RESULTS AND DISCUSSION
Recombinantly expressed GmCHS1 was purified using nickel affinity chromatography with an N -terminal His6-tag, and by using size-exclusion chromatography (Fig. S1). The purified GmCHS1 was crystallized, and the crystal diffracted to 2.5 Å resolution (crystallographic statistics, Table 1; electron density quality, see Fig. S2). GmCHS1 crystallized in space group P212121 with two molecules per asymmetric unit. The structures were solved using a molecular replacement method, with Pinus sylvestris CHS (PDB ID: 6DXA, 85.9% sequence identity) as a search model. The refined structure included all residues from the original sequence of GmCHS1 (residues 1-388). However, the electron densities corresponding toN- terminal extension residues including the His6-tag were not visible owing to the highly disordered structures of the N -terminal regions. The overall structure of GmCHS1 exhibits the αβαβα thiolase fold with a pseudo-symmetric dimer with 12 α-helices and 11 β-strands (Fig. 1A) per monomer, as previously reported for the crystal structure of CHS from Medicago sativa(alfalfa, PDB ID: 1BI5).14 In fact, our size-exclusion chromatography experiment suggests that GmCHS1 mainly exists as a homodimer in solution (Fig. S1). The homodimeric GmCHS1 contains two active sites in the middle of the thiolase fold (Figure 1A, one is highlighted in a dotted circle). When viewed in surface representation, there is a tunnel-like structure, which has been termed the CoA-binding tunnel in previous studies, making a connection between the external solvent environment and the active site in each monomer (Fig. 1B). Further analysis of the interior cavity, including the Co-A binding tunnel, indicates that there is a large cavity around the active site wherein the condensation reactions should occur (Fig. 1C). It is noteworthy that the large cavity forms another tunnel downward to the external solvent (Fig. 1C, indicated by a dotted arrow). Although the mechanism for the product export of CHS is not completely understood, it will be of interest in the future to evaluate whether this additional tunnel functions as a metabolic channel. A close-up view of the active site showed that the catalytic residues (Cys164, His302 and Asn335) are located in similar orientations to those of other CHS structures, with the electron densities for Cys164 sidechains being in their doubly oxidized forms (Fig. 1D). The doubly oxidized catalytic cysteine sulfinic acid is widely observed in the crystal structures of CHS from euphyllophytes (ferns and seed plants), but not in the CHS from either lycophyte or moss species. Because of the clarity of electron densities at the positions 164 and the fact that soybean belongs to euphyllophytes, we modeled S-cysteinesulfinic acid (ligand ID: CSD) molecules for the corresponding positions. Weng et al.hypothesized that the oxidation of catalytic Cys residues was due to the highly reactive environment around Cys during the generation of nucleophilic thiolate anion that was assisted by the conserved His residue (His302 in GmCHS1).15 In the GmCHS1 structure, distances between the sulfur atom of CSD164 and the Nεof His302 are 3.55 and 3.69 Å for molecule A and B in the asymmetric unit, respectively (Fig. 1D). In addition to the oxidation of Cys164, the electron densities for Met70 sidechains are observed in their oxidized forms. Although these Met oxidations are not reported for any other CHS structures, we modeled the methionine sulfoxide (ligand ID: SME) for the positions 70. Further biochemical or physiological investigations will be needed to decipher the importance of the Met oxidation.
To better understand the catalytic mechanism of GmCHS1, newly purified GmCHS1 was crystallized, and the apo-crystals were soaked with naringenin, a product analog of the CHS-catalyzed reaction. The naringenin-soaked crystals diffracted to 1.8 Å resolution, and the structure was determined in a similar procedure to that of the apo-structure. The refined complex structure includes a dimer of GmCHS1, two naringenin molecules, as well as three of acetate ions and three citrate ions, which are contained in the mother liquor. Moreover, SME and CSD are modeled for the positions 70 and 164, respectively as well as the apo-structure. In the complex structure, naringenin molecules are located in the vicinity of the catalytic cysteine sulfinic acids (the residue 164, Fig. 2A). The positions of naringenin molecules were found to be almost equivalent to those of previously reported for the crystal structure of M. sativa CHS (Fig. S3).14 In addition to the CSD164, Phe215 and Phe264 are located in the vicinity of naringenin. The importance of the conserved Phe215 was previously proposed as an involvement in the orientation of substrates and reaction intermediates at the active site.14 We found that one of the citrate ions was located at the end of the CoA binding tunnel, also in the vicinity of Phe264 (Fig. 2A, 2B). The corresponding position of the citrate ion of our structure was occupied by the CoA moiety of malonyl-CoA or 1,4-piperazine di-ethane sulfonic acid (C8H18N2O6S2) in the complex structure of the M. sativa CHS Cys164Ala mutant (Fig. S3, PDB ID: 1CML), suggesting that a negative charge of the citrate ion could potentially mimic that of either CoA or 1,4-piperazine di-ethane sulfonic acid. Although citric acid has not been studied as a potential inhibitor of CHS, our complex structure suggests the possibility of inhibitor binding upon the existence of the product at the active site. Superposition of apo- and naringenin-bound structures of GmCHS1 shows that the Phe265 is likely to change its orientation through the CHS-catalyzed reaction (Fig. 2C). This is supported by the fact that the average temperature factor for the Phe264 sidechain is higher than that of Phe215 (44.2 Å2 and 33.4 Å2 for the Phe264 and the Phe215, respectively). It is an intriguing question whether the flexibility or the bulkiness of position 264 is related to the catalytic efficiency of GmCHS1. In conclusion, the crystal structures of GmCHS1 provide a structural basis for further improvement of our understanding of the catalytic mechanism of CHS. Further efforts will include structural characterization of a complex between GmCHS1 and other components of the soybean isoflavonoid metabolon to better understand how GmCHS1 has a distinct role in the formation of the isoflavonoid metabolon. Coordinates and structural factors of apo-GmCHS1 and GmCHS1/naringenin have been deposited in the Protein Data Bank under accession codes 7BUS and 7BUR, respectively.