Octarellin VII A New Generation of de Novo Designed (β/α)8-Barrel Proteins

Abstract

De novo protein design is a branch in the field of protein engineering that started at the end of the ’70s. Following many successful cases it developed quickly, and nowadays more than 100 papers are published every year on the subject. The number of publications per year on “de novoprotein design” according to PubMed is reported in Figure 1.1.

Despite its large diffusion, it is quite difficult to describe and explain what exactly protein design is. In general, everything that brings something new (from Latin, de novo) in the field of protein science can be classified in this branch: the creation of a new protein topology [7], a new enzymatic activity [8] or an amino acid sequence that is not present in nature [9]. However, the boundaries between protein design and protein engineering are blurred. For example, a single point mutation in a protein might yield a sequence that is not present in nature, but it will not be defined as de novo design. The lack of clear definitions and boundaries is mainly due to its recent development. Moreover the definition of “de novo”is time-dependent: designs and methodologies of the ’80s may be considered obsolete nowadays thanks to the acquired knowledge and the development of new technologies.

1.1.1 The origins of the de novo design

The scientific community generally attributes to Brend Gutte the beginning of the de novodesign era. In fact, he is considered as the father of the field since he published three publications between the end of the ’70s and the beginning of the ’80s.

The first article, published in 1975, reports the synthesis by solid phase method of an analog of Ribonuclease S [10]. The wild-type protein is 124 residue long and presents a wide loop on its surface. In order to study the importance of loops in protein folding, Gutte reduced Ribonuclease S to a 70-residue analog, which misses 5 loops not involved in the enzymatic activity. Surprisingly, despite the removal of 54 residues, the first “de novo” protein retained 4% of the enzymatic activity and specificity of the wild type enzyme.

The second article was published in 1979 [11]. The group wanted to design an artificial peptide with nucleic acid binding activity. They designed a 34-residue peptide in two steps: first they determined that the minimal structure for DNA-binding should contain a β-strand, a reverse turn, an anti-parallel β-strand, another reverse turn and an α-helix (shown in Figure 1.2). Following the design of the backbone model, they applied the rules for secondary structure prediction in order to find the best amino acid sequence to fit the model.

1.1. DE NOVO DESIGN

The peptide was produced by solid phase synthesis and showed binding to single strand DNA. Furthermore, the dimeric peptide also showed ribonuclease activity. Despite the lack of a 3D structure to confirm the design, the results that were obtained at that time are extremely remarkable, especially considering that: 1- computers were not used, 2- the number of available pdb structures was limited (42 according to RCBS-PDB) and 3- the advances in molecular biology, such as recombinant DNA, cloning, sequencing or DNA synthesis were not well-established.