure to be a possible target protein of Lipatinib with significant Z-score = 9.55 (rank 20th in the output table after the energy minimization step). As shown in Figure 5, the Lipatinib in X-ray structure (pink sticks) has the same location as Lipatinib predicted by 3Dclick (gray sticks) and both their CL atoms overlap and point the same direction. Interestingly, our model has the 3D structures of all atoms for Lipatinib while the Lipatinib in X-ray structure of 3bbtB does not have the structure of the atoms CS(=O)(=O)CCN.

  • The study of Davis et al., Nature Biotechnology, 2011 using competition binding assays for comprehensive analysis of kinase inhibitor selectivity determined EGFR, ERBB-4, ERBB-2, ERBB-3, RIPK2, MEK5, LOK, SLK, MKK7, PIK4CB, PIK3C2B to be target proteins of Lipatinib with the binding results (dissociation constant Kd) from 2.4 nM to 5500 nM. Among these eleven targets, 3Dclick identifies ten proteins of EGFR, ERBB-4, ERBB-2, ERBB-3, RIPK2, MEK5, LOK, SLK, MKK7, PIK4CB as targets of Lipatinib (as shown in the output table after the energy minimization step). Figures 6-15 show the complex models of these ten target proteins with Lipatinib on our 3Dclick server.

  • The study of Davis et al., Nature Biotechnology, 2011 determined phosphatidylinositol 4-phosphate 3-kinase C2 domain-containing subunit beta (PIK3C2B) to be a target protein of Lipatinib. The binding result (dissociation constant Kd) of PIK3C2B and Lipatinib is 670 nM. However, PIK3C2B is not in the list of possible targets of Lipatinib using 3Dclick (as shown in the output table after the energy minimization step of Lipatinib on our 3Dclick server). The possible reason is that PIK3C2B is a very large kinase with 1634 residues (https://www.uniprot.org/uniprotkb/O00750/entry), and currently there is no X-ray, NMR, cryo-EM structure of PIK3C2B in the PDB except for only one AlphaFold structure (https://alphafold.ebi.ac.uk/entry/O00750) with some regions with low and very low confident prediction especially in the activation loop on the binding pocket from the position 1211 to 1237 (https://www.uniprot.org/uniprotkb/O00750/entry#family_and_domains). Figure 16 shows the possible complex model of PIK3C2B and Lipatinib using 3Dclick before filtering by the number of ugly contacts. As shown in Figure 17 for the superimposition of PIK3C2B (blue ribbons) on the binding pocket of epidermal growth factor receptor (EGFR) (PDB code: 1xkkA, green ribbons) with Lipatinib (gray sticks), 3Dclick identifies the correct binding pocket of PIK3C2B as the DFG (D1213, F1214, G1215) motif of the activation loop of PIK3C2B (red label) matches and occupies the same location in the 3D structures with the DFG (D855, F856, G857) motif of EGFR (black label). However, the sidechain and aromatic ring of F1214 of the AlphaFold structure of PIK3C2B are possibly not correct location, and thus aromatic ring of F1214 significantly clashes and attaches to Lipatinib (Figure 17). As a result, 3Dclick filters PIK3C2B out of the list of possible targets because of the number of ugly contacts between PIK3C2B and Lipatinib in their complex model (Figure 17).

  • Figure 3: 3Dclick superimposition of the receptor tyrosine-protein kinase (ERBB-4) (blue ribbons) (PDB code: 3bbtB, Qiu et al., Structure, 2008 , rank 20th in the output table after the energy minimization step) on the binding pocket of epidermal growth factor receptor (EGFR) (green ribbons) (PDB code: 1xkkA) with the drug Lapatinib (gray sticks) using Pymol (https://pymol.org). The DFG (D836, F837, G838) motif of the activation loop of ERBB-4 (red label) occupies the same location in the 3D structures with the DFG (D855, F856, G857) motif of EGFR (black label).

    Figure 4: 3Dclick superimposition of the receptor-interacting serine/threonine-protein kinase 2 (RIPK2) (magenta ribbons) (PDB code: 5ng3A, Pellegrini et al., PLoS One, 2017 , rank 35th in the output table after the energy minimization step) on the binding pocket of epidermal growth factor receptor (EGFR) (green ribbons) (PDB code: 1xkkA) with the drug Lapatinib (gray sticks). The DFG (D164, F165, G166) motif of the activation loop of ERBB-4 (cyan label) occupies the same location in the 3D structures with the DFG (D855, F856, G857) motif of EGFR (black label).

    Figure 5: The Lipatinib in X-ray structure of ERBB-4 kinase (pink sticks) has the same location in 3D structure as Lipatinib predicted by 3Dclick (gray sticks) as well as both their CL atoms overlap and point the same direction. The DFG (D836, F837, G838) is the motif of activation loop of ERBB-4 (black label) on the binding pocket with Lipatinib.

    Figure 6: 3Dclick displays the visualization of 3D structure for the complex of epidermal growth factor receptor (EGFR) (PDB code: 2rgpA) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target 2rgpA (rank 2nd) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of EGFR and Lipatinib is 2.4 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 7: 3Dclick displays the visualization of 3D structure for the complex of receptor tyrosine-protein kinase (ERBB-4) (PDB code: 2r4bA) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target 2r4bA (rank 19th) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of ERBB-4 and Lipatinib is 54 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 8: 3Dclick displays the visualization of 3D structure for the complex of receptor tyrosine-protein kinase (ERBB-3) (PDB code: 3lmgA) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target 3lmgA (rank 24th) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of ERBB-3 and Lipatinib is 5500 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 9: 3Dclick displays the visualization of 3D structure for the complex of receptor tyrosine-protein kinase (ERBB-2) (PDB code: 7jxhA) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target 3lmgA (rank 26th) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of ERBB-2 and Lipatinib is 7 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 10: 3Dclick displays the visualization of 3D structures for the complex of receptor-interacting serine/threonine-protein kinase 2 (RIPK2) (PDB code: 5ng3A) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target 5ng3A (rank 35th) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of RIPK2 and Lipatinib is 3600 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 11: 3Dclick displays the visualization of 3D structure for the complex of dual specificity mitogen-activated protein kinase kinase 5 (MEK5) (AlphaFold structure: AF-Q13163-F1) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target AF-Q13163-F1 (rank 36th) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of MEK5 and Lipatinib is 1100 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 12: 3Dclick displays the visualization of 3D structures for the complex of serine/threonine-protein kinase 10 (LOK) (PDB code: 5ajqA) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target 5ajqA (rank 51th) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of LOK and Lipatinib is 4400 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 13: 3Dclick displays the visualization of 3D structures for the complex of STE20-like serine/threonine-protein kinase (SLK) (PDB code: 6hvdA) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target 6hvdA (rank 55th) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of SLK and Lipatinib is 3300 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 14: 3Dclick displays the visualization of 3D structures for the complex of dual specificity mitogen-activated protein kinase kinase 7 (MKK7) (PDB code: 6yg6A) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target 6yg6A (rank 70th) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of MKK7 and Lipatinib is 4400 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 15: 3Dclick displays the visualization of 3D structure for the complex of phosphatidylinositol 4-kinase beta (PIK4CB) (AlphaFold structure: AF-Q9UBF8-F1) and Lipatinib using NGL (https://nglviewer.org/). The detailed visualization is shown when users click on "3D view" of the possible target AF-Q9UBF8-F1 (rank 76th) in the output table after the energy minimization step on our 3Dclick server. The binding result (dissociation constant Kd) of PIK4CB and Lipatinib is 940 nM (Davis et al., Nature Biotechnology, 2011).

    Figure 16: The possible complex of phosphatidylinositol 4-phosphate 3-kinase C2 domain-containing subunit beta (PIK3C2B) (AlphaFold structure: AF-O00750-F1, blue ribbons) and Lipatinib (gray sticks) using 3Dclick before filtering by the number of ugly contacts.

    Figure 17: 3Dclick superimposition of phosphatidylinositol 4-phosphate 3-kinase C2 domain-containing subunit beta (PIK3C2B) (blue ribbons) (PDB code: AF-O00750-F1) on the binding pocket of epidermal growth factor receptor (EGFR) (green ribbons) (PDB code: 1xkkA) with the drug Lapatinib (gray sticks). The DFG (D1213, F1214, G1215) motif of the activation loop of PIK3C2B (red label) occupies the same location in the 3D structures with the DFG (D855, F856, G857) motif of EGFR (black label). This supperimposed result suggests that 3Dclick identifies the correct binding pocket of PIK3C2B with Lapatinib.

    [back to top]