Noel CHI (letter) - Nature

13 downloads 0 Views 629KB Size Report
Rhizobium nodulation genes, and antimicrobial phy- toalexins. The 1.85 Å resolution crystal structure of alfalfa. CHI in complex with (2S)-naringenin reveals a ...
© 2000 Nature America Inc. • http://structbio.nature.com

letters Structure and mechanism of the evolutionarily unique plant enzyme chalcone isomerase Joseph M. Jez1, Marianne E. Bowman1, Richard A. Dixon2 and Joseph P. Noel1

© 2000 Nature America Inc. • http://structbio.nature.com

1Structural Biology Laboratory, The Salk Institute for Biological Studies, 10010 N. Torrey Pines Road, La Jolla, California 92037, USA. 2Plant Biology Division, Samuel Roberts Noble Foundation, P.O. Box 2180, Ardmore, Oklahoma 73402, USA.

Chalcone isomerase (CHI) catalyzes the intramolecular cyclization of chalcone synthesized by chalcone synthase (CHS) into (2S)-naringenin, an essential compound in the biosynthesis of anthocyanin pigments, inducers of Rhizobium nodulation genes, and antimicrobial phytoalexins. The 1.85 Å resolution crystal structure of alfalfa CHI in complex with (2S)-naringenin reveals a novel openfaced β-sandwich fold. Currently, proteins with homologous primary sequences are found only in higher plants. The topology of the active site cleft defines the stereochemistry of the cyclization reaction. The structure and mutational analysis suggest a mechanism in which shape complementarity of the binding cleft locks the substrate into a constrained conformation that allows the reaction to proceed with a second-order rate constant approaching the diffusion controlled limit. This structure raises questions about the evolutionary history of this structurally unique plant enzyme. Plants use flavonoids for protection against overexposure to UV light, as floral pigments for attraction of pollinators, as inducers of Rhizobium nodulation genes, and as antimicrobial phytoalexins1–4. Many flavonoids also exhibit medicinal properties and are common constituents in human diets4,5. As one of the best-characterized natural product pathways in plants, the flavonoid biosynthetic enzymes are attractive targets for metabolic engineering to provide enhanced feed crops, food sources, and medicinal agents6–9. One strategy for the generation of novel specificity in flavonoid biosynthesis uses protein engineering methods specifically targeting the active site residues of these biosynthetic enzymes. This approach requires detailed structural knowledge of enzymes within the targeted pathway6,10. In flavonoid biosynthetic pathways, chalcone isomerase (CHI, E.C. 5.5.1.6) catalyzes the cyclization of chalcone (4,2′,4′,6′-tetrahydroxychalcone) and 6′-deoxychalcone (4,2′,4′-trihydroxychalcone), both of which are synthesized by the upstream enzyme chalcone synthase (CHS, E.C. 2.3.1.74), into (2S)-naringenin (5,7,4′-trihydroxyflavanone) and (2S)-5deoxyflavanone (7,4′-dihydroxyflavanone), respectively (Fig. 1a). Since both chalcone and 6′-deoxychalcone spontaneously cyclize in solution to give enantiomeric mixtures, CHI guarantees formation of the biologically active (S)-isomer. (2S)-Naringenin is the precursor of anthocyanin pigments, and mutations in the gene encoding CHI are linked to changes in floral pigmentation11. (2S)-Naringenin and other flavonoids also activate bacterial transcription regulators governing expression of Rhizobium genes involved in root nodulation12. 786

Mechanistically, CHI catalyzes the cyclization of chalcone with an apparent 100,000:1 preference for synthesis of the S-isomer over the R-isomer. The second-order rate constant (kcat / Km) for conversion of 6′-deoxychalcone by CHI approaches the diffusion controlled limit, with an enzyme catalyzed rate that exceeds the spontaneous conversion rate by 107-fold (ref. 13). Comparison of the spontaneous and enzyme catalyzed reactions, combined with structural knowledge, provides insight into how an enzyme accelerates the rate of an intramolecular chemical reaction. We determined the 2.5 Å crystal structure of CHI from Medicago sativa (alfalfa) by multiple isomorphous replacement with anomalous scattering (MIRAS) and the 1.85 Å resolution structure of CHI in complex with (2S)-naringenin by difference Fourier analysis. These structures provide a molecular understanding of how CHI recognizes substrates and catalyzes the stereospecific cyclization of chalcones. Overall structure CHI is a functional monomer of ∼220 residues and has been isolated from a variety of higher plants13,14. Expression of alfalfa CHI15 in Escherichia coli yielded active enzyme that was purified and crystallized. The overall structure of CHI resembles an upside-down bouquet that adopts an open-faced βsandwich fold (Fig. 1b,c). A large β-sheet (β3a–β3f) and a layer of α-helices (α1–α7) comprise the core structure with three short β-strands (β1a, β1b, β2) on the opposite side of the large β-sheet. A search of the Protein Data Bank using DALI16 revealed no other structurally homologous folds. In addition, a PSI-BLAST17 search of sequence data bases showed that CHIlike sequences are currently found only in higher plants and that these sequences display no detectable homology with other proteins. These results imply that the CHI three-dimensional fold and enzymatic activity are unique to the plant kingdom. Amino acid sequence comparison of CHIs from a variety of angiosperms reveals high homology (49–82% amino acid sequence identity), with regions of conservation spread uniformly throughout the primary structure (Fig. 1d). The residues spanning β3a, β3b, α4, and α6 in the three-dimensional structure are conserved among CHIs from different species. Notably, these structural elements form the active site on the protein surface. Active site and reaction stereoselectivity The location of (2S)-naringenin in the CHI structure defines the active site (Fig. 2). Although a commercially obtained mixture of (2S)-naringenin and (2R)-naringenin was used for cocrystallization, only the (2S)-isomer bound in the CHI active site. The position of the (2S)-naringenin binding cleft is consistent with inactivation studies suggesting that a cysteine residue (Cys 114 in alfalfa CHI) is proximal to the active site 18. In the CHI structure, Cys 114 is near the binding cleft but does not directly contact (2S)-naringenin. The active site cleft is largely apolar and consists of residues from β3a (Arg 36, Gly 37, Leu 38), β3b (Phe 47, Thr 48, Ile 50), α4 (Tyr 106, Lys 109, Val 110, Asn 113), and α6 (Thr 190, Met 191) (Fig. 2b). The apolar methylene carbons of Arg 36 are positioned by a restraining charge–charge interaction from the δ-guanido group to Glu 200. In addition, the methylene carbons of Lys 109 are fixed by a charge–charge interaction between the amino group and Glu 112. Except for Thr 190 and Met 191, the residues contacting (2S)-naringenin are identical among CHIs from different plants (Fig. 1d). Although van der Waals connature structural biology • volume 7 number 9 • september 2000

© 2000 Nature America Inc. • http://structbio.nature.com

letters a

b

© 2000 Nature America Inc. • http://structbio.nature.com

c

d

Fig. 1 Reaction and structure of CHI. a, Overall reaction catalyzed by CHI involves a Michael-type nucleophilic attack of the 2′-hydroxyl on the α,β-unsaturated double bond. The numbering systems for chalcones (left) and flavanones (right) are shown. b, Schematic ribbon diagram of the overall structure of CHI. The N-terminus and C-terminus are labeled, as are the β-strands (blue) and α-helices (gold) of the structure. The position of (2S)naringenin (aqua) is also indicated. c, Stereo view of the Cα backbone of CHI. This orientation is the same as in (a). Every 10th residue is numbered. The position of (2S)-naringenin (aqua) is also shown. d, Primary and secondary structure of CHI from Medicago sativa (alfalfa; P28012) and sequence alignment of CHIs from Phaseolus vulgaris (bean; P14298), Pisum sativum (pea; P41089), Zea maize (corn; S41579), Vitis vinifera (grape; P51117), Ipomoea purpurea (morning glory; af028238), Petunia hybrida (P11651), and Arabidopsis thaliana (P41088). Legumes have genus names in red. α-Helices (gold rectangles) and β-strands (blue arrows) of CHI are indicated with the numbering of each protein in parentheses. Every 10th position in the alignment is dotted. Residues of the (2S)-naringenin binding cleft (green), residues of the active site hydrogen bond network (pink), and other conserved residues (blue) are highlighted. The residues that may influence substrate preference between chalcone and 6′-deoxychalcone are indicated with an asterisk.

tacts dominate the interactions between CHI and (2S)-naringenin, two hydrogen bonding interactions exist. The first occurs between the side chain hydroxyl moiety of Thr 190 and the 7-hydroxyl group of (2S)-naringenin; the second links a water molecule to the naringenin ketone moiety (Fig. 2b). This water molecule and its associated network of hydrogen bonds occupy the same position in the apoenzyme structure. The overall surface topology of the cleft tightly sequesters the (2S)naringenin molecule (Fig. 2c). The CHI–naringenin complex nature structural biology • volume 7 number 9 • september 2000

explains the stereochemical preference of the cyclization reaction and suggests why CHIs from different species show moderate selectivity for either chalcone or 6′-deoxychalcone as substrates. A model of chalcone in the binding cleft was built based on the position of (2S)-naringenin. This model demonstrates that a slight rotation of the chalcone trihydroxyl ring outward in the direction of the active site opening places the 2′-hydroxyl group in position for nucleophilic attack on the α,β-unsatu787

© 2000 Nature America Inc. • http://structbio.nature.com

letters a

Fig. 2 (2S)-Naringenin binding and structure of the active site cleft of CHI. a, Stereo view of the SIGMAA-weighted |2Fo - Fc| electron density (1.2 σ) for (2S)-naringenin (aqua). b, Stereo view of residues in the active site cleft. (2S)-Naringenin and a water molecule are also shown. Hydrogen bond interactions are indicated with dotted lines (rose). This view is oriented looking into the cleft. c, Stereo view surface representation of the active site cleft showing the fit of (2S)-naringenin and a water molecule. The surface corresponding to Lys 109 and Asn 113 has been removed for clarity.

b erence to the amino acid sequences of different CHIs, suggests that Thr 190 and Met 191 may partially modulate substrate preference. In the CHIs from nonlegumes, a Ser and an Ile replace Thr 190 and Met 191, respectively. These two differences may better accommodate the 6′-hydroxyl moiety of chalcone due to a modest increase in active site volume in the vicinity of the trihydroxyl ring.

© 2000 Nature America Inc. • http://structbio.nature.com

c

Catalytic mechanism CHI catalyzes an intramolecular reaction utilizing a substrate derived nucleophile and a carbon–carbon double bond as a Michael acceptor. Two reaction mechanisms have been proposed for (2S)-naringenin formation by CHI. One involves nucleophilic catalysis by an active site residue that forms a covalent intermediate that is released after an SN2 displacement by the 2′hydroxyl group of chalcone19. The other mechanism invokes general acid-base catalysis employing an enol intermediate20. The structure of CHI clearly supports the latter mechanism. Examination of the structure of the CHI–naringenin complex reveals a hydrogen bond network at the bottom of the binding cleft centered about the water molecule that contacts the ketone of (2S)-naringenin (Fig. 4a,b). Of the five amino acids contributing to this network, only Thr 48 and Tyr 106 are conserved in all CHIs. The position of the water molecule between (2S)-naringenin and Tyr 106 suggests a reaction mechanism in which the tyrosine activates the water, allowing it to serve as a general acid in the cyclization reaction (Fig. 4c).

rated double bond of the coumaroyl moiety (Fig. 3a). This rotation preserves the position of the chalcone backbone and the hydrogen bonds with both Thr 190 and the water molecule at the backside of the binding site. Formation of (2R)-naringenin would require substantial rearrangements in the active site of CHI due to significant steric clashes between the trihydroxyl ring and side chains of residues lining the cleft. Although rotation of the trihydroxyl ring away from the active site entrance could reposition the 2′-hydroxyl group for attack on the opposite face of the α,β-double bond, the side chain of Val 110 sterically prevents this movement from occurring (Fig. 3b). Alternatively, rotation of the coumaroyl moiety outward towards the solvent accessible active site entrance a could allow the 2′-hydroxyl group to attack the opposite side of the α,β-double bond to form (2R)-naringenin. However, Leu 38 and Lys 109 constrain the orientation of the coumaroyl moiety in the binding cleft. Architecturally, the CHI active site limits the substrate’s available conformations to ensure stereospecific product formation. CHIs of different species have subtle variations in substrate preference, as reflected in the Km values for chalcone versus 6′-deoxychalcone14. CHIs from legumes such as alfalfa prefer 6′-deoxychalcone as a substrate, but the enzymes from non-legumes like petunia optimally use chalcone. The b structure of the CHI–naringenin complex, viewed with refFig. 3 Proposed enzyme mediated stereochemical control of the cyclization reaction. The surface of the binding cleft is transparent to show selected residues (gold). The surface associated with Lys 109 and Asn 113 has been removed for clarity. The position of chalcone (aqua) prior to cyclization has been modeled to show the formation of the new bond (dotted orange line). Hydrogen bond interactions are indicated by dotted lines in rose. a, Stereo view of the proposed chalcone conformation leading to (2S)-naringenin. b, Stereo view of the steric clash that prevents (2R)-naringenin formation.

788

nature structural biology • volume 7 number 9 • september 2000

© 2000 Nature America Inc. • http://structbio.nature.com

letters a

c

Fig. 4 Proposed reaction mechanism of CHI. a, View of the active site hydrogen bond network. This view is oriented looking out of the active site cleft. Dotted lines (rose) indicate hydrogen bonds. b, Schematic representation of the active site hydrogen bonds. Distances are indicated in Å. c, Proposed cyclization reaction catalyzed by CHI. Following nucleophilic attack of the 2′-oxyanion on the α,β-unsaturated double bond in a Michael addition, the water molecule stabilized by Tyr 106 acts as a general acid to stabilize the enolate. This results in formation of a flav-3-en-4-ol intermediate that tautomerizes into the reaction product. Rendered figures were prepared with MOLSCRIPT40 or GRASP41 and rendered with POV-Ray42.

© 2000 Nature America Inc. • http://structbio.nature.com

b

In the proposed reaction mechanism, the 2′-oxyanion (pKa ∼7–8) forms in solution as suggested by studies on the spontaneous cyclization of chalcones21. The negatively charged oxygen then attacks the carbon–carbon double bond, utilizing a Michael addition with the water molecule at the backside of the active site acting as the general acid in the transient protonation of the intermediate enolate. To test this reaction mechanism, Tyr 106 was replaced with a Phe and the properties of the mutant CHI compared to the wild type enzyme. The kinetics for the cyclization of 6′-deoxychalcone by wild type CHI (kcat = 4,384 min-1; Km = 15.7 µM; kcat / Km = 2.79 × 108 M-1 min-1) versus those of the reaction catalyzed by the CHI Y106F mutant (kcat = 69.0 min-1; Km = 16.3 µM; kcat / Km = 4.23 × 106 M-1 min-1) demonstrate that the tyrosine residue contributes to stabilizing the transition state. The 100-fold reduction in reaction rate is consistent with the decrease in rate associated with the loss of a general acid22. However, the observed reaction rate with the mutant remains greater than that of the uncatalyzed cyclization reaction. We suggest that the structural complementarity of the binding cleft to the transition state contributes additional levels of catalytic rate enhancement. A major contribution to rate enhancement in enzymatic reactions results from bringing substrates or reactive centers in the same molecule together in space23,24. As described above, the topology of the binding cleft limits the flexibility of chalcone and eliminates catalytically unproductive orientations by spatially defining an optimal geometry for (2S)-naringenin formation. This effectively channels the ground state conformation of the substrate into a catalytically productive conformation. Together with contributions from general acid-base catalysis, shape complementarity between the CHI binding pocket and chalcone accelerates the cyclization reaction 107fold over the spontaneous reaction rate. nature structural biology • volume 7 number 9 • september 2000

Macromolecular complexes in flavonoid biosynthesis Published studies suggest that co-localization of proteins in loosely associated macromolecular complexes is a fundamental component of cellular processes, including flavonoid biosynthesis25–27. CHI and other flavonoid biosynthetic enzymes may associate to provide efficient channeling of substrates and products as shown in Arabidopsis thaliana28. The three short β-strands (β1a, β1b, β2) on the backside of the CHI structure form a relatively flat surface that could serve as an interface for protein–protein interactions. However, both gel filtration and analytical ultracentrifugation experiments failed to detect association of alfalfa CHI and its upstream partner alfalfa CHS2 in vitro (J.M.J. and J.P.N, unpublished results). Fast nonenzymatic isomerization of chalcones to flavanones occurs in solution, but yields a significant fraction of the biologically inactive (2R)isomer. It is tempting to conclude that channeling between CHS and CHI is necessary to prevent formation of mixed isomers. However, the second-order rate constant for flavanone synthesis by CHI approaches the diffusion controlled limit. This rate argues against the need for any close association between CHS and CHI to limit production of the catalytically inactive isomer. On the contrary, the moderate lipophilic character of chalcones and flavanones may necessitate close physical association between CHS, CHI, and downstream enzymes to limit the potential sequestration of biosynthetic intermediates in cellular membranes. Since alfalfa expresses at least four CHS isoforms15, the inability to detect protein–protein interactions may result from the involvement of specific isoforms in complex formation or the requirement for additional protein components in these putative complexes28. Alternatively, differing metabolic requirements between alfalfa and Arabidopsis may necessitate formation of multienzyme complexes in one species and not the other. For instance, alfalfa CHI provides both 5,7,4′-trihydroxyflavanone and 7,4′-dihydroxyflavanone to multiple secondary metabolic pathways, including those involved in pigment, phytoalexin, and nodulation inducer biosynthesis. Close physical association of CHS, CHI, and downstream enzymes at specific subcellular locations could divert metabolite flow from pathways specifically requiring 5,7,4′-trihydroxyflavanone or 7,4′-dihydroxyflavanone. However, in Arabidopsis, which uses 5,7,4′trihydroxyflavanone primarily for pigment production, co-localization of CHS and CHI could efficiently channel intermediate metabolites into anthocyanin biosynthesis and 789

© 2000 Nature America Inc. • http://structbio.nature.com

letters vector30. The CHI Y106F mutant was generated with the QuikChange (Stratagene) PCR method. N-terminal His8-tagged protein was expressed in E. coli BL21(DE3) cells. Tagged CHI was purified from sonicates using a Ni2+-NTA (Qiagen) column. Thrombin digestion removed the His tag, and the protein was passed over a Ni2+-NTA column. Digested CHI was depleted of thrombin using a benzamidine-Sepharose column. Gel filtration on a Superdex-75 (Pharmacia) FPLC column was the final purification step.

© 2000 Nature America Inc. • http://structbio.nature.com

Table 1 Crystallographic data, phasing, and refinement information Wavelength (Å) Resolution range (Å) Observations Unique reflections Completeness1 (%) I / σ1 Rsym1,2 (%) PP3 (acentric / centric) Rcullis4 (iso / ano) Rcryst5 / Rfree6 (%) Protein atoms Water molecules Ligand atoms R.m.s. deviations Bonds (Å) Angles (°) Average B-factors Protein (Å2) Water (Å2)

Native 1.08 38.1–2.50 86,781 27,863 90.7 (61.7) 19.4 (2.0) 5.4 (36.5)

K2OsCl6 0.98 78.0–3.24 44,070 24,966 99.2 (99.6) 9.3 (2.6) 10.3 (31.2) 2.55 / 1.85 0.52 / 0.76

HgCl2 0.98 76.0–3.26 39,193 22,238 95.8 (98.2) 17.5 (7.4) 5.8 (12.0) 1.58 / 1.07 0.72 / 0.81

Naringenin 0.95 46.7–1.85 697,121 70,889 85.5 (60.6) 28.0 (2.0) 4.8 (45.4)

23.5 / 27.3 3,181 94

20.8 / 24.3 3,222 382 40 naringenin, 20 sulfates

0.019 2.0

0.019 2.1

61.2 62.2

46.6 53.6

Numbers in parentheses are for the highest resolution shell. Rsym = Σ|Ih - | / Σ Ih, where is the average intensity over symmetry equivalent reflections. 3Phasing power = , where FH(calc) is the calculated difference and E is the lack of closure. 4R cullis = Σ|E| / Σ|FPH - FP|. 5R-factor = Σ|F - F | / ΣF , where summation is over the data used for refinement. o c o 6R free was calculated using 5% of data excluded from refinement. 1

Enzyme assays. CHI assays were performed at 25 °C in a 0.5 ml reaction volume containing 0.05 M Hepes (pH 7.5), 50 µM 6′-deoxychalcone, and 3% (v/v) ethanol as cosolvent13. Time dependent decreases in 6′-deoxychalcone absorbance (λmax = 390 nm; ε = 29,400 M-1 cm-1) were monitored with a Beckman DU-640 spectrophotometer. Determination of steady-state kinetic constants used the standard assay system with varied concentrations of substrate (2–50 µM) following fitting to the MichaelisMenton equation using Kaleidagraph (Abelbeck Software).

2

away from sequestration in cellular membranes. Evidence suggests that Arabidopsis CHI is covalently modified at a cysteine residue28. Although the identity of this moiety is unresolved, the sensitivity of the modification to sulfhydryl containing reductants suggests that a thioester-linked long-chain fatty acid is present on CHI in vivo. CHI, modified in this way, could localize together with other flavonoid biosynthetic enzymes on the membrane of the endoplasmic reticulum 27. Intriguingly, sequence comparisons of alfalfa and Arabidopsis CHIs reveal that Arabidopsis CHI includes an 11-residue Nterminal extension with a cysteine at position 7 (Fig. 1d). Posttranslational modification of this N-terminal cysteine with a fatty acid could provide an ideal membrane anchor for CHI. Conclusions The three-dimensional structure of CHI provides insight into the enzyme architecture responsible for catalyzing a nearly diffusion controlled cyclization reaction that is primarily driven by entropy and induced fit. The uniqueness of the CHI fold and the complete lack of homology to any other protein raise intriguing questions about the evolutionary lineage of this enzyme and the appearance of flavonoid biosynthetic pathways in higher plants. Given that land plants likely evolved from a freshwater ancestor that resembled green algae29, it is possible that distantly related CHI homologs exist in simpler plants and algae. Surprisingly, very little is currently known about secondary metabolic pathways in more primitive plant species. Methods Expression, mutagenesis, and purification. Alfalfa CHI cDNA15 was PCR amplified and inserted into the pHIS8 expression

790

Crystallization, structure determination, and refinement of the native structure. Crystals of CHI were grown at 4 °C by vapor diffusion using the hanging drop method. A 2 µl drop containing a 1:1 mixture of 25 mg ml-1 CHI and crystallization buffer (25% v/v glycerol, 1.8–2.0 M ammonium sulfate and 0.05 M PIPES, pH 6.5) yielded diffraction quality crystals within a few days at 4 °C. Crystals grew in space group P6522 with unit cell dimensions of a = b = 90.37 Å, c = 352.86 Å, with two molecules per asymmetric unit and a solvent content of 72%. Native CHI diffraction data (105 K) were collected at beamline 7-1 of the Stanford Synchrotron Radiation Source (SSRL 7-1) on a 30 cm MAR imaging plate system. For generation of heavy atom derivatives, CHI crystals were soaked in mother liquor with either 1.2 mM K2OsCl6 or 1 mM HgCl2 for 12–16 hours. Heavy atom data (105 K) were collected at SSRL 9-1 on a 30 cm MAR imaging plate system. All images were indexed and integrated using DENZO31 and the reflections merged with SCALEPACK31. Data reduction was completed using programs from CCP432 (Table 1). Heavy atom sites were located with SOLVE33. Refinement of sites and location of additional sites used MLPHARE34. SHARP35 was used for phase calculation and heavy atom refinement. This set of experimental phases was improved and extended using solvent flipping with SOLOMON36. Model building was performed with O 37. CNS38 was used for refinement. The initial model was subjected to simulated annealing, positional refinement, and group B-factor refinement with strict noncrystallographic symmetry maintained between both molecules in the asymmetric unit. In subsequent rounds of model building and refinement, noncrystallographic constraints were released and water molecules were added using CNS to yield the R-factors shown in Table 1. The final model included residues 4–215 of monomer A, residues 3–38 and 45–215 of monomer B, and 94 water molecules. The quality of the CHI model was checked with PROCHECK39. A total of 89.6% of the residues in CHI are in the most favored regions of the Ramachandran plot and 10.4% are in the additionally allowed region. Crystallization, structure determination, and refinement of the CHI–naringenin complex structure. Crystals of the CHI–naringenin complex (P6522; a = b = 89.47 Å; c = 351.19 Å) were grown as above from a crystallization buffer containing 2.5 mM (2S/2R)naringenin and 5% (v/v) ethanol. Data (105 K) were collected at SSRL 9-2 with a Quantum 4 CCD detector. Images were processed as above. Following rigid body refinement with CNS, electron density resembling naringenin was observed in each monomer

nature structural biology • volume 7 number 9 • september 2000

© 2000 Nature America Inc. • http://structbio.nature.com

letters and modeled as such. In subsequent rounds of refinement and rebuilding, the R-factors converged to those listed in Table 1. The final model includes residues 4–215 of both monomers, 382 water molecules, two naringenin molecules, and five sulfates. Coordinates. Coordinates and structure factors have been deposited in the Protein Data Bank (accession codes 1EYP for the CHI native and 1EYQ for the naringenin complex).

Acknowledgments We thank A. Hirsch (UCLA) for providing the CHI cDNA. The SSRL Biotechnology Program is supported by the NIH, National Center for Research Resources, Biomedical Technology Program, and the DOE, Office of Biological and Environmental Research. This work was supported by a grant from the National Science Foundation awarded to J.P.N. J.M.J. is a NIH Postdoctoral Research Fellow and also received support from the Hoffman Foundation.

© 2000 Nature America Inc. • http://structbio.nature.com

Correspondence and requests for materials should be addressed to J.P.N. email: [email protected] Received 15 May, 2000; accepted 22 June, 2000. 1. Dooner, H.K., Robbins, T.P. & Jorgensen, R.A. Annu. Rev. Genet. 25, 173–199 (1991). 2. Long, S.R. Cell 56, 203–214 (1989). 3. Dixon, R.A. & Paiva, N.L. Plant Cell 7, 1085–1097 (1995). 4. Bohm, B.A. Introduction to Flavonoids (Harcourt, Singapore; 1998). 5. Setchell, K.D.R. & Cassidy, A. J. Nutr. 129, 758S–767S (1999). 6. Dixon, R.A. & Steele, C.J. Trends Plant Sci. 4, 394–400 (1999). 7. Davies, K.M., Bloor, S.J., Spiller, G.B. & Deroles, S.C. Plant J. 13, 259–266 (1998). 8. Jung, W. et al. Nature Biotech. 18, 208–212 (2000). 9. DellaPenna, D. Science 285, 375–379 (1999). 10. Ferrer, J.L., Jez, J.M., Bowman, M.E., Dixon, R.A. & Noel, J.P. Nature Struct. Biol. 6, 775–784 (1999). 11. van Tunen, A.J., Mur, L.A., Recourt, K., Gerats, A.G. & Mol, J.N. Plant Cell 3, 39– 48 (1991).

nature structural biology • volume 7 number 9 • september 2000

12. Firmin, J.L., Wilson, K.E., Rossen, L. & Johnston, A.W.B. Nature 324, 90–92 (1986). 13. Bednar, R.A. & Hadcock, J.R. J. Biol. Chem. 263, 9582–9588 (1988). 14. Dixon, R.A., Blyden, E.R., Robbins, M.P., van Tunen, A.J. & Mol, J.N. Phytochemistry 27, 2801–2808 (1988). 15. McKhann, H.I. & Hirsch, A.M. Plant Mol. Biol. 24, 767–777 (1994). 16. Sander, C. & Schneider, R. Proteins 9, 56–68 (1991). 17. Altschul, S.F. et al. Nucleic Acids Res. 25, 3389–3402 (1997). 18. Bednar, R.A., Fried, W.B., Lock, Y.W. & Pramanik, B. J. Biol. Chem. 264, 14272– 14276 (1989). 19. Hahlbrock, K., Wong, E., Schill, L. & Grisebach, H. Phytochemistry 9, 949–958 (1970). 20. Boland, M.J. & Wong, E. Bioorg. Chem. 8, 1–8 (1979). 21. Rathmell, W.G. & Bendall, D.S. Biochem. J. 127, 125–132 (1972). 22. Jencks, W.P. Catalysis in Chemistry and Enzymology (McGraw-Hill, New York; 1969). 23. Bruice, T.C. & Pandit, U.K. J. Am. Chem. Soc. 82, 5858–5865 (1960). 24. Page, M.I. & Jencks, W.P. Proc. Natl. Acad. Sci. USA 68, 1678–1683 (1971). 25. Winkel-Shirley, B. Physiol. Plant. 107, 142–149 (1999). 26. Hrazdina, G. & Wagner, G.J. Arch. Biochem. Biophys. 237, 88–100 (1985). 27. Hrazina, G., Zobel, A.M. & Hoch, H.C. Proc. Natl. Acad. Sci. USA 84, 8966–8970 (1987). 28. Burbulis, I.E. & Winkel-Shirley, B. Proc. Natl. Acad. Sci. USA 96, 12929–12934 (1999). 29. Graham, L.E., Cook, M.E. & Busse, J.S. Proc. Natl. Acad. Sci. USA 97, 4535–4540 (2000). 30. Jez, J.M., Ferrer, J.-L., Bowman, M.E., Dixon, R.A. & Noel, J.P. Biochemistry 39, 890–902 (2000). 31. Otwinowski, Z. & Minor, W. Methods Enzymol. 276, 307–326 (1997). 32. Collaborative Computational Project, Number 4. Acta Crystallogr. D 53, 240–255 (1994). 33. Terwilliger, T.C. & Berendzen, J. Acta Crystallogr. D 55, 849–861 (1999). 34. Otwinowski, Z. ML-PHARE in Daresbary Study Weekend Proceedings (CCP4, SERCDaresbary Laboratory, Warrington, UK; 1991). 35. de La Fortelle, E. & Bricogne, G. Methods Enzymol. 276, 472–494 (1997). 36. Abrahams, J.P. & Leslie, A.G.W. Acta Crystallogr. D 52, 30–42 (1996). 37. Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. Acta Crystallogr. D 49, 148– 157 (1993). 38. Brünger, A.T. et al. Acta Crystallogr. D 54, 905–921 (1998). 39. Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton, J.M. J. Appl. Crystallogr. 26, 283–291 (1993). 40. Kraulis, P.J. J. Appl. Crystallogr. 24, 946–950 (1991). 41. Nicholls, A., Sharp, K. & Honig, B. Proteins 11, 281–296 (1991). 42. Amundsen, S. et al. POV-Ray: persistence of vision ray-tracer. http://www.povray.org (1997).

791