Structure-based redesign of the binding specificity of anti-apoptotic Bcl-x(L)

Abstract

Many native proteins are multi-specific and interact with numerous partners, which can confound analysis of their functions. Protein design provides a potential route to generating synthetic variants of native proteins with more selective binding profiles. Redesigned proteins could be used as research tools, diagnostics or therapeutics. In this work, we used a library screening approach to reengineer the multi-specific anti-apoptotic protein Bcl-x(L) to remove its interactions with many of its binding partners, making it a high-affinity and selective binder of the BH3 region of pro-apoptotic protein Bad. To overcome the enormity of the potential Bcl-x(L) sequence space, we developed and applied a computational/experimental framework that used protein structure information to generate focused combinatorial libraries. Sequence features were identified using structure-based modeling, and an optimization algorithm based on integer programming was used to select degenerate codons that maximally covered these features. A constraint on library size was used to ensure thorough sampling. Using yeast surface display to screen a designed library of Bcl-x(L) variants, we successfully identified a protein with ~1000-fold improvement in binding specificity for the BH3 region of Bad over the BH3 region of Bim. Although negative design was targeted only against the BH3 region of Bim, the best redesigned protein was globally specific against binding to 10 other peptides corresponding to native BH3 motifs. Our design framework demonstrates an efficient route to highly specific protein binders and may readily be adapted for application to other design problems.

Figure 1

Library design. (A) The computational library design protocol. In stage 1, non-disruptive and specific mutations were predicted at every design position using structural modeling with Rosetta; in stage 2, ILP optimization was performed to select degenerate codons that maximized the inclusion of predicted non-disruptive mutations while enforcing the inclusion of predicted specific mutations, under a library size constraint. (B) The interface between Bcl-x_L and a Bad BH3 peptide (PDB ID: 2BZW), with Bcl-x_L in cyan and the peptide in green sticks (PyMol, Delano Scientific). Designed positions for the first library are in blue and those for the second library are in red. (C) Sequence alignment of the BH3 regions of Bim (residues 142 to 169) and Bad (residues 104 to 131). Non-identical peptide positions used to select Bcl-x_L sites for library design are colored in blue (for the first library) or red (for the second library). BH3 positions were numbered using a heptad convention as shown at the top.

Figure 2

Screening results for the first designed library. (A) Flow cytometry plots showing binding of native Bcl-x_L to Bad (top) and Bim (bottom) at 1 nM. Expression and binding signals are plotted on the horizontal and vertical axes, respectively. (B) Flow cytometry plots showing binding of the first designed library to Bad and Bim at the concentrations indicated. (C) Flow cytometry plots showing binding of the population obtained from 6 rounds of sorting of the first designed library to Bad and Bim at the concentrations indicated. (D) Sequence frequency plot for 21 unique sequences identified as specific for Bad over Bim BH3 in the first designed library (Table S1), with the native Bcl-x_L residue shown below each column. Plots were generated using WebLogo.

Figure 3

Screening results for the second designed library. (A) Flow cytometry plots showing binding of the population obtained from 2 rounds of sorting of the second designed library to Bad (top) and Bim (bottom) at the concentrations indicated. Expression and binding signals are plotted on the horizontal and vertical axes, respectively. (B) Flow cytometry plots showing binding of the population obtained from 7 rounds of sorting of the second designed library to Bad and Bim at the concentrations indicated. (C) Sequence frequency plot for 28 unique sequences (Table S2) identified as specific for Bad over Bim BH3 from the second designed library after two rounds of screening. Note that position 142 was not varied in this library but is a change from wild-type Bcl-x_L.

Figure 4

Fluorescence polarization experiments characterizing Bcl-x_L and its variants binding to BH3 peptides derived from Bim or Bad. (A, B) Competition of Bim-28 or Bad-28 with fBad-23 for binding to native Bcl-x_L (A) or RX1 (B). The average values of two independent measurements are plotted as a function of competitor peptide concentration; error bars are the standard deviations for 2 values. (C) The influence of point mutations in Bcl-x_L on binding to Bim or Bad BH3. Ratios of the K_i values for Bcl-x_L point mutants binding Bim-22 (white) or Bad-22 (gray) to the K_i values for Bcl-x_L binding the same BH3 peptide are shown. (D) Mutational effects in RX1, as in (C). Peptides used were Bim-28 (white) and Bad-28 (for RX1 mutants L105F, I122S, V125Q, G142A) or Bad-22 (for RX1 mutants H101Y, G111Q, A126V, I130L, A146F) (gray). The ratio for RX1-V125Q binding to Bim-28 is marked with an asterisk (*), because only a lower bound on the K_i value could be determined (Table S7). Binding conditions are described in Materials and Methods, and peptide sequences are given in Table S3.

Figure 5

Fluorescence polarization experiments characterizing Bcl-x_L and RX1 binding to 10 BH3 peptides from human proteins. (A, B) Competition of BH3s with fBad-23 for binding to native Bcl-x_L (A) or RX1 (B). The average values as well as the standard deviations of two independent measurements are plotted as a function of competitor peptide concentration. In (B), competition curves are shown only for competitor peptides that bound Bcl-x_L significantly. K_i values are listed in Table S5.