General Methodology to Identify the Minimum Alphabet Size for Heteropolymer Design

Abstract

Understanding how to design the structure of heteropolymers through their monomer sequence will have a significant impact on the creation of novel artificial materials. According to mean‐field theories, the minimum number—or alphabet—of distinct monomers necessary to achieve such designability is directly related to the conformational entropy ω of compact polymer structures. Here, a computational strategy to calculate this conformational entropy is introduced and thus predict the minimum alphabet to achieve designability, for a generalized heteropolymer model. The comparison of the predictions with previous results proves the robustness of the approach. It is quantified for the first time how the number of directional interactions is critical for achieving the designability. The methodology that is introduced can be easily generalized to models representing specific polymers. A comparison between conventional polymers monomers are provided, and it is predicted that polyurea, polyamide, and polyurethane residues are optimal candidates to be functionalized for the experimental synthesis of designable heteropolymers. As such, our method can guide the engineering of new types of self‐assembling modular polymers, that will open new possibilities for polymer‐based materials with unmatched versatility and control.

Identification of Protein Functional Regions

“…The first evidence to support our claim comes from our previous work on heteropolymer design including the Caterpillar design Our work on design showed that provided that a heteropolymer chain is designable (we defined the rules to identify such property) then the 3D structures can be designed with high accuracy independently of the interaction matrix used to define the amino acid interactions . In fact, the same design strategy works for lattice and off‐lattice proteins with implicit or explicit solvent, plus the above mentioned patchy polymers . This result is the first indications that the key correlations that determine the folding do not depend on the particular model used to represent the residue interactions.…”

“…This property means that also that any model or force field capable of generating sequences and refold them into the natural backbone structure would be usable for the purpose of our analysis. The first evidence to support our claim comes from our previous work on heteropolymer design including the Caterpillar design Our work on design showed that provided that a heteropolymer chain is designable (we defined the rules to identify such property) then the 3D structures can be designed with high accuracy independently of the interaction matrix used to define the amino acid interactions . In fact, the same design strategy works for lattice and off‐lattice proteins with implicit or explicit solvent, plus the above mentioned patchy polymers .…”

“…This result is the first indications that the key correlations that determine the folding do not depend on the particular model used to represent the residue interactions. The only requirement, of course, is that the protein structural space is correctly represented and for that, we can bring not only the evidence produced by the Caterpillar model it‐self but also made with its close cousins: the tube model of Maritan et al . and the CamTube model .…”

“…Hence, the design is not sensitive to limits in the experimental resolution below 3 Å which for X‐ray structure prediction is feasible. The second indication that prediction does not depend on the particular choice of the amino‐amino acid interactions (provided that are heterogeneous for instance according to a Gaussian distribution) is given by the DCA it‐self. In fact, the direct couplings are nothing else then specific residue‐residue interactions that reproduce the correlations functions measured over the evolution process.…”

In Silico Evidence That Protein Unfolding is a Precursor of Protein Aggregation

“…The analysis of the 3D version is object of current systematic investigation, with preliminary results that confirm our findings qualitatively in 2D: (i) we checked that the water model in 3D is in agreement with our current understanding of the bulk water phase diagram; (ii) we verified that the protein folding analysis performed in 2D can be extended with similar results in 3D; (iii) we also extended the protein design results to 3D with preliminary results consistent with those in 2D . Since the design and folding in 2D is easier than in 3D because the conformational space is smaller, the unfolding event we discuss in this work should be even easier to find in 3D. We then expect that our results would be not only confirmed in a 3D model but would be even stronger.…”