research papers
Structural insight into the hydrolase and synthase activities of an alkaline α-galactosidase from Arabidopsis from complexes with substrate/product
aLife Science Group, Scientific Research Division, National Synchrotron Radiation Research Cente, Hsinchu 30076, Taiwan, bInstitute of Tropical Plant Sciences and Microbiology, National Cheng Kung University, Tainan City 701, Taiwan, cInstitute for Protein Research, Osaka University, Suita, Osaka 565-0871, Japan, dDepartment of Biotechnology and Bioindustry Sciences, National Cheng Kung University, Tainan City 701, Taiwan, eDepartment of Physics, National Tsing Hua University, Hsinchu 30013, Taiwan, and fDepartment of Biological Science and Technology, National Chiao Tung University, Hsinchu 30010, Taiwan
*Correspondence e-mail: cjchen@nsrrc.org.tw
The alkaline α-galactosidase AtAkαGal3 from Arabidopsis thaliana catalyzes the hydrolysis of α-D-galactose from galacto-oligosaccharides under alkaline conditions. A phylogenetic analysis based on sequence alignment classifies AtAkαGal3 as more closely related to the raffinose family of oligosaccharide (RFO) synthases than to the acidic α-galactosidases. Here, is used to demonstrate that AtAkαGal3 exhibits a dual function and is capable of synthesizing stachyose using raffinose, instead of galactinol, as the galactose donor. Crystal structures of complexes of AtAkαGal3 and its D383A mutant with various substrates and products, including galactose, galactinol, raffinose, stachyose and sucrose, are reported as the first representative structures of an alkaline α-galactosidase. The structure of AtAkαGal3 comprises three domains: an N-terminal domain with 13 antiparallel β-strands, a with an (α/β)8-barrel fold and a C-terminal domain composed of β-sheets that form two Greek-key motifs. The WW box of the N-terminal domain, which comprises the conserved residues FRSK75XW77W78 in the RFO synthases, contributes Trp77 and Trp78 to the +1 subsite to contribute to the substrate-binding ability together with the (α/β)8 barrel of the The C-terminal domain is presumably involved in structural stability. Structures of the D383A mutant in complex with various substrates and products, especially the natural substrate/product stachyose, reveal four complete subsites (–1 to +3) at the catalytic site. A functional loop (residues 329–352) that exists in the alkaline α-galactosidase AtAkαGal3 and possibly in RFO synthases, but not in acidic α-galactosidases, stabilizes the stachyose at the +2 and +3 subsites and extends the catalytic pocket for the transferase mechanism. Considering the similarities in amino-acid sequence, and activity between alkaline α-galactosidases and RFO synthases, the structure of AtAkαGal3 might also serve a model for the study of RFO synthases, structures of which are lacking.
Keywords: Arabidopsis thaliana; alkaline α-galactosidases; raffinose; RFO synthases; transferase mechanism.
1. Introduction
α-Galactosidases (α-D-galactoside galactohydrolases; EC 3.2.1.22), which are widely found in microorganisms, plants, animals and humans, catalyze the removal of the nonreducing α-D-galactose moiety from galacto-oligosacharides, galactolipids and (Henrissat et al., 2001). Fabry disease is a rare genetic disease in humans that is due to a deficiency of α-galactosidase A, which is needed to catabolize the α-D-galactosyl moiety of sphingolipids in the lysosome (Calhoun et al., 1985; Weidemann et al., 2003). The hydrolytic specificity of α-galactosidases enables their application in the food industry to remove raffinose and to increase the yield of sucrose in the sugar beet industry (Shibuya et al., 1995), to improve the gelling properties of galactomannans for use as food thickeners and to degrade the raffinose family of (RFOs) in food and feed materials and in enzymotherapy (Guimarães et al., 2001).
Based on their glycoside hydrolase (GH) signatures, α-galactosidases and RFO synthases belong to GH families 27 and 36, which are characterized by two conserved motifs: KxD and RxxxD (Henrissat & Romeu, 1995; Henrissat, 1991; Carmi et al., 2003; Lee et al., 2004; https://www.cazy.org/). Plant α-galactosidases can be further divided into acidic and alkaline forms, according to their optimal pH value for enzymatic activity. Alkaline α-galactosidases have optimal activities at neutral–alkaline pH values, whereas acidic α-galactosidases have optimal pH values of 3–6.5 (Lee et al., 2004). Another major difference between these two types of α-galactosidase is that the total number of amino acids in alkaline α-galactosidases is about twice that in acidic α-galactosidases. A phylogenic analysis showed that the amino-acid sequences of alkaline α-galactosidases share a higher similarity to those of RFO synthases than to those of acidic α-galactosidases (Lee et al., 2004). Both alkaline α-galactosidases and RFO synthases are unique to plants. Most studies have focused on acidic α-galactosidases (Keller & Pharr, 1996; Dey & Pridham, 1972; Herman & Shannon, 1985). Acidic α-galactosidases are involved in the mobilization of seed reserves of RFOs and the degradation of cell-wall galactomannan during seed germination (Corchete & Guerra, 1987). It has been suggested that these enzymes also play a role in cell-wall modification during fruit ripening (Kang & Lee, 2001; Nunan et al., 2001).
The first alkaline α-galactosidase was uncovered in the Cucurbitaceae family (Gaudreault & Webb, 1986; Gao & Schaffer, 1999). High alkaline α-galactosidase activity was mostly found in roots and immature leaves in the early development stage; the activity declined with increasing leaf maturity in Cucurbita pepo (Gaudreault & Webb, 1986). High activity in an immature organ is due to the translocation of RFOs, the major galactosyl-sucrose from their source organs to developing organs such as leaves (Gaudreault & Webb, 1986) and fruits (Gao & Schaffer, 1999). Alkaline α-galactosidases can be found in more than one form in the Cucurbitaceae family, with varied substrate specificities. Alkaline α-galactosidases from Cucurbita pepo, Cucumis sativus and Cucurbita maxima prefer stachyose to raffinose, whereas in Cucumis melo fruit alkaline α-galactosidases of forms I and II exhibit a higher activity towards raffinose and stachyose, respectively (Gao & Schaffer, 1999). The maximum activity of maize alkaline α-galactosidase was detected in mature dehydrated, germinating and germinated seeds (Zhao et al., 2006); ZmAGA1, with maximum catalysis at pH 7.5, was solely involved in seed germination. Elevated levels of alkaline α-galactosidase gene expression were detected when seed germination in maize was interrupted by heat, cold or dehydration (Zhao et al., 2006) and when New Zealand spinach was grown under drought conditions (Hara et al., 2008). A rice chloroplast alkaline α-galactosidase (OsAkαGal) has been suggested to exhibit a function in degrading digalactosyl diacylglycerol, a major glycolipid in the chloroplast thylakoid membrane, during leaf senescence (Lee et al., 2004, 2009).
The Arabidopsis thaliana (Arabidopsis) genome contains genes for three alkaline α-galactosidases (At1g55740, At3g57520 and At5g20250), two RFO synthases (At4g01970 and At5g40390) and four acidic α-galactosidases (At3g26380, At5g08370, At5g08380 and At3g56310). Here, the proteins encoded by At1g55740, At3g57520 and At5g20250 are referred to as AtAkαGal1, AtAkαGal2 and AtAkαGal3, respectively (Lee et al., 2004). A BLAST search showed that AtAkαGal3 is related to probable galactinol–sucrose galactosyltransferases from various species, such as Arabidopsis lyrata subsp. lyrate (XP_020876652), Camelina sativa (XP_010493135), Capsella rubella (XP_006287127), Raphanus sativus (XP_018444981), Brassica oleracea var. oleracea (XP_013612073.1) and Prunus persica (XP_020424207), with 76–94% sequence identity. Moreover, AtAkαGal3 shares sequence identities of less than 30% with all acidic α-galactosidases from Arabidopsis and other plants.
To date, only structures of acidic α-galactosidases have been reported, including rice acidic α-galactosidase (PDB entry 1uas; Fujimoto et al., 2003), chicken α-N-acetylgalactosaminidase (PDB entry 1ktb; Garman et al., 2002) and human α-galactosidase (PDB entry 1r46; Garman & Garboczi, 2004). The structures of rice acidic α-galactosidase and the closely related chicken α-N-acetylgalactosaminidase revealed a double-displacement catalytic mechanism and the mode of substrate binding (Fujimoto et al., 2003; Garman et al., 2002). However, no structure of any alkaline α-galactosidase or RFO synthase was available prior to this work. Here, we report structures of AtAkαGal3 and its D383A mutant in complex with various substrates and products, including the natural substrate/product stachyose, revealing four complete substrate-binding subsites (–1 to +3) at the catalytic site for catalysis. The dual-functional AtAkαGal3 was also discovered to exhibit stachyose synthase activity in this work; the new structure of AtAkαGal3 thus potentially serves as a model for the study of raffinose/stachyose synthases.
2. Methods
2.1. Cloning, expression and purification
Materials and chemicals were obtained from Sigma unless specified otherwise. A. thaliana L. Herynh plants (Col-0) were used in this study. AtAkαGal3 full-length open frame sequences (2273 bp) were obtained using database searches and sequence analyses using OsAkαGal as a query (Lee et al., 2004). The S3 stage of senescent leaves containing about 30–50% chlorophyll compared with full-grown green rosette leaves (100% chlorophyll) was used for total RNA extraction. Chlorophyll determination was carried out as described previously (Lee & Collins, 2001). Total RNA was isolated using the method described previously (Chang et al., 1993), and an Agilent 2100 Bioanalyzer (ThermoFisher, USA) was used to assess its quality and quantity. First-strand cDNA was synthesized from 1 µg of total RNA using the SuperScript III First-Strand Synthesis System, as described by the manufacturer (ThermoFisher, USA). A DNase (Promega, USA) digestion step was conducted prior to reverse transcription in a total volume of 20 µl. AtAkαGal3 full-length open frame sequences (2273 bp) containing KpnI and SalI sites were amplified from 1 µl cDNA solution by 30 cycles of at 94°C for 30 s, annealing at 55°C for 30 s and extension at 72°C for 1 min in 20 µl PCR reaction mixture. The primer pair used was At5gKpnIF2 (GATCTGGGTACCATGACGATTAAACCGGCGGT) and At5gnostopSalIR (CTGCAGGTCGACTAACTCAACTTGGATCAGA). The AtAkαGal3 full-length open frame sequences were cloned into the expression vector pET-30a(+) (Novagen) at KpnI and XhoI sites with a 6×His tag at the N-terminus and were transformed into Escherichia coli BL21(DE3) cells for expression. The bacteria were cultured at 37°C until the OD600 reached ∼0.6 prior to induction with isopropyl β-D-thiogalactopyranoside (IPTG; final concentration 0.15 mM) in Luria–Bertani (LB) medium at 20°C for 20 h with shaking, and were then harvested by centrifugation at 8000g for 25 min at 4°C. The cell pellet from 1 l culture was suspended in lysis buffer (35 ml) consisting of 50 mM Tris–HCl pH 8.0 and was subjected to cell disruption by ultrasonication using a pulse cycle of 2 s on and 3 s off with a total duration of 20 min at 40% energy on ice. The soluble protein extract was collected by centrifugation at 12 000g for 30 min at 4°C. The target AtAkαGal3 was purified using a nickel-immobilized metal ion-affinity (Ni-IMAC) column with a low imidazole concentration (50 mM), thus avoiding protein contamination at higher concentrations that affects crystallization, in 20 mM Tris–HCl buffer pH 7.8 containing 500 mM NaCl and was further dialyzed against 20 mM Tris–HCl buffer pH 8.2 containing 1 mM dithiothreitol (DTT) and 1 mM galactose overnight. The eluted AtAkαGal3 was concentrated and subjected to on an S200 column previously equilibrated with the abovementioned dialysis buffer. The target protein was then purified using the same buffer condition as the dialysis buffer to improve the purity.
Two AtAkαGal3 mutants, D383A and D447A, were generated by a site-directed mutagenesis method (QuikChange kit, Stratagene) using the pET-30a-AtAkαGal3 recombinant vector as a template. The forward primers for the mutants are as follows: D383A, 5′-GGACGGTGTGAAAGTGGCTGTGCAGTGTGTATTGG-3′; D447A, 5′-TGATTAGAGCATCAGATGCTTTCTATCCACGGGATCC-3′. The D383A and D447A mutants were confirmed by sequencing before expression. After Ni-IMAC purification, both mutants were dialyzed against the same dialysis buffer as wild-type AtAkαGal3 protein with and without galactose for various substrate and product complex studies. Selenomethionine-labeled AtAkαGal3 (SeMet-AtAkαGal3) protein was prepared following the method of Van Duyne (Doublié, 2007). Briefly, a single colony of recombinant AtAkαGal3 in E. coli BL21(DE3) cells was inoculated into 5 ml LB culture at 37°C for 12 h and then inoculated into 100 ml M9 medium in a 250 ml flask and cultured overnight. The bacterial cells from this overnight culture were collected and inoculated into 1 l M9 medium; a mixture of amino acids (100 mg lysine, phenylalanine and threonine; 50 mg isoleucine, leucine and valine; 60 mg SeMet) was added when the OD reached 0.6. After bacterial growth for a further 15 min, 0.15 mM IPTG was added to induce protein production. The rice raffinose synthase from O. sativa L. var. Nipponbare (XP_015621501) was overexpressed, purified and characterized using the protocol described previously (Li et al., 2007).
2.2. Activity assay using thin-layer chromatography
The purified proteins, wild-type AtAkαGal3 and its D383A and D447A mutants, as well as rice raffinose synthase, at equal amounts of 4 × 10−12 mol were incubated with the substrate raffinose at a final concentration of 10 mM at 35°C for 1 h. Each protein reaction and standard (galactose, sucrose, raffinose and stachyose) were spotted (2 µl) onto a silica thin-layer plate (Merck Millipore); the sample was developed with a solvent consisting of a mixture of chloroform, acetic acid and water in a 6:7:1 ratio by volume (Farag, 1978). The results of the reaction were examined by spraying ethanol containing sulfuric acid (10%) onto the plate and baking it on a hotplate until brown spots developed.
2.3. Crystallization and collection of X-ray data
The purified SeMet-AtAkαGal3, wild-type and mutant AtAkαGal3 protein fractions from were concentrated to 40–50 mg ml−1 in a buffer consisting of 20 mM Tris–HCl pH 8.2, 1 mM DTT with or without 1 mM galactose before crystallization, which was performed using a microbatch method. The protein crystals appeared in 2–3 months using the best crystallization condition (0.1 M ammonium sulfate, 0.3 M sodium formate, 0.1 M Tris pH 7.8, 10% PEG 2000) at 18°C. Among the crystallization trials with various AtAkαGal3 proteins, crystals of SeMet-AtAkαGal3, wild-type AtAkαGal3 and the D383A mutant (referred to in the following as D383A) in various substrate or product complexes could be obtained. No satisfactory crystals of the D447A mutant were obtained. In addition, without galactose neither wild-type nor mutant crystals could be obtained. Crystals of D383A–substrate complexes were obtained by co-crystallization and subsequent crystal soaking with various substrates, including galactinol, galactose, raffinose and stachyose (final concentration 5 mM), in the reservoir solution at 25°C for 10 min.
The crystals were harvested, transferred to reservoir solution containing glycerol (20%) as a cryoprotectant and immediately cooled with liquid nitrogen for data collection. The X-ray diffraction experiments were performed at various wavelengths ranging from 0.9 to 1.0 Å on synchrotron beamlines TPS 05A (with a Rayonix MX300HS detector) and TLS 15A (with a Rayonix MX300HE detector) at the National Synchrotron Radiation Research Center (NSRRC) in Taiwan and on BL44XU (with a Dectris EIGER X 16M detector) at SPring-8 in Japan. The data sets for SeMet-AtAkαGal3, AtAkαGal3–galactose, D383A–galactose, D383A–stachyose and D383A–galactose–sucrose collected on the TPS 05A and TLS 15A beamlines were processed using HKL-2000 (Otwinowski & Minor, 1997) and the D383A–galactinol and D383A–rafffinose data sets collected on the BL44XU beamline were processed with XDS (Kabsch, 2010).
2.4. and refinement
The initial αGal3 was determined by the selenomethionine multi-wavelength anomalous diffraction (Se-MAD) phasing method. 29 selenomethionine sites were first located using SOLVE (Terwilliger & Berendzen, 1999) to generate reasonable initial phases with a figure of merit of 0.34 for data in the resolution range 30–3.0 Å. The phases were further calculated and improved with AutoSol (Zhao et al., 2006) in Phenix and the direct phase selection method (Chen et al., 2014). A model of most of the structure was autobuilt with ARP/wARP (Langer et al., 2008), and further model building and adjustment were undertaken with Coot (Emsley et al., 2010). The structures of the D383A mutant in its various complexes were determined by with MOLREP (Vagin & Teplyakov, 2010) from the CCP4 suite (Winn et al., 2011) using the previously solved structure of wild-type AtAkαGal3 as a model. All refinements were performed with REFMAC (Murshudov et al., 2011) from the CCP4 suite. The structural models of galactose, galactinol, raffinose and starchyose were obtained from eLBOW (Moriarty et al., 2009) in Phenix before building and refinement.
of wild-type AtAk2.4.1. Model validation
The final refined structures of wild-type AtAkαGal3 and the various D383A–substrate and D383A–product complexes were validated with MolProbity (Williams et al., 2018). The ligand-binding interaction at the catalytic binding site of the AtAkαGal3 structures was analyzed by PLIP (Salentin et al., 2015).
2.4.2. Accession codes
All of the coordinates and structural factors in this work have been deposited in the Protein Data Bank as PDB entries 7exf, 7exg, 7exh, 7exj, 7exr and 7exq for WT AtAkαGal3–galactose, D383A–galactose, D383A–galactinol, D383A–rafffinose, D383A–stachyose and D383A–galactose–sucrose, respectively.
3. Results and discussion
3.1. Protein purification
Both the recombinant wild-type AtAkαGal3 and the D383A mutant were passed through an Ni-IMAC column and a size-exclusion column (S200) for purification. All purified AtAkαGal3 and mutant proteins showed a single band corresponding to a molecular mass of ∼82 kDa on SDS–PAGE. The target proteins eluted from at a molecular mass in the range 80–100 kDa, indicating that the biological unit of AtAkαGal3 is a monomer.
3.2. Hydrolytic activity assay
A hydrolytic reaction mixture containing AtAkαGal3 and its natural substrate raffinose was incubated and subsequently analyzed by (TLC; Supplementary Fig. S1). For a comparison of activity, a similar assay was simultaneously performed with purified recombinant rice raffinose synthase (O. sativa L. var. Nipponbare; Li et al., 2007). After the hydrolysis reactions with rice raffinose synthase and with AtAkαGal3, the TLC plate showed that the raffinose substrate was hydrolyzed to galactose and sucrose (lane 5) and to galactose, sucrose and stachyose (lane 6), respectively, while excess raffinose substrate remained. The results indicate that the hydrolase activity of AtAkαGal3 is greater than that of rice raffinose synthase towards raffinose as a substrate. Besides hydrolase activity, AtAkαGal3 exhibits stachyose synthase activity by transferring a galactose moiety to the raffinose substrate, in which both the galactose donor and acceptor are raffinose. Moreover, the TLC results showed that AtAkαGal3 appears to be incapable of synthesizing raffinose as a raffinose synthase according to an activity comparison between AtAkαGal3 and rice raffinose synthase with galactinol as the galactose donor and sucrose as the acceptor (data not shown). The activity assay thus shows that alkaline α-galactosidase exhibits a dual function of hydrolase and stachyose synthase activities.
3.3. Overall structure of Arabidopsis alkaline α-galactosidase
The αGal3 was first solved ab initio by the selenomethionine multi-wavelength anomalous diffraction (Se-MAD) method at a resolution of 2.8 Å and was subsequently refined to a high resolution of 2.17 Å using the native crystal with the best diffraction quality, which contains two molecules in the (Supplementary Fig. S2a). The crystal structures of complexes of the D383A mutant with galactose, galactinol, raffinose and stachyose were subsequently determined using the wild-type AtAkαGal3 structure as a search model by the molecular-replacement method (Table 1). The of AtAkαGal3 comprises 724 of a total of 749 amino acids that fold into three domains: N-terminal, catalytic and C-terminal (Supplementary Fig. S2b). The first four residues are absent due to a lack of electron density. The N-terminal domain (residues 5–187) consists of 13 antiparallel β-strands connected by variable loops with an α-helix at the end. The (residues 206–518) folds into an (α/β)8-barrel structure with eight parallel β-strands placed around a central axis and surrounded by eight α-helices, which can be seen in α-galactosidases of both the GH27 and GH36 families. The C-terminal domain comprises 16 β-strands that form two Greek-key motifs comprising residues 537–653 and 654–749, respectively, which are presumably involved in structural stability. Two regions of residues, 103–119 at the N-terminal domain and 254–263 at the are disordered without electron density, possibly due to loop flexibility on the protein surface (Supplementary Fig. S2b).
of wild-type AtAk
|
3.4. Protein sequence alignment
A BLAST search with the amino-acid sequence of AtAkαGal3 shows that the highest sequence similarities of 75–94% correspond to various probable galactinol–sucrose galactosyltransferases from different species, such as A. lyrata subsp. lyrate (94%), C. rubella (92%), C. sativa (91%), R. sativus (87%), O. sativa L. var. Nipponbare, B. oleracea (87%) and P. persica (76%). However, Lee et al. (2009) analyzed 31 plant α-galactosidases and RFO synthases based on sequence alignments and identities and found that they could be classified into acidic α-galactosidases, alkaline α-galactosidases and RFO synthases, among which AtAkαGal3 belongs to the alkaline α-galactosidase group. Some structures of acidic α-galactosidases have been reported, including α-galactosidase from Thermotoga maritima (Ren et al., 2018), chicken α-N-acetylgalactosaminidase (Garman et al., 2002), rice α-galactosidase (Fujimoto et al., 2003), human α-galactosidase (Garman & Garboczi, 2004) and Trichoderma reesei α-galactosidase (Golubev et al., 2004), which share low sequence identities of 32%, 29%, 27%, 25% and 24%, respectively, with AtAkαGal3, whereas other α-galactosidases from Nicotiana benthamiana (Kytidou et al., 2018), Bacteroides fragilis (Joint Center for Structure Genomics, unpublished work) and Saccharomyces cerevisiae (Fernández-Leiro et al., 2010) show no reasonable identity to AtAkαGal3.
3.5. The catalytic residues and binding sites for various natural substrates
A sequence alignment of acidic α-galactosidases from rice (Fujimoto et al., 2003), chicken (Garman et al., 2002), human (Garman & Garboczi, 2004), N. benthamiana (Kytidou et al., 2018), B. fragilis, a Gram-negative bacterium (Joint Center for Structure Genomics, unpublished work), S. cerevisiae (Fernández-Leiro et al., 2010) and T. reesei (Golubev et al., 2004) with known structures indicates the presence of two conserved motifs KXD and RxxxD, where X can be Y, F, L or I and x is any variety of amino acid. The aspartic acid residues in the KXD and RxxxD motifs serve as the and the catalytic acid (Fujimoto et al., 2003; Garman et al., 2002), respectively (Supplementary Fig. S3). A sequence comparison of AtAkαGal3 with acidic α-galactosidases further confirmed that Asp383 and Asp447 are presumably critical residues that serve as the and the catalytic acid/base, respectively. In addition, a sequence alignment of AtAkαGal3 with alkaline α-galactosidases from tomato (AF512549) and melon (AY114164), seed imbibition protein from Arabidopsis (AtAkαGal1) and RFO synthases from japonica rice (XP_015621501), cucumber (AAD02832), soybean (XP_006576826), yoshino cherry (PQQ02596) and clementine (XP_006444535) shows these two residues, Asp383 and Asp447 in AtAkαGal3, to be conserved in these motifs: 381KVD383, except for raffinose synthase from cucumber where the valine is replaced by isoleucine, and 443RxxDD447 in the Based on the alignment, Arg443 and Ala444 of the 443RASDD447 motif in AtAkαGal3 and most alkaline α-galactosidases are evidently conserved compared with those in RFO synthases (the RxxxD motifs contain the and catalytic acid; Fujimoto et al., 2003; Garman et al., 2002; Supplementary Fig. S4).
To confirm that these two residues, Asp383 and Asp447 in AtAkαGal3, play important roles in catalysis, we generated two single mutants D383A and D447A by replacing Asp383 and Asp447 with alanine by site-directed mutagenesis. The TLC results showed that the hydrolase activity of both mutants towards raffinose was abolished (data not shown).
3.6. Interactions of galactose at the binding site of wild-type AtAkαGal3 and the D383A mutant
Before the crystallization of wild-type AtAkαGal3, galactose was added to the protein solution via dialysis in order to obtain suitable crystals. Our first solved structure of wild-type AtAkαGal3 was thus in complex with galactose at a resolution of 2.17 Å. The galactose moiety, with a chair conformation, is located in the −1 subsite of the catalytic binding site (Fig. 1a). The structure of the AtAkαGal3–galactose complex shows that the OH group at C1, the anomeric C atom, of the galactose moiety is in a β-anomeric form and points towards the bottom side of the catalytic binding site, forming a direct hydrogen bond (distance 3.06 Å) to the Asp383 and to the OD1 group of the acid/base catalytic residue Asp447 through a water molecule, with distances of 2.66 and 2.87 Å, respectively. Moreover, the OD1 group of Asp447 makes a strong hydrogen bond (distance 2.54 Å) to the OH2 group of the galactose. The other hydroxyl groups of the galactose are also stabilized by numerous hydrogen-bond interactions with various surrounding side chains of the residues at the −1 binding site, such as Arg443, Lys381, Asp243, Asp244 and Trp307. Furthermore, three water molecules help to stabilize the galactose moiety through water-mediated hydrogen-bonding interactions with Trp78, Lys381, Arg443, Asp447 and Asp479.
The structure of the D383A mutant in complex with galactose was also determined at a resolution of 2.05 Å (Fig. 1b). Interestingly, superposition of the two galactose moieties at the catalytic binding sites of the wild-type AtAkαGal3–galactose and D383A–galactose complex structures shows that the galactose moiety in the D383A mutant exhibits a chair form with a slight distortion; the O1 hydroxyl group at the C1 carbon is in an α-anomeric form (Supplementary Fig. S5). This particular structural form thus places the O1 hydroxyl group of the galactose in a position nearer the OD2 group of Asp447 to form a direct hydrogen-bond interaction with a distance of 2.87 Å. This effect might be due to the loss of an interaction with the carboxyl group of Asp383, which is replaced by alanine with a short side chain. Except for O1 at the anomeric C atom, which exhibits a conformational alteration from a β-form to an α-form, the positions of the remaining C atoms and hydroxyl groups of galactose are conserved and make unchanged interactions around the catalytic site compared with the wild-type enzyme. In the D383A structure, two additional water molecules interacting with O1 and O6 of galactose help to stabilize the galactose moiety compared with the wild-type AtAkαGal3 structure. Using the Cremer–Pople calculator, the galactose molecules complexed in the D383A and D383D mutants show Cremer–Pople parameters φ, θ, Q of 88.829°, 24.021°, 0.713 and 123.014°, 6.867°, 0.646, respectively, which might indicate an induced fit of the substrate to push it towards the transition state.
The residues involved in product-binding interactions with the galactose moiety at the −1 subsite of AtAkαGal3 are similar to those in rice acidic α-galactosidase (Fujimoto et al., 2003), except for Trp78 in AtAkαGal3, which is structurally related to Trp164 of rice acidic α-galactosidase and to Trp65 of T. maritima α-galactosidase (Ren et al., 2018), which positions its side chain farther away from the galactose moiety. Trp164 of rice acidic α-galactosidase, which is located on the loop at the end of the β5 strand, directly interacts with the O1 hydroxyl group of the galactose at the −1 subsite and is a conserved residue among acidic α-galactosidases from plants, bacteria and yeast, suggesting that it plays a role in recognizing the galactose substrate moiety (Fujimoto et al., 2003). In contrast, Trp78 of AtAkαGal3, which is located in a short β-turn (residues 75–78) between two β-strands of the N-terminal domain, contributes part of the catalytic binding site together with the Except for water-mediated interactions with the O1 hydroxyl group of the galactose and the side chain of Asp447, Trp78 seems to make no direct interactions with the galactose moiety at the −1 subsite in AtAkαGal3, but rather to make a major interaction at the +1 subsite, which is discussed in the following section.
3.7. Interactions of galactinol in the binding site of the D383A mutant
The structure of D383A in complex with galactinol was determined at a resolution of 2.63 Å from D383A crystals by a combination of co-crystallization and soaking with galactinol as described in Section 2. The 5R-OH group (O6′) of the myo-inositol moiety of galactinol at the +1 subsite is stabilized by the NE1 group of Trp78 together with OD1 of Asp446 by hydrogen bonds. In addition, 4S-OH is stabilized by the NZ group of Lys75 by hydrogen bonds (Fig. 2a). The conformation of the galactosyl moiety of galactinol differs from that of galactose bound to wild-type AtAkαGal3 and the D383A mutant (Fig. 2b), but the galactosyl moiety remains stabilized at the −1 subsite by mostly the same residues, including Asp447, Arg443, Lys381 and Asp244, but not Asp243 and Trp307, making interactions with galactose. Two water molecules are located in the catalytic binding site, one of which interacts with the galactosyl moiety and the other with the myo-inositol moiety and the side chain of Asp346, to help stabilize the galactinol. This Asp346 residue seems to play a major role in galactinol binding at the +1 subsite (Fig. 2a).
3.8. Interactions of raffinose in the binding site of the D383A mutant
The ). The NE1 group of Trp78 and OD1 of Asp446 form hydrogen bonds to O4 of the hydroxyl group with distances of 3.08 and 3.47 Å, respectively. NZ of the Lys75 side chain stabilizes O3 of the hydroxyl group through a hydrogen bond with a distance of 3.54 Å. Tyr449 stabilizes the glucose at the +1 subsite and the fructose moiety at the +2 subsite of raffinose via hydrogen bonds from its side-chain OH group to the O3 hydroxyl groups of both galactose and fructose with distances of 3.52 Å (Fig. 3). One water molecule that forms a hydrogen bond (distance of 2.70 Å) to the side chain of Asp447 seems to be involved in stabilization of the fructose moiety through weak hydrogen bonding. The interactions of the residues at the −1 subsite with the galactosyl moiety of raffinose are similar to those with the galactosyl moiety of galactinol bound in the D383A complex.
of D383A in complex with raffinose was determined at a resolution of 2.47 Å. Lys75, Tyr78, Asp446 and Tyr449 help to stabilize the glucose moiety of the raffinose substrate at the +1 subsite (Fig. 33.9. Interactions of the galactose and sucrose products of hydrolyzed raffinose in the binding site
During the crystallization of the D383A–raffinose complex, co-crystallization of the D383A mutant and the raffinose substrate was occasionally set up for over two months. Unexpectedly, we observed both the galactose and sucrose products in the catalytic binding site and an additional cavity in the a). The D383A mutant seems to be capable of gradually hydrolyzing the raffinose substrate into galactose and sucrose products that remain in the catalytic site. The D383A mutant in complex with hydrolyzed raffinose reveals a secondary product-binding site for sucrose; this site is stabilized and formed in part by a unique short loop (residues 329–352) composed of three short helices at the surface of the catalytic pocket, which equip Thr342 and Asp346 to interact with the glucose and fructose moieties of sucrose through hydrogen bonds (Fig. 4b). This distinct extra loop (residues 329–352) in AtAkαGal3 can also be found in raffinose/stachyose synthases based on sequence alignment (Fig. 4c), but not in acid α-galactosidase-related proteins. Lys75 from a β-turn (residues 75–78) in the N-terminal domain also helps to stabilize the glucose and fructose moieties of sucrose by hydrogen bonds with a distance of 2.80 Å. In addition, Trp78 and Asp447 help to stabilize both the sucrose product and the galactose moiety at the −1 subsite through a water molecule (Fig. 4b).
after a sufficient reaction time (Fig. 43.10. Interactions of stachyose in the binding site of the D383A mutant
The αGal3 (Fig. 5). The galactose at the +1 subsite of stachyose is stabilized by hydrogen bonds to both NE of the Lys75 side chain and NH1 of the Trp78 side chain with distances of ∼3 Å. Even though Trp77 makes no direct interaction with galactose at the +1 subsite, it helps to stabilize galactose through a water molecule. The extensive water network through the side chain of Arg451 helps to stabilize the glucose moiety at the +2 subsite through a water molecule. In addition, water-mediated hydrogen-bond interactions greatly stabilize both the glucose and fructose moieties at the +2 and +3 subsites.
of the complex of D383A with stachyose determined at a resolution of 2.0 Å further reveals the complete subsites (–1, +1, +2 and +3) of AtAk3.11. Structural and sequence comparison explain the dual function of AtAkαGal3
The GH36 family contains enzymes, including RFO synthases and alkaline α-galactosidases, that generally exhibit transglycosylation and hydrolase activities. The first reported α-galactosidase structure was that of T. maritima α-galactosidase (Ren et al., 2018). Structural alignment of T. maritima α-galactosidase with the related GH27 family enzymes shows that T. maritima α-galactosidase and rice acidic α-galactosidase (Fujimoto et al., 2003) share the greatest structural similarity. A structural comparison of the (α/β)8-barrel domain of our alkaline α-galactosidase AtAkαGal3 (residues 201–531) with those of rice acidic α-galactosidase (residues 6–271) and T. maritima α-galactosidase (residues 165–481) shows poor structural similarity, with r.m.s.d.s of 3.74 and 11.37 Å, respectively. The structure of wild-type AtAkαGal3 shows that the catalytic residues Asp383 and Asp447 are separated by approximately 6.7 Å, which is also generally found for the double-replacement retaining mechanism in members of clan D of the GH27 enzymes, with an average distance of 6.5 Å, and in T. maritima α-galactosidase, with a distance of 6.3 Å (Comfort et al., 2007). A structural alignment of the (α/β)8-barrel domains of AtAkαGal3, rice acidic α-galactosidase and T. maritima α-galactosidase based on the AtAkαGal3 structure shows that most catalytic binding residues at the −1 subsite are conserved. In addition, Trp78 of AtAkαGal3 is not part of the (α/β)8-barrel of the like the corresponding Trp of rice acidic α-galactosidase, but is located in the N-terminal domain like Trp65 of T. maritima α-galactosidase. Moreover, Lys75 and Trp78 in the N-terminal domain share hydrogen-bond interactions that stabilize the second galactose moiety in the +1 subsite (Table 2).
|
A sequence alignment (ESPript3.0; Robert & Gouet, 2014) of plant alkaline α-galactosidases with plant raffinose and stachyose synthases shows that the FRSK75xW77W78 region (referring to AtAkαGal3) located at the β-turn between two β-strands of the N-terminal domain is conserved among these families (Fig. 6). Furthermore, two residues, Arg451 and Tyr449 of AtAkαGal3, which are also found in raffinose and stachyose synthases, help to stabilize the binding of the galactose moiety at the +2 subsite through water molecules and directed interaction, respectively. These observations from both structural and sequence alignments lead to the possibility of a dual function of the alkaline α-galactosidase AtAkαGal3, as the identified residues involved in product and substrate binding at the −1, +1 and +2 subsites of the alkaline α-galactosidase AtAkαGal3 are identical to the functional residues in both the GH27 and GH36 families as well as the raffinose and stachyose synthase families. However, the structure of D383A complexed with stachyose shows that water molecules seem to be important to help to stabilize the fructose moiety at the +3 subsite (Table 2).
3.12. N-terminal domain and catalytic site of AtAkαGal3
Our work confirms that the N-terminal domain of the alkaline α-galactosidase AtAkαGal3 is important for protein function because it contributes part of the core (α/β)8-barrel catalytic site for substrate binding. The C-terminal domain helps to stabilize this structural architecture through interactions with both the N-terminal domain and the (α/β)8-barrel domain. Sequence alignment indicates that the RFO synthases might also use their N-terminal domain for a similar function. The stachyose bound in the catalytic pocket of AtAkαGal3 shows that the depth and dimensions of the binding pocket can accommodate four sugar moieties, whereas the catalytic binding sites of the acidic α-galactosidases rice α-galactosidase and T. maritima α-galactosidase are relatively shallow and are likely to accommodate only one galactose. Moreover, the extra loop (residues 329–352) that only exists in the alkaline α-galactosidase AtAkαGal3 provides two key residues Asp346 and Thr342 to help stabilize the fructose and glucose moieties of the sucrose product in the +2 and +3 subsites, respectively (Figs. 4 and 7).
Our D383A mutant can gradually hydrolyze raffinose into galactose and sucrose. A structural superimposition of D383A in complex with raffinose, with stachyose and with hydrolyzed raffinose in the catalytic binding pocket shows that the wider binding-site pocket of AtAkαGal3 provides a secondary product-binding site to stabilize the sucrose after raffinose has been hydrolyzed into galactose and glucose. This product-binding site contains Thr342 and Asp346 on a short loop (residues 329–352) that interact with the glucose and fructose moieties of sucrose through hydrogen bonds. Lys75 in the N-terminal domain also helps to stabilize the glucose moiety of sucrose by hydrogen bonding (Fig. 8).
3.13. Dual function of AtAkαGal3
We compared the hydrolase activities of the alkaline α-galactosidase AtAkαGal3 and rice (O. sativa L. var. Nipponbare) raffinose synthase towards only the natural substrate raffinose, without galactinol, which serves as a galactose donor for raffinose or stachyose synthase activity. The TLC assay showed that AtAkαGal3 exhibits a greater hydrolase activity towards the raffinose substrate considering the TLC spot intensities of the galactose and sucrose products. AtAkαGal3 can also produce a stachyose product without galactinol as a galactose donor. This observation indicates that AtAkαGal3 seems to act with the transferase mechanism observed for fructosyl and xyloglucan endotransferases. Our crystal structures of wild-type AtAkαGal3 and its D383A mutant in complex with galactose and with hydrolyzed raffinose, i.e. galactose and sucrose, reveal one water molecule at the conserved position near the anomeric C atom and Asp447. This water is replaced by O4 of galactose in the +1 subsite of raffinose and stachyose in the complexes of D383A with raffinose and stachyose. The only difference between the hydrolase and transferase reactions is the acceptor type after the removal of the terminal glycosyl moiety. The acceptor for a hydrolase is a water molecule, whereas a transferase uses a saccharide or an oligosaccharide, such as sucrose, raffinose or stachyose, as the acceptor (Henrissat et al., 2001; Carmi et al., 2003; Chuankhayan et al., 2010). In AtAkαGal3, the conserved water molecule identified above might thus serve as an acceptor in the hydrolase step, releasing galactose as a product, while sucrose and raffinose act as an acceptor for the transferase mechanism, releasing raffinose and stachyose as products. In addition, the unique extra loop (residues 329–352) in AtAkαGal3, which contributes to the depth enhancement of the catalytic pocket, presumably supports the transferase mechanism to help to stabilize the oligomer products.
3.14. Catalytic and synthetic mechanism
The structure of D383A in complex with hydrolyzed raffinose reveals the products galactose at the −1 subsite and sucrose at the secondary binding site. Unexpectedly, after raffinose has been hydrolyzed between galactose and glucose at subsites −1 and +1, respectively, the sucrose shifts and rotates by 180°. The glucose moiety of sucrose in the +1 subsite is thus replaced by a fructose moiety and is bound in a different orientation and position to the related moieties of raffinose and stachyose in other D383A–substrate complexes (Fig. 8). We specify this newly discovered binding position as the secondary product-binding site. A 180° rotation of the sucrose product after the hydrolysis of raffinose is possible because the catalytic pocket is wide, which provides sufficient space for flipping and binding (Fig. 9). After rotation, Thr342 and Asp346 help to stabilize the glucose and fructose of the sucrose product, respectively. Furthermore, Lys75, which helps to stabilize the sugar moiety in the +1 subsite, also makes a strong interaction with the glucose of the sucrose product in the +3 subsite, indicating that the conserved residues Lys75 and Trp78 in the WW box region are important for stabilization of both the substrate and the product in the +1 subsite, according to the proposed mechanism in Fig. 10. Lys381 might be the key during the catalytic process.
In summary, the alkaline α-galactosidase AtAkαGal3 from A. thaliana exhibits a dual function: it not only catalyzes the hydrolysis of α-D-galactose from galacto-oligosaccharides under alkaline conditions but can also synthesize stachyose using raffinose, instead of galactinol, as the galactose donor. Structural analyses of AtAkαGal3 and its D383A mutant in complex with various substrates and products, including galactose, galactinol, raffinose, stachyose and sucrose, elucidate four complete subsites (–1 to +3) and a new secondary product-binding site. The AtAkαGal3 structure in this study is the first representative structure of an alkaline α-galactosidase and may also serve a structural model for the raffinose family of RFO synthases.
Supporting information
Supplementary Figures. DOI: https://doi.org/10.1107/S2059798323000037/ji5027sup1.pdf
Acknowledgements
We are indebted to the staff of beamlines TPS 05A and TLS 15A at the National Synchrotron Radiation Research Center (NSRRC) in Taiwan, Eiki Yamashita at BL44XU and the staff of the Taiwan beamline BL12B2 at SPring-8 in Japan for technical assistance. We are grateful for travel support from the International Collaborative Research Program of the Institute for Protein Research, Osaka University. We are also grateful to Ming-Hong Lai for the preliminary crystallization conditions of AtAkαGal3 and to Chien-Ming Chiang for discussion. We thank Shinya Fushinobu for the online Cremer–Pople parameter calculator.
Funding information
This work was supported by National Science and Technology Council (NSTC) grants 105-2311-B-213-001-MY3, 107-2923-B-213-001-MY3, 108-2311-B-213-001-MY3 and 111-2311-B-213-001 and NSRRC grants to C.-J. Chen.
References
Calhoun, D. H., Bishop, D. F., Bernstein, H. S., Quinn, M., Hantzopoulos, P. & Desnick, R. J. (1985). Proc. Natl Acad. Sci. USA, 82, 7364–7368. CrossRef CAS PubMed Web of Science Google Scholar
Carmi, N., Zhang, G., Petreikov, M., Gao, Z., Eyal, Y., Granot, D. & Schaffer, A. A. (2003). Plant J. 33, 97–106. Web of Science CrossRef PubMed CAS Google Scholar
Chang, S., Puryear, J. & Cairney, J. (1993). Plant Mol. Biol. Rep. 11, 113–116. CrossRef CAS Google Scholar
Chen, C.-D., Huang, Y.-C., Chiang, H.-L., Hsieh, Y.-C., Guan, H.-H., Chuankhayan, P. & Chen, C.-J. (2014). Acta Cryst. D70, 2331–2343. Web of Science CrossRef IUCr Journals Google Scholar
Chuankhayan, P., Hsieh, C.-Y., Huang, Y.-C., Hsieh, Y.-Y., Guan, H.-H., Hsieh, Y.-C., Tien, Y.-C., Chen, C.-D., Chiang, C.-M. & Chen, C.-J. (2010). J. Biol. Chem. 285, 23251–23264. Web of Science CrossRef CAS PubMed Google Scholar
Comfort, D. A., Bobrov, K. S., Ivanen, D. R., Shabalin, K. A., Harris, J. M., Kulminskaya, A. A., Brumer, H. & Kelly, R. M. (2007). Biochemistry, 46, 3319–3330. Web of Science CrossRef PubMed CAS Google Scholar
Corchete, M. P. & Guerra, H. (1987). Phytochemistry, 26, 927–932. CrossRef CAS Web of Science Google Scholar
Dey, P. M. & Pridham, J. B. (1972). Adv. Enzymol. Relat. Areas Mol. Biol. 36, 91–130. CAS PubMed Web of Science Google Scholar
Doublié, S. (2007). Methods Mol. Biol. 363, 91–108. PubMed Google Scholar
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). Acta Cryst. D66, 486–501. Web of Science CrossRef CAS IUCr Journals Google Scholar
Farag, S. (1978). J. Am. Soc. Sugar Beet Technol. 20, 251–254. CrossRef Google Scholar
Fernández-Leiro, R., Pereira-Rodríguez, A., Cerdán, M. E., Becerra, M. & Sanz-Aparicio, J. (2010). J. Biol. Chem. 285, 28020–28033. Web of Science PubMed Google Scholar
Fujimoto, Z., Kaneko, S., Momma, M., Kobayashi, H. & Mizuno, H. (2003). J. Biol. Chem. 278, 20313–20318. Web of Science CrossRef PubMed CAS Google Scholar
Gao, Z. & Schaffer, A. A. (1999). Plant Physiol. 119, 979–988. Web of Science CrossRef PubMed CAS Google Scholar
Garman, S. C. & Garboczi, D. N. (2004). J. Mol. Biol. 337, 319–335. Web of Science CrossRef PubMed CAS Google Scholar
Garman, S. C., Hannick, L., Zhu, A. & Garboczi, D. N. (2002). Structure, 10, 425–434. Web of Science CrossRef PubMed CAS Google Scholar
Gaudreault, P. R. & Webb, J. A. (1986). Plant Sci. 45, 71–75. CrossRef CAS Web of Science Google Scholar
Golubev, A. M., Nagem, R. A. P., Brandão Neto, J. R., Neustroev, K. N., Eneyskaya, E. V., Kulminskaya, A. A., Shabalin, K. A., Savel'ev, A. N. & Polikarpov, I. (2004). J. Mol. Biol. 339, 413–422. Web of Science CrossRef PubMed CAS Google Scholar
Guimarães, V. M., de Rezende, S. T., Moreira, M. A., de Barros, E. G. & Felix, C. R. (2001). Phytochemistry, 58, 67–73. Web of Science PubMed Google Scholar
Hara, M., Tokunaga, K. & Kuboi, T. (2008). Plant Biotechnol. 25, 497–501. Web of Science CrossRef CAS Google Scholar
Henrissat, B. (1991). Biochem. J. 280, 309–316. CrossRef PubMed CAS Web of Science Google Scholar
Henrissat, B., Coutinho, P. M. & Davies, G. J. (2001). Plant Mol. Biol. 47, 55–72. Web of Science CrossRef PubMed CAS Google Scholar
Henrissat, B. & Romeu, A. (1995). Biochem. J. 311, 350–351. CrossRef CAS PubMed Web of Science Google Scholar
Herman, E. M. & Shannon, L. M. (1985). Plant Physiol. 77, 886–890. CrossRef PubMed CAS Web of Science Google Scholar
Kabsch, W. (2010). Acta Cryst. D66, 125–132. Web of Science CrossRef CAS IUCr Journals Google Scholar
Kang, H.-C. & Lee, S.-H. (2001). Phytochemistry, 58, 213–219. Web of Science CrossRef PubMed CAS Google Scholar
Keller, F. & Pharr, D. M. (1996). Photoassimilate Distribution in Plants and Crops: Source–Sink Relationships, edited by E. Zamski & A. A. Scharffer, pp. 157–184. New York: Marcel Dekker. Google Scholar
Kytidou, K., Beekwilder, J., Artola, M., van Meel, E., Wilbers, R. H. P., Moolenaar, G. F., Goosen, N., Ferraz, M. J., Katzy, R., Voskamp, P., Florea, B. I., Hokke, C. H., Overkleeft, H. S., Schots, A., Bosch, D., Pannu, N. & Aerts, J. M. F. G. (2018). J. Biol. Chem. 293, 10042–10058. Web of Science CrossRef CAS PubMed Google Scholar
Langer, G., Cohen, S. X., Lamzin, V. S. & Perrakis, A. (2008). Nat. Protoc. 3, 1171–1179. Web of Science CrossRef PubMed CAS Google Scholar
Lee, D. W. & Collins, T. M. (2001). Int. J. Plant Sci. 162, 1141–1153. Web of Science CrossRef CAS Google Scholar
Lee, R.-H., Hsu, J.-H., Huang, H.-J., Lo, S.-F. & Chen, S.-C. G. (2009). New Phytol. 184, 596–606. Web of Science CrossRef PubMed CAS Google Scholar
Lee, R.-H., Lin, M.-C. & Chen, S.-C. G. (2004). Plant Mol. Biol. 55, 281–295. Web of Science CrossRef PubMed CAS Google Scholar
Li, S., Li, T., Kim, W. D., Kitaoka, M., Yoshida, S., Nakajima, M. & Kobayashi, H. (2007). Biotechnol. Lett. 29, 635–640. Web of Science CrossRef PubMed CAS Google Scholar
Moriarty, N. W., Grosse-Kunstleve, R. W. & Adams, P. D. (2009). Acta Cryst. D65, 1074–1080. Web of Science CrossRef CAS IUCr Journals Google Scholar
Murshudov, G. N., Skubák, P., Lebedev, A. A., Pannu, N. S., Steiner, R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011). Acta Cryst. D67, 355–367. Web of Science CrossRef CAS IUCr Journals Google Scholar
Nunan, K. J., Davies, C., Robinson, S. P. & Fincher, G. B. (2001). Planta, 214, 257–264. Web of Science CrossRef PubMed CAS Google Scholar
Otwinowski, Z. & Minor, W. (1997). Methods Enzymol. 276, 307–326. CrossRef CAS PubMed Web of Science Google Scholar
Ren, W., Pengelly, R., Farren-Dai, M., Shamsi Kazem Abadi, S., Oehler, V., Akintola, O., Draper, J., Meanwell, M., Chakladar, S., Świderek, K., Moliner, V., Britton, R., Gloster, T. M. & Bennet, A. J. (2018). Nat. Commun. 9, 3243. Web of Science CrossRef PubMed Google Scholar
Robert, X. & Gouet, P. (2014). Nucleic Acids Res. 42, W320–W324. Web of Science CrossRef CAS PubMed Google Scholar
Salentin, S., Schreiber, S., Haupt, V. J., Adasme, M. F. & Schroeder, M. (2015). Nucleic Acids Res. 43, W443–W447. Web of Science CrossRef CAS PubMed Google Scholar
Shibuya, H., Kobayashi, H., Park, G. G., Komatsu, Y., Sato, T., Kaneko, R., Nagasaki, H., Yoshida, S., Kasamo, K. & Kusakabe, I. (1995). Biosci. Biotechnol. Biochem. 59, 2333–2335. CrossRef CAS PubMed Web of Science Google Scholar
Terwilliger, T. C. & Berendzen, J. (1999). Acta Cryst. D55, 849–861. Web of Science CrossRef CAS IUCr Journals Google Scholar
Vagin, A. & Teplyakov, A. (2010). Acta Cryst. D66, 22–25. Web of Science CrossRef CAS IUCr Journals Google Scholar
Weidemann, F., Breunig, F., Beer, M., Sandstede, J., Turschner, O., Voelker, W., Ertl, G., Knoll, A., Wanner, C. & Strotmann, J. M. (2003). Circulation, 108, 1299–1301. Web of Science CrossRef PubMed CAS Google Scholar
Williams, C. J., Headd, J. J., Moriarty, N. W., Prisant, M. G., Videau, L. L., Deis, L. N., Verma, V., Keedy, D. A., Hintze, B. J., Chen, V. B., Jain, S., Lewis, S. M., Arendall, W. B. III, Snoeyink, J., Adams, P. D., Lovell, S. C., Richardson, J. S. & Richardson, D. C. (2018). Protein Sci. 27, 293–315. Web of Science CrossRef CAS PubMed Google Scholar
Winn, M. D., Ballard, C. C., Cowtan, K. D., Dodson, E. J., Emsley, P., Evans, P. R., Keegan, R. M., Krissinel, E. B., Leslie, A. G. W., McCoy, A., McNicholas, S. J., Murshudov, G. N., Pannu, N. S., Potterton, E. A., Powell, H. R., Read, R. J., Vagin, A. & Wilson, K. S. (2011). Acta Cryst. D67, 235–242. Web of Science CrossRef CAS IUCr Journals Google Scholar
Zhao, T. Y., Corum, J. W. III, Mullen, J., Meeley, R. B., Helentjaris, T., Martin, D. & Downie, B. (2006). Seed Sci. Res. 16, 107–121. Web of Science CrossRef CAS Google Scholar
This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.