Crystal structure of Arabidopsis thaliana cytokinin dehydrogenase

Cytokinins are a class of plant hormones regulating cell division and a variety of developmental events.1 Cytokinin homeostasis is important in many aspects of plant growth, and their control has economic significance in agriculture.2 These hormones are chemically adenine derivatives with a variable N6 substituent. Most of their catabolism is regulated by cytokinin dehydrogenase (CKX, EC 1.5.99.12),3 which is involved in oxidative cleavage of cytokinins to the adenine and the substituent. CKX has originally been thought to be an oxidase transferring electrons to molecular oxygen.4 However, the enzyme has since been found to use other electron acceptors and therefore reclassified as a dehydrogenase.5, 6 It has been reported that CKX is posttranslationally modified by glycosylation7 and uses a covalently linked FAD as a cofactor.8 Since first discovered in Zea mays,7 CKX genes have been identified in many plants including rice and Arabidopsis thaliana, which possesses CKX homologues (AtCKX1–AtCKX7).3 So far, the three-dimensional structure of only Z. mays CKX (ZmCKX1) has been determined.9 The crystal structures of ZmCKX1 have been solved in the native state, and in complex with reaction products and a slowly reacting substrate. The structures revealed four glycosylated asparagine residues and a histidine residue covalently linked to FAD. Combined with the structural information, recent biochemical analyses of ZmCKX1 concluded that the final products of the reaction, adenine and a side chain aldehyde, are formed by nonenzymatic hydrolytic cleavage of cytokinin imine products resulting directly from CKX catalysis.10 Here, we report the crystal structure of AtCKX7 (gene locus At5g21482.1, UniProt code Q9FUJ1). The CKX gene was cloned and selenomethionyl protein was purified following the standard Center for Eukaryotic Structural Genomics (CESG) pipeline protocol for cloning,11 protein expression,12 protein purification,13 and overall information management.14 Yellow crystals were grown by the sitting-drop method from 10 mg/mL protein solution in buffer (50 mM NaCl, 3 mM NaN3, 0.3 mM TCEP, 5 mM BisTris pH 7.0) mixed with an equal amount of reservoir solution containing 24% (w/v) MEPEG 5K, 0.05M magnesium sulfate, 0.1M HEPPS pH 8.5 at 4°C. The crystals were soaked in reservoir solutions supplemented with increasing amounts of ethylene glycol up to 20% (v/v) and were flash-frozen in liquid nitrogen. Diffraction data were collected at Southeast Regional Collaborative Access Team (SER-CAT) 22-ID beamline at the Advanced Photon Source (APS), Argonne National Laboratory at 100 K. The diffraction images were processed with HKL2000.15 The selenium substructure of the crystal was determined by using SHELXD16 and HySS from PHENIX,17, 18 and the selenium positions were used for multiple wavelength anomalous diffraction phasing in AUTOSHARP.19 The initial model was built by the automatic tracing procedure of ARP/wARP,20 and the structure was completed using alternate cycles of manual building in COOT21 and refinement in REFMAC5.22 The stereochemical quality of the final model was assessed using MolProbity.23 The final coordinates were deposited in the RCSB Protein Data Bank24 with accession number 2EXR. The crystal structure of AtCKX7 was determined to a resolution of 1.7 Å using multiple wavelength anomalous diffraction. Data collection, phasing and refinement statistics are summarized in Table I. The asymmetric unit of the crystal contains a single AtCKX7 monomer (residues 34–524), one FAD molecule and 633 water molecules. The 33 N-terminal residues were not included in the model due to insufficient electron density. In the ZmCKX1 structure, N-terminal residues are also missing. It has been suggested that these residues are cleaved upon processing newly synthesized protein in cells.9 However, the mass spectrometric analyses of AtCKX7 sample used for crystallization indicated a full-length protein (data not shown), suggesting that the N-terminal region is disordered in the crystal lattice, rather than removed prior to crystallization. This is consistent with the previous analysis that unlike other CKX proteins AtCKX7 has no signal peptide at its N-terminal end.3 There is no evidence of glycosylation in the AtCKX7 structure, which is expected for recombinant proteins expressed from bacteria. AtCKX7, like ZmCKX1, displays two-domain topology found in the members of the vanillyl-alcohol oxidase family [Fig. 1(A)].9, 25 The FAD-binding domain [residues 34–237 and 480–514, cyan in Fig. 1(A)] forms a binding site for the cofactor, and comprises one mixed β-sheet, one antiparallel β-sheet and flanking α-helices. The second domain [residue 238–479, red in Fig. 1(A)] consists of one antiparallel β-sheet and surrounding α-helices. In the structures of ZmCKX1 complexes, the corresponding domain is involved in substrate binding.9 Although the AtCKX7 structure has no bound substrate, we will refer this domain as the "substrate-binding domain" because the AtCKX7 and ZmCKX1 structures share high similarity (see below). Comparison of the two CKX structures using DaliLite results in a root mean square deviation of 1.9 Å, a Z-score of 51.4 and 43% sequence identity for 467 (out of 491 residues) structurally equivalent Cα atom pairs.26 Three of the four asparagines residues which are glycosylated in the ZmCKX1 structure are not conserved in AtCKX7. Interestingly, structurally divergent regions between the two structures include residues in the vicinity of the glycosylation sites in the ZmCKX1 structure. Crystal structure of AtCKX7. (A) Ribbon diagram of AtCKX7. The FAD-binding and the substrate-binding domains are shown in cyan and red, respectively. The FAD molecule is represented as a stick model in yellow. (B) Covalent bond between His96 and the isoalloxazine ring of the FAD molecule. The FAD is shown in yellow and the His residue is colored by atom types (carbon: green; oxygen: red; nitrogen: blue). The 2Fobs − Fcalc map is contoured at 1.5σ. The crystal structure of AtCKX7 revealed a bound cofactor, FAD (yellow in Fig. 1). The ADP portion of the cofactor is embedded in the FAD-binding domain whereas its isoalloxazine ring is located at the interface of the two domains. The C8M atom of the flavin ring is covalently linked to the ND1 atom of the conserved His96 with a distance of 1.4 Å [Fig. 1 (B)]. The details of the FAD binding in the AtCKX7 structure are essentially the same as in the ZmCKX1 structure. Residues whose side chains are interacting with the cofactor are conserved in both CKX proteins. Although no substrate was bound in the AtCKX7 structure, it is possible to infer a role of the substrate-binding domain based on the comparison with the ZmCKX1 structure. The substrate-binding cavity in this domain of the ZmCKX1 structures is structurally conserved in the AtCKX7 structure. Most of residues participating in substrate binding in ZmCKX1 are also conserved in AtCKX7. They include the active site carboxylate–carboxylate pair (Asp162 and Glu275 in AtCKX7), which has been suggested to be a key element in catalysis.9 Therefore, it is most likely that this domain is in fact involved in substrate binding, and that AtCKX7 and ZmCKX1 share a conserved catalytic mechanism. We thank the members of the CESG. Data were collected at SER-CAT 22-ID beamline at the APS, Argonne National Laboratory. Supporting institutions may be found at www.ser-cat.org/members.html.