De novo design of helical bundles as models for understanding protein folding and function

Abstract

De novo protein design has proven to be a powerful tool for understanding protein folding, structure, and function. In this Account, we highlight aspects of our research on the design of dimeric, four-helix bundles. Dimeric, four-helix bundles are found throughout nature, and the history of their design in our laboratory illustrates our hierarchic approach to protein design. This approach has been successfully applied to create a completely native-like protein. Structural and mutational analysis allowed us to explore the determinants of native protein structure. These determinants were then applied to the design of a dinuclear metal-binding protein that can now serve as a model for this important class of proteins.

FIGURE 1

Hypothetical free energy diagram for a protein. Each line or bar represents a distinct conformational state, with the native state as the lowest energy state and the unfolded states a densely populated, nearly isoenergetic ensemble. Between these extremes are non-native folded states, oftentimes referred to as the molten globule ensemble. These non-native folded states exhibit extensive secondary structure but lack well-defined tertiary structure. The population of each state is dictated by the Boltzmann distribution. For native protein structure the free energy gap, Δ, must be large enough to significantly populate a distinct native state. For natural proteins Δ has evolved to be much larger than in de novo-designed proteins. Thus, a single mutation in designed proteins offers the advantage of being able to access the non-native states in a manner not often observed with natural proteins.

FIGURE 2

A dimeric, four-helix bundle can adopt six distinct topologies. The topology diagrams shown are for the three possible clockwise-turning helical bundles. Counterclockwise-turning helical bundles are also possible and have interfacial interactions different from their clockwise-turning correlates.

FIGURE 3

Amino acid sequences of the α₂ family, reflecting the hierarchical approach to protein design. Each peptide is comprised of 35 residues with the N-terminus acetylated and the C-terminus amidated. α₂B is comprised solely of leucine residues in the hydrophobic core positions. α₂C is comprised of a more diverse set of nonpolar and aromatic residues in the hydrophobic core positions. α₂D has three additional changes at positions 7, 26, and 30.

FIGURE 4

The bisecting U motif is a common structural motif. α₂D was the first protein in which the bisecting U motif was recognized. (A backbone overlay of the 10 lowest energy NMR structures is shown.) Subsequently, this motif has been identified in a wide variety of proteins, including other dimeric helical bundles (FIS DNA-binding protein and Rop mutant68), intramolecularly folded helical bundles (FAS death domain and N-terminal domain of the δ subunit of the ATPsynthase118), and β-sheet proteins (the IgG fold119), and multimeric proteins with both α/β structures (p53 tetramerization domain120).

FIGURE 5

α₂D has three nonequivalent interfaces: (a) Interdigitated Leu residues stabilize the interface between helices 1 and 1′. (b) A diverse collection of aromatic and hydrophobic residues stabilize the interface between helices 1 and 2′. (c) Hydrogen-bonded clusters of His residues stabilize the interface between helices 2 and 2′.

FIGURE 6

Glu7 is not involved in stabilizing the native state of α₂D. Superposition of the 10 lowest energy NMR-derived structures (left panel) shows that Glu7 adopts multiple conformations typical of solvent-exposed residues (hydrophobic core residues in gray). Replacing Glu7 with other polar or nonpolar residues has little effect on the overall stability (ΔG) of the protein but has a large effect on the conformational specificity (Δ) of the protein (right panel). For example, replacing Glu7 with a polar residue (as in E7H) results in a protein with a Δ large enough to adopt a native structure. However, replacing Glu7 with a hydrophobic residue (as in E7V) results in a protein that has lost its ability to adopt a native structure. Note that the number of conformations that E7V adopts cannot be too large because this would result in a large entropic stabilization of this ensemble, which is not observed. In conclusion, Glu7 is involved in “negative design” by increasing Δ through the destabilization of the non-native folded states relative to the native state. We assume that the energy level of the unfolded ensemble is not significantly affected by these substitutions.

FIGURE 7

Hydrogen-bonding within a His cluster is essential for the native state of α₂D. An interior His (H26) forms a hydrogen bond across the interface with an exterior His (H30′) on the other monomer (left panel). A symmetrical interaction also occurs between H26′ and H30. Substitution of residues that are unable to form hydrogen bonds (as in H30K) results in a protein that has lost all ability to adopt the native structure. Substitution of residues that are capable of forming hydrogen bonds (as in H30D) results in a protein that maintains the native structure (right panel). Note that the non-native states of H30K are depicted slightly lower in energy than those of α₂D because Lys is known have a higher helical propensity than His. In conclusion, the His cluster is involved in “positive design” by increasing Δ through the direct stabilization of the native state relative to the non-native folded states.

FIGURE 8

X-ray structure of the di-Zn(II) form of DF1 (2.5 Å resolution), which is nearly identical to the intended design. The backbone of the structure plus the ligands are shown in the two views. At left, a Tyr phenolic group hydrogen bonds to a Glu carboxylate (a second, symmetry-related Tyr-Glu interaction is not shown for clarity). At right, hydrogen bonds between Asp carboxylates and His side chains are shown.