Context-Dependent Folding: Sequence-Encoded Strategies for Stabilizing a Protein Subdomain in Isolation

Abstract

Many proteins are made up of smaller structural units called subdomains. Dissecting proteins into these small units is important for protein engineering and design. In part, the ability to create novel fusion proteins relies on a subdomain's ability to fold autonomously. Subdomains are usually identified as regions that have high contact density, in other words, they have more residue-residue contacts within a specified region than they do with other areas of the protein. Due to their high contact density, subdomains are often assumed to be able to fold in isolation. We evaluate this hypothesis, by examining the foldability of a subdomain from T4 Lysozyme (T4L). In full-length T4L, the N-terminal subdomain has high contact density yet its sequence does not fold in isolation. This suggests that high contact density is not always enough for subdomain folding. To explore this, we have extended the boundary to the complete the sequence of a helix found in the full-length protein that has no contacts with the rest of the subdomain. We find that even without increasing predicted contacts, the sequence is able to fold autonomously. To determine if the sequence of the helix is important, we created a fusion protein that appends an alanine-based helix, which is able to fold in isolation, to the original N-terminal subdomain sequence. Surprisingly, we find that this structural mimicry allows the sequence to fold. In total, our findings point to the importance of considering not just considering the contact density when choosing boundaries of subdomains. We also demonstrate strategies to stabilize the folding of otherwise unfolded subdomain sequences.