Structural insights for engineering binding proteins based on non-antibody scaffolds

Abstract

Engineered binding proteins derived from non-antibody scaffolds constitute an increasingly prominent class of reagents in both research and therapeutic applications. The growing number of crystal structures of these 'alternative' scaffold-based binding proteins in complex with their targets illustrate the mechanisms of molecular recognition that are common among these systems and those unique to each. This information is useful for critically assessing and improving/expanding engineering strategies. Furthermore, the structural features of these synthetic proteins produced under tightly controlled, directed evolution deepen our understanding of the underlying principles governing molecular recognition.

Figure 1. General flow of a protein-engineering project using a non-antibody scaffold

A schematic is shown depicting a positive feedback loop linking structural characterization and combinatorial library design.

Figure 2. Non-Antibody Scaffold Systems and Library Designs with Available 3D Structures

On the left, scaffolds are depicted with positions diversified in combinatorial libraries indicated by red spheres. On the right, structures are shown that depict distinct modes of target interaction. Targets are shown as gray surface models. The target protein name, and PDB ID for each structure are indicated. In the DARPin system, scaffolds based on three (left top) or two (left bottom) diversified ankyrin repeat units have been used [4]. Example structures of binders derived from both scaffold types are shown. In the Anticalin system, libraries directed at the production of protein binders [19] (left top) or small molecule binders [17, 33] (left bottom) have been used. A representative example of each type of complex is shown. In the monobody system, two library design strategies have been used [9, 12]. (B) Structures of the Monobody and DARPin scaffolds are shown with positions colored according to the frequency with which they contact targets in the available structures Contact is defined by burial of > 1 Ų of surface on complex formation. Positions diversified in typical Monobody and DARPin libraries are indicated by spheres. For monobodies only binders selected from conventional loop-based libraries were considered (PDB IDs 3QHT, 3RZW, 3K2M, 3CSB and 3T04). Also, because monobody loop lengths vary, positions from each variable loop were considered and colored as a group. For DARPins only binders based on the 3-ankyrin repeat scaffold were considered (PDB IDs: 2P2C, 2BKK, 2J8S, 1SVX and 2V5Q). Surface burial calculated using the Protorp server [37].

Figure 3. Amino Acid Usage in Target Contacts and Structural Parameters for the Target-binding Surfaces

(A) Bar graphs of the amino acid compositions in typical naïve libraries (white bars), target contacting positions in the structures of binder/target complexes (black bars), and target contacting positions that were also diversified in the library (gray bars). Contact is defined by burial of > 1 Ų of surface on complex formation. Naïve library compositions shown are those encoded by the mutagenic oligonucleotides used to create typical libraries. Data are shown for the Monobody, DARPin and Affibody scaffolds. In the case of DARPins, 2XZT (point mutant of 2XZD), 3NOC and 3NOG (recognize a nearly identical epitope to 2J8S) were excluded to avoid introducing artificial bias. In the case of monobodies, 3CSG was excluded because it was isolated from a Tyr/Ser-only library. 2OTK was excluded in the case of affibodies (highly atypical structure). For 2B87, the lowest energy model (model 1) was used for analysis. Surface burial calculated using the Protorp server [37] (B) Scatter plot of the surface area buried by binders upon complex formation vs. the K_d values reported for those complexes. Surface burial calculated using the Protorp server [37] (C) Scatter plot of the surface complementarity (S_C) value for the interfaces of binder-target complexes vs. the K_d values measured for those complexes. S_C values calculated using the SC program in the CCP4 suite [23, 38]. For monobodies all structures except 2OCF were considered (no reported Kd value). For DARPins all structures except for 2XZD and 2XZT were considered (no reported Kd values). For anticalins, only the 2BX7 structure was considered since all other structures are complexes with small molecules. For affibodies, all structures were considered and for 2B87, the lowest energy model (model 1) was used for analysis.