Specific coiled coil assembly through size-complementarity

Abstract

The α-helical coiled coil is a ubiquitous protein structural motif formed by the supercoiling of two or more α-helices. The high regularity in amino acid sequence renders it particularly attractive from a molecular recognition standpoint. Specifically, coiled coil primary structure exhibits a three-four hydrophobic repeat known as the heptad repeat, with residues designated abcdefg. The hydrophobic core is comprised of nonpolar a and d residues while core-flanking e and g positions contribute to electrostatic interactions at the hydrophilic interface. Positions b, c, and f are solvent exposed and generally contain amino acids with high helix propensity. Hydrophobic contacts are widely considered the main driving force in strand association. We sought to establish an approach toward specificity in the hydrophobic core using a non-natural amino acid, cyclohexylalanine (Chx), as a steric complement to two opposing alanine (Ala) residues. A specific 2:1 heterotrimer was demonstrated where a central core layer juxtaposed these residues, thereby specifying both strand orientation and stoichiometry. The isolated Chx peptide, however, displayed higher stability than the designed complex, which would become a problem when more demanding assembly problems were addressed. To circumvent this, a cooperative approach was employed whereby three consecutive core residues were modified to either Ala or Chx such that each strand of a 1:1:1 heterotrimer donates a single large Chx residue to the core structure. All competing assemblies suffer from at least one destabilizing all-alanine core layer. Analysis of the designed system and all component strands revealed the 1:1:1 heterotrimer as the most stable species formed. With a well-behaved core structure in hand, efforts turned toward manipulation of hydrophilic interface residues which could serve as an orthogonal recognition domain. Persubstitution of these positions with either glutamic acid (Glu), lysine (Lys), or a combination of the two resulted in two stable complexes at neutral pH. The first contained a core alignment as described above with electrostatically matched hydrophilic interfaces. The second bore the same core structure but suffered from Lys-Lys interactions at a single interface. At lower pH, a similar complex could be obtained bearing a Glu-Glu interface. Subjection of the latter two complexes to varying pH revealed differential stability at pH extremes which prompted us to consider the possibility of specific strand exchange. Starting with the Lys-Lys complex, it was shown that adding an appropriate acidic peptide at low pH could effect formation of a specific Glu-Glu complex by ejection of a basic component strand. We then demonstrated sequential exchange of either two or all three component strands from an initial assembly. The exchange strategy was then exploited in both the formation and dissolution of a set of crosslinked trimers. Addition of a preformed disulfide-linked dimer to a susceptible trimer at varied pH produced a pentamer containing the newly introduced linkage. Reversal of the process was achieved by addition of monomeric peptide to a preformed pentamer at appropriate pH. This process represents a novel approach to chemical crosslinking where the covalent bond is introduced via an exchange process. This level of control is only achievable through incorporation of orthogonal recognition elements. Having demonstrated size complementarity as a viable assembly mechanism, we sought to extend complexation-control to include orientation state specificity. By simply reversing the sequence of a peptide from previous dual-interface work, we were able to successfully design an antiparallel heterotrimer. Equilibrium disulfide-exchange experiments provided strong evidence for the expected strand arrangement. Direct competition between analogous parallel and antiparallel structures resulted in an initial kinetic preference for the antiparallel orientation. However, given sufficient equilibration time, a slight thermodynamic preference for the parallel arrangement appeared. With this work, we were now able to specify both oligomerization and orientation through simple steric matching of hydrophobic core side chains, lending access to the entire series of native-like, trimeric coiled coil structures. More recently, we have focused efforts on understanding coiled coil surface-binding events. Several viral entry mechanisms have been shown to include helix bundle formation through peptide binding at a hydrophobic pocket presented on a coiled coil exterior. Our designed heterotrimer would serve as an appropriate model system to test various design requirements to effectively inhibit this event. The N-terminal portion of gp41, an HIV envelope protein, forms a parallel, trimeric coiled coil. Binding of three C-terminal strands to the exterior of the trimer results in a six-helix bundle thought to facilitate host-cell infection. We designed an N-terminal trimer mimic which contained a single hydrophobic binding interface to study C-terminal peptide affinity. Analysis of N- and C- terminal peptide mixtures resulted in strong evidence suggesting our heterotrimer as an appropriate gp41 mimic. Control experiments verified a stable and specific binding event between a short C-terminal gp41 peptide and the exterior of our N-terminal gp41 trimer mimic. With initial characterization of this complex completed, future experiments aimed at uncovering important interactions required for coiled coil surface-binding may be executed.