Single amino acid mutation altered substrate specificity for l-glucose and inositol in scyllo-inositol dehydrogenase isolated from Paracoccus laeviglucosivorans

ABSTRACT

scyllo-inositol dehydrogenase, isolated from Paracoccus laeviglucosivorans (Pl-sIDH), exhibits a broad substrate specificity: it oxidizes scyllo- and myo-inositols as well as l-glucose, converting l-glucose to l-glucono-1,5-lactone. Based on the crystal structures previously reported, Arg178 residue, located at the entry port of the catalytic site, seemed to be important for accepting substrates. Here, we report the role of Arg178 by using an alanine-substituted mutant for kinetic analysis as well as to determine the crystal structures. The wild-type Pl-sIDH exhibits the activity for scyllo-inositol most preferably followed by myo-inositol and l-glucose. On the contrary, the R178A mutant abolished the activities for both inositols, but remained active for l-glucose to the same extent as its wild-type. Based on the crystal structures of the mutant, the side chain of Asp191 flipped out of the substrate binding site. Therefore, Arg178 is important in positioning Asp191 correctly to exert its catalytic activities.

Abbreviations: IDH: inositol dehydrogenase; LB: Luria-Bertani; k_cat: catalyst rate constant; K_m: Michaelis constant; NAD: nicotinamide dinucleotide; NADH: nicotinamide dinucleotide reduced form; PDB; Protein Data Bank; PDB entry: 6KTJ, 6KTK, 6KTL

R178A mutation of scyllo-inositol dehydrogenase (isolated from Paracoccus laeviglucosivorans) turned it to an l-glucose substrate-specific enzyme.

Chiral conformation, enzyme kinetics, sugar metabolism, X-ray crystallography

d-glucose is the main energy source of life. As opposed to d-glucose, l-glucose, which is the enantiomer of d-glucose, is rarely found in nature; therefore, it has been thought that it has not been used by living organisms for a long time. In 1940, Rudney investigated whether L-form of glucose was assimilated by several organisms, and reported that no organism could utilize it [1]. Then, however, Sasajima et al. reported that D-threo-aldose dehydrogenase from Pseudomonas caryophylli was able to oxidize l-glucose [2]; however, no information on l-glucose assimilation was provided.

Previously, a microorganism capable of assimilating l-glucose, Paracoccus laeviglucosivorans, has been isolated, and its l-glucose catabolic pathway was elucidated [3,4]. The first enzyme in this pathway showed a unique property in substrate specificity: it oxidizes l-glucose, converting to l-glucono-1,5-lactone in an NAD⁺ dependent manner, and it also exhibited a strong scyllo-inositol dehydrogenase activity. Furthermore, it also catalyzes the oxidation of myo-inositol and d-glucose (Figure 1). As the most preferable substrate was scyllo-inositol, and its gene was located in a cluster of a putative inositol catabolic genes in P. laeviglucosivorans genome, this enzyme was recognized as scyllo-inositol dehydrogenase (Pl-sIDH) and its physiological function was assumed to assimilate scyllo-inositol. To reveal the substrate recognition mechanism of Pl-sIDH, we have analyzed the crystal structures of the ternary complexes of Pl-sIDH with scyllo-inosose/NAD⁺, myo-inositol/NAD⁺, and l-glucono-1,5-lactone/NADH [5]. Pl-sIDH belongs to the glucose-fructose oxidase/inositol dehydrogenase/microbial rhizopine-catabolizing (GFO/IDH/MocA) family (Pfam ID: PF01408) [6], and accordingly, Pl-sIDH adopted a homotetramer structure that was similar to the enzymes in the GFO/IDH/MocA family. Based on the crystal structures, two conserved catalytic residues, Lys106 and His195, were also involved in the binding of substrates. When the structure of Pl-sIDH was compared between the inositols and the lactone complexes, the loop structure close to the active site showed structural differences in open as well as in close conformations as a lid, and the Arg178 residue located in the loop showed different interactions with substrates (Figure 2). Therefore, this arginine residue was assumed to play an important role in the substrate-binding process. Then, this arginine residue was mutated to alanine, and the enzyme activities for l-glucose and scyllo-inositol were measured. As a result, the activity remained for l-glucose but not for scyllo-inositol [5]. In this study, to elucidate the role of the arginine residue in detail, further kinetic and structural analyses were conducted by using the mutant.

Figure 1.

Structures of Pl-sIDH substrates and products.

Figure 2.

Conformational change of the loop structure over the active site between the lactone and myo-inositol bound Pl-sIDHs.

The crystal structures of the lactone bound (PDB entry 5yap) and myo-inositol bound (PDB entry 5ya8) complexes are superposed and colored in green and red, respectively. Lac and INS denote l-glucono-1,5-lactone and myo-inositol, respectively. The R178A mutant of the enzyme was also superposed on the lactone bound wild type colored in cyan. Dotted lines show hydrogen bonds with the distance less than 3.1Å.

Materials and methods

Bacterial strains, plasmids and a mutant enzyme

An overexpression system and alanine-substituted mutant reported previously [5] were used in this study. Briefly, the lgdA gene, encoding Pl-sIDH, derived from the P. laeviglucosivorans strain cloned in the pET21a(+) vector was introduced into the Escherichia coli BL21(DE3) strain. The C-terminus of the lgdA gene region was fused with a 6 × His-tag in the vector. The alanine-substituted mutation for 178th arginine residue (R178A) of Pl-sIDH was introduced by using pET21a(+)-lgdA as a template.

Enzyme production and purification

An E. coli strain BL21(DE3) containing pET21a(+)-lgdA was cultured in LB medium supplemented with 1.7% of preculture seed and ampicillin (final concentration of 50 μg/mL) and incubated for 6 h at 37°C under rotary shaking. Then, isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to the medium at a final concentration of 200 μM followed by cultivation at 25°C overnight.

Cultured cells were collected by centrifugation, resuspended in buffer A (20 mM Tris-HCl, pH 7.5, 10% glycerol) and disrupted by ultrasonication on ice. After centrifugation, the soluble fraction was loaded onto a Ni-NTA agarose column equilibrated with buffer A. The column was washed with buffer A containing 20 mM imidazole. The bound protein was eluted with buffer A containing 300 mM imidazole, and dialyzed for 3 h with buffer A followed by further overnight dialysis after a buffer exchange. The purified protein was concentrated to 10 mg/mL by using Amicon Ultra 100K.

Enzyme activity assay

The reaction mixture was composed of 100 mM Tris-HCl, pH 9.0, 20 mM NAD⁺, and a substrate with a series of concentrations in a volume of 900 μL. Substrates and concentrations used for the assay were 10, 20, 40, 60, and 70 mM (WT), and 12.5, 25, 50, 75, and 100 mM (R178A) for l-glucose: 20, 40, 60, 80, and 100 mM (WT), and 80, 100, 200, 400, and 800 mM (R178A) for d-glucose: 5, 10, 20, 30, and 40 mM (WT), and 6.25, 12.5, 25, 50, and 90 mM (R178A) for scyllo-inositol: 5, 10, 20, 40, and 80 mM (WT), and 80, 160, 240, 320, and 400 mM (R178A) for myo-inositol. The purified enzyme of 0.2 mg/mL in a total 50 μL solution volume was added to initiate the reaction, which was measured by monitoring NAD⁺ absorption at 340 nm using a V-750 spectrophotometer (Jasco corp., Tokyo, Japan). Kinetic parameters were calculated by the Michaelis-Menten equation using the software supplied with the spectrophotometer.

Crystallization, data collection, and structure determination

Crystallization was performed by a hanging drop vapor diffusion method at 20°C. The protein solution and a crystallization buffer containing 0.1 M sodium acetate, pH 5.2, and 20% (w/v) polyethylene glycerol (PEG) 3350 were mixed in a 1:1 ratio. For the determination of the Pl-sIDH structure in complexes with its substrates, crystals were soaked for 30 s with 25 mM of NAD⁺ (Sigma-Aldrich Co., MO) and 100 mM l-glucose or myo-inositol (Sigma-Aldrich Co., MO) in a buffer containing 0.1 M sodium acetate, pH 5.2, and 30% PEG3350.

Crystals were flash frozen in liquid N₂ in nylon loops without cryoprotectant solutions. The diffraction data were collected on beamlines BL-5A equipped with a Pilatus3 S6M detector and BL-17A equipped with an Eiger X16M detector at Photon Factory for determination of apo- and complexed form crystal structures, respectively. Indexing, integration, scaling, and merging were carried out by the softwares HKL2000 [7] and XDS [8] for the apo- and complexed form crystals, respectively. The initial structures were determined by the molecular replacement method using the software MOLREP [9] in the CCP4 suite [10] with the structure of wild type Pl-sIDH of apo-form (PDB entry 5yab) as a search model. Further iterative refinement and model building were conducted with REFMAC5 [11] and Coot [12], respectively. Data collection and refinement statistics are summarized in Table 1.

Table 1.

Data collection and refinement statistics.

	R178A apo-form	R178A/NADH/lactone	R178A/NAD+/myo-inositol
Data collection
Beamline	PF BL-5A	PF BL-17A	PF BL-17A
Wavelength (Å)	1.000	0.980	0.980
Detector	Pilatus3 S6M	Eiger X16M	Eiger X16M
Space group	P212121	P212121	P212121
Unit-cell parameters (Å)	a = 91.7, b = 128.8, c = 138.5	a = 90.6, b = 127. 6, c = 137. 6	a = 90.6, b = 128.2, c = 137.4
Resolution range (Å)	50 – 2.10 (2.14–2.10)	50 – 1.65 (1.69–1.65)	50 – 1.65 (1.69–1.65)
Unique reflections	94,977 (4688)	190,486 (13,395)	191,340 (13,572)
Redundancy	3.3 (3.3)	6.5 (4.5)	6.6 (4.4)
Rmerge (%)	8.9 (46.2)	5.9 (37.2)	7.3 (55.0)
I/σI	14.3 (1.9)	18.9 (3.3)	16.0 (2.4)
Completeness (%)	99.3 (99.6)	99.6 (95.7)	99.6 (96.3)
CC (1/2)	0.986 (0.811)	0.999 (0.898)	0.999 (0.784)
Refinement
Resolution range (Å)	49.2 − 2.1 (2.16–2.1)	48.4–1.65 (1.69–1.65)	48.9–1.65 (1.69–1.65)
No. of reflections	90,166 (6568)	180,883 (12,679)	181,721 (12,893)
Rwork/Rfree	0.181/0.207 (0.244/0.259)	0.160/0.171 (0.213/0.218)	0.164/0.179 (0.249/0.261)
Rmsd from ideal
Bond (Å)	0.011	0.006	0.004
Angle (°)	1.410	1.437	1.281
No. of non-H atoms
Protein	11,175	11,239	11,260
Water	404	930	1171
Small molecule	16 (ACT)	26 (PG4)/48 (8S0)/176 (NAI)	12 (INS)/12 (ACT)/176 (NAD)
Average B factors (Å2)
Protein	31.7	18.7	18.3
Water	34.5	25.2	28.7
Small molecule	34.1 (ACT)	36.9 (PG4)/24.0 (8S0)/24.3 (NAI)	30.2 (INS)/26.4 (ACT)/22.3 (NAD)
Ramachandran plot
Favored (%)	98.5	98.8	98.8
Allowed (%)	1.5	1.2	1.2
Disallowed (%)	0.0	0.0	0.0

ACT, PG4, INS, 8S0, NAD and NAI refer to three letter codes in PDB as acetate ion, tetraethylene glycol, myo-inositol, L-glucono-1,5-lactone, NAD⁺ and NADH, respectively. Ramachandran plot was calculated by Rampage in the CCP4 suite.

Figures for protein structures were generated using the software PyMOL (Schrödinger, Inc., MA). The atomic coordinates and structure factors for the R178A mutant of the apo-form, l-glucono-1,5-lactone/NADH, and myo-inositol/NAD⁺ complexes have been deposited in the Protein Data Bank under the accession codes 6KTJ, 6KTK, and 6KTL, respectively.

Results and discussion

Catalytic activities of Pl-sIDH R178A mutant

Since the structural studies reported previously suggested the involvement of Arg178 for the enzyme activity, the alanine-substituted mutant (R178A) has been prepared. The kinetic study for l-glucose and scyllo-inositol revealed the importance of Arg178 for the activity [5]; therefore, further assay with glucoses and inositols was carried out at first in this study to analyze the substrate specificity in more detail. Representative s-v plots are shown in Figure 3, and kinetic parameters of the wild type and R178A mutant are summarized in Table 2.

Table 2.

Kinetic parameters of the Pl-sIDH wild-type and R178A mutant.

Substrate	L-glucose		D-glucose		scyllo-inositol		myo-inositol
Pl-sIDH	WT	R178A	WT	R178A	WT	R178A	WT	R178A
Km (mM)	49.2 ± 4.3	39.4 ± 14	58.7 ± 6.8	3.97 x 102 ± 27	8.22 ± 0.33	42.1 ± 5.1	32.5 ± 5.5	2.71 x 102 ± 44
Vmax (U/mg)	1.00 x 102 ± 19	65.6 ± 9.3	5.46 ± 0.52	0.96 ± 0.14	1.02 x 102 ± 1.8	4.23 ± 0.26	40.7 ± 4.7	1.76 ± 0.17
kcat (min−1)	3.95 x 103 ± 7.4 x 102	5.37 x 103 ± 7.6 x 102	4.42 x 102 ± 42	37.5 ± 5.5	4.02 x 103 ± 72	1.66 x 102 ± 10	1.60 x 103 ± 1.8 x 102	75.8 ± 7.1
kcat/Km (min−1mM−1)	79.8 ± 8.6	1.45 x 102 ± 27	7.55 ± 0.35	0.09 ± 0.01	4.90 x 102 ± 24	3.98 ± 0.26	49.7 ± 3.4	0.28 ± 0.02

WT denotes wild type.

Figure 3.

Enzyme reaction of Pl-sIDH.

Representative s-v plots of the forward reactions of the wild type (WT) and mutant (R178A) with l-glucose and scyllo-inositol as substrates. The lines indicate a curve fitting calculated by the Michaelis-Menten equation. Detailed procedures are described in the Materials and Methods, and kinetic parameters are summarized in Table 2.

The wild type Pl-sIDH showed the highest k_cat/K_m value for scyllo-inositol among substrates tested as reported previously. On the contrary, the R178A mutant dramatically decreased its activity for myo- and scyllo-inositols. Interestingly, for l-glucose, the k_cat/K_m value of the R178A mutant did not change compared to that of wild type. Therefore, the single amino acid mutation of R178A was considered to have altered the substrate specificity of Pl-sIDH. The K_m values of scyllo- and myo-inositols increased about five- and eight-folds and the k_cat values decreased about 24- and 21-folds in the mutant in comparison to that of wild type, respectively. Some of values in parameters showed discrepancies with those in the previous reports [3,5] although the tendency of the catalytic activities between the wild type and mutant were not changed. These differences might have been arose partly by the different assay system [3], but no clear reasons were identified. Since Arg178 interacted with inositols in the active site in the wild type, Ala mutation apparently affected the K_m values for inositols. Furthermore, the mutation affected k_cat values as well as K_m values for inositols. However, with the lactone, Arg178 did not involve in mediating the substrate binding directly in the wild type; therefore, the mutation did not affect the kinetics for the lactone. To evaluate the effect of altered catalytic activities on the mutant, the structure determinations of the mutants were conducted by X-ray crystallography.

Overall structure of Pl-sIDH R178A mutant

In this study, we have determined three structures of the R178A mutant of Pl-sIDH, an apo-form, and two ternary complexes with l-glucono-1,5-lactone/NADH and myo-inositol/NAD⁺. Initially, we tried to obtain the structure of R178A with an l-glucose complex; however, the lactone complex was obtained as described in the next section. We also tried to obtain the scyllo-inositol or scyllo-inosose complex of the R178A mutant; however, no bound structures were obtained.

All of the structure belonged to the space group P2₁2₁2₁ with almost identical unit cell parameters containing four subunits in an asymmetric unit, which are the same as the case of wild type. The structures of apo-form and ternary complexes with the lactone and myo-inositol were refined at 2.1 Å, 1.65 Å, and 1.65 Å resolution, respectively. As a result, overall structures of the wild type and the R178A mutant were almost identical as shown in Figure 4 for the lactone/NADH complex as a representative, where the superposition of chain A of two structures, the wild type and R178A mutant complexed with the lactone by the software Chimera gave root mean square deviation (r.m.s.d.) value of 0.348 Å for chain A (366 residues) and 1.2 Å for all the chains (1460 residues).

Figure 4.

Overall structure of the R178A mutant of Pl-sIDH.

Tetramer conformations of wild type and R178A mutant of Pl-sIDH complexed with l-glucono-1,5-lactone and NADH are shown. The wild type and mutant structures were superposed by the software Chimera [20] on chain As colored in cyan and pink for the wild type and the mutant, respectively. The lactone and NADH are represented as a stick model.

Substrate recognition by Pl-sIDH R178A mutant

To obtain the ternary complex structure of l-glucose and NAD⁺, the R178A mutant crystal was soaked with l-glucose and NAD⁺. After the structure determination, the electron density of the substrate was clearly observed, and based on the shape of the density, l-glucono-1,5-lactone was considered to bind to the active site as was the case in the wild type reported previously (Figure 5(a)) [5]. Accordingly, NAD⁺ was considered to be converted to NADH. In the crystal structure, the lactone was considered to be bound to the active site although the crystal was soaked with the excessive amount of l-glucose. In the case of NADH-dependent L-lactate dehydrogenase and levoglucosan dehydrogenase, the reactions are highly biased toward the reducing direction with the equilibrium constant of 2.76 × 10⁻¹² M [13] and (1.21 ± 0.02) x 10⁻¹³ M [14], respectively. Since Pl-sIDH is highly homologous to levoglucosan dehydrogenase [14] as described below, the crystal structure of Pl-sIDH complexed with l-glucose was expected. According to our observations so far as we tried, the reverse reactions (reducing) have been detected for myo- and scyllo-inositols [5] but not for l-glucose. This might have led to the capture of the lactone in the active site although we concluded the bound compound as the lactone based on only the shape of the electron density.

Figure 5.

The binding mode of the lactone in the active site of the R178A mutant.

(a) The electron density of the Fo-Fc omit map for the lactone in the R178A mutant contoured at 3σ when calculated by CNS [21,22] by excluding the molecule. The bottom figure represents the structure above it when viewed from the direction indicated by the orange arrow. (b) The binding mode of l-glucono-1,5-lactone in the R178A mutant and wild type are compared by the superposition, colored in cyan and red, respectively. Dotted lines show hydrogen bonds with distances less than 2.9Å except for Y163 with that of 3.1 Å

As shown in Figure 5(b), the superposition of subunits between the wild type and the R178A mutant resulted in almost identical substrate binding mode between the two structures, where side chains of six residues, Lys106, Tyr135, Tyr163, Glu165, His195, and His318 formed hydrogen bonds to the lactone. In the wild type, Arg178 did not interact with the lactone directly, and the superposition revealed that the Arg178 and Ala178 located at the almost identical positions. This, therefore, is in good agreement with the results from the catalytic activity assay of the enzyme, in which no activity changes were observed between the catalytic activity of the wild type and the R178A mutant for l-glucose.

We also obtained the structure of the R178A mutant with the myo-inositol complex by the crystal soaking method. After the iterative model building and refinement, the structure was determined at 1.65 Å-resolution. Although the partial electron density were observed in the active sites, among the four subunits of the enzyme, myo-inositol was considered to bind to one subunit (Figure 6(a)). The concentration of myo-inositol for soaking was the same as that for the wild type as reported previously; therefore, the faint binding of the inositol in the crystal structure of the R178A mutant was considered to reflect the larger K_m value than that of the wild type (Figure 6(b)). As described above, the equilibrium constant of Pl-sIDH was assumed to be so small that the reaction would be biased for the reducing direction, and Pl-sIDH catalyzes L-epi-2-inosose to myo-inositol as the reverse reaction [5], the bound compound, therefore, was considered to be myo-inositol although the electron density was not clearly observed.

Figure 6.

The binding mode of myo-inositol in the active site of the R178A mutant.

(a) The electron density of the Fo-Fc omit map for the lactone in the R178A mutant contoured at 1.5σ calculated by CNS [21,22] by excluding the molecule. An asterisk shows the C3 atom in the top figure. The bottom figure represents the structure above it when viewed from the direction indicated by the orange arrow. (b) The binding mode of myo-inositol in the R178A mutant and wild type are compared by the superposition, colored in cyan and red, respectively. The numbers colored in red and cyan indicate the C3 and C4 atoms of myo-inositol (INS) in each complex. The C4 atom of myo-inositol in the wild type complex, where a hydrogen atom is at an axial position, and the C3 atom in the mutant complex, where a hydrogen atom positions at axial conformation but directs an opposite side to the nicotinamide ring of NAD⁺. Dotted lines show hydrogen bonds with distances less than 3.1Å except for 3.3 Å (K106-INS in R178A), 3.2 Å (Y135-INS in R178A) and 3.4 Å (Y163-INS in wild type).

The binding of myo-inositol to the mutant active site was formed by hydrogen bonds consisting of six residues, Lys106, Tyr135, Tyr163, Glu165, His195, and His318, but not Asp191 observed in the case of the wild-type. Superposition of the structures between the wild-type and the mutant revealed the structural differences in binding of myo-inositol to the wild-type and mutant Pl-sIDH in detail. Above 6 residues of both the R178A mutant and the wild-type fit well upon superposition. On the contrary, although the Cα atoms of Asp191 fit well in the two structures, the side chain of this residue in the mutant was flipped outside from the active site, thus not forming hydrogen bonds with the inositol. This orientation of the side chain was similar to those observed in structures of the NAD⁺-bound wild-type (PDB entry 5yab) and apo-R178A mutant (this study, Table 1), and similar to the lactone-bound complexes of the wild-type (PDB entry 5yap) and the R178A mutant (Figure 5). Furthermore, this conformational change of the side chain seemed to affect the binding mode of the inositol. In the R178A mutant, the C3 atom in the inositol is located below the C4 atom (NC4) of the nicotinamide ring, and the hydrogen atom at the C3 atom adopted equatorial conformation, whereas, in the wild-type, the axial hydrogen at the C4 atom is located below the NC4 atom. Thus, the inositol was bound to the active site with counter-clockwise rotation at an almost one-carbon unit in the mutant (Figure 6). This change in the binding mode of myo-inositol might affect the catalytic activity of the mutant enzyme, the hydrogen transfer to the NC4 atom from inositol, where a hydrogen atom of the inositol below the NC4 atom adopt an axial and equatorial conformation in the wild type and mutant, respectively. Since Asp191 interacts directly with myo-inositol in the wild-type Pl-sIDH (PDB entry 5ya8), replacing Arg178 with alanine might have induced a conformational change in the side chain of Asp191 and therefore affecting the binding of the inositol in the active site (Figure 6). As a result, this shift of binding mode might have caused about 180-fold reduction of the k_cat/K_m value (Table 2). Based on the crystal structures of Pl-sIDH studied so far, the loop structure in which Arg178 locates adopted two conformations as open and close. The Cα atoms of Arg178 in the substrate-unbound and lactone complexed structures [5], and Ala178 in the mutant structures exactly fitted on the same positions (Figure 2). Therefore, on binding of inositols, Arg178 moves toward them to induce the closed form of the loop. Accordingly, Arg178 may play an important role to set Asp191 in a correct position to induce the correct binding of myo-inositol.

It is noted that the R178A mutant abolished the catalytic activity for myo-inositol, and in the crystal structure, the sidechain of Asp191 flipped out from the inositol, that is the main difference from the wild-type structure. On the other hand, the mutant as well as wild type retain the activities for l-glucose, but in the crystal structures complexed with the lactone, the sidechain of Asp191 flipped out from the lactone in the both structures, much like the R178A mutant bound with myo-inositol. A structural difference between l-glucose and inositols is mainly the protruding C6 hydroxyl group in the glucose, and this C6 group is facing to the sidechain of Asp191 in the wild-type structure. Therefore, the conformational change of the sidechain upon binding of l-glucose in the wild type may be induced by a steric hindrance between the sidechain and the C6 group. van Straaten et al., reported that inositol dehydrogenase from Bacillus subtilis contains a catalytic triad Lys, Asp and His as proposed for 1,5-anhydro-D-fructose reductase (AFR) and other members of the GFO/IDH/MocA family [15–17]. Asp191 in Pl-sIDH corresponds to the one in the triad and is highly conserved among the family, and the D191A mutant of Pl-sIDH completely abolished the enzymatic activities for l-glucose as well as inositols, as demonstrated previously [5]. However, the side chain of Asp191 flipped out from the substrate in the crystal structure despite that Pl-sIDH exhibits the catalytic activity for l-glucose. On the other hands, in the GFO/IDH/MocA family, catalytic residues of aldose-aldose oxidoreductase from Caulobacter crescentus (CcAAOR) and GFO from Zymomonas mobilis are dyad of Lys104 and Tyr189, and Lys181 and Tyr269, respectively, [18,19]. These two residues correspond to Lys106 and His195 in Pl-sIDH, where His195 is considered to be the acid-base catalyst [6]. Therefore, Asp191 may be critical for the substrate binding but not for catalysis. Further study may need to solve the role of Asp191 in detail.

As described above, Pl-sIDH belongs to GFO/IDH/MocA family, and in our previous study, based on the primary sequence homology and crystal structures deposited in PDB, a subfamily including Pl-sIDH was revealed, that is characterized with a tetramer conformation, a swap loop from neighbor subunit involving the substrate binding, and a flexible loop at the active site entrance [5]. As shown in Figure 7, the sequence alignment by ClustalW among the subfamily shows that the position at Arg178 in Pl-sIDH is semi-conserved with either Arg or Lys. Several structures in the subfamily are deposited in PDB; however, only few have been functionally characterized so far. The crystal structure and functional characterization of bacterial levoglucosan dehydrogenase (LGDH) was reported recently [14], demonstrating that Arg176 and Asp189 in LGDH, which correspond to Arg178 and Asp191 in Pl-sIDH, respectively, are involved in the substrate binding. Although Pl-sIDH and LGDH do not share common substrates, the conserved arginine residue could affect the substrate specificities and catalytic activities of enzymes in the subfamily.

Figure 7.

Sequence and structure comparisons between members of the IDH subfamily.

A part of the amino acid sequences among the subfamily of IDHs were aligned by the software ClustalW. PDB entries, 5YA8 underlined in red as Pl-sIDH, 4FB5 as an oxidoreductase from Rhizobium etii, 6A3I as levoglucosan dehydroganase, 4H3V as an oxidoreductase from Kribbella flavida, 4GQA as an oxidoreductase from Klebsiella pneumoniae and 3DTY as an oxidoreductase from Pseudomonas syringae were compared. An orange rectangle shows the position of Arg178 in Pl-sIDH, and a blue rectangle shows the conserved catalytic residues of Asp and His. Superposition of the wild-type Pl-sIDH complexed with myo-inositol (INS) (PDB entry 5ya8) colored in green and LGDH complexed with levoglucosan (LG) (PDB entry 6a3i) colored in red is shown. Arginine and aspartic acid residues presented in stick-models are conserved structurally as well as in amino acid sequences.

Acknowledgments

We thank the beamline staff at the Photon Factory for the technical support during the data collection

Author Contributions

SI, YS, AN, and SY conceived and supervised the study; MS, KF, SI, YS, AN, and SY designed the experiments; MS, KK, MT, and KF performed the experiments; MS, KF, MT, SI, YS, AN, and SY analyzed the data; MS, AN, and SY wrote the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.