Living cells are densely crowded systems. Their interior contains a high density of proteins, nucleic acids, and other macromolecules, and it is divided into numerous compartments enclosed by lipid membranes. These cellular compartments, or organelles, form microscopic chemical reactors in which life-sustaining enzymatic reactions take place. Therefore, insight into the behavior of molecules in these complex objects is needed in order to understand living organisms at the microscopic level. Moreover, cellular compartments have recently become an inspiration to the fields of biotechnology and biomedicine. Novel biological nanoreactors encapsulating therapeutic enzymes are designed to replace malfunctioning organelles in the human body or to deliver drugs into cells. For rational design of such constructs, it is critical to understand how enzymatic reactions proceed when confined to crowded microscopic volumes, i.e., to conditions which are very different from those in standard biochemical assays.
Computer simulations can offer unprecedented insights into biomolecular motions and interactions, often going beyond what can be accessed by experimental measurements. However, simulating large constructs, such as a whole nanoreactor, remains a considerable challenge for conventional atomistic molecular dynamics (MD) simulations because the computational costs become prohibitive with the growing size of the system.
In this project, we employed a novel computational approach to simulate the first detailed model of a biological nanoreactor, representing constructs investigated in previous experimental studies. The nanoreactor had a form of a lipid vesicle (8500 lipids) with a diameter of 34 nm. The vesicle was filled with five enzyme molecules (a-chymotrypsin) in the presence of several crowders (bovine serum albumin) and multiple oligopeptide molecules, forming the substrate for the enzymatic reaction. The lattice Boltzmann molecular dynamics (LBMD) simulation technique, which we utilized, allows for describing proteins and lipids with a high level of molecular detail. At the same time, the technique significantly reduces the computational cost since it captures hydrodynamic effects, which strongly influence motions of molecules in a confined space, in an efficient yet accurate way using a lattice-based approach.
(Left) streaming of the fluid around the nanoreactor, (center) a cross section of the nanoreactor showing a cut through the lipid membrane and revealing the various molecules confined inside, (right) examples of our quantitative results on the distributions and motions of molecules in the nanoreactor; the dashed lines in the lower graph show a comparison with dilute conditions without confinement.
Our simulation, which we performed using up to 3375 processor cores at a time, allowed us to characterize the spatial distributions of the different molecules inside the vesicle as well as their diffusion coefficients. These factors govern the probability that an enzyme will meet a substrate and thus they affect the rate of the chemical reaction. Notably, we observed a tendency of proteins to diffuse along- and interact with the inner wall of the vesicle, which highlights the important role of the membrane surface in the nanoreactor. Moreover, we identified interaction patterns of the molecules inside the nanoreactor, and we characterized the sizes of transient protein clusters which formed during the simulation. In addition to the fully detailed model, we performed several simulations with a simplified representation of the membrane as a rigid wall, which allowed us to investigate the effects of switching between protein–wall attraction and repulsion.
The results of our project confirm the capacity of the LBMD technique to simulate protein diffusion in large and realistic systems consisting of many biomolecules surrounded by a membrane. Thus, the project opens the way for simulations of entire organelles or therapeutic constructs to capture the complex biological processes occurring inside them.