X-ray structure of Arthrobacter globiformis M30 ketose 3-epimerase for the production of D-allulose from D-fructose

Abstract

The X-ray structure of ketose 3-epimerase from Arthrobacter globiformis M30, which was previously reported to be a D-allulose 3-epimerase (AgD-AE), was determined at 1.96 Å resolution. The crystal belonged to the hexagonal space group P6₅22, with unit-cell parameters a = b = 103.98, c = 256.53 Å. The structure was solved by molecular replacement using the structure of Mesorhizobium loti L-ribulose 3-epimerase (MlL-RE), which has 41% sequence identity, as a search model. A hexagonal crystal contained two molecules in the asymmetric unit, and AgD-AE formed a homotetramer with twofold symmetry. The overall structure of AgD-AE was more similar to that of MlL-RE than to the known structures of D-psicose (alternative name D-allulose) 3-epimerases (D-PEs or D-AEs), although AgD-AE and MlL-RE have different substrate specificities. Both AgD-AE and MlL-RE have long helices in the C-terminal region that would contribute to the stability of the homotetramer. AgD-AE showed higher enzymatic activity for L-ribulose than D-allulose; however, AgD-AE is stable and is a unique useful enzyme for the production of D-allulose from D-fructose.

Keywords: Arthrobacter globiformis; X-ray structure; d-allulose; ketose 3-epimerase; β/α-barrel.

Figure 1

The overall structure of AgD-AE. (a) Tetramer structure of AgD-AE showing the front view, side view and top view (left to right). (b) Monomer structure (molecule A) of AgD-AE.

Figure 2

Structure comparisons amongst related enzymes. The monomer structure and the front view, side view and top view of the overall structure are represented from the left to right for each enzyme. (a) A. globiformis d-allulose 3-epimerase (AgD-AE), (b) M. loti l-ribulose 3-epimerase (MlL-RE), (c) P. cichorii d-tagatose 3-­epimerase (PcD-TE), (d) C. cellulolyticum d-psicose 3-epimerase (CcD-PE), (e) A. tumefaciens d-psicose 3-epimerase (AtD-PE). In (a) and (e) one dimer is shown as a ribbon model to show the position of α8 in the homotetramer.

Figure 3

Active-site structure comparisons of AgD-AE, MlL-RE and PcD-TE. The active-site structures of (a) AgD-AE, (b) MlL-RE and (c) PcD-TE in complex with d-fructose are shown. Manganese ions are shown as blue or purple spheres and water molecules are shown as red or pink spheres. The bound d-­fructose in PcD-TE is presented as a stick model with the C atoms numbered. (d) The active-site structures in (b) and (c) were superimposed onto that in (a). Most of the residues in the active sites are highly conserved. The only difference in residues between AgD-AE and MlL-RE is Met110 in AgD-AE, which corresponds to His111 in MlL-RE. In AgD-AE the water molecules (W1, W2 and W3) correspond to the positions of O1, O2 and O3 of the bound d-­fructose in PcD-TE, respectively.

Figure 4

Structural difference between AgD-AE and MlL-RE. The N-terminal region (α1–β2–α2) of AgD-AE (gray model) is shifted slightly towards the entrance to the active site, including His6 and Ser63 (stick models), compared with the corresponding region (α1–β2–α2) of MlL-RE (pink model). In AgD-AE, β2 (Leu33-Val34-Glu35-Phe36-Pro37) connects to α2 (Asp45–Gly57) via the distorted region of α2 (Asp40-Pro41-Phe42-Ser43-Phe44). In MlL-RE, the corresponding region (Asp41-Pro42-Ala43-Asp44-Val45) does not form a distorted helix between the shorter β2 (Leu34-Ile35-Glu36-Phe37) and α2 (Asp46–Gly58). Because the active-site structure is highly conserved in both enzymes, a subtle shift in this region may affect the entrance leading to the active site and be responsible for the difference in substrate specificity.

Figure 5

Sequence alignment of AgD-AE and MlL-RE. The residues colored red are coordinated to the metal ion. A difference in the residues in the active site between AgD-AE and MlL-RE is indicated in blue.