>Bcl-xL 1-209 with 9 mutations (underlined)
BH3 peptide complex partners [GK-should we list others?]
>human BAD (28-mer; ~1 nM affinity)
>human BID (20-mer; >2 μM affinity)
(human, 1-26,83-209): Bcl-xL/foldamer (2YJ1, 3FDM, 4A1U, 4A1W), Bcl-xL/Bim (3FDL), BclxL/BimL12F(3IO8) [Note: this construct forms artificial domain-swapped dimers] (human, loop 45-84 deletion): Bcl-xL/Bax (3PL7) [Note: this structure forms an unusual ternary complex with 2:1 Bcl-xL:Bax]
(human, loop 27-82 deletion): Bcl-xL/Beclin (2P1L), Bcl-xL/Soul (3R85) [Note: this construct forms artificial domain-swapped dimers]
(mouse, 1-211) Bcl-xL/Bad (2BZW), missing loop in structure
(mouse, 1-196) Bcl-xL/Bim (1PQ1), missing loop in structure
(human, loop 45-84 deletion): Bcl-xL/Beclin (2PON)
(human? loop 45-84 deletion): Bcl-xL/Bak (1BXL)
(human? loop 49-88 deletion): Bcl-xL/Bad (1G5J)
The anti-apoptotic Bcl-2 proteins have a globular, helical fold and bind to short helical peptides derived from pro-apoptotic proteins, here called BH3 peptides. These interactions are highly selective and play an important role in cellular apoptosis. A comprehensive understanding of the structural basis behind the selectivity of these interactions are of great interest for developing reagents for studying apoptosis or therapeutics for targeting relevant diseases. The anti-apoptotic protein Bcl-xL interacts strongly with BH3 peptides including Bim, Bad, Bid, Hrk, Bik, Bak, PUMA and Bmf.
Combining structural modeling with experimental library screening, we redesigned Bcl-xL such that it retained tight binding to Bad (Kd < 1 nM) but lost strong interactions to other BH3 peptides (Kd > 1 μM) with the exception of PUMA (Kd ~ 40 nM). Compared to native Bcl-xL, the designed protein showed increases in specificity of > 1,000 fold
for binding Bad over several other BH3s [GK- by what means were these binding affinities obtained?]. A total of 9 mutations were present in the designed protein. Interestingly, analysis of binding behaviors of 9 individual point mutants made in the sequence context of Bcl-xL revealed that their effects on binding Bad were highly non-additive.
Examining available structures of complexes between Bcl-xL and Bad or other BH3s did not provide obvious reasons for the non-additive contributions. However, the α2-α3 regions (two Bcl-xL helices at the interface connected by a short loop) were shown to be highly variable among these different structures, and several of the important specificity mutations are located in this region. Based on these observations, it is highly probable that the designed protein adopts a unique conformation in this region that specifically favors binding Bad over other BH3s. This can help explain the non-additive contributions from individual residues when measured in the Bcl-xL sequence context. A solved structure between the designed protein and Bad can therefore not only help rationalize the molecular mechanism behind the unique global specificity profile for our designed protein, but also advance our understanding of the capacity of conformational flexibility at the Bcl-xL interface. Such knowledge could shed light on how flexibility at protein interfaces can help accommodate different peptide ligands to achieve novel binding specificities. Published crystal structures of human Bcl-xL have a loop deletion that leads to formation of a domain-swapped dimer or other unusual interactions. We are considering trying to solve an x-ray structure, but NMR seems the safer route to examining the designed protein complex.|