Combining spatial and chemical information for clustering pharmacophores

Abstract

Background: A pharmacophore model consists of a group of chemical features arranged in three-dimensional space that can be used to represent the biological activities of the described molecules. Clustering of molecular interactions of ligands on the basis of their pharmacophore similarity provides an approach for investigating how diverse ligands can bind to a specific receptor site or different receptor sites with similar or dissimilar binding affinities. However, efficient clustering of pharmacophore models in three-dimensional space is currently a challenge.

Results: We have developed a pharmacophore-assisted Iterative Closest Point (ICP) method that is able to group pharmacophores in a manner relevant to their biochemical properties, such as binding specificity etc. The implementation of the method takes pharmacophore files as input and produces distance matrices. The method integrates both alignment-dependent and alignment-independent concepts.

Conclusions: We apply our three-dimensional pharmacophore clustering method to two sets of experimental data, including 31 globulin-binding steroids and 4 groups of selected antibody-antigen complexes. Results are translated from distance matrices to Newick format and visualised using dendrograms. For the steroid dataset, the resulting classification of ligands shows good correspondence with existing classifications. For the antigen-antibody datasets, the classification of antigens reflects both antigen type and binding antibody. Overall the method runs quickly and accurately for classifying the data based on their binding affinities or antigens.

Figure 1

2D molecular structures and names of the 31 globulin binding steroids.

Figure 2

ICP application to two antigens from PDB entries 1ADQ_P2[33]and 3GBN [34]. 1ADQ_P2 is shown in blue and is the reference model. Green points represent 3GBN before application of ICP. Red points correspond to 3GBN after ICP transformation based on 1ADQ_P2.

Figure 3

Workflow of the ICP aided pharmacophore clustering method.

Figure 4

Clustering of the 31 globulin binding steroids. This dendrogram is showing the clustering of the 31 globulin binding steroids derived using a combination of 3D and chemical distances.

Figure 5

Clustering of the 31 globulin binding steroids. This dendrogram is showing the clustering of the 31 globulin binding steroids derived using the group average clustering method by Rodriguez [16].

Figure 6

Dendrogram showing the clustering of the 31 globulin binding steroids derived using Ward's clustering method by Rodriguez [16]. This dendrogram is showing the clustering of the 31 globulin binding steroids derived using Ward's clustering method by Rodriguez [16].

Figure 7

Clustering of 41 antibody-antigen complexes based on combined distance. This dendrogram is showing the clustering of four groups of antibody-antigen complexes based on a combination of 3D and chemical distances. Complexes with antigen GP41 are shown with a green background. Complexes with antigen GP120 are shown with a yellow background. Complexes with antibody 17B, are shown with their PDB ID colored blue. Complexes with antibody 2F5 are shown with their PDB ID colored red. Complexes with antibody ANTI-HIV-1 V3 FAB 2557 are shown with their PDB ID colored black.

Figure 8

Clustering of 41 antibody-antigen complexes based solely on 3D distance. This dendrogram is derived from a classification of antigens in a set of 41 antibody-antigen complexes based solely on 3D distance. 1U8H (yellow highlight) is structurally different from other similar complexes and is clustered separately from the other 1U8* complexes (black highlights).

Figure 9

Clustering of 41 antibody-antigen complexes based solely on chemical distance. This is a dendrogram based on a classification of antigens in a set of 41 antibody-antigen complexes based solely on chemical differences between pharmacophores. 1U8L (yellow highlight) chemically differs from other similar complexes and is clustered away from the other 1U8* complexes (black highlights).

Figure 10

Clustering of 41 antibody-antigen complexes based on combined distance. This section of a dendrogram calculated from a distance measure combining 3D structural distance and chemical distance between pharmacophores. 1U8L and 1U8H (highlighted in yellow) are correctly identified based on antigen.