32152 Drugs as cargo: development of advanced path sampling simulations to investigate how peptides can carry drugs through cell membranes
Begeleider(s): Samaneh Davoudi en Wouter Vervust
Richtingen: Master of Science in Biomedical Engineering, Master of Science in Engineering Physics

Probleemstelling:

Viruses may enter and infect the cells by membrane fusion. This process involves fusion peptides, which are short amino acid chains (peptides), that interact with the membrane to initiate the fusion. As an example, the fusion peptides of viral spike protein assist the corona virus to enter the host cell. This behavior has inspired a new approach in drug design. It is often a challenge to bring the less soluble drug molecules into the interior of the cells. A particularly interesting route of thought is then to attach the drug molecule as 'cargo' to a cell penetrating peptide (CPP), a short amino acid chain (peptide) that can permeate membranes. Potential clinical applications of this facility have prompted enormous effort, both experimental and theoretical, to better understand how CPPs manage to overcome the prodigious thermodynamic cost of membrane permeation. Current use is limited by a lack of cell specificity, meaning that drugs-loaded CPP cannot distinguish for instance between healthy and cancerous cells. There is currently an insufficient understanding of the modes of their uptake, which hinders developments to improve the specificity e.g. by varying the amino acid sequence, charges, and chain length in a guided manner. Modeling at the molecular scale can aid in understanding the process.

Figure 1.  A cell penetrating peptide (CPP) is a short chain of amino acids that can enter the cell and that can carry a drug across the membrane. The route that will be investigated in this thesis is the direct transit through the membrane (translocation).

 

Figure 2. A protein (light blue) is attached to a membrane (horizontal in figure) consisting of phospho-groups (blue and red atoms) and hydrocarbon chains (light grey). Several peptide chains (dark grey and orange) are immersed in the membrane. Water atoms are not shown.

 


Doelstelling:

The primary goal of the thesis is to model peptide transport through  phospholipid bilayers, also called peptide translocation, in order to understand its process from a biophysics point of view. In particular, the student will tackle the challenge of the time scales. Indeed, when using molecular dynamics simulations to create a trajectory of the peptide, the transition from one side of the membrane to the other might simply not occur during the simulation time, as it is a slow and rare event. Therefore, a new methodology will be developed based on transition path sampling, where multiple trajectories are created by a combination of molecular dynamics and Monte Carlo [1]. The result is a statistical ensemble of many short trajectories of which many are transition trajectories. A re-weighing with probabilities needs to occur to retract the original true 'rareness' of the translocation event.

The above method is called Replica Exchange Transition Interface Sampling (RETIS) and is an advanced method in statistical physics that allows for the extraction of not only thermodynamic equilibrium properties but also kinetics information. RETIS is being tested for small molecules like O2 and H2O at Ghent University using the PyRetis code that was developed by our Norwegian partner at the Norwegian University of Science and Technology. In addition, the RETIS approach will be held against a recently biasing scheme developed in collaboration with the National Institute of Health of the USA. Firstly, a large part of the thesis will be to adapt the current methodology to be suitable for peptide chains that are much larger than oxygen molecules or water molecules. In particular, new Monte Carlo moves will be designed to improve the sampling, for instance by letting the paths loose some of their memory. This requires a combination of skills in algorithmic thinking and physics. A particular nice property of the peptides is that the peptide chain can be varied in length and chemistry, making this an ideal tunable system in terms of computational complexity. Secondly, the student can work on understanding the translocation event and its optimal conditions. The results can be then be interpreted in terms of drug delivery applications. This links the theoretical work with real-world drug design.

This thesis is embedded in a close collaboration between Ghent University, the Norwegian University of Science and Technology (NTNU Trondheim), and the National Institutes of Health (NIH, USA).

references:

[1] Foundations and latest advances in replica exchange transition interface sampling, J. Chem. Phys., 147, 152722 (2017), Raffaela Cabriolu, Kristin M. Skjelbred Refsnes, Peter G. Bolhuis, and Titus S. van Erp