- Email: [email protected]
The Structures of l -Rhamnose Isomerase from Pseudomonas stutzeri in Complexes with l -Rhamnose and d -Allose Provide Insights into Broad Substrate Specificity
J. Mol. Biol. (2007) 365, 1505–1516
doi:10.1016/j.jmb.2006.11.004
The Structures of L-Rhamnose Isomerase from Pseudomonas stutzeri in Complexes with L-Rhamnose and D-Allose Provide Insights into Broad Substrate Specificity Hiromi Yoshida 1 , Mitsugu Yamada 1 , Yuya Ohyama 2 , Goro Takada 2 Ken Izumori 2 and Shigehiro Kamitori 1 ⁎ 1
Molecular Structure Research Group, Information Technology Center and Faculty of Medicine, Kagawa University, 1750-1 Ikenobe, Miki-cho, Kita-gun, Kagawa 761-0793, Japan 2
Department of Biochemistry and Food Science, Faculty of Agriculture and Rare Sugar Research Center, Kagawa University, Miki-cho, Kagawa 761-0795, Japan
Pseudomonas stutzeri L-rhamnose isomerase (P. stutzeri L-RhI) can efficiently catalyze the isomerization between various aldoses and ketoses, showing a broad substrate specificity compared to L-RhI from Escherichia coli (E. coli L-RhI). To understand the relationship between structure and substrate specificity, the crystal structures of P. stutzeri L-RhI alone and in complexes with L-rhamnose and D-allose which has different configurations of C4 and C5 from L-rhamnose, were determined at a resolution of 2.0 Å, 1.97 Å, and 1.97 Å, respectively. P. stutzeri L-RhI has a large domain with a (β/α)8 barrel fold and an additional small domain composed of seven α-helices, forming a homo tetramer, as found in E. coli L-RhI and D-xylose isomerases (D-XIs) from various microorganisms. The β1–α1 loop (Gly60–Arg76) of P. stutzeri L-RhI is involved in the substrate binding of a neighbouring molecule, as found in D-XIs, while in E. coli L-RhI, the corresponding β1–α1 loop is extended (Asp52–Arg78) and covers the substrate-binding site of the same molecule. The complex structures of P. stutzeri L-RhI with L-rhamnose and D-allose show that both substrates are nicely fitted to the substrate -binding site. The part of the substrate-binding site interacting with the substrate at the 1, 2, and 3 positions is equivalent to E. coli L-RhI, and the other part interacting with the 4, 5, and 6 positions is similar to D-XI. In E. coli L-RhI, the β1–α1 loop creates an unique hydrophobic pocket at the the 4, 5, and 6 positions, leading to the strictly recognition of L-rhamnose as the most suitable substrate, while in P. stutzeri L-RhI, there is no corresponding hydrophobic pocket where Phe66 from a neighbouring molecule merely forms hydrophobic interactions with the substrate, leading to the loose substrate recognition at the 4, 5, and 6 positions. © 2006 Elsevier Ltd. All rights reserved.
*Corresponding author
Keywords: X-ray structure; rhamnose isomerase; rare sugars; Pseudomonas stutzeri
Abbreviations used: P. stutzeri L-RhI, Pseudomonas stutzeri L-rhamnose isomerase; E. coli L-RhI, Escherichia coli L-rhamnose isomerase; S. rubiginosus D-XI, Streptomyces rubiginosus D-xylose isomerase; A. missouriensis D-XI, Actinoplanes missouriensis D-xylose isomerase; T. thermophilus D-XI, Thermus thermophilus D-xylose isomerase; P. stutzeri L-RhI/L-rhamnose, P. stutzeri L-RhI/ L-rhamnose complex; P. stutzeri L-RhI/D-allose, P. stutzeri L-RhI/D-allose complex; P. stutzeri SeMet L-RhI, selenomethionine substituted P. stutzeri L-RhI. E-mail address of the corresponding author: [email protected]
Introduction L-Rhamnose isomerase, which catalyzes the reversible isomerization of L-rhamnose to L-rhamnulose, has been found to be involved in the metabolism of 1,2 L-rhamnose in Escherichia coli and also to be present in some other microorganisms like Lactobacillus, 3 Salmonella, 4 and Pseudomonas. 5 L-Rhamnose isomerase from Pseudomonas stutzeri (P. stutzeri L-RhI; Swiss Prot accession number Q75WH8, 430 amino acid residues, 46,975 Da)6 shows a broader substrate specificity than L-RhI
0022-2836/$ - see front matter © 2006 Elsevier Ltd. All rights reserved.
1506 from Escherichia coli (E. coli L-RhI; Swiss Prot accession number P32170, 419 amino acid residues, 47,199 Da)2, catalyzing the isomerization between various aldoses and ketoses, as well as between L-rhamnose and L-rhamnulose, in the presence of appropriate metal ions.5–7 Leang et al. reported that a recombinant His-tagged P. stutzeri L-RhI efficiently catalyzed the isomerization between L-rhamnose and L-rhamnulose, L-mannose and L-fructose, L-lyxose and L-xylulose, D-ribose and D-ribulose, and D-allose and D-psicose, as listed in Figure 1(a) with the enzyme activities. Since some of them are so-called “rare sugars” which exist in small amounts in nature, P. stutzeri L-RhI is exploited for industrial applications in rare sugar production.8–10 However, little is known about the mechanism responsible for the broader substrate specificity of the enzyme. Korndörfer et al. reported the X-ray structure of E. coli L-RhI (PDB codes 1D8W, 1DE5 and 1DE6)11 and a structural comparison with the D-xylose isomerase from Streptomyces rubiginosus (S. rubiginosus D-XI; Swiss Prot accession number P24300,
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
387 amino acid residues, 43,096 Da, PDB code 4XIS). 12,13 Despite the poor sequence identity (13%) between them, they have some similarities in structure. Both have a large domain with a (β/α)8 barrel and an additional small domain composed of α-helices, to form a homo tetramer, and each subunit has two adjacent metal ions at the substrate-binding site. It was also proposed that the aldose-ketose isomerization reactions of E. coli L-RhI are catalyzed by the metal-mediated hydride shift mechanism (Figure 1(b)), which is available to D-XIs.11 To elucidate the substrate-recognition mechanism responsible for the broad substrate specificity of P. stutzeri L-RhI, understanding the three-dimensional structure of P. stutzeri L-RhI in complexes with its substrates is important. Sequence alignment between P. stutzeri L-RhI and E. coli L-RhI indicates the amino acid residues involved in the metal binding are strictly conserved, but the sequence identity between them is only 17%,6,7 suggesting that P. stutzeri L-RhI has an unique three-dimensional structure at the substrate-binding site. The crystallization and preliminary X-ray analysis of P. stutzeri L-RhI have been reported.14 Here we report the crystal structure of P. stutzeri L-RhI alone and in complexes with L-rhamnose (P. stutzeri L-RhI/Lrhamnose) and D-allose (P. stutzeri L-RhI/D-allose).
Results and Discussion Quality of structures
Figure 1. (a) Chemical structures of L-rhamnose, L-mannose, L-lyxose, D-ribose, D-allose, D-xylose, D-glucose, and D-sorbitol are shown in Fischer projection formulas. The enzyme activities of P. stutzeri L-RhI for them, Kcat/Km (μM−1 min−1),7 are given in parentheses. (b) Schematic diagram of the proposed metal-mediated hydride shift reaction of L-RhI.
The structures have been refined to an R-factor of 0.155 (P. stutzeri L-RhI), 0.170 (P. stutzeri L-RhI/ L-rhamnose) and 0.172 (P. stutzeri L-RhI/D-allose), respectively, with good chemical geometries, as indicated in Table 1. In Ramachandran plots15 90.6% of all residues in P. stutzeri L-RhI, 91.1% in P. stutzeri L-RhI/L-rhamnose, and 90.7% in P. stutzeri LRhI/D-allose are shown to be in the most favoured regions, with no residue in the disallowed regions as determined by the program PROCHECK.16 In all X-ray structures, there are four protein molecules (Mol-A, B, C and D), forming a homotetramer, in an asymmetric unit. The final (2Fo–Fc) electron density map of all X-ray structures shows that almost all atoms of protein molecules, solvent molecules, and metal ions fitted nicely. Poor or invisible electron density is found in several residues at the N terminus (Met1–Glu3), the loop region residues (Glu313–Phe320), and the C terminus (Ala421–Ile430) with 6-His, suggesting that these regions are mobile and highly disordered. Two bound metal ions (M-1 and M-2) at the substratebinding site were found in each protein molecule. The electron density of M-1 is very strong, while that of M-2 is relatively weak, suggesting the lower occupancy of M-2. In E. coli L-RhI, M-1 and M-2 are considered as to be Zn2+ and Mn2+ , respectively; however, identifications of metal ions, M-1 and M-2, of P. stutzeri L-RhI are difficult from current X-ray
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
1507
Table 1. X-ray data collection and refinement statistics Datase Beam line Wavelength (Å) Temperature (K) Resolution (Å) No. of measured refs No. of unique refs Completeness (%) Rmerge Io/σ(Io) Space group Cell dimensions a(Å) b(Å) c(Å) β(°) Resolution (Å) Completeness (%) Rfactor Rfree r.m.s.d. bond lengths (Å) r.m.s.d. bond angles (deg.) No. of amino acids
P. stutzeri SeMet L-RhI
P. stutzeri Native L-RhI
P. stutzeri L-RhI with L-rhamnose
P. stutzeri L-RhI with D-allose
KEK PF BL-6A 0.97870 100 50–2.10 711,721 97,228 100.0 (100.0) 0.062 (0.123) 7.0 (4.7) P21
KEK PF BL-6A 1.0000 100 50–2.00 647,726 99,909 94.9 (95.0) 0.074 (0.298) 13.3 (7.5) P21
KEK PF BL-5A 1.0000 100 50–1.97 876,965 117,301 99.5 (95.1) 0.104 (0.242) 9.5 (5.5) P21
KEK PF BL-5A 1.0000 100 50–1.97 876,565 117,529 99.9 (98.7) 0.115 (0.234) 10.2 (6.6) P21
74.63 104.79 113.65 107.86
74.25 103.97 107.02 106.81 48.79–2.00 92.3 0.155 0.192 0.005 1.1 Mol-A 421 Mol-B 420 Mol-C 421 Mol-D 417 1783 Zn 8
74.62 104.76 114.29 107.96 33.25–1.97 97.6 0.170 0.205 0.005 1.1 Mol-A 421 Mol-B 421 Mol-C 428 Mol-D 419 1364 Zn 8 L-Rhamnose 4
74.65 104.73 114.53 108.00 33.25–1.97 97.7 0.172 0.204 0.005 1.1 Mol-A 419 Mol-B 421 Mol-C 426 Mol-D 419 1236 Zn 8 D-Allose 4
19.1 29.9 50.6 19.7 30.4
22.1 34.6 31.9 42.0 23.1 28.4
20.4 34.9 30.3 40.7 21.7 25.7
90.6/ 8.9/ 0.5
91.1/ 8.4/ 0.5
90.6/ 8.9/ 0.5
No. of solvent molecules No. of ligand B-factor (Å2) Protein Ligand M-1 M-2 β1−α1 region solvent Ramachandran plot (%) Favoured/additional allowed/generously allowed
Rmerge = ∑∑ | Ii- | / ∑. Values in parentheses are of the high resolution bin (2.18–2.10 Å for SeMet, 2.07–2.00 Å for native, 2.04–1.97 Å for L-rhamnose and 2.04–1.97 Å for D-allose).
diffraction data. Therefore, the bound metal ions of P. stutzeri L-RhI were refined as Zn2+ with occupancies of 1.0 for M-1 and 0.5 for M-2, since the atomic absorption spectrum of purified P. stutzeri L-RhI showed the presence of metal ions, Zn2+, Mn2+, and Ni2+, at a ratio of 4:1:1 (data not shown). The bound substrates were found in each protein molecule in the enzyme–substrate complexes. Simulated annealing omit maps for the bound metal ions and substrates are shown in Figure 2.17,18 Since the four molecules are almost identical in structure with r.m.s. deviations for main-chain atoms being 0.69 Å– 0.85 Å in P. stutzeri L-RhI, 0.54 Å–0.81 Å in P. stutzeri L-RhI/L-rhamnose, and 0.53 Å–0.78 Å in P. stutzeri L-RhI/D-allose, the structural description here concentrates on Mol-A (if not stated otherwise). Overall structure of P. stutzeri L-RhI The overall subunit structure of P. stutzeri L-RhI is shown in Figure 3. P. stutzeri L-RhI has 15 α-helices with 54% of the amino acid residues and eight
β-strands with 8% of the amino acid residues. The structure can be divided into a large domain (Phe50–Val356) and an additional small domain (N terminus-Lys49, and Asp357-C terminus). The large domain forms the core of the enzyme with the (β/α)8 barrel fold. The bound metal ions are located at the centre of the barrel, and several His and Asp residues in β-strands direct their side-chain groups toward the inside of the barrel to coordinate with metal ions. The large domain of P. stutzeri L-RhI is similar in structure to that of E. coli L-RhI and S. rubiginosus D-XI, in spite of low sequence identities between them. The r.m.s. deviations for Cα atoms and Z-scores of DALI search19 for the large domain of P. stutzeri L-RhI are 2.0 Å and 28.5 with E. coli L-RhI (1DE5, 19% identity), and 1.9 Å and 33.3 with S. rubiginosus D-XI (4XIS, 19% identity). The additional small domain is composed of two α-helices (α0 and α9′) in the N-terminal region and five α-helices (α9–α13) in the C-terminal region. The α-helices are packed pairwise against each other to form a bundle of α-helices. The α0 is largely kinked to form
1508
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
Figure 2. Simulated annealing omit map at the 4.0σ contours level for the bound metal ion and (a) L-rhamnose, and (b) D-allose illustrated by the programs X-fit17 and Raster3D.18 Metal ions and the C1 and C6 of substrates are labeled.
hydrophobic interactions with α11 and α12. The additional small domain composed of α-helices has been also observed in E. coli L-RhI and D-XIs; however, its structure varies depending on the enzyme, leading to an unique inter-subunit surface in each enzyme, as described next. The homo-tetramer of P. stutzeri L-RhI has dimensions of 100 Å × 90 Å × 80 Å with 222 symmetry, as shown in Figure 4. The four subunits are
arranged so that the loop regions between the βstrands and α-helices of the large domain are located at the centre of the tetramer, and the outer surface of the tetramer is mostly covered by α-helices. Inter-subunit contact areas calculated using the program AREAIMOL in the CCP4 program suite20 are 1943 Å2 between Mol-A and Mol-B, 2303 Å2 between Mol-A and Mol-C, and 1625 Å2 between Mol-A and Mol-D, respectively (Table 2). A similar
Figure 3. Overall subunit structure of P. stutzeri L-RhI illustrated by the program PyMol. The structure of P. stutzeri L-RhI is comprised of a large domain with the (β/α)8 barrel fold shown in gray (α-helices) and yellow (β-strands), and an additional small domain shown in dark green. In the large domain, α-helices are numbered from α1 to α8, and β-strands are numbered from β1 to β8. Two α-helices in the N-terminal region of the additional small domain are numbered as α0 and α0′, and five α-helices of the Cterminal region are numbered α9, α10, α11, α12 and α13. The small blue spheres indicate metal ions [http://www.pymol.org].
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
Figure 4. Overall tetrameric structure of P. stutzeri L-RhI illustrated by the program PyMol in (a) top view and (b) side view. The four molecules are coloured in light green (Mol-A), light blue (Mol-B), light magenta (Mol-C) and light yellow (Mol-D). The dark coloured part of each molecule represents the additional small domain. The small blue spheres indicate metal ions [http://www. pymol.org].
homo-tetramer is also found in E. coli L-RhI11 and other D-XIs from S. rubiginosus 13 , Actinoplanes missouriensis (A. missouriensis D-XI; Swiss Prot accession number P12851, 393 amino acid residues, 43,368 Da, PDB code 2XIN) 21,22 and Thermus thermophilus (T. thermophilus D-XI; Swiss Prot accession number P26997, 387 amino acid residues, 43,907 Da, PDB code 1BXB).23,24
1509 The structure of Mol-A and Mol-B in the tetramer is illustrated in Figure 5(a). Although the intersubunit contact area is 1943 Å2, the dimer structure of Mol-A and Mol-B is thought to be relatively unstable, because all of the inter-subunit interactions are formed by the loop regions of two molecules, suggesting that the formation of a tetramer is indispensable to maintaining the presented dimer structure of Mol-A and Mol-B. The structure of the inter-subunit between Mol-A and Mol-B is very important for the enzyme, because the loop between β1 and α1 (the β1–α1 loop) approaches the substrate-binding site for a neighbouring molecule to interact with a substrate as shown by red in Figure 5(a), and because Mol-A and Mol-B are expected to form the accessible surface for substrate binding; the bound metal ions can be visible from this molecular surface. In A. missouriensis D-XI, the β1–α1 loop partly makes up the substrate-binding site for a neighbouring molecule and an extensively opened accessible surface is formed between Mol-A and Mol-B, as found in P. stutzeri L-RhI (Figure 5(b)). By contrast, in E. coli L-RhI, the β1–α1 loop covers the bound metal ions in the same molecule, and there is a clear boundary between Mol-A and Mol-B, giving a relatively small contact area of 977 Å2, as shown in Figure 5(c). The inter-subunit interactions between Mol-A and Mol-B are closely related to the difference in substrate specificity between P. stutzeri L-RhI and E. coli L-RhI, as described in detail below. Between Mol-A and Mol-C, α4, α5, α6 and α7 are involved in the inter-subunit interaction, forming the wide inter-subunit contact area. In A. missouriensis D-XI, along with α4, α5, α6 and α7, the additional small domain is also involved in the intersubunit contacts, giving the extensively wide contact area of 4857 Å2, suggesting that the additional small domain contributes greatly to the stability of the subunit association in A. missouriensis D-XI. Between Mol-A and Mol-D, the additional small domains are mainly involved at the inter-subunit surface, with a contact area of 1625 Å2, which does not vary much in E. coli L-RhI (1928 Å2) and A. missouriensis D-XI (1640 Å2), in spite of the difference in structure of the additional small domains. Structure of substrate-binding site P. stutzeri RhI requires metal ions for its catalytic activity. Two putative metal ions were observed in the P. stutzeri RhI and P. stutzeri RhI/substrate complexes. Since no metal ions were added in the
Table 2. Inter-subunit contact areas mol A–B, mol A–C, and A–D of P. stutzeri L-RhI, E.coli RhI, and A. missouriensis D-XI Contact areas [Å2] P. stutzeri L-RhI E. coli L-RhI A. missouriensis D-XI
Mol A–B
Mol A–C
Mol A–D
1943 977 1698
2303 2102 4857
1625 1928 1640
1510
Figure 5. Subunit interfaces between Mol-A and MolB of (a) P. stutzeri RhI/D-allose, (b) A. missouriensis D-XI/Dsorbitol (PDB code 2XIN), and (c) E. coli RhI/L-rhamnose (PDB code 1DE6) illustrated by the program PyMol. The β1–α1 loops are indicated in red. Zn2+ in P. stutzeri RhI/Dallose are shown as small blue spheres. Co2+ in A. missouriensis D-XI/D-sorbitol are shown as orange spheres. Zn2+ and Mn2+ in E. coli RhI/L-rhamnose are shown as small blue spheres and large purple spheres, respectively. The bound D-allose, D-sorbitol and L-rhamnose are represented by pink, red and green stick models, respectively, and the amino acid residues in the β1–α1 loop interacting with a substrate are represented by a blue stick model [http://www.pymol.org].
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
cultivation and purification process, it is possible that metal ions from the culture medium were retained in the enzyme. With respect to D-XIs and E. coli RhI, which also require a divalent metal ion for their activity, one or two metal ions were observed. One is “structural” (M-1) to help the substrate binding, and the other is “catalytic” (M-2) to help the hydride shift. The structural metal ion is reported to bind tightly with a lower dissociation constant than the catalytic metal ion.11 Also in P. stutzeri RhI, a low occupancy of catalytic metal ions is found. As shown in Figure 6(a) the metal ions in P. stutzeri LRhI are coordinated in a distorted octahedral form with six coordination bonds; they are Glu219, Asp254, His281, Asp327, and two water molecules (W-1 and W-2) for M-1, and His257, Asp289 and four water molecules (W-2, W-3, W-4 and W-5) for M-2. The distance between M-1 and M-2 is 4.1 Å, and the coordination water molecule of W-2 bridges the metal ions. The structure of the substrate-binding site of P. stutzeri L-RhI with the bound L-rhamnose is shown in Figure 6(b). The O1, O2, and O3 atoms of L-rhamnose coordinate with M-1 and M-2, instead of W-3, W-2, and W-1 in the unliganded form, respectively. O1 and O3 make a hydrogen bond with Lys221 and Glu219, respectively, helping to fix the substrate in the proper conformation for the hydride shift between the C1 and C2 of substrate. There are still two water molecules (W-4 and W-5) coordinating to M-2 with the hydrogen bonds of Asp291 (W-4), and of Asp254 and Asp291 (W-5). It is likely that W-4 is a key water molecule responsible for the hydride shift in the proposed catalytic mechanism, because it possibly makes hydrogen bonds with the O1 and O2 of L-rhamnose. O4 and O5 make a hydrogen bond with Asp327 and His101, respectively, recognizing the hydroxyl groups at the 4 and 5 positions of L-rhamnose, and C6 undergoes hydrophobic interactions with Trp57, Phe131, and *Phe66 (Phe66 from a neighbouring molecule). On the basis of structure of P. stutzeri L-RhI/L-rhamnose, the enzyme–substrate interactions of L-mannose and L-lyxose can be deduced. Since L-mannose and L-lyxose have the same configurations of asymmetric carbon atoms as L -rhamnose, substrate–enzyme interactions are thought to be almost identical between them, but enzyme activities for them are significantly different, as listed in Figure 1(a). The structural variation between these substrates is found at 6 position, where L-mannose has an additional hydroxyl group and L-lyxose has no group. The simple modelling to generate the bound L-mannose by adding a hydroxyl group to L-rhamnose at the 6 position indicated that the O6 atom can be located in a staggered conformation without any unusual short contacts with amino acid residues; the torsion angle of C4-C5-C6-O6 is −60 °. In the resultant model of the bound L-mannose, O6 is in hydrophobic environment created by *Phe66, Trp104, and Phe131, and it may decrease the affinity of substrate to enzyme, as described in detail below. In the binding of L-lyxose, which has no group at the 6 position, a lack of hydrophobic interactions
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
of C6 with the enzyme found in P. stutzeri L-RhI/ L-rhamnose may lead the lower enzyme activity for L-lyxose. The structure of the substrate-binding site of P. stutzeri L-RhI with the bound D-allose is shown in Figure 6(c). The interactions between O1, O2, and O3 of D-allose and the enzyme including metal ions are almost identical to those found in P. stutzeri L-RhI/L-rhamnose, because D-allose has the same configurations of C1, C2, and C3 as L-rhamnose. O4 makes a hydrogen bond with His101, and O5 with Asp327, different from the bound L-rhamnose, where O4 bonds with Asp327 and O5 with His101. This means that P. stutzeri L-RhI can recognize substrates with different configurations of C4 and C5, by using His101 and Asp327, and vice versa. However, the substrate recognition at the 4 and 5 positions of L-rhamnose is thought to be more favourable for P. stutzeri L-RhI than that of D-allose, because the enzyme activity for D-allose is significantly lower than that for L-rhamnose. C6 makes hydrophobic interactions with Trp57, and also O6 is in a hydrophobic environment created by *Phe66, Trp104, and Phe131 without any hydrogen bond with amino acid residues. The hydroxyl group, O6, in hydrophobic environment increases structural energy of the substrate–enzyme complex, leading low affinity of substrate to the enzyme. This is consistent with the fact that enzyme activity for Dribose, which has no group at the 6 position is higher than that for D-allose.7 Comparison with E. coli RhI and A. missouriensis D-XI A superimposed substrate-binding structure of E. coli RhI with the bound L-rhamnose and P. stutzeri L-RhI with the bound L-rhamnose and D-allose is illustrated in Figure 7(a). The amino acid residues involved in metal-binding are conserved, and also the interactions between O1, O2 and O3 of L-rhamnose and the enzyme found in P. stutzeri L-RhI are very similar to those in E. coli RhI, suggesting that the catalytic mechanism of the metal-mediated hydride shift proposed in E. coli RhI (Figure 1(b)) can be assumed for P. stutzeri L-RhI. However, remarkable structural differences are found in the recognition of the atoms at the 4, 5 and 6 positions of a substrate. The β1–α1 loop (Gly60–Arg76) of P. stutzeri L-RhI is involved in the substrate binding of a neighbouring molecule, while in E. coli L-RhI, the corresponding β1–α1 loop is extended (Asp52– Arg78) and covers the substrate-binding site of the same molecule, as mentioned above (Figure 5). In E. coli L-RhI, Val53, Leu63, and Ile67 from the β1–α1 loop create a unique hydrophobic pocket to make many van der Waals contacts with the substrate. There is no corresponding hydrophobic pocket in P. stutzeri L-RhI, where *Phe66 from a neighbouring molecule merely forms hydrophobic interactions with the substrate. In addition, Phe336 of E. coli L-RhI partly attends to the formation of this hydrophobic pocket by changing its side-chain
1511 conformation on the binding of a substrate to make favourable interactions with the substrate.11 In P. stutzeri L-RhI, the corresponding residue is Ser329, 8.8 Å from C6 of L-rhamnose, and there is a large space between Ser329 and the bound substrate. The unique hydrophobic pocket around the methyl group of C6 formed by residues Val53, Leu63, Ile67 from the β1–α1 loop and Phe336 in E. coli L-RhI strictly recognizes L-rhamnose as the most suitable substrate, leading to the high substrate specificity of E. coli L-RhI. In fact, unusual short contacts between Ile67 and Phe336 of E. coli L-RhI and the bound D-allose in P. stutzeri L-RhI are observed (Figure 7(a)). On the other hand, the substratebinding site of P. stutzeri L-RhI without the hydrophobic pocket loosely recognizes both L-rhamnose and D-allose as substrates. A superimposed substrate binding structure of A. missouriensis D-XI with the bound inhibitor Dsorbitol (Figure 1(a)) and P. stutzeri L-RhI with the bound L-rhamnose is illustrated in Figure 7(b). Since D-glucose, as well as D-xylose, is also a suitable substrate of D-XI, D-sorbitol having exactly the same configuration of asymmetric carbon atoms as D-glucose acts as an inhibitor of D-XI. The amino acid residues involved in the metal binding are not conserved between them (His281/Asp245 and Asp254/Glu217), and O1, O2, and O4 of D-sorbitol coordinate with metal ions in A. missouriensis D-XI, instead of O1, O2, and O3 of L-rhamnose in P. stutzeri L-RhI. Since the substrates of D-XI, D-xylose and D-glucose (as an inhibitor, D-sorbitol), have a different configuration at C3 from the substrate of L-RhI, L-rhamnose, the variations in metal-binding site structure and substrate–metal interactions are expected to be closely related to the difference in substrate specificity between D-XI and L-RhI. Interestingly, the amino acid residues involved in the interactions with the substrate at the 4, 5 and 6 positions are conserved between P. stutzeri L-RhI and A. missouriensis D-XI (Trp57/16, His101/54, Phe131/94, Trp179/137, Asp327/292 and *Phe66/26), except Ser329/Lys294, suggesting that A. missouriensis D-XI loosely recognizes a substrate at these positions, as well as P. stutzeri L-RhI. It is notable that Streptomyces albus D-XI (Swiss Prot accession number P24299, 390 amino acid residues, 43,289 Da, PDB code 6XIA)25 having 64% homology with A. missouriensis D-XI was reported to have for D-allose and for L-rhamnose, 38% and 13%, respectively, of the activity it has for D-xylose,26 and also that glucose isomerase/xylose isomerase was mentioned to be able to isomerize a wide variety of substrates including L-rhamnose and D-allose in a review article.27 In D-XI, His has been considered as the residue involving in ring opening by assistance of an acidic amino acid residue, Asp,28 and they are His54 and Asp57 in A. missouriensis D-XI, as shown in Figure 7(b). In E. coli L-RhI, the corresponding residues are His103 and Tyr106, and in P. stutzeri L-RhI, the corresponding residues are His101 and Trp104. Therefore, E. coli and P. stutzeri L-RhIs can not
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
1512 adopt the above ring opening mechanism. Lambeir et al. and Allen et al. proposed the different ring opening mechanism that Lys182 or a metal-liganded water molecule involves in ring opening in D-XI,29–31
and Korndörfer et al. supposed that the similar ring opening mechanism could be expected for E. coli L-RhI.11 For P. stutzeri L-RhI, the similar ring opening mechanism seems to be possible; however,
Figure 6 (legend on next page)
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
1513
Figure 6. Stereo view of (a) the metal-binding site of P. stutzeri L-RhI, and the substrate-binding sites for (b) and (c) D-allose illustrated by the program PyMol. The metal ions are shown as blue spheres and the water molecules as red spheres. The bound L-rhamnose and D-allose are shown in blue and orange, respectively. The selected interactions among amino acid residues, metal ions, and water molecules are indicated by dotted lines. *Phe66 from a neighbouring molecule is shown as thin stick bonds. Carbon atoms of substrates are numbered from 1 to 6 [http://www. pymol.org]. L-rhamnose
there is no clear evidence to elucidate the ring opening mechanism in P. stutzeri L-RhI, so far.
Conclusion The presented X-ray structures of P. stutzeri L-RhI show that the part of the substrate-binding site interacting with the substrate at the 1, 2, and 3 positions is equivalent to E. coli L-RhI, and the other part interacting with the 4, 5, and 6 positions is similar to D-XI, leading to the loose substrate recognition at the 4, 5, and 6 positions. Because O1, O2 and O3 of a substrate coordinate with metal ions, configurations of C2, C3 of a substrate are very important for the enzyme activity. This is consistent with the fact that suitable substrates for P. stutzeri L-RhI, L-rhamnose, L-mannose, L-lyxose, D-ribose and D-allose, have the same configurations of C2 and C3. The configurations of C4 and C5 of a substrate do not affect the enzyme activity as much. P. stutzeri L-RhI with D-XI-like substrate-binding site at the 4 and 5 positions can recognize substrates with different configurations of C4 and C5, by using His101 and Asp327, and vice versa. This is why P. stutzeri L-RhI can recognize D-ribose and D-allose as a substrate. In SCOP (structural classification of proteins),32 E. coli L-RhI and D-XI are classified into different
families, the L-rhamnose isomerase family and the xylose isomerase family, respectively, because of the variation in the metal-binding site structure. Based on the metal-binding site structure, P. stutzeri L-RhI should also be placed in the L-rhamnose isomerase family; however, the substrate-binding site structure around the 4, 5 and 6 positions and the position of the β1–α1 loop of P. stutzeri L-RhI is similar to those of DXI, meaning that P. stutzeri L-RhI has structural features of both L-RhI and D-XI. This may strongly support the assumption by Korndörfer et al. that L-RhI and D-XI are derived from common precursor.11
Materials and Methods Protein preparation and crystallization The purification and crystallization of P. stutzeri L-RhI have been reported.14 The seleno-methionine-substituted P. stutzeri L-RhI (P. stutzeri SeMet L-RhI) was prepared in the same manner as that of P. stutzeri L-RhI except that the E.coli B834 cells were grown in LeMaster broth.33 A crystal of P. stutzeri L-RhI was grown by the vapour diffusion method using a protein solution (20 mg/ml P. stutzeri L-RhI in 20 mM Tris–HCl (pH 8.0)) and a reservoir solution (7–8% (w/v) polyethylene glycol 20,000, 50 mM Mes buffer (pH 6.3)). A crystal of P. stutzeri SeMet L-RhI was obtained under the same conditions. Crystals of P. stutzeri L-RhI/L-rhamnose and L-RhI/D-allose were obtained by a
1514
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
Figure 7. Stereo view of the substrate-binding sites of (a) E. coli L-RhI (cyan, 1DE6) with rhamnose (blue) and (b) A. missouriensis D-XI (green, 2XIN) with D-sorbitol (magenta) illustrated by the program PyMol. The substrate-binding site of P. stutzeri L-RhI (yellow) with L-rhamnose (thin blue bonds) and D-allose (thin orange bonds) are superimposed. Zn2+ in P. stutzeri L-RhI is shown as blue small spheres, Zn2+ and Mn2+ in E. coli L-RhI are shown as a small gray sphere and a large purple sphere, respectively. The Co2+ in A. missouriensis D-XI is shown as orange spheres. Blue and green labels indicate the residues belonging to E. coli L-RhI and A. missouriensis D-XI, respectively [http://www.pymol.org].
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
1515
soaking method, by incubation for 24 to 31 h with an additional 0.5 μl of 100 mM substrate solution. Data collection and structure determination Crystals were flash-cooled in liquid nitrogen at 100 K and X-ray diffraction data were collected using an ADSC Quantum 315 detector system on the BL-5A beam line and an ADSC Quantum 4R CCD detector system on the BL-6A beam line in the Photon Factory (Tsukuba, Japan). Diffraction data were processed using the programs HKL200034 and the CCP4 program suite.20 The collected data and scaling results are listed in Table 1. Initial SAD phasing at 2.1 Å of P. stutzeri SeMet L-RhI was performed by locating 24 selenium sites in the peak dataset using the program SOLVE.35 After electron density modification, 90% of amino acid residues could be located in the resultant electron density map, using the program RESOLVE.36,37 Further model building was performed with the program X-fit17 in the XtalView program system,38 and the structure was refined using the program CNS39 with Engh & Huber stereochemical parameters.40 Using the structure of P. stutzeri SeMet L-RhI, the structure of P. stutzeri L-RhI and that of P. stutzeri L-RhI/substrates were determined by molecular replacement method and refined at a resolution of 2.0 Å for P. stutzeri L-RhI, 1.97 Å for P. stutzeri L-RhI/L-rhamnose and 1.97 Å for P. stutzeri L-RhI/D-allose, using the program CNS. Water molecules were gradually introduced if the peaks above 3.5σ in the (Fo–Fc) electron density map were in the range of a hydrogen bond.
4.
5.
6.
7.
8.
9.
10.
Protein Data Bank accession number The atomic coordinates and structure factors of P. stutzeri L-RhI (PDB code 2HCV), P. stutzeri L-RhI/L-rhamnose (PDB code 2I56) and P. stutzeri L-RhI/D-allose (PDB code 2I57) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ.
11.
12.
Acknowledgements
13.
This study was supported in part by the National Project on Protein Structural and Functional Analyses and the “intellectual-cluster” from the Ministry of Education, Culture, Sports, Science and Technology of Japan. This research was performed with the approval of the Photon Factory Advisory Committee and the National Laboratory for High Energy Physics, and SPring-8, Japan.
14.
References
16.
1. Wilson, D. M. & Ajl, S. (1957). Metabolism of L-rhamnose by Escherichia coli. I. L-rhamnose isomerase. J. Bacteriol. 73, 410–414. 2. Moralejo, P., Egan, S. M., Hidalgo, E. & Aguilar, J. (1993). Sequencing and characterization of a gene cluster encoding the enzymes for L-rhamnose metabolism in Escherichia coli. J. Bacteriol. 175, 5585–5594. 3. Domagk, G. F. & Zech, R. (1963). Über den Abbau der
15.
17. 18.
Desoxyzucker durch Bakterienenzyme. I. L-RhamnoseIsomerase aus Lactobacillus plantarum. Biochem. Z. 339, 145–153. Englesberg, E. & Baron, L. S. (1959). Mutation to L-rhamnose resistance and transduction to L-rhamnose utilization in Salmonella typhosa. J. Bacteriol. 78, 675–686. Bhuiyan, S. H., Itami, Y. & Izumori, K. (1997). Isolation of an L-rhamnose isomerase-constitutive mutant of Pseudomonas sp. strain LL172: purification and characterization of the enzyme. J. Ferment. Bioeng. 84, 319–323. Leang, K., Takada, G., Ishimura, A., Okita, M. & Izumori, K. (2004). Cloning, nucleotide sequence, and overexpression of the L-rhamnose isomerase gene from Pseudomonas stutzeri in Escherichia coli. Appl. Environ. Microbiol. 70, 3298–3304. Leang, K., Takada, G., Fukai, Y., Morimoto, Y., Granstrom, T. B. & Izumori, K. (2004). Novel reactions of L-rhamnose isomerase from Pseudomonas stutzeri and its relation with D-xylose isomerase via substrate specificity. Biochim. Biophys. Acta, 1674, 68–77. Bhuiyan, S. H., Itami, Y. & Izumori, K. (1997). Immobilization of L-rhamnose isomerase and its application in L-mannose production from L-fructose. J. Ferment. Bioeng. 84, 558–562. Bhuiyan, S. H., Itami, Y., Rokui, Y., Katayama, T. & Izumori, K. (1998). D-Allose production from Dpsicose using immobilized L-rhamnose isomerase. J. Ferment. Bioeng. 85, 539–541. Bhuiyan, S. H., Itami, Y., Takada, G. & Izumori, K. (1999). Preparation of L-Talose and D-Gulose from L-Tagatose and D-Sorbose, respectively, using immobilized L-Rhamnose Isomerase. J. Ferment. Bioeng. 88, 567–570. Korndörfer, I. P., Fessner, W. D. & Matthews, B. W. (2000). The structure of rhamnose isomerase from Escherichia coli and its relation with xylose isomerase illustrates a change between inter and intra-subunit complementation during evolution. J. Mol. Biol. 300, 917–933. Wong, H. C., Ting, Y., Lin, H. C., Reichert, F., Myambo, K., Watt, K. W. et al. (1991). Genetic organization and regulation of the xylose degradation genes in Streptomyces rubiginosus. J. Bacteriol. 173, 6849–6858. Whitlow, M., Howard, A. J., Finzel, B. C., Poulos, T. L., Winborne, E. & Gilliland, G. L. (1991). A metalmediated hydride shift mechanism for xylose isomerase based on the 1.6 Å Streptomyces rubiginosus structures with xylitol and d-xylose. Proteins: Struct. Funct. Genet. 9, 153–173. Yoshida, H., Waynoon, P., Takada, G., Izumori, K. & Kamitori, S. (2006). Crystallization and preliminary X-ray diffraction studies of l-rhamnose isomerase from Pseudomonas stutzeri. Acta Crystallog. sect. F, 62, 550–552. Ramachandran, G. N. & Sasisekharan, V. (1968). Conformation of polypeptides and protein. Advan. Protein Chem. 23, 283–437. Laskowski, R. A., MacArthur, M. W., Moss, D. S. & Thornton, J. M. (1992). PROCHECK v.2: Programs to Check the Stereochemical Quality of Protein Structures. Oxford Molecular Ltd, Oxford, England. McRee, D. E. (1999). XtalView/Xfit: a versatile program for manipulating atomic coordinate and electron density. J. Struct. Biol. 125, 156–165. Merritt, E. A. & Bacon, D. J. (1997). Raster3D: photorealistic molecular graphics. Methods Enzymol. 277, 505–524.
1516
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
19. Holm, L. & Sander, C. (1993). Protein structure comparison by alignment of distance matrices. J. Mol. Biol. 233, 123–138. 20. Collaborative Computational Project 4. (1994). The CCP4 Suite: programs for protein crystallography. Acta Crystallog. sect. D, 50, 760–763. 21. Amore, R. & Hollenberg, C. P. (1989). Xylose isomerase from Actinoplanes missouriensis: primary structure of the gene and the protein. Nucl. Acids Res. 17, 7515. 22. Jenkins, J., Janin, J., Rey, F., Chiadmi, M., van Tilbeurgh, H., Lasters, I. et al. (1992). Protein engineering of xylose (glucose) isomerase from Actinoplanes missouriensis. 1. Crystallography and site-directed mutagenesis of metal binding sites. Biochemistry, 31, 5449–5458. 23. Dekker, K., Yamagata, H., Sakaguchi, K. & Udaka, S. (1991). Xylose (glucose) isomerase gene from the thermophile Thermus thermophilus: cloning, sequencing, and comparison with other thermostable xylose isomerases. J. Bacteriol. 173, 3078–3083. 24. Chang, C., Park, B. C., Lee, D.-S. & Suh, S. W. (1999). Crystal structures of thermostable xylose isomerases from Thermus caldophilus and Thermus thermophilus: possible structural determinants of thermostability. J. Mol. Biol. 288, 623–634. 25. Dauter, Z., Terry, H., Witzel, H. & Wilson, K. S. (1990). Refinement of glucose isomerase from Streptomyces albus at 1.65 Å with data from an imaging plate. Acta Crystallog. sect. B, 46, 833–841. 26. Sanchez, S. & Smiley, K. L. (1975). Properties of D-xylose isomerase from Streptomyces albus. Appl. Microbiol. 29, 745–750. 27. Bhosale, S. H., Rao, M. B. & Deshpande, V. V. (1996). Molecular and industrial aspects of glucose isomerase. Microbiol. Rev. 60, 280–300. 28. Collyer, C. A., Henrick, K. & Blow, D. M. (1990). Mechanism for aldose-ketose interconversion by D-xylose isomerase involving ring opening followed by a 1,2hydride shift. J. Mol. Biol. 212, 211–235. 29. Lambeir, A. M., Lauwereys, M., Stanssens, P., Mrabet, N. T., Snauwaert, J., van Tilbeurgh, H. et al. (1992). Protein engineering of xylose (glucose) isomerase
from Actinoplanes missouriensis. 2. Site-directed mutagenesis of the xylose binding site. Biochemistry, 31, 5459–5466. Allen, K. N., Lavie, A., Farber, G. K., Glasfeld, A., Petsko, G. A. & Ringe, D. (1994). Isotopic exchange plus substrate and inhibition kinetics of D-xylose isomerase do not support a proton-transfer mechanism. Biochemistry, 33, 1481–1487. Allen, K. N., Lavie, A., Glasfeld, A., Tanada, T. N., Gerrity, D. P., Carlson, S. C. et al. (1994). Role of the divalent metal ion in sugar binding, ring opening, and isomerization by D-xylose isomerase: replacement of a catalytic metal by an amino acid. Biochemistry, 33, 1488–1494. Murzin, A. G., Brenner, S. E., Hubbard, T. & Chothia, C. (1995). SCOP: a structural classification of proteins database for the investigation of sequences and structures. J. Mol. Biol. 247, 536–540. Hendrickson, W. A., Horton, J. R. & LeMaster, D. M. (1990). Selenomethionyl proteins produced for analysis by multiwavelength anomalous diffraction (MAD): a vehicle for direct determination of three-dimensional structure. EMBO J. 9, 1665–1672. Otwinowski, Z. & Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode. Macromolecular Crystallography, part A. Methods Enzymol. 276, 307–326. Terwilliger, T. C. & Berendzen, J. (1999). Automated MAD and MIR structure solution. Acta Crystallog. sect. D, 55, 849–861. Terwilliger, T. C. (2000). Maximum likelihood density modification. Acta Crystallog. sect. D, 56, 965–972. Terwilliger, T. C. (2002). Automated main-chain model-building by template-matching and iterative fragment extension. Acta Crystallog. sect. D, 59, 34–44. McRee, D. E. (1993). In (McRee, D. E., ed.), Academic Press Inc., New York. Brunger, A. T. (1993). X-PLOR 3.1: A System for X-ray Crystallography and NMR. Yale University Press, New Haven and London. Engh, R. A. & Huber, R. (1991). Accurate bond and angle parameters for X-ray protein-structure refinement. Acta Crystallogr. sect. A, 47, 392–400.
30.
31.
32.
33.
34.
35. 36. 37. 38. 39. 40.
Edited by I. Wilson (Received 31 August 2006; received in revised form 27 October 2006; accepted 1 November 2006) Available online 6 November 2006
doi:10.1016/j.jmb.2006.11.004
The Structures of L-Rhamnose Isomerase from Pseudomonas stutzeri in Complexes with L-Rhamnose and D-Allose Provide Insights into Broad Substrate Specificity Hiromi Yoshida 1 , Mitsugu Yamada 1 , Yuya Ohyama 2 , Goro Takada 2 Ken Izumori 2 and Shigehiro Kamitori 1 ⁎ 1
Molecular Structure Research Group, Information Technology Center and Faculty of Medicine, Kagawa University, 1750-1 Ikenobe, Miki-cho, Kita-gun, Kagawa 761-0793, Japan 2
Department of Biochemistry and Food Science, Faculty of Agriculture and Rare Sugar Research Center, Kagawa University, Miki-cho, Kagawa 761-0795, Japan
Pseudomonas stutzeri L-rhamnose isomerase (P. stutzeri L-RhI) can efficiently catalyze the isomerization between various aldoses and ketoses, showing a broad substrate specificity compared to L-RhI from Escherichia coli (E. coli L-RhI). To understand the relationship between structure and substrate specificity, the crystal structures of P. stutzeri L-RhI alone and in complexes with L-rhamnose and D-allose which has different configurations of C4 and C5 from L-rhamnose, were determined at a resolution of 2.0 Å, 1.97 Å, and 1.97 Å, respectively. P. stutzeri L-RhI has a large domain with a (β/α)8 barrel fold and an additional small domain composed of seven α-helices, forming a homo tetramer, as found in E. coli L-RhI and D-xylose isomerases (D-XIs) from various microorganisms. The β1–α1 loop (Gly60–Arg76) of P. stutzeri L-RhI is involved in the substrate binding of a neighbouring molecule, as found in D-XIs, while in E. coli L-RhI, the corresponding β1–α1 loop is extended (Asp52–Arg78) and covers the substrate-binding site of the same molecule. The complex structures of P. stutzeri L-RhI with L-rhamnose and D-allose show that both substrates are nicely fitted to the substrate -binding site. The part of the substrate-binding site interacting with the substrate at the 1, 2, and 3 positions is equivalent to E. coli L-RhI, and the other part interacting with the 4, 5, and 6 positions is similar to D-XI. In E. coli L-RhI, the β1–α1 loop creates an unique hydrophobic pocket at the the 4, 5, and 6 positions, leading to the strictly recognition of L-rhamnose as the most suitable substrate, while in P. stutzeri L-RhI, there is no corresponding hydrophobic pocket where Phe66 from a neighbouring molecule merely forms hydrophobic interactions with the substrate, leading to the loose substrate recognition at the 4, 5, and 6 positions. © 2006 Elsevier Ltd. All rights reserved.
*Corresponding author
Keywords: X-ray structure; rhamnose isomerase; rare sugars; Pseudomonas stutzeri
Abbreviations used: P. stutzeri L-RhI, Pseudomonas stutzeri L-rhamnose isomerase; E. coli L-RhI, Escherichia coli L-rhamnose isomerase; S. rubiginosus D-XI, Streptomyces rubiginosus D-xylose isomerase; A. missouriensis D-XI, Actinoplanes missouriensis D-xylose isomerase; T. thermophilus D-XI, Thermus thermophilus D-xylose isomerase; P. stutzeri L-RhI/L-rhamnose, P. stutzeri L-RhI/ L-rhamnose complex; P. stutzeri L-RhI/D-allose, P. stutzeri L-RhI/D-allose complex; P. stutzeri SeMet L-RhI, selenomethionine substituted P. stutzeri L-RhI. E-mail address of the corresponding author: [email protected]
Introduction L-Rhamnose isomerase, which catalyzes the reversible isomerization of L-rhamnose to L-rhamnulose, has been found to be involved in the metabolism of 1,2 L-rhamnose in Escherichia coli and also to be present in some other microorganisms like Lactobacillus, 3 Salmonella, 4 and Pseudomonas. 5 L-Rhamnose isomerase from Pseudomonas stutzeri (P. stutzeri L-RhI; Swiss Prot accession number Q75WH8, 430 amino acid residues, 46,975 Da)6 shows a broader substrate specificity than L-RhI
0022-2836/$ - see front matter © 2006 Elsevier Ltd. All rights reserved.
1506 from Escherichia coli (E. coli L-RhI; Swiss Prot accession number P32170, 419 amino acid residues, 47,199 Da)2, catalyzing the isomerization between various aldoses and ketoses, as well as between L-rhamnose and L-rhamnulose, in the presence of appropriate metal ions.5–7 Leang et al. reported that a recombinant His-tagged P. stutzeri L-RhI efficiently catalyzed the isomerization between L-rhamnose and L-rhamnulose, L-mannose and L-fructose, L-lyxose and L-xylulose, D-ribose and D-ribulose, and D-allose and D-psicose, as listed in Figure 1(a) with the enzyme activities. Since some of them are so-called “rare sugars” which exist in small amounts in nature, P. stutzeri L-RhI is exploited for industrial applications in rare sugar production.8–10 However, little is known about the mechanism responsible for the broader substrate specificity of the enzyme. Korndörfer et al. reported the X-ray structure of E. coli L-RhI (PDB codes 1D8W, 1DE5 and 1DE6)11 and a structural comparison with the D-xylose isomerase from Streptomyces rubiginosus (S. rubiginosus D-XI; Swiss Prot accession number P24300,
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
387 amino acid residues, 43,096 Da, PDB code 4XIS). 12,13 Despite the poor sequence identity (13%) between them, they have some similarities in structure. Both have a large domain with a (β/α)8 barrel and an additional small domain composed of α-helices, to form a homo tetramer, and each subunit has two adjacent metal ions at the substrate-binding site. It was also proposed that the aldose-ketose isomerization reactions of E. coli L-RhI are catalyzed by the metal-mediated hydride shift mechanism (Figure 1(b)), which is available to D-XIs.11 To elucidate the substrate-recognition mechanism responsible for the broad substrate specificity of P. stutzeri L-RhI, understanding the three-dimensional structure of P. stutzeri L-RhI in complexes with its substrates is important. Sequence alignment between P. stutzeri L-RhI and E. coli L-RhI indicates the amino acid residues involved in the metal binding are strictly conserved, but the sequence identity between them is only 17%,6,7 suggesting that P. stutzeri L-RhI has an unique three-dimensional structure at the substrate-binding site. The crystallization and preliminary X-ray analysis of P. stutzeri L-RhI have been reported.14 Here we report the crystal structure of P. stutzeri L-RhI alone and in complexes with L-rhamnose (P. stutzeri L-RhI/Lrhamnose) and D-allose (P. stutzeri L-RhI/D-allose).
Results and Discussion Quality of structures
Figure 1. (a) Chemical structures of L-rhamnose, L-mannose, L-lyxose, D-ribose, D-allose, D-xylose, D-glucose, and D-sorbitol are shown in Fischer projection formulas. The enzyme activities of P. stutzeri L-RhI for them, Kcat/Km (μM−1 min−1),7 are given in parentheses. (b) Schematic diagram of the proposed metal-mediated hydride shift reaction of L-RhI.
The structures have been refined to an R-factor of 0.155 (P. stutzeri L-RhI), 0.170 (P. stutzeri L-RhI/ L-rhamnose) and 0.172 (P. stutzeri L-RhI/D-allose), respectively, with good chemical geometries, as indicated in Table 1. In Ramachandran plots15 90.6% of all residues in P. stutzeri L-RhI, 91.1% in P. stutzeri L-RhI/L-rhamnose, and 90.7% in P. stutzeri LRhI/D-allose are shown to be in the most favoured regions, with no residue in the disallowed regions as determined by the program PROCHECK.16 In all X-ray structures, there are four protein molecules (Mol-A, B, C and D), forming a homotetramer, in an asymmetric unit. The final (2Fo–Fc) electron density map of all X-ray structures shows that almost all atoms of protein molecules, solvent molecules, and metal ions fitted nicely. Poor or invisible electron density is found in several residues at the N terminus (Met1–Glu3), the loop region residues (Glu313–Phe320), and the C terminus (Ala421–Ile430) with 6-His, suggesting that these regions are mobile and highly disordered. Two bound metal ions (M-1 and M-2) at the substratebinding site were found in each protein molecule. The electron density of M-1 is very strong, while that of M-2 is relatively weak, suggesting the lower occupancy of M-2. In E. coli L-RhI, M-1 and M-2 are considered as to be Zn2+ and Mn2+ , respectively; however, identifications of metal ions, M-1 and M-2, of P. stutzeri L-RhI are difficult from current X-ray
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
1507
Table 1. X-ray data collection and refinement statistics Datase Beam line Wavelength (Å) Temperature (K) Resolution (Å) No. of measured refs No. of unique refs Completeness (%) Rmerge Io/σ(Io) Space group Cell dimensions a(Å) b(Å) c(Å) β(°) Resolution (Å) Completeness (%) Rfactor Rfree r.m.s.d. bond lengths (Å) r.m.s.d. bond angles (deg.) No. of amino acids
P. stutzeri SeMet L-RhI
P. stutzeri Native L-RhI
P. stutzeri L-RhI with L-rhamnose
P. stutzeri L-RhI with D-allose
KEK PF BL-6A 0.97870 100 50–2.10 711,721 97,228 100.0 (100.0) 0.062 (0.123) 7.0 (4.7) P21
KEK PF BL-6A 1.0000 100 50–2.00 647,726 99,909 94.9 (95.0) 0.074 (0.298) 13.3 (7.5) P21
KEK PF BL-5A 1.0000 100 50–1.97 876,965 117,301 99.5 (95.1) 0.104 (0.242) 9.5 (5.5) P21
KEK PF BL-5A 1.0000 100 50–1.97 876,565 117,529 99.9 (98.7) 0.115 (0.234) 10.2 (6.6) P21
74.63 104.79 113.65 107.86
74.25 103.97 107.02 106.81 48.79–2.00 92.3 0.155 0.192 0.005 1.1 Mol-A 421 Mol-B 420 Mol-C 421 Mol-D 417 1783 Zn 8
74.62 104.76 114.29 107.96 33.25–1.97 97.6 0.170 0.205 0.005 1.1 Mol-A 421 Mol-B 421 Mol-C 428 Mol-D 419 1364 Zn 8 L-Rhamnose 4
74.65 104.73 114.53 108.00 33.25–1.97 97.7 0.172 0.204 0.005 1.1 Mol-A 419 Mol-B 421 Mol-C 426 Mol-D 419 1236 Zn 8 D-Allose 4
19.1 29.9 50.6 19.7 30.4
22.1 34.6 31.9 42.0 23.1 28.4
20.4 34.9 30.3 40.7 21.7 25.7
90.6/ 8.9/ 0.5
91.1/ 8.4/ 0.5
90.6/ 8.9/ 0.5
No. of solvent molecules No. of ligand B-factor (Å2) Protein Ligand M-1 M-2 β1−α1 region solvent Ramachandran plot (%) Favoured/additional allowed/generously allowed
Rmerge = ∑∑ | Ii- | / ∑. Values in parentheses are of the high resolution bin (2.18–2.10 Å for SeMet, 2.07–2.00 Å for native, 2.04–1.97 Å for L-rhamnose and 2.04–1.97 Å for D-allose).
diffraction data. Therefore, the bound metal ions of P. stutzeri L-RhI were refined as Zn2+ with occupancies of 1.0 for M-1 and 0.5 for M-2, since the atomic absorption spectrum of purified P. stutzeri L-RhI showed the presence of metal ions, Zn2+, Mn2+, and Ni2+, at a ratio of 4:1:1 (data not shown). The bound substrates were found in each protein molecule in the enzyme–substrate complexes. Simulated annealing omit maps for the bound metal ions and substrates are shown in Figure 2.17,18 Since the four molecules are almost identical in structure with r.m.s. deviations for main-chain atoms being 0.69 Å– 0.85 Å in P. stutzeri L-RhI, 0.54 Å–0.81 Å in P. stutzeri L-RhI/L-rhamnose, and 0.53 Å–0.78 Å in P. stutzeri L-RhI/D-allose, the structural description here concentrates on Mol-A (if not stated otherwise). Overall structure of P. stutzeri L-RhI The overall subunit structure of P. stutzeri L-RhI is shown in Figure 3. P. stutzeri L-RhI has 15 α-helices with 54% of the amino acid residues and eight
β-strands with 8% of the amino acid residues. The structure can be divided into a large domain (Phe50–Val356) and an additional small domain (N terminus-Lys49, and Asp357-C terminus). The large domain forms the core of the enzyme with the (β/α)8 barrel fold. The bound metal ions are located at the centre of the barrel, and several His and Asp residues in β-strands direct their side-chain groups toward the inside of the barrel to coordinate with metal ions. The large domain of P. stutzeri L-RhI is similar in structure to that of E. coli L-RhI and S. rubiginosus D-XI, in spite of low sequence identities between them. The r.m.s. deviations for Cα atoms and Z-scores of DALI search19 for the large domain of P. stutzeri L-RhI are 2.0 Å and 28.5 with E. coli L-RhI (1DE5, 19% identity), and 1.9 Å and 33.3 with S. rubiginosus D-XI (4XIS, 19% identity). The additional small domain is composed of two α-helices (α0 and α9′) in the N-terminal region and five α-helices (α9–α13) in the C-terminal region. The α-helices are packed pairwise against each other to form a bundle of α-helices. The α0 is largely kinked to form
1508
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
Figure 2. Simulated annealing omit map at the 4.0σ contours level for the bound metal ion and (a) L-rhamnose, and (b) D-allose illustrated by the programs X-fit17 and Raster3D.18 Metal ions and the C1 and C6 of substrates are labeled.
hydrophobic interactions with α11 and α12. The additional small domain composed of α-helices has been also observed in E. coli L-RhI and D-XIs; however, its structure varies depending on the enzyme, leading to an unique inter-subunit surface in each enzyme, as described next. The homo-tetramer of P. stutzeri L-RhI has dimensions of 100 Å × 90 Å × 80 Å with 222 symmetry, as shown in Figure 4. The four subunits are
arranged so that the loop regions between the βstrands and α-helices of the large domain are located at the centre of the tetramer, and the outer surface of the tetramer is mostly covered by α-helices. Inter-subunit contact areas calculated using the program AREAIMOL in the CCP4 program suite20 are 1943 Å2 between Mol-A and Mol-B, 2303 Å2 between Mol-A and Mol-C, and 1625 Å2 between Mol-A and Mol-D, respectively (Table 2). A similar
Figure 3. Overall subunit structure of P. stutzeri L-RhI illustrated by the program PyMol. The structure of P. stutzeri L-RhI is comprised of a large domain with the (β/α)8 barrel fold shown in gray (α-helices) and yellow (β-strands), and an additional small domain shown in dark green. In the large domain, α-helices are numbered from α1 to α8, and β-strands are numbered from β1 to β8. Two α-helices in the N-terminal region of the additional small domain are numbered as α0 and α0′, and five α-helices of the Cterminal region are numbered α9, α10, α11, α12 and α13. The small blue spheres indicate metal ions [http://www.pymol.org].
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
Figure 4. Overall tetrameric structure of P. stutzeri L-RhI illustrated by the program PyMol in (a) top view and (b) side view. The four molecules are coloured in light green (Mol-A), light blue (Mol-B), light magenta (Mol-C) and light yellow (Mol-D). The dark coloured part of each molecule represents the additional small domain. The small blue spheres indicate metal ions [http://www. pymol.org].
homo-tetramer is also found in E. coli L-RhI11 and other D-XIs from S. rubiginosus 13 , Actinoplanes missouriensis (A. missouriensis D-XI; Swiss Prot accession number P12851, 393 amino acid residues, 43,368 Da, PDB code 2XIN) 21,22 and Thermus thermophilus (T. thermophilus D-XI; Swiss Prot accession number P26997, 387 amino acid residues, 43,907 Da, PDB code 1BXB).23,24
1509 The structure of Mol-A and Mol-B in the tetramer is illustrated in Figure 5(a). Although the intersubunit contact area is 1943 Å2, the dimer structure of Mol-A and Mol-B is thought to be relatively unstable, because all of the inter-subunit interactions are formed by the loop regions of two molecules, suggesting that the formation of a tetramer is indispensable to maintaining the presented dimer structure of Mol-A and Mol-B. The structure of the inter-subunit between Mol-A and Mol-B is very important for the enzyme, because the loop between β1 and α1 (the β1–α1 loop) approaches the substrate-binding site for a neighbouring molecule to interact with a substrate as shown by red in Figure 5(a), and because Mol-A and Mol-B are expected to form the accessible surface for substrate binding; the bound metal ions can be visible from this molecular surface. In A. missouriensis D-XI, the β1–α1 loop partly makes up the substrate-binding site for a neighbouring molecule and an extensively opened accessible surface is formed between Mol-A and Mol-B, as found in P. stutzeri L-RhI (Figure 5(b)). By contrast, in E. coli L-RhI, the β1–α1 loop covers the bound metal ions in the same molecule, and there is a clear boundary between Mol-A and Mol-B, giving a relatively small contact area of 977 Å2, as shown in Figure 5(c). The inter-subunit interactions between Mol-A and Mol-B are closely related to the difference in substrate specificity between P. stutzeri L-RhI and E. coli L-RhI, as described in detail below. Between Mol-A and Mol-C, α4, α5, α6 and α7 are involved in the inter-subunit interaction, forming the wide inter-subunit contact area. In A. missouriensis D-XI, along with α4, α5, α6 and α7, the additional small domain is also involved in the intersubunit contacts, giving the extensively wide contact area of 4857 Å2, suggesting that the additional small domain contributes greatly to the stability of the subunit association in A. missouriensis D-XI. Between Mol-A and Mol-D, the additional small domains are mainly involved at the inter-subunit surface, with a contact area of 1625 Å2, which does not vary much in E. coli L-RhI (1928 Å2) and A. missouriensis D-XI (1640 Å2), in spite of the difference in structure of the additional small domains. Structure of substrate-binding site P. stutzeri RhI requires metal ions for its catalytic activity. Two putative metal ions were observed in the P. stutzeri RhI and P. stutzeri RhI/substrate complexes. Since no metal ions were added in the
Table 2. Inter-subunit contact areas mol A–B, mol A–C, and A–D of P. stutzeri L-RhI, E.coli RhI, and A. missouriensis D-XI Contact areas [Å2] P. stutzeri L-RhI E. coli L-RhI A. missouriensis D-XI
Mol A–B
Mol A–C
Mol A–D
1943 977 1698
2303 2102 4857
1625 1928 1640
1510
Figure 5. Subunit interfaces between Mol-A and MolB of (a) P. stutzeri RhI/D-allose, (b) A. missouriensis D-XI/Dsorbitol (PDB code 2XIN), and (c) E. coli RhI/L-rhamnose (PDB code 1DE6) illustrated by the program PyMol. The β1–α1 loops are indicated in red. Zn2+ in P. stutzeri RhI/Dallose are shown as small blue spheres. Co2+ in A. missouriensis D-XI/D-sorbitol are shown as orange spheres. Zn2+ and Mn2+ in E. coli RhI/L-rhamnose are shown as small blue spheres and large purple spheres, respectively. The bound D-allose, D-sorbitol and L-rhamnose are represented by pink, red and green stick models, respectively, and the amino acid residues in the β1–α1 loop interacting with a substrate are represented by a blue stick model [http://www.pymol.org].
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
cultivation and purification process, it is possible that metal ions from the culture medium were retained in the enzyme. With respect to D-XIs and E. coli RhI, which also require a divalent metal ion for their activity, one or two metal ions were observed. One is “structural” (M-1) to help the substrate binding, and the other is “catalytic” (M-2) to help the hydride shift. The structural metal ion is reported to bind tightly with a lower dissociation constant than the catalytic metal ion.11 Also in P. stutzeri RhI, a low occupancy of catalytic metal ions is found. As shown in Figure 6(a) the metal ions in P. stutzeri LRhI are coordinated in a distorted octahedral form with six coordination bonds; they are Glu219, Asp254, His281, Asp327, and two water molecules (W-1 and W-2) for M-1, and His257, Asp289 and four water molecules (W-2, W-3, W-4 and W-5) for M-2. The distance between M-1 and M-2 is 4.1 Å, and the coordination water molecule of W-2 bridges the metal ions. The structure of the substrate-binding site of P. stutzeri L-RhI with the bound L-rhamnose is shown in Figure 6(b). The O1, O2, and O3 atoms of L-rhamnose coordinate with M-1 and M-2, instead of W-3, W-2, and W-1 in the unliganded form, respectively. O1 and O3 make a hydrogen bond with Lys221 and Glu219, respectively, helping to fix the substrate in the proper conformation for the hydride shift between the C1 and C2 of substrate. There are still two water molecules (W-4 and W-5) coordinating to M-2 with the hydrogen bonds of Asp291 (W-4), and of Asp254 and Asp291 (W-5). It is likely that W-4 is a key water molecule responsible for the hydride shift in the proposed catalytic mechanism, because it possibly makes hydrogen bonds with the O1 and O2 of L-rhamnose. O4 and O5 make a hydrogen bond with Asp327 and His101, respectively, recognizing the hydroxyl groups at the 4 and 5 positions of L-rhamnose, and C6 undergoes hydrophobic interactions with Trp57, Phe131, and *Phe66 (Phe66 from a neighbouring molecule). On the basis of structure of P. stutzeri L-RhI/L-rhamnose, the enzyme–substrate interactions of L-mannose and L-lyxose can be deduced. Since L-mannose and L-lyxose have the same configurations of asymmetric carbon atoms as L -rhamnose, substrate–enzyme interactions are thought to be almost identical between them, but enzyme activities for them are significantly different, as listed in Figure 1(a). The structural variation between these substrates is found at 6 position, where L-mannose has an additional hydroxyl group and L-lyxose has no group. The simple modelling to generate the bound L-mannose by adding a hydroxyl group to L-rhamnose at the 6 position indicated that the O6 atom can be located in a staggered conformation without any unusual short contacts with amino acid residues; the torsion angle of C4-C5-C6-O6 is −60 °. In the resultant model of the bound L-mannose, O6 is in hydrophobic environment created by *Phe66, Trp104, and Phe131, and it may decrease the affinity of substrate to enzyme, as described in detail below. In the binding of L-lyxose, which has no group at the 6 position, a lack of hydrophobic interactions
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
of C6 with the enzyme found in P. stutzeri L-RhI/ L-rhamnose may lead the lower enzyme activity for L-lyxose. The structure of the substrate-binding site of P. stutzeri L-RhI with the bound D-allose is shown in Figure 6(c). The interactions between O1, O2, and O3 of D-allose and the enzyme including metal ions are almost identical to those found in P. stutzeri L-RhI/L-rhamnose, because D-allose has the same configurations of C1, C2, and C3 as L-rhamnose. O4 makes a hydrogen bond with His101, and O5 with Asp327, different from the bound L-rhamnose, where O4 bonds with Asp327 and O5 with His101. This means that P. stutzeri L-RhI can recognize substrates with different configurations of C4 and C5, by using His101 and Asp327, and vice versa. However, the substrate recognition at the 4 and 5 positions of L-rhamnose is thought to be more favourable for P. stutzeri L-RhI than that of D-allose, because the enzyme activity for D-allose is significantly lower than that for L-rhamnose. C6 makes hydrophobic interactions with Trp57, and also O6 is in a hydrophobic environment created by *Phe66, Trp104, and Phe131 without any hydrogen bond with amino acid residues. The hydroxyl group, O6, in hydrophobic environment increases structural energy of the substrate–enzyme complex, leading low affinity of substrate to the enzyme. This is consistent with the fact that enzyme activity for Dribose, which has no group at the 6 position is higher than that for D-allose.7 Comparison with E. coli RhI and A. missouriensis D-XI A superimposed substrate-binding structure of E. coli RhI with the bound L-rhamnose and P. stutzeri L-RhI with the bound L-rhamnose and D-allose is illustrated in Figure 7(a). The amino acid residues involved in metal-binding are conserved, and also the interactions between O1, O2 and O3 of L-rhamnose and the enzyme found in P. stutzeri L-RhI are very similar to those in E. coli RhI, suggesting that the catalytic mechanism of the metal-mediated hydride shift proposed in E. coli RhI (Figure 1(b)) can be assumed for P. stutzeri L-RhI. However, remarkable structural differences are found in the recognition of the atoms at the 4, 5 and 6 positions of a substrate. The β1–α1 loop (Gly60–Arg76) of P. stutzeri L-RhI is involved in the substrate binding of a neighbouring molecule, while in E. coli L-RhI, the corresponding β1–α1 loop is extended (Asp52– Arg78) and covers the substrate-binding site of the same molecule, as mentioned above (Figure 5). In E. coli L-RhI, Val53, Leu63, and Ile67 from the β1–α1 loop create a unique hydrophobic pocket to make many van der Waals contacts with the substrate. There is no corresponding hydrophobic pocket in P. stutzeri L-RhI, where *Phe66 from a neighbouring molecule merely forms hydrophobic interactions with the substrate. In addition, Phe336 of E. coli L-RhI partly attends to the formation of this hydrophobic pocket by changing its side-chain
1511 conformation on the binding of a substrate to make favourable interactions with the substrate.11 In P. stutzeri L-RhI, the corresponding residue is Ser329, 8.8 Å from C6 of L-rhamnose, and there is a large space between Ser329 and the bound substrate. The unique hydrophobic pocket around the methyl group of C6 formed by residues Val53, Leu63, Ile67 from the β1–α1 loop and Phe336 in E. coli L-RhI strictly recognizes L-rhamnose as the most suitable substrate, leading to the high substrate specificity of E. coli L-RhI. In fact, unusual short contacts between Ile67 and Phe336 of E. coli L-RhI and the bound D-allose in P. stutzeri L-RhI are observed (Figure 7(a)). On the other hand, the substratebinding site of P. stutzeri L-RhI without the hydrophobic pocket loosely recognizes both L-rhamnose and D-allose as substrates. A superimposed substrate binding structure of A. missouriensis D-XI with the bound inhibitor Dsorbitol (Figure 1(a)) and P. stutzeri L-RhI with the bound L-rhamnose is illustrated in Figure 7(b). Since D-glucose, as well as D-xylose, is also a suitable substrate of D-XI, D-sorbitol having exactly the same configuration of asymmetric carbon atoms as D-glucose acts as an inhibitor of D-XI. The amino acid residues involved in the metal binding are not conserved between them (His281/Asp245 and Asp254/Glu217), and O1, O2, and O4 of D-sorbitol coordinate with metal ions in A. missouriensis D-XI, instead of O1, O2, and O3 of L-rhamnose in P. stutzeri L-RhI. Since the substrates of D-XI, D-xylose and D-glucose (as an inhibitor, D-sorbitol), have a different configuration at C3 from the substrate of L-RhI, L-rhamnose, the variations in metal-binding site structure and substrate–metal interactions are expected to be closely related to the difference in substrate specificity between D-XI and L-RhI. Interestingly, the amino acid residues involved in the interactions with the substrate at the 4, 5 and 6 positions are conserved between P. stutzeri L-RhI and A. missouriensis D-XI (Trp57/16, His101/54, Phe131/94, Trp179/137, Asp327/292 and *Phe66/26), except Ser329/Lys294, suggesting that A. missouriensis D-XI loosely recognizes a substrate at these positions, as well as P. stutzeri L-RhI. It is notable that Streptomyces albus D-XI (Swiss Prot accession number P24299, 390 amino acid residues, 43,289 Da, PDB code 6XIA)25 having 64% homology with A. missouriensis D-XI was reported to have for D-allose and for L-rhamnose, 38% and 13%, respectively, of the activity it has for D-xylose,26 and also that glucose isomerase/xylose isomerase was mentioned to be able to isomerize a wide variety of substrates including L-rhamnose and D-allose in a review article.27 In D-XI, His has been considered as the residue involving in ring opening by assistance of an acidic amino acid residue, Asp,28 and they are His54 and Asp57 in A. missouriensis D-XI, as shown in Figure 7(b). In E. coli L-RhI, the corresponding residues are His103 and Tyr106, and in P. stutzeri L-RhI, the corresponding residues are His101 and Trp104. Therefore, E. coli and P. stutzeri L-RhIs can not
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
1512 adopt the above ring opening mechanism. Lambeir et al. and Allen et al. proposed the different ring opening mechanism that Lys182 or a metal-liganded water molecule involves in ring opening in D-XI,29–31
and Korndörfer et al. supposed that the similar ring opening mechanism could be expected for E. coli L-RhI.11 For P. stutzeri L-RhI, the similar ring opening mechanism seems to be possible; however,
Figure 6 (legend on next page)
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
1513
Figure 6. Stereo view of (a) the metal-binding site of P. stutzeri L-RhI, and the substrate-binding sites for (b) and (c) D-allose illustrated by the program PyMol. The metal ions are shown as blue spheres and the water molecules as red spheres. The bound L-rhamnose and D-allose are shown in blue and orange, respectively. The selected interactions among amino acid residues, metal ions, and water molecules are indicated by dotted lines. *Phe66 from a neighbouring molecule is shown as thin stick bonds. Carbon atoms of substrates are numbered from 1 to 6 [http://www. pymol.org]. L-rhamnose
there is no clear evidence to elucidate the ring opening mechanism in P. stutzeri L-RhI, so far.
Conclusion The presented X-ray structures of P. stutzeri L-RhI show that the part of the substrate-binding site interacting with the substrate at the 1, 2, and 3 positions is equivalent to E. coli L-RhI, and the other part interacting with the 4, 5, and 6 positions is similar to D-XI, leading to the loose substrate recognition at the 4, 5, and 6 positions. Because O1, O2 and O3 of a substrate coordinate with metal ions, configurations of C2, C3 of a substrate are very important for the enzyme activity. This is consistent with the fact that suitable substrates for P. stutzeri L-RhI, L-rhamnose, L-mannose, L-lyxose, D-ribose and D-allose, have the same configurations of C2 and C3. The configurations of C4 and C5 of a substrate do not affect the enzyme activity as much. P. stutzeri L-RhI with D-XI-like substrate-binding site at the 4 and 5 positions can recognize substrates with different configurations of C4 and C5, by using His101 and Asp327, and vice versa. This is why P. stutzeri L-RhI can recognize D-ribose and D-allose as a substrate. In SCOP (structural classification of proteins),32 E. coli L-RhI and D-XI are classified into different
families, the L-rhamnose isomerase family and the xylose isomerase family, respectively, because of the variation in the metal-binding site structure. Based on the metal-binding site structure, P. stutzeri L-RhI should also be placed in the L-rhamnose isomerase family; however, the substrate-binding site structure around the 4, 5 and 6 positions and the position of the β1–α1 loop of P. stutzeri L-RhI is similar to those of DXI, meaning that P. stutzeri L-RhI has structural features of both L-RhI and D-XI. This may strongly support the assumption by Korndörfer et al. that L-RhI and D-XI are derived from common precursor.11
Materials and Methods Protein preparation and crystallization The purification and crystallization of P. stutzeri L-RhI have been reported.14 The seleno-methionine-substituted P. stutzeri L-RhI (P. stutzeri SeMet L-RhI) was prepared in the same manner as that of P. stutzeri L-RhI except that the E.coli B834 cells were grown in LeMaster broth.33 A crystal of P. stutzeri L-RhI was grown by the vapour diffusion method using a protein solution (20 mg/ml P. stutzeri L-RhI in 20 mM Tris–HCl (pH 8.0)) and a reservoir solution (7–8% (w/v) polyethylene glycol 20,000, 50 mM Mes buffer (pH 6.3)). A crystal of P. stutzeri SeMet L-RhI was obtained under the same conditions. Crystals of P. stutzeri L-RhI/L-rhamnose and L-RhI/D-allose were obtained by a
1514
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
Figure 7. Stereo view of the substrate-binding sites of (a) E. coli L-RhI (cyan, 1DE6) with rhamnose (blue) and (b) A. missouriensis D-XI (green, 2XIN) with D-sorbitol (magenta) illustrated by the program PyMol. The substrate-binding site of P. stutzeri L-RhI (yellow) with L-rhamnose (thin blue bonds) and D-allose (thin orange bonds) are superimposed. Zn2+ in P. stutzeri L-RhI is shown as blue small spheres, Zn2+ and Mn2+ in E. coli L-RhI are shown as a small gray sphere and a large purple sphere, respectively. The Co2+ in A. missouriensis D-XI is shown as orange spheres. Blue and green labels indicate the residues belonging to E. coli L-RhI and A. missouriensis D-XI, respectively [http://www.pymol.org].
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
1515
soaking method, by incubation for 24 to 31 h with an additional 0.5 μl of 100 mM substrate solution. Data collection and structure determination Crystals were flash-cooled in liquid nitrogen at 100 K and X-ray diffraction data were collected using an ADSC Quantum 315 detector system on the BL-5A beam line and an ADSC Quantum 4R CCD detector system on the BL-6A beam line in the Photon Factory (Tsukuba, Japan). Diffraction data were processed using the programs HKL200034 and the CCP4 program suite.20 The collected data and scaling results are listed in Table 1. Initial SAD phasing at 2.1 Å of P. stutzeri SeMet L-RhI was performed by locating 24 selenium sites in the peak dataset using the program SOLVE.35 After electron density modification, 90% of amino acid residues could be located in the resultant electron density map, using the program RESOLVE.36,37 Further model building was performed with the program X-fit17 in the XtalView program system,38 and the structure was refined using the program CNS39 with Engh & Huber stereochemical parameters.40 Using the structure of P. stutzeri SeMet L-RhI, the structure of P. stutzeri L-RhI and that of P. stutzeri L-RhI/substrates were determined by molecular replacement method and refined at a resolution of 2.0 Å for P. stutzeri L-RhI, 1.97 Å for P. stutzeri L-RhI/L-rhamnose and 1.97 Å for P. stutzeri L-RhI/D-allose, using the program CNS. Water molecules were gradually introduced if the peaks above 3.5σ in the (Fo–Fc) electron density map were in the range of a hydrogen bond.
4.
5.
6.
7.
8.
9.
10.
Protein Data Bank accession number The atomic coordinates and structure factors of P. stutzeri L-RhI (PDB code 2HCV), P. stutzeri L-RhI/L-rhamnose (PDB code 2I56) and P. stutzeri L-RhI/D-allose (PDB code 2I57) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ.
11.
12.
Acknowledgements
13.
This study was supported in part by the National Project on Protein Structural and Functional Analyses and the “intellectual-cluster” from the Ministry of Education, Culture, Sports, Science and Technology of Japan. This research was performed with the approval of the Photon Factory Advisory Committee and the National Laboratory for High Energy Physics, and SPring-8, Japan.
14.
References
16.
1. Wilson, D. M. & Ajl, S. (1957). Metabolism of L-rhamnose by Escherichia coli. I. L-rhamnose isomerase. J. Bacteriol. 73, 410–414. 2. Moralejo, P., Egan, S. M., Hidalgo, E. & Aguilar, J. (1993). Sequencing and characterization of a gene cluster encoding the enzymes for L-rhamnose metabolism in Escherichia coli. J. Bacteriol. 175, 5585–5594. 3. Domagk, G. F. & Zech, R. (1963). Über den Abbau der
15.
17. 18.
Desoxyzucker durch Bakterienenzyme. I. L-RhamnoseIsomerase aus Lactobacillus plantarum. Biochem. Z. 339, 145–153. Englesberg, E. & Baron, L. S. (1959). Mutation to L-rhamnose resistance and transduction to L-rhamnose utilization in Salmonella typhosa. J. Bacteriol. 78, 675–686. Bhuiyan, S. H., Itami, Y. & Izumori, K. (1997). Isolation of an L-rhamnose isomerase-constitutive mutant of Pseudomonas sp. strain LL172: purification and characterization of the enzyme. J. Ferment. Bioeng. 84, 319–323. Leang, K., Takada, G., Ishimura, A., Okita, M. & Izumori, K. (2004). Cloning, nucleotide sequence, and overexpression of the L-rhamnose isomerase gene from Pseudomonas stutzeri in Escherichia coli. Appl. Environ. Microbiol. 70, 3298–3304. Leang, K., Takada, G., Fukai, Y., Morimoto, Y., Granstrom, T. B. & Izumori, K. (2004). Novel reactions of L-rhamnose isomerase from Pseudomonas stutzeri and its relation with D-xylose isomerase via substrate specificity. Biochim. Biophys. Acta, 1674, 68–77. Bhuiyan, S. H., Itami, Y. & Izumori, K. (1997). Immobilization of L-rhamnose isomerase and its application in L-mannose production from L-fructose. J. Ferment. Bioeng. 84, 558–562. Bhuiyan, S. H., Itami, Y., Rokui, Y., Katayama, T. & Izumori, K. (1998). D-Allose production from Dpsicose using immobilized L-rhamnose isomerase. J. Ferment. Bioeng. 85, 539–541. Bhuiyan, S. H., Itami, Y., Takada, G. & Izumori, K. (1999). Preparation of L-Talose and D-Gulose from L-Tagatose and D-Sorbose, respectively, using immobilized L-Rhamnose Isomerase. J. Ferment. Bioeng. 88, 567–570. Korndörfer, I. P., Fessner, W. D. & Matthews, B. W. (2000). The structure of rhamnose isomerase from Escherichia coli and its relation with xylose isomerase illustrates a change between inter and intra-subunit complementation during evolution. J. Mol. Biol. 300, 917–933. Wong, H. C., Ting, Y., Lin, H. C., Reichert, F., Myambo, K., Watt, K. W. et al. (1991). Genetic organization and regulation of the xylose degradation genes in Streptomyces rubiginosus. J. Bacteriol. 173, 6849–6858. Whitlow, M., Howard, A. J., Finzel, B. C., Poulos, T. L., Winborne, E. & Gilliland, G. L. (1991). A metalmediated hydride shift mechanism for xylose isomerase based on the 1.6 Å Streptomyces rubiginosus structures with xylitol and d-xylose. Proteins: Struct. Funct. Genet. 9, 153–173. Yoshida, H., Waynoon, P., Takada, G., Izumori, K. & Kamitori, S. (2006). Crystallization and preliminary X-ray diffraction studies of l-rhamnose isomerase from Pseudomonas stutzeri. Acta Crystallog. sect. F, 62, 550–552. Ramachandran, G. N. & Sasisekharan, V. (1968). Conformation of polypeptides and protein. Advan. Protein Chem. 23, 283–437. Laskowski, R. A., MacArthur, M. W., Moss, D. S. & Thornton, J. M. (1992). PROCHECK v.2: Programs to Check the Stereochemical Quality of Protein Structures. Oxford Molecular Ltd, Oxford, England. McRee, D. E. (1999). XtalView/Xfit: a versatile program for manipulating atomic coordinate and electron density. J. Struct. Biol. 125, 156–165. Merritt, E. A. & Bacon, D. J. (1997). Raster3D: photorealistic molecular graphics. Methods Enzymol. 277, 505–524.
1516
X-ray Structure of P. stutzeri L-Rhamnose Isomerase
19. Holm, L. & Sander, C. (1993). Protein structure comparison by alignment of distance matrices. J. Mol. Biol. 233, 123–138. 20. Collaborative Computational Project 4. (1994). The CCP4 Suite: programs for protein crystallography. Acta Crystallog. sect. D, 50, 760–763. 21. Amore, R. & Hollenberg, C. P. (1989). Xylose isomerase from Actinoplanes missouriensis: primary structure of the gene and the protein. Nucl. Acids Res. 17, 7515. 22. Jenkins, J., Janin, J., Rey, F., Chiadmi, M., van Tilbeurgh, H., Lasters, I. et al. (1992). Protein engineering of xylose (glucose) isomerase from Actinoplanes missouriensis. 1. Crystallography and site-directed mutagenesis of metal binding sites. Biochemistry, 31, 5449–5458. 23. Dekker, K., Yamagata, H., Sakaguchi, K. & Udaka, S. (1991). Xylose (glucose) isomerase gene from the thermophile Thermus thermophilus: cloning, sequencing, and comparison with other thermostable xylose isomerases. J. Bacteriol. 173, 3078–3083. 24. Chang, C., Park, B. C., Lee, D.-S. & Suh, S. W. (1999). Crystal structures of thermostable xylose isomerases from Thermus caldophilus and Thermus thermophilus: possible structural determinants of thermostability. J. Mol. Biol. 288, 623–634. 25. Dauter, Z., Terry, H., Witzel, H. & Wilson, K. S. (1990). Refinement of glucose isomerase from Streptomyces albus at 1.65 Å with data from an imaging plate. Acta Crystallog. sect. B, 46, 833–841. 26. Sanchez, S. & Smiley, K. L. (1975). Properties of D-xylose isomerase from Streptomyces albus. Appl. Microbiol. 29, 745–750. 27. Bhosale, S. H., Rao, M. B. & Deshpande, V. V. (1996). Molecular and industrial aspects of glucose isomerase. Microbiol. Rev. 60, 280–300. 28. Collyer, C. A., Henrick, K. & Blow, D. M. (1990). Mechanism for aldose-ketose interconversion by D-xylose isomerase involving ring opening followed by a 1,2hydride shift. J. Mol. Biol. 212, 211–235. 29. Lambeir, A. M., Lauwereys, M., Stanssens, P., Mrabet, N. T., Snauwaert, J., van Tilbeurgh, H. et al. (1992). Protein engineering of xylose (glucose) isomerase
from Actinoplanes missouriensis. 2. Site-directed mutagenesis of the xylose binding site. Biochemistry, 31, 5459–5466. Allen, K. N., Lavie, A., Farber, G. K., Glasfeld, A., Petsko, G. A. & Ringe, D. (1994). Isotopic exchange plus substrate and inhibition kinetics of D-xylose isomerase do not support a proton-transfer mechanism. Biochemistry, 33, 1481–1487. Allen, K. N., Lavie, A., Glasfeld, A., Tanada, T. N., Gerrity, D. P., Carlson, S. C. et al. (1994). Role of the divalent metal ion in sugar binding, ring opening, and isomerization by D-xylose isomerase: replacement of a catalytic metal by an amino acid. Biochemistry, 33, 1488–1494. Murzin, A. G., Brenner, S. E., Hubbard, T. & Chothia, C. (1995). SCOP: a structural classification of proteins database for the investigation of sequences and structures. J. Mol. Biol. 247, 536–540. Hendrickson, W. A., Horton, J. R. & LeMaster, D. M. (1990). Selenomethionyl proteins produced for analysis by multiwavelength anomalous diffraction (MAD): a vehicle for direct determination of three-dimensional structure. EMBO J. 9, 1665–1672. Otwinowski, Z. & Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode. Macromolecular Crystallography, part A. Methods Enzymol. 276, 307–326. Terwilliger, T. C. & Berendzen, J. (1999). Automated MAD and MIR structure solution. Acta Crystallog. sect. D, 55, 849–861. Terwilliger, T. C. (2000). Maximum likelihood density modification. Acta Crystallog. sect. D, 56, 965–972. Terwilliger, T. C. (2002). Automated main-chain model-building by template-matching and iterative fragment extension. Acta Crystallog. sect. D, 59, 34–44. McRee, D. E. (1993). In (McRee, D. E., ed.), Academic Press Inc., New York. Brunger, A. T. (1993). X-PLOR 3.1: A System for X-ray Crystallography and NMR. Yale University Press, New Haven and London. Engh, R. A. & Huber, R. (1991). Accurate bond and angle parameters for X-ray protein-structure refinement. Acta Crystallogr. sect. A, 47, 392–400.
30.
31.
32.
33.
34.
35. 36. 37. 38. 39. 40.
Edited by I. Wilson (Received 31 August 2006; received in revised form 27 October 2006; accepted 1 November 2006) Available online 6 November 2006