5.4. Structure of Δ3-Δ2-enoyl-CoA isomerase complexed with octanoyl–CoA (IV)
The liganded structure of the yeast enoyl-CoA isomerase was determined from crystals co-crystallized with octanoyl-CoA, an inhibitor of the enoyl-CoA isomerase. This structure is also referred to as the neutral pH form on the basis of its crystallization condition. Molecular replacement calculations detected the orientations and positions of two subunits of the isomerase trimer in the space group P41212. The solution for the third subunit could not be found using AmoRe (Navaza 1994). Instead, the coordinates were obtained by superimposing the unliganded enoyl-CoA isomerase trimer, also referred to as the low pH form, onto the two assembled AMoRe solutions with the lsq option in O (Jones et al. 1991) and by extracting the coordinates of the superimposed third subunit. The rigid body refinement (Murshudov et al. 1997) of the new trimer resulted in a model with R- and Rfree-factors of 40.1 % and 41.0 %, respectively. The octanoyl-CoA molecules were built into the active sites of the subunits, and the structure was refined at 3.3 Å resolution to the final R-factor of 21.8 % and free R-factor of 29.2 % with good geometry statistics. During refinement, the residues 74-88, which were disordered in the unliganded structure (the low pH form), became visible in the electron density map, although they still continued to refine with high B-factors. This region, located just before helix H2, adopted a helical conformation and was named helix H2A. Part of helix H2 also had a different, slightly upward bent conformation compared to the low pH form of isomerase. In addition to the structural changes in helix H2, H9 and H10 also adopted somewhat different conformations when the two structures were compared (see “Discussion”, Fig. 13B). The N- and C-termini were also disordered in the liganded structure. A few more residues could, however, be built in each terminus. In the current structure, the subunits A, B and C consist of the residues 4-272, 4-274 and 5-270, respectively, leaving the peroxisomal targeting sequence still disordered.
The monomers of the liganded, neutral pH form, the yeast enoyl-CoA isomerase, assemble into a tight trimer similarly to the unliganded, low pH form. In addition, both of the pH forms crystallize as hexamers with 32 symmetry. However, the mode of assembly is rather different. In contrast to the loose intertrimer contacts in the low pH form, the trimers in the neutral pH form interact closely with each other, having 159 contacts across the intertrimer space. The subunits are also 5.6 Å closer to each other. Moreover, the second trimeric disk appears to have rotated 25° around its threefold axis when the two pH forms are compared. The different hexameric packing and the loose intertrimer contacts in the low pH form suggested that the yeast isomerase could be a trimer at low pH. The ultracentrifugation sedimentation velocity runs indeed suggested that, at pH 5.6, the yeast enoyl-CoA isomerase could occur as a trimer, whereas at pH 7.2, a hexamer was found in good agreement with the gel filtration and dynamic light scattering as well as crystallographic studies.
The mode of octanoyl-CoA binding to enoyl-CoA was very much like expected on the basis of the superposition studies. The CoA moiety is bound in a bent conformation against the β -sheet formed by regions B1, B2, B3 and B4. The entrance to the catalytic site is covered by the helices H2A and H10. The octanoyl moiety is bound near the catalytic residue Glu158 and points upwards towards Gly35, Tyr38, Phe97 and Arg100 (Figs 12, 13B). The thioester oxygen is hydrogen-bonded to the main chain NH groups of Ala70 and Leu126, thus forming an oxyanion hole (Fig. 12), as speculated on the basis of the superposition studies. The only structural difference near the catalytic site in a comparison of the unliganded and octanoyl-CoA-complexed crystal forms is a small shift of approximately 1 Å in the Glu158 side chain towards the ligand.
