Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Bioinformatics, Computational and Systems Biology

Bioinformatics, Computational and Systems Biology - Poster Session B

Poster A3 - Computationally investigating the effect of point mutations on drug targets for Alzheimer’s disease

Thursday, October 24, 2024

2:45 PM - 3:45 PM EST

Location: Exhibit Hall E, F & G

Presenting Author(s)

Nida Shah (she/her/hers)

Undergraduate Student
University of Maryland College Park
Silver Spring, Maryland, United States

Primary Investigator(s)

JK

Jeffery Klauda, PhD

Primary Investigator
University of Maryland College Park, United States

Last Author(s)

TT

Tejaswi Tammareddy

Graduate Student
University of Maryland College Park, United States

Introduction: The protein cyclin-dependent kinase (CDK5) has many biological functions in the neuronal cell. Under certain pathological conditions CDK5 is activated by the protein p25, leading to the hyperphosphorylation of various proteins (tau and amyloid precursor proteins among them). This results in the formation of plaques and tangles in the brain, which can lead to the onset of Alzheimer’s disease. However, when the man-made peptide p5 developed by Shukla et al. (2012) binds to CDK5, the CDK5-p25 binding complex can be inhibited. This makes the regulation of CDK5 by p5 a crucial area of study for therapeutics designed to combat Alzheimer’s disease. Recent studies on CDK5 performed by Tammareddy et al. at the University of Maryland (UMD) identified and proposed interactions governing the p5 inhibition of CDK5 hyperactivity. However, there is still limited research on the pharmacophore p5, particularly on how mutations in its amino acid sequence can affect its binding to CDK5. This research provides a promising method for preventing Alzheimer’s disease by inhibiting the formation of the CDK5-p25 binding complex. Our study uses a computational modeling system to introduce point mutations in the amino acid sequence of p5 to deepen our understanding of how p5 binds to CDK5 and inspire therapeutics aimed at preventing Alzheimer’s disease.

Materials and

Methods: The first step in our approach was visually analyzing the 24 base pair amino acid sequence of p5, to determine which sites would be optimal for our mutations. The goal is to disrupt hydrophobic interactions between p5 and CDK5, as we hypothesized that these are essential for stable binding. We therefore created two mutation variants of p5, in which variant one mutates the amino acids ILE 12 to HIS 12 and LEU 20 to ARG 20, and variant two mutates the amino acids PHE 4 to ARG 4 and VAL 11 to HIS 11. From here our methods have a two part software approach. One set of softwares is used to create point mutations in p5, and another set of softwares is used to dock p5 to CDK5. A crucial step in our design process was evaluating several softwares to determine which make the most accurate mutations and docking poses. We determined that VMD most accurately represents point mutations in p5, and Autodock 4 most accurately represents the docking of p5 to CDK5. Future steps include generating files of wild type and mutated sequences of p5, and using these to generate docking poses of wild type and mutated p5 to CDK5. Once these docking poses are obtained, we can use a script we have created to run docking poses in NAMD, which will tell us if the binding of our mutated p5 sequences to CDK5 is stable.

Results, Conclusions, and Discussions: We have currently examined three softwares for making mutations to the p5 peptide: Alphafold2, CHARMM-GUI, and VMD. We obtained visual representations of wild type p5 and mutations in Alphafold2, but found that p5 was represented as a single alpha helix. Our figures did not match the ensemble of secondary structures for p5 produced by Tammareddy et. al (2022) using a method based on configurational bias Monte Carlo (MC) sampling, so we explored other methods for making our point mutations. Both CHARMM-GUI and VMD provided visual representations of p5 and CDK5 that accurately represented the secondary structure of p5 in accordance with Tammareddy et al. (2022). We opted to use VMD for our point mutations due to the ease of software use and visualization. In regards to generating our docking poses, we began with examining Autodock Vina. When we represented the docking of wild-type p5 to CDK5, we found that p5 was represented as an unstructured peptide without any of the alpha helices or beta sheets that were shown in Tammareddy et al. 's Monte Carlo sampling visualization of p5. Our next steps involve generating docking poses in Autodock 4, a more powerful version of Autodock Vina in which users can fix a ligand structure so it is not distorted during docking. Our work aims to serve as a guide to experimentalists hoping to utilize the inhibitory properties of p5 to create therapeutics that can prevent the onset of Alzheimer’s disease. Treatment options for Alzheimer’s disease are limited to symptom management, so utilizing p5 as a drug treatment can provide a novel therapeutic that targets Alzheimer’s at its source. Additionally, current structure prediction for the p5 peptide is limited to configurational bias Monte Carlo (MC) sampling utilized by Tammareddy et al. (2022), which is a lengthy process that requires vast resources. Modeling p5 using our proposed system is a much shorter process and more user friendly, thus encouraging more researchers to model and examine the p5 peptide as a potential drug treatment for Alzheimer’s.

Acknowledgements (Optional): Jeffery Klauda, Tejaswi Tammareddy, Klauda Lab at the University of Maryland