Ivan Coluzza

My research aims at the understanding of the design and the folding of proteins. With computer simulations I study what allows proteins to use just about 20 building blocks to assemble the most variable and complex structures in nature. Proteins are also sophisticated molecular machines, that can walk push pull and even trap molecules and other proteins. An important goal of my work is to understand the function of specific proteins and design new molecular machines. Finally, I work on the construction of artificial chains of particles that can be designed and fold just like proteins, but are much simpler to control and synthesized

Is the Caterpillar model a solution to the protein folding problem?

In a previous publication we demonstrated that the caterpillar model was capable of correctly fold designed protein sequences into naturally occurring protein structures. Now we refined the model under the condition that the natural sequences were contained in the ensemble of designed sequences. With the new parameters we not only produced sequences for 120 proteins with similar energies and hydrophobic/philic profiles than the natural ones, but we also correctly refolded over 15 proteins to their natural native structure. By refolded we mean that the natural structure corresponded to the minimum of the free energy along the collective variable distance root mean square deviation (DRMSD) (see to figure). In other words the PDB structure is the most probable. We have now a quantitative model for protein folding and design and we have now strong evidence that the minimum constraint principle is a universal condition for heteropolymer designability (see patchy polymer publications). In other words we can fold real proteins, we can reproduce the natural sequences, and we can produce artificial proteins (see patchy polymer publications), so the caterpillar model is a possible answer to the protein folding problem. This work has just been submitted.