Is it possible to design proteins with specific characteristics

Stabilization of thermophilic proteins is achieved by negative and positive design working together, i. In particular, our recent analysis of complete bacterial proteomes [ 8 ] revealed that proteomes of thermophilic bacteria are enriched in both hydrophobic residues IVYLW and charged ones ER , while all polar residues are suppressed.

Discrepancies between different hydrophobicity scales [ 42 ], the statistical nature of knowledge-based Miyazawa—Jernigan potential [ 22 ], and limitations of the lattice model make it impossible to quantitatively compare the content of individual amino acids in lattice and natural proteomes or exactly predict the amino acid composition of thermophilic proteomes with very high accuracy from lattice model calculations.

The knowledge-based Miyazawa—Jernigan potentials, derived from native structures of proteins, are certainly a crude approximation to real protein energetics [ 43 ]. A question arises as to whether our observations are generic or are due to the specific potential used to design model proteins.

A detailed comparison of several potentials—all atom and group-based derived by different methods—was carried out recently in our lab [ 44 ]. Remarkably, we found that despite differences in detail all these potentials reflect the same dominant contributions to protein stabilization. It appears that dominant contributions to energy gaps in proteins come principally from two types of interactions: hydrophobic interactions and electrostatics [ 44 ].

Further, it was found that knowledge-based potentials derived using structures of meso- and thermophilic proteins are virtually indistinguishable KZ and ES, unpublished data. While positive design [ 45 ] is universally used in experiments, the role and omnipresence of negative design are still under discussion [ 46 ]. The main challenge in the study of negative design stems from the difficulties in the modeling of relevant misfolded conformations and energetic effects of mutations that destabilize them [ 46 ].

It was shown that charged residues can be effectively used in negative design [ 30 ]. Another indirect evidence of the contribution of charged residues to negative design emerges from site-directed mutagenesis, where mutations of polar groups to charged ones on the surface of a protein lead to protein stabilization even in the absence of salt—bridge partners of the mutated group [ 47 — 49 ].

In a series of experiments [ 47 , 48 , 50 ], surface electrostatic interactions were shown to provide a marginal contribution to stability of the native structure, hence their possible importance for making unfavorable high-energy contact in decoys.

An alternative view, proposed recently by Makhatadze et al. Our simulations and proteomic analysis point to a possible role of some surface charged residues as contributing to destabilization of misfolded structures through a negative design mechanism. Positive and negative elements of design affect the evolution of protein sequences. The dependence of substitution rates in sequences of natural proteins BLOSUM62 substitution matrix on interaction energies according to knowledge-based Miyazawa—Jernigan potential has a peculiar nonmonotonic shape showing elevated substitution rates between residues that attract each other as well as between residues that repel each other.

Upon substitutions, energy of attractive contacts in native states should be preserved as well as energies of specific repulsive contacts in misfolded conformations. Apparently both these factors act in concert to preserve the energy gap in proteins. Our study deepens an understanding of correlated mutations in proteins. With regard to native contacts, the fact that amino acids making strongly attractive native interactions should exhibit correlated mutations had been realized long ago.

Several authors proposed to use correlated mutations as a tool to determine possible native contacts from multiple sequence alignment [ 33 — 36 ]. However, this suggestion is complicated by the observation that correlated mutations are often found between residues that have no obvious functional role and are distant in structure [ 33 , 38 , 52 , 53 ]. Using the double mutant technique, Horovitz et al. In this work, we developed a simple exact model of thermophilic adaptation and discovered fundamental statistical—mechanical rules that Nature uses in her quest to enhance protein stability.

While many other factors, including dependence of hydrophobic and other interactions on temperature, certainly play a role in protein stabilization, the action of positive and negative design found and described here in a minimalistic model appears to be a basic universal principle determining evolution of sequences of thermostable proteins.

A better understanding of fundamental principles of protein design and stability makes it possible to decipher peculiar signals that emerge in the analysis of meso- and thermophilic genomes and proteomes [ 8 ] and in many studies of correlated mutations in proteins [ 33 , 35 , 53 ]. The residues interact with each other via the Miyazawa—Jernigan pairwise contact potential [ 22 ]. Note that if the energy spectrum E i is sparse enough at low energies, the value of P nat is determined chiefly by the energy gap E 1 — E 0 between the native state and the closest decoy structure that has no structural relation to the native state.

To design lattice proteins, we use here a Monte-Carlo procedure P-design, [ 14 , 15 ] that maximizes the Boltzmann probability P nat of the native state by introducing mutations in the amino acid sequence and accepting or rejecting them according to the Metropolis criterion. As this procedure takes the environmental temperature T env as an input physical parameter, and generates amino acid sequences designed to be stable at T env , it is an obvious choice for modeling the thermophilic adaptation.

At each Monte-Carlo step, a random mutation of one amino acid in a sequence is attempted and P nat of the mutated protein is determined. The native structure is determined at every step of the simulation; generally, the native state changes upon mutation of the sequence.

The design procedure is stopped after 2, Monte-Carlo iterations. Such length of design runs is sufficient to overcome any possible effects of the initial composition of the sequences, so the amino acid composition of the designed sequences depends only on the environmental temperature T env. To relate the trends in amino acid composition with the physical properties and interaction energies of individual amino acids, we use hydrophobicity as a generic parameter characterizing an amino acid [ 42 ].

To characterize the hydrophobicity of amino acids in the simulations, we make use of the fact that the Miyazawa—Jernigan interaction energy matrix is very well approximated by its spectral decomposition [ 43 ]. In this representation, hydrophobic residues have the largest values of q, while hydrophilic charged residues correspond to small q.

All sequences of TIM—barrel folds with length less than amino acid residues were extracted according to the SCOP database description [ 56 ]. Identical sequences were excluded from further consideration. Remaining sequences total 39 were aligned against the sequence of the triosephosphate isomerase 7tim.

The CRASP program gives the correlation coefficient between the values of physicochemical parameters at a pair of positions of sequence alignment. We chose hydropathy [ 39 ] as a physicochemical characteristic appropriate for establishing correlated mutations of interest. Only significant correlations, with the correlation coefficient higher than the critical threshold 0.

See Table S1 for optimal growth temperatures and references. We thank George Makhatadze for useful correspondence.

Abstract The aim of this work is to elucidate how physical principles of protein design are reflected in natural sequences that evolved in response to the thermal conditions of the environment. Author Summary What mechanisms does Nature use in her quest for thermophilic proteins?

Abbreviations: OGT, optimal growth temperature. Introduction Despite recent advances in computational protein design [ 1 ], there is no complete understanding of basic principles that govern design and selection of naturally occurring proteins [ 2 ]. Results We design lattice model proteins with selected thermostability as a first step toward modeling thermal adaptation of organisms.

Download: PPT. Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Schematic Illustration of the Concept of Mutations by Swaps. Figure 9. Figure Discussion Stabilization of thermophilic proteins is achieved by negative and positive design working together, i. Supporting Information. Figure S1. Figure S2. Just as water flows downhill until it finds the lowest point, proteins tend to fold into their lowest energy conformation.

If we understand 75 to 80 percent of the problem, we focus on those parts. Years of tackling increasingly difficult computational problems have helped IPD to reach a major inflection point. In the near-term, the collaboration will seek to apply de novo design to expand the versatility of traditional protein-based medicines. If IPD can create a new methodology to display these loops in their native shape, we could use that to design better immunogens and make more effective antibodies against some very promising drug targets.

Ray Deshaies left and David Baker. The two once worked in the same room when they were both grad students at the University of California at Berkeley. In this method the status of each amino acid is predicted according to its place in the type of secondary structure and the certainty of this prediction is displayed with numbers from zero to nine. Figure 2 shows an example of the second structure of a recombinant protein.

In determining the tertiary structure of recombinant proteins, first to determine the pattern, blasts against pdb from phyre server is done which at this website, the structure with the greatest homology is selected as a template. Then the recombinant protein is matched on a natural protein. This number is between zero and one, and as the amount of RMSD is closer to zero, it indicates that the mutation has not changed the overall structure of recombinant protein and these proteins are matched on each other [8].

Second structures of proteins are created by Sai and Phi angles and Ramachandran diagram shows the authorized status of each angle for protein structures. As Absolute magnitude of Z-Score is closer to 10, the structure has a higher quality and the more negative energy profile shows a higher structure quality [2]. For predicting the function of recombinant protein, the interaction of various ligands with the recombinant and natural protein is investigated using MVD Molegro Virtual Docker software [11].

As a result of this ligand-protein connection by providing the MolDock score, interaction is estimated. MolDock, Escore is defined by the following energy expressions:. In this regard Einter is the protein-ligand interaction energy and Eintra shows the ligand internal energy. The internal energy of ligand here is the same for each ligand and the effective energy in this equation depends on the energy of ligand-protein interactions and as this energy is less more negative the stability of the substrate at the active site is done better and substrate proteins are more stable.

In studying the recombinant protein interactions with various ligands, it reveals that according to predictions done, how much is the required energy for recombinant and natural protein. Thus, we can predict the recombinant proteins function comparing to natural proteins Especially in the case of enzymes. Today, knowledge of enzymology has created profound changes in biotechnology industries. Until the 60s, the income of industrial enzymes was only a few thousand dollars a year, but by the growth of this industry in recent years, this income has increased.

Today, most of enzymes are prepared by fermentation of bio-based materials. At the University of Birmingham, UK, for instance, chemist Anna Peacock studies metallopeptides — miniature proteins that bind metal ions. But other metals could enable different chemistry. Peacock has used de novo proteins as scaffolds to create molecules capable of binding gadolinium, complexes of which are commonly used as contrast agents for magnetic resonance imaging.

She is also crafting enzymes that can use metals such as platinum or iridium to explore reactions not found in nature. As each design goal is achieved, it becomes easier for others to emulate them. The Baker lab has even developed an online gaming interface to Rosetta, called FoldIt, that challenges players few of whom are scientists to create proteins in silico.

In a study this year analysing their work 7 , the players delivered. Few scientists have the time or expertise to design a protein from the ground up, of course; for them, de novo designs are foundations to build upon.

But in the Baker lab, the design work continues. With each success, the lab celebrates. Kuhlman, B. Science , — PubMed Article Google Scholar. Hosseinzadeh, P. Chen, Z. Marcandalli, J. Cell , — Dou, J. Nature , — Mravic, M. Koepnick, B. Download references. Article 03 NOV Article 27 OCT Research Highlight 22 OCT Research Highlight 02 JUN Technology Feature 03 MAR Article 10 NOV University of Vienna.

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.