Stability and interface organization of a large homodimeric enzyme, and engineering of allosteric communication between its two active sites

One of the biggest challenges in structural biology is to understand how the primary sequence of a protein codes for its three-dimensional structure. The folding problem becomes extremely complex for multimeric proteins because the polypeptide chains not only need to fold properly, but also form a network of intersubunit interactions. Most protein folding and protein/protein interactions thermodynamic studies have been limited to very small systems. Relatively few average size protein complexes have been characterized in depth. It is generally accepted that the structural mechanisms that stabilize small complexes are also valid for larger assemblies; however, it is necessary to validate this hypotheses with extensive empirical data.;The average size Escherichia coli aspartate aminotransferase (eAATase) homodimer (87 kDa) contains multiple structural elements that are not generally found in small systems: two pyridoxal 5'-phosphate (PLP) dependent intersubunit active sites, multiple domains (2 per monomer), and a non-continuous interface composed of two small and one large surfaces. The guanidine hydrochloride (GdnHCl) mediated denaturation pathway for the apoenzyme includes a partially folded monomeric intermediate, M* [Herold, M., and Kirschner, K. (1990) Biochemistry 29, 1907-1913; Birolo, L., Dal Piaz, F., Pucci, P., and Marino, G. (2002) J. Biol. Chem. 277, 17428-17437]. The present investigation of the urea mediated denaturation of eAATase finds no evidence for an M* species, but uncovers a partially denatured dimeric form, D* that is unpopulated in GdnHCl. Thus, the unfolding process is a function of the employed denaturant. D* retains less than 50% of the native secondary structure (circular dichroism), conserves significant quaternary and tertiary interactions, and unfolds cooperatively at ∼ 5 M urea. Therefore, the following equilibria obtain in the denaturation of apo-eAATase: D &rlarr2; 2M &rlarr2; 2M* &rlarr2; 2U in GdnHCl and D &rlarr2; D* &rlarr2; 2U in urea (D = native dimer, M = folded monomer and U = unfolded state). The free energy of unfolding of apo-eAATase (D &rlarr2; 2U) is 36 +/- 3 kcal mol-1, while that for the D* &rlarr2; 2U transition is 24 +/- 2 kcal mol-1, both at 1 M standard state and pH 7.5.;D* is also observed in the urea mediated unfolding pathway of the holo-enzyme, where PLP dissociates during the D &rlarr2; D* transition. Reductive trapping of the cofactor to a non-dissociable derivative (PPL-eAATase) precludes the formation of D*. A novel monomeric intermediate (M'-PPL) with 70% of the native secondary structure (circular dichroism) was identified in the unfolding pathway of PPL-eAATase: D-PPL 2 &rlarr2; 2M'-PPL &rlarr2; 2U-PPL. The combined results define two structural regions with distinct stabilities: the active site region (ASR) and, the generally more stable, dimerization region (DMR). The DMR is responsible for the multimeric nature of D*, and its disorder leads to dimer dissociation. Selective strengthening of the ASR-cofactor interactions by cofactor trapping reverses the relative stabilities of the two regions (from DMR>ASR in the apoenzyme to ASR>DMR in PPL-eAATase) and results in a reordering of the eAATase denaturation pathway.;Disruption of any single intersubunit side-chain/side-chain interaction by site directed mutagenesis results in a > 2.6 kcal mol-1 decrease of the native state stability, independently of its locations (large or small interface) or nature (hydrophobic, hydrogen bond, or salt bridge). However, the stability of D* respect to U is the same for all mutants. The stability of the eAATase interface cannot be accounted for by the contribution of a few hot-spots, neither by the accumulation of a large number of weak contacts, but only by the presence of multiple important and interconnected interactions. It is proposed that a "molten surface" structure, flexible enough to accommodate point mutations, accounts for the stability of D*, and that nuclei of tertiary structure not involved in native intersubunit contacts provide a scaffold for the unstructured interface of D*. (Abstract shortened by UMI.).