Ayaka Harada

Structural Basis of the γ-Lactone-Ring Formation in Ascorbic Acid Biosynthesis by the Senescence Marker Protein-30/Gluconolactonase

The senescence marker protein-30 (SMP30), which is also called regucalcin, exhibits gluconolactonase (GNL) activity. Biochemical and biological analyses revealed that SMP30/GNL catalyzes formation of the γ-lactone-ring of l-gulonate in the ascorbic acid biosynthesis pathway. The molecular basis of the γ-lactone formation, however, remains elusive due to the lack of structural information on SMP30/GNL in complex with its substrate. Here, we report the crystal structures of mouse SMP30/GNL and its complex with xylitol, a substrate analogue, and those with 1,5-anhydro-d-glucitol and d-glucose, product analogues. Comparison of the crystal structure of mouse SMP30/GNL with other related enzymes has revealed unique characteristics of mouse SMP30/GNL. First, the substrate-binding pocket of mouse SMP30/GNL is designed to specifically recognize monosaccharide molecules. The divalent metal ion in the active site and polar residues lining the substrate-binding cavity interact with hydroxyl groups of substrate/product analogues. Second, in mouse SMP30/GNL, a lid loop covering the substrate-binding cavity seems to hamper the binding of l-gulonate in an extended (or all-trans) conformation; l-gulonate seems to bind to the active site in a folded conformation. In contrast, the substrate-binding cavities of the other related enzymes are open to the solvent and do not have a cover. This structural feature of mouse SMP30/GNL seems to facilitate the γ-lactone-ring formation.

The crystal structure of a new O‐demethylase from Sphingobium sp. strain SYK‐6

In the cell, tetrahydrofolate (H4folate) derivatives with a C1 unit are utilized in various ways, such as for the synthesis of amino acids and nucleic acids. While H4folate derivatives with the C1 unit are typically produced in the glycine cleavage system, Sphingobium sp. strain SYK‐6, which can utilize lignin‐derived aromatic compounds as a sole source of carbon and energy, lacks this pathway, probably due to its unique nutrient requirements. In this bacterium, H4folate‐dependent O‐demethylases in catabolic pathways for lignin‐derived aromatic compounds seem to be involved in the C1 metabolism. LigM is one of the O‐demethylases and catalyzes a C1‐unit transfer from vanillate (VNL) to H4folate. As the primary structure of LigM shows a similarity to T‐protein in the glycine cleavage system, we hypothesized that LigM has evolved from T‐protein, acquiring its unique biochemical and biological functions. To prove this hypothesis, structure‐based understanding of its catalytic reaction is essential. Here, we determined the crystal structure of LigM in apo form and in complex with substrates and H4folate. These crystal structures showed that the overall structure of LigM is similar to T‐protein, but LigM has a few distinct characteristics, particularly in the active site. Structure‐based mutational analysis revealed that His60 and Tyr247, which are not conserved in T‐protein, are essential to the catalytic activity of LigM and their interactions with the oxygen atom in the methoxy group of VNL seem to facilitate a methyl moiety (C1‐unit) transfer to H4folate. Taken together, our structural data suggest that LigM has evolved divergently from T‐protein.

Complete pyridine-nucleotide-specific conversion of an NADH-dependent ferredoxin reductase

The coenzyme specificity of enzymes is one of the critical parameters for the engineered production of biological compounds using bacteria. Since NADPH is produced abundantly in photosynthetic organisms, conversion of an NADH-specific enzyme into an NADPH-specific one is a useful approach for the efficient carbon-neutral production of biological compounds in photosynthetic organisms. In the present study, an NADH-specific ferredoxin reductase component, BphA4 of biphenyl dioxygenase BphA from Acidovorax sp. strain KKS102, was changed to an NADPH-dependent form using a method combining structure-based systematic mutations and site-directed random mutagenesis. The resultant CRG mutant, in which Glu175-Thr176-Gln177 of an NADH-recognition loop in the wild-type BphA4 was replaced with Cys175-Arg176-Gly177, was highly specific and active for NADPH, and its biochemical and structural properties for NADPH were nearly the same as those of the wild-type BphA4 for NADH. In addition, this mutation project was assessed by a semi-empirical prediction method of mutation effects, and the results suggested that the CRG mutant was one of the best NADPH-specific mutants.

Characterization of Heme Orientational Disorder in a Myoglobin Reconstituted with a Trifluoromethyl-Group-Substituted Heme Cofactor

The orientation of a CF3-substituted heme in sperm whale myoglobin and L29F, H64L, L29F/H64Q, and H64Q variant proteins has been investigated using 19F NMR spectroscopy to elucidate structural factors responsible for the thermodynamic stability of the heme orientational disorder, i.e., the presence of two heme orientations differing by a 180° rotation about the 5-15 meso axis, with respect to the protein moiety. Crystal structure of the met-aquo form of the wild-type myoglobin reconstituted with 13,17-bis(2-carboxylatoethyl)-3,8-diethyl-2,12,18-trimethyl-7-trifluoromethylporphyrinatoiron(III), determined at resolution of 1.25 Å, revealed the presence of the heme orientational disorder. Alterations of the salt bridge between the heme 13-propionate and Arg45(CD3) side chains due to the mutations resulted in equilibrium constants of the heme orientational disorder ranging between 0.42 and 1.4. Thus, the heme orientational disorder is affected by the salt bridge associated with the heme 13-propionate side chain, confirming the importance of the salt bridge in the heme binding to the protein.

Overcoming a hemihedral twinning problem in tetrahydrofolate-dependent O-demethylase crystals by the microseeding method.

A tetrahydrofolate-dependent O-demethylase, LigM, from Sphingobium sp. SYK-6 was crystallized by the hanging-drop vapour-diffusion method. However, the obtained P3121 or P3221 crystals, which diffracted to 2.5-3.3 Å resolution, were hemihedrally twinned. To overcome the twinning problem, microseeding using P3121/P3221 crystals as microseeds was performed with optimization of the reservoir conditions. As a result, another crystal form was obtained. The newly obtained crystal diffracted to 2.5-3.0 Å resolution and belonged to space group P21212, with unit-cell parameters a = 102.0, b = 117.3, c = 128.1 Å. The P21212 crystals diffracted to better than 2.0 Å resolution after optimizing the cryoconditions. Phasing using the single anomalous diffraction method was successful at 3.0 Å resolution with a Pt-derivative crystal. This experience suggested that microseeding is an effective method to overcome the twinning problem, even when twinned crystals are utilized as microseeds.

Laser processing of protein crystals for native SAD data collection

Native SAD phasing uses anomalous scattering signals from light atoms such as sulfur and phosphorus in protein crystals. The anomalous signals from these atoms are, however, much weaker than those from heavy atoms that are frequently utilized in protein crystallography. Therefore, high quality data collection is essential for native SAD phasing. Lower energy X-ray is favorable to enhance the anomalous signals but on the other hand X-ray absorption by sample, solvent, and air hinders the diffraction data collection. Here, we present a method minimizing the X-ray absorption; the solvent portion of a mounted frozen crystal is removed or the mounted crystal is spherically shaped by the deep UV laser processing technique. In this study, crystals of BphA4 and LigM proteins with various size and shape were used. Diffraction data for native SAD phasing were collected with the crystals in two times before and after the laser processing using the X-ray of 3.7 keV and 4.5 keV at BL-1A of Photon Factory (Tsukuba, Japan), equipped with a helium beam path and a helium cold stream to reduce X-ray absorption by air. The data statistics of the laser-processed crystals were much better: the values of I/sigma(I) and SigAno were significantly increased. The improvement are expected to work advantageously in phase determination by native-SAD method.

Exploiting wavelengths longer than 3 Å for native SAD phasing

Native SAD (Single-wavelength anomalous dispersion) is a crystallographic phasing method which utilizes anomalous signals from naturally-included light atoms in macromolecules. It has been known as an attractive method because no derivative crystals are required for de-novo structural solutions. The method appeared more than 30 years ago [1], however, it is still uncommon even at the latest synchrotron MX beamlines due to experimental difficulties in using long wavelength X-ray required to enhance weak anomalous signals [2]. Using longer wavelength is preferred to efficiently detect anomalous signals from light atoms, on the other hand, causes severe beam absorption by the sample or air on the X-ray beam path. In addition, the large diffracting angle makes it difficult to obtain high resolution data with a flat area detector. We have implemented a dedicated data collection environment for native SAD phasing at the long-wavelength MX beamline BL1A at the Photon Factory. The diffractometer is equipped with a standing chamber enclosing both the goniometer and the X-ray detector, a helium cold stream recycling system, and a specially designed sample changer. The beam path is completely covered by helium allowing data collection with low background. Under such optimized experimental environment we showed using the wavelength longer than 2.7 Å was advantageous in native SAD phasing against using the wavelength of 1.9 Å, by which most native SAD data collection has been performed so far [3]. In exploiting further longer wavelength than 3 Å, two flat area detectors (Eiger X4M, Dectris Ltd.) were installed which can be configured in V-shape in the helium chamber. The problem of large diffracting angle was mitigated by the V-shape configuration: the data obtained with the V-shape configuration was superior in terms of not only the availability of higher resolution but also the data statistics at high resolution range when compared with the normal (single) detector configuration of very short sample-todetector distance. One drawback of the V-shape configuration is the lower completeness of data which can be recovered by the mini-kappa goniometer installed at the beamline. We will present the feasibility of using the wavelength longer than 3 Å for native SAD phasing based on the systematic data collections and analyses with different conditions such as wavelength, detector configuration and sample size.

Case studies of sulfur SAD phasing with LigM from Sphingobium sp. SYK-6

Sphingobium sp. SYK-6 grows on a lignin-related biphenyl compound as the sole carbon and energy source and was initially isolated from a pond waste liquor from a kraft pulp mill. In SYK-6, 5-CH3-H4folate is synthesized from aromatic compounds such as a vanillate by a O-demethylase, LigM. The 5-CH3-H4folate is then converted to 5,10-CH2-H4folate, which is utilized for syntheses of DNA, repair DNA, and methylate DNA as well as to act as a cofactor in certain biological reactions, by another enzyme, MetF. In other bacterial speceis, 5,10-CH2-H4folate is directly synthesized by Tand H-proteins that are enzymes in Glycine Cleavage System. It is considered that SYK-6 has evolved to acquire this unique pathway for the 5,10-CH2-H4folate production, in order to survive in extreme environmental condition. To elucidate the molecular mechanisim of this pathway, we have carried out the structural analysis of LigM. LigM was purified by using IMPACT system provided from NEB, which use intein and affinity chitin-binding tag. After crystallization screening, a reservoir solution of 0.2 M Mg acetate, 0.1 M Acetate buffer pH 4.6, and 20 %(w/v) PEG8000 gave a needle crystals with approximate dimensions of 0.3×0.1×0.01 mm3. A diffraction data set was collected with 1.1 Å wavelength at BL1-A in the Photon Factory. However, phasing trials via molecular replacement (using a model with 19% sequence identity) failed. Because LigM is a 53 kDa protein and contains fourteen sulfur atoms, LigM is an interesting candidate for SAD phasing with sulfur (S-SAD). Diffraction data sets of LigM crystals were collected with 1.9 and 2.7 Å wavelengths, reaching a maximum resolution of 2.3Å. Preliminary results are promising for solving the phase problem via S-SAD. This study is also of methodological interest as the phasing capability of two different wavelengths can be compared. A thorough analysis of the diffraction data is in progress.