Structures of endo-1,5-α-l-arabinanase mutants from Bacillus thermodenitrificans TS-3 in complex with arabino-oligosaccharides

Crystal structures of D27A and D147N mutants of the endo-1,5-α-l-arabinanase from Bacillus thermodenitrificans TS-3 in complex with arabinohexaose were determined.

Abstract

The thermostable endo-1,5-α-l-arabinanase from Bacillus thermodenitrificans TS-3 (ABN-TS) hydrolyzes the α-1,5-l-arabinofuranoside linkages of arabinan. In this study, the crystal structures of inactive ABN-TS mutants, D27A and D147N, were determined in complex with arabino-oligosaccharides. The crystal structures revealed that ABN-TS has at least six subsites in the deep V-shaped cleft formed across one face of the propeller structure. The structural features indicate that substrate recognition is profoundly influenced by the remote subsites as well as by the subsites surrounding the active center. The ‘open’ structure of the substrate-binding cleft of the endo-acting ABN-TS is suitable for the random binding of several sugar units in polymeric substrates.

1. Introduction  

Hemicellulose is the second most abundant polymer in plant biomass after cellulose. This renewable polysaccharide has attracted attention as an important resource for biotechnological conversion to biofuels and other value-added materials in various fields (Saha, 2003 ▸; Jørgensen et al., 2007 ▸; Liu et al., 2017 ▸). Arabinan, which is distributed in hemicellulose, comprises an α-1,5-linked backbone of l-arabinofuranosyl residues, some of which are substituted with α-1,2- and/or α-1,3-linked arabinofuranoside side chains (Bacic et al., 1988 ▸). Arabinan-degrading enzymes are classified by their mode of action and their substrate specificity (Beldman et al., 1997 ▸). α-l-Arabinanases (Abns), which include the endo-type (EC 3.2.1.99) and exo-type (EC 3.2.1.–), hydrolyze α-1,5-l-arabinofuranoside linkages, and α-l-arabinofuranosidases (Abfs; EC 3.2.1.55) cleave the arabinose side chains. Based on amino-acid sequence similarity according to the Carbohydrate-Active Enzymes (CAZy) classification (http://www.cazy.org/; Cantarel et al., 2009 ▸), Abns have been classified into glycoside hydrolase (GH) families 43 and 93, and Abfs into GH3, GH43, GH51, GH54 and GH62. They act synergistically to reduce arabinan to l-arabinose and/or arabino-oligosaccharides (Kaneko et al., 1998 ▸). l-Arabinose is poorly absorbed by the intestine and inhibits its sucrase activity, suppressing the increase in plasma glucose levels, and can potentially be used as a bioactive noncaloric sweetener (Seri et al., 1996 ▸; Osaki et al., 2001 ▸). Arabinan-degrading enzymes have gained interest in recent years because of their potential applications in diverse fields including food technology, nutrition research, bio-energy production, organic synthesis and the medicinal industry (Beldman et al., 1997 ▸; Numan & Bhosle, 2006 ▸; Jørgensen et al., 2007 ▸).

In previous work, we determined the crystal structure of the thermostable endo-1,5-α-l-arabinanase from Bacillus thermodenitrificans TS-3 (ABN-TS) at 1.9 Å resolution and clarified the structural features responsible for its thermostability (Yamaguchi et al., 2005 ▸). ABN-TS, which belongs to GH43, hydrolyzes linear arabinan almost exclusively to arabinobiose using an endo mechanism with inversion of configuration at the anomeric C atom (Takao, Akiyama et al., 2002 ▸). The enzyme molecule has the five-bladed β-propeller fold that is common to GH43 enzymes. The deep V-shaped cleft of ABN-TS is formed across one face of the propeller and is presumed to be the substrate-binding site. The putative catalytic residues, Asp27, Asp147 and Glu201, which are conserved at equivalent positions in all enzymes belonging to GH families 32, 43, 62 and 68 (Pons et al., 2004 ▸), are located in the central part of the cleft.

Currently, only three structures of complexes of endo-type Abns with the short-chain arabino-oligosaccharides arabinobiose and arabinotriose have been reported (Alhassid et al., 2009 ▸; de Sanctis et al., 2010 ▸), although it is considered that the open extended cleft of endo-type Abns possesses at least five sugar-binding subunits. Thus, we attempted to determine the structure of the complex of ABN-TS with the long-chain arabino-oligosaccharide arabinohexaose. Here, we report the crystal structures of ABN-TS with mutations of the catalytic residues in complex with arabinohexaose and arabinotriose. A detailed study of the substrate-binding structures should provide some information about the substrate-recognition mechanism of endo-type Abn enzymes.

2. Materials and methods  

2.1. Macromolecule production  

The gene for ABN-TS was amplified by PCR using the plasmid pUBabn containing the ABN-TS gene (Takao, Yamaguchi et al., 2002 ▸) as a template with the cloning primers shown in Table 1 ▸. The PCR fragment was inserted into the NdeI–XhoI sites of the pET-30a(+) vector (Novagen) containing a C-terminal His tag to construct a new plasmid, pET-ABN-TS. Site-directed mutagenesis was performed with a QuikChange Site-Directed Mutagenesis Kit (Stratagene) using pET-ABN-TS as a template. Expression plasmids containing the genes for the D27A and D147N mutants of ABN-TS were generated using the mutation primer pairs shown in Table 1 ▸. DNA sequences were verified by sequencing using a dye terminator cycle sequencing kit (Beckman Coulter) and a CEQ2000 fragment-analysis system (Beckman Coulter).

Table 1. Macromolecule-production information.

Source organism	B. thermodenitrificans strain TS-3
DNA source	pUBabn
Cloning, forward primer	5′-TCCCATATGGTTCATTTCCACCCCTTTGGC-3′
Cloning, reverse primer	5′-TACCTCGAGCAAATACGGCCACCCTTCATC-3′
Mutation primer 1 for D27A	5′-TATGGGCTCATGCACCAGTCATCGC-3′
Mutation primer 2 for D27A	5′-GCGATGACTGGTGCATGAGCCCATA-3′
Mutation primer 1 for D147N	5′-ATGCGATAAACCCGAATGTAG-3′
Mutation primer 2 for D147N	5′-CTACATTCGGGTTTATCGCAT-3′
Expression vector	pET-30a(+)
Expression host	E. coli BL21(DE3)
Complete amino-acid sequences of the constructs produced†
ABN-TS D27A mutant	MVHFHPFGNVNFYEMDWSLKGDLWAHAPVIAKEGSRWYVFHTGSGIQIKTSEDGVHWENMGRVFPSLPDWCKQYVPEKDEDHLWAPDICFYNGIYYLYYSVSTFGKNTSVIGLATNRTLDPRDPDYEWKDMGPVIHSTASDNYNAIDPNVVFDQEGQPWLSFGSFWSGIQLIQLDTETMKPAAQAELLTIASRGEEPNAIEAPFIVCRNGYYYLFVSFDFCCRGIESTYKIAVGRSKDITGPYVDKNGVSMMQGGGTILDAGNDRWIGPGHCAVYFSGVSAILVNHAYDALKNGEPTLQIRPLYWDDEGWPYLLEHHHHHH
ABN-TS D147N mutant	MVHFHPFGNVNFYEMDWSLKGDLWAHDPVIAKEGSRWYVFHTGSGIQIKTSEDGVHWENMGRVFPSLPDWCKQYVPEKDEDHLWAPDICFYNGIYYLYYSVSTFGKNTSVIGLATNRTLDPRDPDYEWKDMGPVIHSTASDNYNAINPNVVFDQEGQPWLSFGSFWSGIQLIQLDTETMKPAAQAELLTIASRGEEPNAIEAPFIVCRNGYYYLFVSFDFCCRGIESTYKIAVGRSKDITGPYVDKNGVSMMQGGGTILDAGNDRWIGPGHCAVYFSGVSAILVNHAYDALKNGEPTLQIRPLYWDDEGWPYLLEHHHHHH

^†The C-terminal 6×His fusion tag is underlined.

The recombinant ABN-TS mutants were expressed in Escherichia coli BL21(DE3) cells as fusion proteins with a His tag at the C-terminus. E. coli BL21(DE3) cells harboring the expression plasmid were cultured in 1 l LB medium containing 100 µg ml⁻¹ kanamycin at 310 K. Expression was induced by the addition of 0.5 mM isopropyl β-d-1-thiogalactopyranoside when the absorbance at 600 nm reached 0.6; cultivation was continued for a further 4 h at 310 K. The cells were harvested, resuspended in 30 ml 20 mM sodium phosphate pH 7.4 and sonicated. The cellular debris was removed by centrifugation and the supernatant was purified by chromatography using HisTrap HP (GE Healthcare), UNO Q (Bio-Rad) and Superdex 75 (GE Healthcare) columns on an FPLC system from GE Healthcare at 277 K. The fractions containing the recombinant proteins were pooled and the protein purity was confirmed by SDS–PAGE. The purified protein was dialysed against 20 mM Tris–HCl buffer pH 7.5 and concentrated using an Amicon Ultra centrifugal filter with a molecular-mass cutoff of 10 kDa (Merck Millipore). The catalytic activities of the ABN-TS mutants were measured using previously described methods (Takao, Akiyama et al., 2002 ▸; Yamaguchi et al., 2005 ▸).

2.2. Crystallization  

Crystallization trials were conducted using the sitting-drop vapor-diffusion method at 293 K. Crystallization conditions for the ABN-TS D27A and D147N mutants in complex with the natural substrate arabinohexaose (D27A-A6 and D147N-A6) were initially searched for using the Crystal Screen and Crystal Screen 2 kits (Hampton Research). Arabinohexaose (>95% purity) was purchased from Megazyme International. After several optimization steps, well formed crystals of the D27A-A6 and D147N-A6 complexes were obtained in two weeks under the conditions shown in Table 2 ▸.

Table 2. Crystallization.

	D27A-A6	D147N-A6
Method	Sitting-drop vapor diffusion	Sitting-drop vapor diffusion
Plate type	24-well plate	24-well plate
Temperature (K)	293	293
Protein concentration (mg ml−1)	15	20
Buffer composition of protein solution	20 mM Tris–HCl pH 7.5	20 mM Tris–HCl pH 7.5
Composition of reservoir solution	20%(w/v) PEG 8000, 0.2 M magnesium chloride, 0.1 M Tris–HCl pH 8.5	16%(w/v) PEG 8000, 0.2 M magnesium chloride, 0.1 M MES pH 5.5
Volume and ratio of drop	1 µl:1 µl	1 µl:1 µl
Volume of reservoir (µl)	400	400

2.3. Data collection and processing  

Prior to data collection at cryogenic temperature, crystals of the D27A-A6 and D147N-A6 complexes were soaked in a cryoprotectant solution containing a small amount of arabino­hexaose and 20% glycerol. The crystals were mounted in LithoLoops (Molecular Dimensions) and directly flash-cooled in a nitrogen-gas stream at 100 K. X-ray diffraction data sets were collected on beamline BL38B1 at SPring-8 for the D27A-A6 complex and on beamline BL6A at Photon Factory for the D147N-A6 complex. The data sets were processed and scaled using HKL-2000 (Otwinowski & Minor, 1997 ▸). The crystal of the D27A-A6 complex belonged to the orthorhombic space group P2₁2₁2₁ with one protein monomer in the crystallo­graphic asymmetric unit, whereas the crystal of the D147N-A6 complex belonged to the monoclinic space group P2₁ with two protein monomers in the crystallographic asymmetric unit. The data statistics are summarized in Table 3 ▸.

Table 3. Data collection and processing.

Values in parentheses are for the outer shell.

	D27A-A6	D147N-A6
Diffraction source	BL38B1, SPring-8	BL6A, Photon Factory
Wavelength (Å)	1.000	0.978
Temperature (K)	100	100
Detector	Jupiter 210	Quantum 4
Crystal-to-detector distance (mm)	150	150
Rotation range per image (°)	1	1
Total rotation range (°)	270	180
Exposure time per image (s)	15	10
Space group	P212121	P21
a, b, c (Å)	40.13, 77.03, 88.70	45.89, 92.26, 78.67
α, β, γ (°)	90, 90, 90	90, 91.52, 90
Mosaicity (°)	0.902	0.781
Resolution range (Å)	50.00–1.65 (1.68–1.65)	50.00–1.90 (1.93–1.90)
Total No. of reflections	330245	175322
No. of unique reflections	33864 (1677)	51538 (2522)
Completeness (%)	99.9 (99.6)	99.2 (97.6)
Multiplicity	9.8 (8.2)	3.7 (3.6)
〈I/σ(I)〉	34.4 (3.95)	13.9 (2.36)
R p.i.m.	0.019 (0.178)	0.055 (0.324)
Overall B factor from Wilson plot (Å2)	18.5	15.8

2.4. Structure solution and refinement  

The crystal structures of the D27A-A6 and D147N-A6 complexes were solved by the molecular-replacement method with MOLREP (Vagin & Teplyakov, 2010 ▸) as implemented in the CCP4 suite (Winn et al., 2011 ▸) using the structure of wild-type ABN-TS (PDB entry 1wl7; Yamaguchi, 2005 ▸) as a search model. Subsequently, refinement was performed using REFMAC5 (Murshudov et al., 2011 ▸) followed by manual rebuilding of the structure in Coot (Emsley et al., 2010 ▸). The stereochemistry of the final model was analyzed using PROCHECK (Laskowski et al., 1993 ▸). The final refinement statistics are summarized in Table 4 ▸. The coordinates and structure factors have been deposited in the Protein Data Bank with accession codes 6a8h and 6a8i. All structural figures were generated using PyMOL (http://pymol.org/).

Table 4. Structure refinement.

Values in parentheses are for the outer shell.

	D27A-A6	D147N-A6
Resolution range (Å)	50.00–1.65 (1.69–1.65)	50.00–1.90 (1.95–1.90)
Completeness (%)	99.9 (99.4)	99.3 (98.0)
No. of reflections
Working set	32149 (2324)	48588 (3527)
Test set	1658 (107)	2625 (199)
Final R cryst (%)	14.7	18.1
Final R free (%)	19.5	23.3
No. of non-H atoms
Protein	2523	5112
Ion	2	2
Ligand	56	110
Water	266	461
Total	2847	5685
R.m.s. deviations
Bonds (Å)	1.00	0.97
Angles (°)	1.03	0.96
Average B factors (Å2)
Protein	24.8	22.9
Ion	28.5	42.7
Ligand	49.4	42.8
Water	36.4	33.4
Ramachandran plot (%)
Favored regions	96.2	94.7
Additionally allowed	3.5	4.8
PDB entry	6a8h	6a8i

3. Results and discussion  

Based on sequence homology, three acidic residues, Asp27, Asp147 and Glu201, which are located in the central part of the substrate-binding cleft, are estimated to be critical for the catalytic activity of ABN-TS, with Asp27 as the general base, Asp147 as the pK _a modulator and Glu201 as the general acid (Alhassid et al., 2009 ▸). Prior to the X-ray crystallographic experiments, we confirmed that the ABN-TS mutants D27A, D147A, D147N and E201A did not show enzyme activity with debranched arabinan as a substrate on incubation at 343 K for 30 min (data not shown) according to the method reported previously (Yamaguchi et al., 2005 ▸).

The crystal structures of the D27A-A6 and D147N-A6 complexes were solved at 1.65 and 1.90 Å resolution, respectively. The asymmetric unit comprised one enzyme molecule in the crystal structure of the D27A-A6 complex and two enzyme molecules, with the same fold, in the crystal structure of the D147N-A6 complex. The overall structures of the ABN-TS D27A and D147N mutants had a five-bladed β-propeller fold (Figs. 1 ▸ a and 1 ▸ b), which was identical to that in wild-type ABN-TS, as indicated by the r.m.s.d. values of 0.21 and 0.38 Å for equivalent C^α atoms between the wild-type enzyme and the D27A and D147N mutants, respectively. The F _o − F _c difference map of the D147N-A6 complex structure showed continuous electron density corresponding to an arabino­hexaose molecule in the deep V-shaped cleft formed across one face of the propeller fold (Fig. 1 ▸ d). There were no significant differences in conformation and binding mode between the two arabinohexaose molecules bound to the two crystallo­graphically independent ABN-TS D147N mutant molecules. The subsites occupied by the arabinose units of arabinohexaose were identified as −3 to +3 (Fig. 1 ▸ d). On the other hand, the electron-density distribution in the cleft was cut at the active center for the D27A-A6 complex structure. Thus, we assigned two arabinotriose molecules to the electron-density distribution (Fig. 1 ▸ c). We considered that the arabino­triose was formed as a reaction product during long-term storage for crystallization with a high concentration of substrate and the ABN-TS D27A mutant, which is less active than the wild-type enzyme but retains catalytic activity. The two arabinotriose molecules occupied subsites −3 to −1 and subsites +1 to +3, respectively.

Figure 1.

Crystal structures of ABN-TS mutants in complex with arabino-oligosaccharides. (a, b) The overall structures of D27A-A6 (a) and D147N-A6 (b) are represented as cartoon models colored from the N-terminus to the C-terminus (blue to red). The catalytic residues and mutated residues are shown as green sticks. The arabinose units are shown as yellow sticks. (c, d) Closer views of the substrate-binding clefts of D27A-A6 (c) and D147N-A6 (d). The F _o − F _c maps omitting the arabino-oligosaccharides are shown as a cyan mesh contoured at 1.0σ for stick models of arabinotriose and arabinohexaose, respectively.

The D147N-A6 complex structure revealed the amino-acid residues participating in substrate recognition (Fig. 2 ▸ a). At the −1 subsite, extensive interaction and tight binding was observed between the enzyme and arabinohexaose. The position of the arabinose unit in the −1 subsite is maintained by hydrogen bonds to the three catalytic residues, Asp27, Glu201 and Asn147, which substitutes for the pK _a modulator Asp147. The Asp27 O^δ2 atom is located 2.9 Å from the C5 atom, forming a weak hydrogen bond (Gu et al., 1999 ▸). The Glu201 O^∊1 and Asn147 O^δ1 atoms are 3.1 Å from the 2-OH group and 2.8 Å from the 3-OH group, respectively, and are also maintained through hydrophobic interactions with Ile146 and Trp84. The arabinohexaose also showed high affinity for the +1 subsite, like the −1 subsite. The arabinose unit occupying the +1 subsite forms hydrogen bonds to Gly105 NH (2.8 Å), Ser164 O^γ (2.8 Å) and Asn144 N^δ2 (3.0 Å). A notable characteristic is the bridge-like structure formed by the side chain of Phe104 and the disulfide bond of the adjacent Cys221 and Cys222 (Fig. 2 ▸ b). The phenylalanine and cysteine residues form hydrophobic interactions with the C4 and C5 atoms of the arabinose unit at the +1 subsite; therefore, these residues seem to be responsible for substrate recognition (Miyanaga et al., 2004 ▸).

Figure 2.

Structure of the substrate-binding cleft of ABN-TS D147N complexed with arabinohexaose. (a) Amino-acid residues (green) that interact with arabinohexaose (yellow) are shown as stick models. The hydrogen bonds are shown as dashed lines. (b) Surface representation of the bridge-like structure at the active site.

The arabinose unit at the +2 subsite is exposed to the solvent region and does not form short contacts with the enzyme. The +2 subsite seems to have no potential for furanose-ring recognition. In contrast, the arabinose unit is held tightly in the +3 subsite through hydrogen bonds to Glu196 O^∊1 (2.6 Å) and Trp166 NH (2.9 Å), and hydrophobic interaction with Phe165. A stacking interaction of Trp166 was formed in the +3 subsite, where the indole ring of Trp166 is nearly parallel to the arabinose ring. The arabinose unit of the nonreducing end is not held tightly in the −3 subsite and only forms hydrophobic interactions with His26, Thr42 and His82, which are arranged between the −3 and −2 subsites. At the −2 subsite, the arabinose unit is bound to the enzyme by a hydrogen bond to His26 N^∊2 (2.7 Å) and hydrophobic interactions with His26, His82 and Phe104. Notably, Trp84 also forms a stacking interaction with the arabinose unit through its indole ring in a manner similar to that of Trp166. These features indicate that interactions at the remote −2 and +3 subsites can profoundly influence the specificity of arabinanase and that two tryptophan residues, Trp84 and Trp166, play an important role in substrate recognition.

The most remarkable feature of the D147N-A6 complex structure is that the side chain of Trp24 moves out of the cleft to a solvent-exposed position (Fig. 3 ▸). In the structure of wild-type ABN-TS the side chain of Trp24 points into of the −3 subsite. In addition, the Trp24 side chain of the D27A-A6 complex structure, in which the arabinotriose molecule is bound to the −3 to −1 subsites, is also located in the same position as in the wild type. Interestingly, in the D27A-A6 complex structure the arabinose unit of the nonreducing end is located in a position that avoids steric hindrance with the Trp24 residue. These findings seem to suggest that the Trp24 residue is involved in regulation of the catalytic activity of ABN-TS through conformational alteration. The shift of the Trp24 side chain towards the solvent region contributes to enhance the binding affinities of the long-chain arabino-oligosaccharides for the active site and the shift towards the −3 subsite destabilizes product binding, allowing its release from the active site. The crystal structure of the GH43 arabinanase from Geobacillus stearothermophilus in complex with arabino­octaose has recently been released in the Protein Data Bank (PDB entry 5ho9; S. Lansky, R. Salama, O. Shwartstien, Y. Shoham & G. Shoham, unpublished work). However, there is no information available about its enzymatic characteristics at present. Detailed information on the structure and function of the enzyme will be expected to give more insight into the substrate-recognition mechanism for arabinanases.

Figure 3.

Structure of Trp24 around the −3 subsite. Superposition of the substrate-binding cleft of wild-type ABN-TS (cyan) and the D27A-A6 (green) and D147N-A6 (pink) mutants.

We previously reported that the structure of the substrate-binding cleft of ABN-TS, which is a typical endo-acting enzyme, differs from that of the exo-acting enzyme CjArb43A, which has very a similar topology to that of ABN-TS (Yamaguchi et al., 2005 ▸). To further elucidate the characteristic of the endo-type and exo-type substrate-binding clefts, the structures of the D147N-A6 complex and the CjArb43A–arabinohexaose complex (PDB entry 1gye; Nurizzo et al., 2002 ▸) were compared (Fig. 4 ▸). The substrate-binding cleft of CjArb43A is blocked at one end by a large loop structure composed of 13 amino-acid residues between the third and fourth propeller blades, and the substrate is bent by the loop, as shown in Figs. 4 ▸(b) and 4 ▸(d). At the −3 subsite of CjArb43A, Asp35 and Gln316 interact with the arabinose unit of the nonreducing end through hydrogen bonds and are considered to be involved in recognizing the nonreducing end of the substrate (Proctor et al., 2005 ▸). On the other hand, ABN-TS has only a small loop composed of four residues (Glu194–Asn197) and both sides of the cleft are open, as shown in Figs. 4 ▸(a) and 4 ▸(c). It was found that the substrate was located along the deep V-shaped cleft formed across one face of the propeller fold in the D147N-A6 complex structure and the arabinose unit of the nonreducing end made no specific interactions with the enzyme. This ‘open’ structure of the cleft of ABN-TS will allow the random binding of several sugar units in long-chain arabino-oligosaccharides.

Figure 4.

Comparison of the structures of the substrate-binding clefts of ABN-TS and CjArb43A. (a, b) The cleft structures of ABN-TS (a) and CjArb43A (b) are shown as cartoon models colored from the N-terminus to the C-terminus (blue to red). The loop structures are colored magenta. (c, d) Surface representations of the substrate-binding clefts of D147N-A6 (c) and CjArb43A complexed with arabinohexaose (d). The surfaces of the ABN-TS loop and the CjArb43A loop are colored pink and cyan, respectively. Amino-acid residues and arabinohexaose are shown as stick models.

Supplementary Material

PDB reference: endo-1,5-α-l-arabinanase, D27A mutant, complex with arabinotriose, 6a8h

PDB reference: D147N mutant, complex with arabinohexaose, 6a8i