A snapshot taken of slingshot homolog2 (SSH-2)-a member of the dual specific phosphatase (DSP) family-using the visualization program UCSF Chimera. Vicky Hwang, Olivia Yang and Joshua Wei were looking for ligands that would inhibit SSH-2 without binding to other members of the DSP family.
Grid computing has become a popular method for drug discovery, and this project builds the PRIME 2011 project of Brian Tsui. Rather than test each chemical compound in wet lab experiments, virtual screening using grid computational resources can simulate the interactions between the compounds (often in a database library) and proteins. In addition, the use of grid computing shares hardware resources and facilitates the performing of multiple computational simulations to determine optimal bindings of compounds to proteins (i.e. docking) at the same time, expediting the process. This project focused on the discovery of inhibitors that would bind specifically to SSH-2, a member of the dual-specific phosphatase (DSP) family. The DSP family of proteins has roles in various cellular processes, and the proteins have been linked to cancer and Alzheimer’s disease. The challenge in identifying possible inhibitors of SSH-2 is that its overall structure is similar to the structures of other family members; thus there is a need to identify compounds that only interact with SSH-2, not the other members.
Using two previously established PRAGMA grid-enabled programs—MODELLER, which approximates the three-dimensional structure of a protein, and DOCK6, which calculates and scores the binding strength between a protein and chemical compound–this project enhances previous research. To optimize the protein models and analyze how well the resultant protein structure models were folded, the online program MolProbity was used.
The protein structure was prepared for docking and the active site–where the protein-compound interaction would occur–was specifically targeted for docking. In order to provide a higher level of accuracy, two scoring systems were used to assess the binding strength: an energy grid scoring system and the more time-intensive AMBER scoring system. The resulting scores were used to obtain a final consensus score and rank of the best to worst binding compounds. During the project, protein structure models were generated for SSH-1, SSH-3, and DSP-21. Consensus scoring and compound ranking were completed for SSH-1, SSH-3, and DSP-21. The findings suggest that modeled protein structures can be used in virtual screening experiments. Also, the results show promising chemical compounds that may specifically bind to SSH-2, with minimal interactions with the remaining DSP family members. These compounds will need to be tested in actual cellular experiments to determine if they can specifically inhibit SSH-2 in vitro.
PARTICIPATING RESEARCHERS: Osaka University: Susumu Date, Shinji Shimojo; UCSD: Jason H. Haga