Functional Expression and Purification

Accepted ManuscriptFunctional Expression and Purification of CYP93C20a Plant Membrane-As‐sociated Cytochrome P450 from Medicago truncatulaZhenzhan Chang, Xiaoqiang Wang, Risheng Wei, Zhouying Liu, Hong Shan,Guizhen Fan, Hongli HuPII:S1046-5928(10)00322-0DOI:10.1016/j.pep.2010.11.012Reference:YPREP 3835To appear in:Protein Expression and PurificationReceived Date:7 April 2010Revised Date:16 November 2010Please cite this article as: Z. Chang, X. Wang, R. Wei, Z. Liu, H. Shan, G. Fan, H. Hu, Functional Expression and Purification of CYP93C20a Plant Membrane-Associated Cytochrome P450 from Medicago truncatula, Protein Expression and Purification (2010), doi: 10.1016/j.pep.2010.11.012This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.Functional Expression and Purification of CYP93C20 a PlantMembrane-Associated Cytochrome P450 from Medicago truncatulaZhenzhan Chang , Xiaoqiang Wang, Risheng Wei, Zhouying Liu, Hong Shan,Guizhen Fan, Hongli HuDepartment of Biophysics, Peking University Health Science Center, 38 Xue Yuan Road, Beijing 100191, ChinaAddress correspondence to Dr. Zhenzhan Chang: Department of Biophysics, Peking University Health Science Center, Xue Yuan Road 38, Beijing 100191, China. Tel. +86 10 82805369; Fax: +86 10 82801444; E-mail: changz@Abbreviations used: IFS, isoflavone synthase; ALA, 5-aminolevulinic acid; IPTG, isopropyl -D-1-thiogalactopyranoside; MALDI-TOF-MS, matrix-assisted laser desorption ionization-time of flight-mass spectrometry; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; NADPH, nicotinamide adenine dinucleotide phosphate reduced form; HPLC, high performance liquid chromatography.Abstract Plants possess very large numbers of biosynthetic cytochrome P450 enzymes. In spite of the importance of these enzymes for the synthesis of bioactive plant secondary metabolites, only two plant P450 structures has been obtained to date. Isoflavone synthase (IFS) is a membrane-associated cytochrome P450 enzyme catalyzing the entry-point reaction into isoflavonoid biosynthesis. IFS from the model legume Medicago truncatula (CYP93C20) was engineered by deleting the membrane-spanning domain and inserting a hydrophilic polypeptide in the N-terminus and a four histidine tag at the C-terminus. The truncated form exhibited dramatically enhanced expression and solubility. The engineered enzyme was expressed in Escherichia coli XL1-blue cells and was purified by Ni 2+-NTA affinity chromatograph and size-exclusion chromatograph. The purified enzyme was characterized by enzyme assay, reduced carbon monoxide difference spectroscopy and peptide mass fingerprinting. The engineered soluble enzyme exhibited the same activity as the full length membrane-associated enzyme expressed in yeast. These studies suggest an approach for engineering plant membrane-associated P450s with enhanced expression and solubility for mechanistic and structural studies. Keywords : cytochrome P450; isoflavone synthase; functional expression; purification Introduction异黄酮蒺藜状苜蓿Cytochromes P450 (P450s) are external monooxygenases encoded by a superfamily of genes ubiquitously distributed in different organisms from all biological kingdoms. The P450s reactions are extremely diverse and contribute to the biotransformation of drugs, the异型生物质bioconversion of xenobiotics, the biosynthesis of physiologically important compounds such as steroids, fatty acids, fat-soluble vitamins and bile acids, the metabolism of chemical carcinogens, as well as the degradation of herbicides and insecticides [1]. Eukaryotic P450s are expressed at low levels, sometimes as P420 forms, the biologically inactive forms which possess ferrous CO Soret absorption at 420 nm [2]. Therefore their crystallization and structure determination are challenging. Progress on structural studies of mammalian P450 enzymes have been significantly advanced by the use of E. coli as a heterologous expression system. Efforts to decrease the hydrophobicity and thus increase the solubility of mammalian P450s have primarily focused on deletion of the N-terminal membrane anchor with addition of a His-tag at the C-terminus to aid in protein purification. Williams et al.crystallized and determined the structure of the first membrane-bound P450 (2C5). They expressed a soluble monomeric form compatible with growing diffraction-quality crystals by truncating the N-terminus and introducing mutations into a peripheral membrane-binding site in the F-G loop region [3]. Success in crystallographic studies of other mammalian P450s, (e.g. 2B4, 3A4, 2C8 and 2A6) [4, 5, 6, 7] and plant P450 CYP74A [8], suggested that removal of the N-terminal transmembrane domain with addition of a C-terminal His-tag is sufficient to crystallize these enzymes, without need for additional mutations in the F-G loop region.Several widely used expression systems have been developed to produce catalytically active plant P450s in bacteria, yeast and insect cells, with the yeast expression system being themost frequently used [9]. Plant P450 expression studies in E. coli have focused on the CYP74 [10, 11, 12, 13, 14] and CYP79 [15, 16, 17] families. Arabidopsis CYP90B1 [18], flavonoid 3’, 5’-hydroxylase (F 3’5’H) [19] and carotene hydroxylases (CYP97 family) [20] have also been expressed in E. coli. However, only a few plant P450s (eg. CYP74A1, CYP74C3, CYP74A2, CYP74A) have been produced in E. coli at the milligram scale necessary for crystallization studies [13, 14, 21, 8] and so far only two crystal structures of plant P450s (CYP74A2 and CYP74A) have been determined [22, 8].Here we describe the membrane-associated CYP93C20 IFS from the model legume Medicago truncatula utilizing molecular engineering techniques. We were able to obtain soluble CYP93C20 protein in milligram quantities that is functionally characterized, highly active, homogeneous and detergent-free. The CYP93C20 protein obtained is prepared for crystallographic studies. Our results describe a general methodology for molecular engineering and purification of membrane-associated plant P450 enzymes.Materials and methodsChemicalsOligonucleotides for cloning, site-directed mutagenesis and DNA sequencing were purchased from IDT (Coralville, IA, USA). Liquiritigenin was purchased from Indofine chemical company, Inc. (Hillsborough, NJ, USA), NADPH-P450 oxidoreductase (Human, recombinant) from Calbiochem-EMD Biosciences (La Jolla, CA, USA), NADPH, ALA and sodium cholate hydrate from Sigma-Aldrich (St. Louis, MO, USA), IPTG from Promega (Madison, WI, USA) and Ni2+-NTA affinity resin from Qiagen (Valencia, CA, USA).Generation of recombinant CYP93C20 constructsAlignment of amino acid sequences of CYP93C20 from Medicago truncatula with mammalian P4502C5 and 2B4 [3, 4] was performed using ClustalW (Fig. 1). Based on the sequence alignment analysis, four truncated forms of CYP93C20 were designed and generated using the templates and primers listed in Table 1. CYP93C20 cDNA encoding M. truncatula IFS (GenBank: AY167424) served as an initial template for polymerase chain reaction (PCR). The forward 5’-primer was designed to introduce an Nde I restriction site and to delete codons for residues 2-18 followed by a sequence corresponding to codons for residues 19-24 of CYP93C20. The reverse 3’-primer includes an Xba I restriction site, the termination codon TTA, codons for four histidines, and a sequence complementary to codons for residues 522-517 of CYP93C20. The PCR products were digested with Nde I and Xba I, and subcloned into pCW ori+ expression vector. This cloning strategy generated the construct that we call pCW ori+ 2-18ifsH which contained 4xHis tag at the C-terminus and deleted the N-terminal 17 amino acids.The construct pCW ori+ 2-34ifsH (with further deletion of N-terminal amino acids 19-34) was generated using the QuikChange site-directed mutagenesis strategy (Stratagene) with pCW ori+ 2-18ifsH as template and the set of complementary primers listed in Table 1.To generate constructs with insertion of a short charged, hydrophilic polypeptide at the N-terminus, site-directed mutagenesis was performed using pCW ori+ 2-34ifsH as template with the primers shown in Table 1. This strategy replaced the N-terminal transmembrane signal anchor domain (residues 1-34) with an optimized sequence encodingthe peptide sequence MAKKTSSKGKL and inserted four codons for histidine upstream of the stop codon to facilitate purification of the enzyme. The resulting plasmid was termed pCW ori+ 2-34 10ifsH.In order to determine how the C-terminus sequence of CYP93C20 affects the expression and purification of CYP93C20, we deleted amino acids ARAGVADKLLSS utilizing site-directed mutagenesis with 2-34 10ifsH as the template with primers in Table 1 to generate a construct pCW ori+ 2-34 10ifs 12H.Optimization of expression of recombinant CYP93C20 in E. coliThe construct pCW ori+ 2-34 10ifsH was transformed into BL21(DE3) competent cells to screen expression conditions for recombinant CYP93C20. One colony of BL21(DE3)/ pCW ori+ 2-34 10ifsH was inoculated into 50 ml LB medium containing 100 g/ml ampicillin and incubated at 37o C and 250 rpm; 100 ml overnight cultures were started by inoculating with 0.5 ml aliquots of overnight culture. Induction of recombinant CYP93C20 was initiated at OD600 = 0.5 by adding IPTG to a final concentration of 0.5 mM or 1 mM; ALA was supplemented as a precursor for heme synthesis to a final concentration of 0.5 mM or 1 mM. Incubation was continued with shaking at 190 rpm at 16, 28 or 30o C. The expression level was assessed at 24, 32, 48 and 72 h by SDS-PAGE analysis.To select the best E. coli strains for expression of recombinant CYP93C20, the construct pCW ori+ 2-34 10ifsH was transformed into different competent cell lines. One liter batches of LB medium fortified with suitable antibiotics for BL21(DE3) cells (100 g/ml ampicillin ), XL1-blue cells (100 g/ml ampicillin and 12.5 g/ml tetracycline), JM109 cells(100 g/ml ampicillin), Rosetta (DE3) cells (50 g/ml chloramphenicol) and DH5 cells (100 g/ml ampicillin) were started in 2-liter culture flasks at 37o C and 250 rpm. At OD600= 0.4~0.6, IPTG was added to a final concentration of 1mM to induce expression. ALA was supplemented at a final concentration of 0.5 mM and imidazole was added to a final concentration of 5 mM at the time of induction to stabilize holoprotein. Cultures were grown for an additional 48 h at 28o C and 190 rpm. The cultures were placed on ice for 15 min and the cells harvested by centrifugation at 5000×g for 10 min at 4°C. Enzyme was purified by Ni-NTA affinity chromatography from individual culture extracts. The expression level was compared by SDS-PAGE analysis.The construct pCW ori+ 2-34 10ifs 12H was also transformed into XL1-blue competent cells for determining its expression level in E. coli.UV-visible and carbon monoxide difference spectra of recombinant CYP93C20 UV-visible and carbon monoxide difference spectroscopy was performed using a dual-beam scanning Shimadzu UV-visible spectrophotometer (Model UV-1601; Shimadzu, Milton Keynes, U.K.) A 200 ml culture of XL1-blue cells transformed with pCW ori+ 2-34 10ifsH was prepared as described above. The cells were harvested and resuspended in 10 ml cell lysis buffer (500 mM potassium phosphate pH 7.4, 250 mM KCl, 1 mM PMSF, 10% glycerol), followed by sonication for 10 X 20 s with 40 s intervals on ice and centrifugation at 29000 g for 45 min.The reduced-CO difference spectrum of the supernatant containing CYP93C20 was obtained according to a published method [23]. Briefly, 3 ml of the supernatant wasreduced with a few mg of sodium dithionite, and equal volumes of this preparation were distributed into sample and reference quartz cuvettes (0.7 ml) for baseline correction. CO was gently bubbled through the sample cuvette for 1 min prior to recording the difference spectrum from 500 to 400 nm.Peptide mass fingerprintingIn-gel digestion, MALDI-TOF-MS and database searching were performed according to a reported method [24]. Briefly, protein bands were excised from the SDS-PAGE gel, washed and destained completely with 50% acetonitrile in 50 mM ammonium bicarbonate. The gel pieces were then dehydrated with acetonitrile and dried in a vacuum centrifuge. Digestion was performed overnight with bovine trypsin (10 ng/ l), stopped by addition of 10% formic acid, and the supernatant recovered. The gel pieces were extracted once with 25 l 50% acetonitrile and once with 100% acetonitrile. The supernatant was combined and dried in a vacuum centrifuge. Peptides were suspended in a 1:1 (v/v) solution of 2% formic acid in acetonitrile, mixed 1:1 (v/v) with matrix (10 mg ml-1 –cyano-4-hydroxycinnamic acid in the same solution), and spotted on a MALDI plate. MALDI-TOF-MS peptide maps were analyzed by a database search against NCBInr (release January 9, 2005) using MS-Fit () with the following parameters: Mass accuracy, 100 ppm; missed cleavage, 1; possible modifications, oxidation of Met.Purification of recombinant CYP93C20Large scale (usually 4 liters) expression with XL1-blue or BL21(DE3) cells transformed with pCW ori+ 2-34 10ifsH was carried out to produce recombinant CYP93C20.A 5 ml overnight culture was inoculated into 1 liter of LB medium containing 100 g/ml ampicillin and 12.5 g/ml tetracycline, maintaining the same expression conditions as used to screen competent cells.Red cell pellets were resuspended in lysis buffer (500 mM potassium phosphate, 0.25 M NaCl, 10 mM -mercaptoethanol, 10% glycerol, 10 mM imidazole, 0.25% sodium cholate, pH 7.4). Lysozyme (0.2 mg/ml), Dnase I (1.2 g/ml), MgSO4 (2 mM) and phenylmethanesulfonyl fluoride (1 mM) were added. The cells were incubated and stirred for 30 min at 4°C and homogenized by passing through a French press cell homogenizer, followed by centrifugation at 29,000 g for 1 h. The supernatant containing the CYP93C20 was applied to a Ni2+-NTA column previously equilibrated with 20 column volumes of equilibration buffer/wash buffer A (500 mM potassium phosphate, pH 7.4, 0.5 M NaCl, 10 mM -mercaptoethanol, 10% glycerol). The column was washed with: (1), 100 column volumes of equilibration buffer/wash buffer A; (2), 100 column volumes of wash buffer B (10 mM potassium phosphate, pH 7.4, 0.5 M NaCl, 10 mM -mercaptoethanol, 20 mM imidazole, 10% glycerol) and (3), 20 column volumes of wash buffer C (10 mM potassium phosphate, pH 7.4, 0.5 M NaCl, 10 mM -mercaptoethanol, 40 mM imidazole, 10% glycerol). The protein was then eluted with elution buffer (10 mM potassium phosphate, pH7.4, 0.5 M NaCl, 10 mM -mercaptoethanol, 250 mM imidazole, 10% glycerol). Fractions with the highest concentration of protein were pooled. Finally the protein was purified by gel filtration on a HiLoad 16/60 Superdex 200 column eluting in gel filtration buffer (10 mM potassiumphosphate, pH 7.2, 0.5 M NaCl, 1 mM DTT, 0.2 mM EDTA, 10% glycerol). The red-colored fractions were pooled and concentrated with a Vivaspin 20 concentrator (30 kDa molecular-weight cut-off, Stonehouse, UK).Assay of enzyme activity of recombinant CYP93C20Transformation of yeast WAT11 cells with the empty pYeDP60 vector and pYeDP60-IFS construct 51865 and preparation of microsomes were conducted as previously described [23]. E. coli-expressed recombinant CYP93C20 (20 M), NADPH (1 mM) and liquiritigenin (80 M) were mixed in reaction buffer (0.1 M K2HPO4, 0.4 M sucrose, 0.5 mM glutathione, pH 8.0) with approximately 1 mg of microsomes containing Arabidopsis thaliana(ATR1) NADPH-P450 reductase prepared from yeast WAT11 cells transformed with pYeDP60 vector, or 0.7 units of human NADPH-P450 oxidoreductase, and incubated overnight at 16 o C. Microsomes (1 mg) prepared from yeast WAT11 cells transformed with pYeDP60-IFS construct 51865 were incubated with liquiritigenin in reaction buffer as positive control.Following incubation, reactions were stopped and extracted three times with one volume of ethyl acetate. Ethyl acetate extracts were combined, dried under a stream of N2, and the residues resuspended in 100 l methanol. HPLC analysis was performed on an Agilent HP1100 HPLC equipped with an auto sampler, quaternary pump, and diode array detector. Solvent A was 1% aqueous phosphoric acid and solvent B was acetonitrile. Reaction samples (40 l) were applied to an ODS reverse-phase column (5 m particle size, 4.6 x 250 mm) and eluted in 1% (v/v) phosphoric acid with an increasing gradient ofacetonitrile (0-5 min, 5%; 5-10 min, 5-10%; 10-25 min, 10-17%; 25-30 min, 17-23%; 30-65 min, 23-50%; 65-69 min, 50-100%) at a flow rate of 1 ml min-1. The eluants were monitored at 235, 254 and 270 nm. Isoflavonoid products were identified by comparing retention times and UV spectra with those of authentic standards.Results and discussionEffect of N-terminal and C-terminal modifications on protein expressionThe first eukaryotic P450 structure solved was that of mammalian P450 2C5 (Williams, 2000), which was engineered to delete the N-terminal membrane-spanning helix ( 3-21) and to mutate a peripheral membrane–binding site to facilitate protein expression. For solving the structure of rabbit P450 2B4, the N-terminal transmembrane domain was truncated ( 3-21) and several N-terminal residues were mutated to facilitate expression.CYP93C20 from M. truncatula is also a membrane-associated P450, and modification of its amino acid sequence is also required to enhance solubility and achieve high-level expression level. To design truncation forms of CYP93C20, its amino acid sequence was aligned with those of two structurally characterized mammalian P450s (2C5 and 2B4) [25, 3, 4, 26]. The alignment showed that the amino acid sequence identity between M. truncatula CYP93C20 and the two mammalian P450s had about 20% homology, but they share many conserved features such as the heme binding motif (Fig. 1). A conserved proline-rich region with a PXXP motif is also present in the N-terminus between the signal-anchor and catalytic domain, and previous work on cytochrome P450 2C2 and others suggested its importance inthe formation of a functional cytochrome P450 across species from humans to plants [27, 28, 29]. Based on the sequence alignment, three truncated forms of CYP93C20 were designed to delete the putative transmembrane domain ( 2-18, 2-24, 2-29) and keep the proline-rich motif. All constructs had low level or no expression in E. coli. Additionally, reduced-CO difference spectra shown the proteins existed predominately as inactive cytochrome P420 (data not shown).The construct 2-34ifsH was designed to remove the N-terminal residues 2-34. Subsequently, 2-34 10ifsH was generated to incorporate the small peptide AKKTSSKGKL in the N-terminus of 2-34ifsH by referring to the method for generateing the construct which was successful in solubilization of cytochrome 2B4 [26]. 2-34 10ifsH was transformed into BL21(DE3) cells for expression. The expression conditions were optimized with respect to temperature (28o C), time (48h), aeration (190 rpm), and concentrations of IPTG (1 mM) and ALA (0.5 mM).The optimized growth conditions were applied to cultures of different E. coli strains transformed with 2-34 10ifsH. SDS-PAGE analysis (Fig. 2) indicated that XL1-blue, BL21 (DE3), DH5 and Rosetta (DE3) cells gave higher level expression of the target protein than JM109 cells. However, Rosetta (DE3) and DH5 cells expressed more non-target proteins than with the other cells. XL1-blue and BL21 (DE3) cells were therefore chosen for large scale expression of CYP93C20.To analyze the effects of the C-terminal sequence on protein expression/stability, the C-terminally truncated form pCW ori+ 2-34 10ifs 12H was expressed in XL1-blue cells. The cells were less pink in color than those transformed with pCW ori+ 2-34 10ifsH withoutthe C-terminal truncation. Further purification trials showed that the protein did not bind well to Ni2+-NTA beads, suggesting that residues at the C-terminus of CYP93C20 are important for the enzyme to form the correct conformation.Identification of recombinant CYP93C20 by peptide mass fingerprintingTo confirm that the protein expressed and purified from E. coli was recombinant CYP93C20, peptide mass fingerprint analysis was performed with a mass spectrometry with bands excised from the Coomassie-stained SDS-PAGE gel. High quality MALDI-TOF-MS peptide maps were obtained (Fig. 3), and analyzed by a database search against NCBInr. Two proteins were successfully identified. Band 1 from the SDS-PAGE (Fig.2) was identified as chaperone Hsp70 with 34 peptides and sequence coverage of 59%. Band 2, the major band in the SDS-PAGE, was identified as isoflavone synthase from Medicago truncatula with 13 peptides and sequence coverage of 29%.Spectral properties of recombinant engineered CYP93C20The UV-visible spectrum of CYP93C20 showed a Soret band at 391 nm and major features at 508 and 545 nm (Fig. 4). The reduced-CO difference spectrum had a characteristic peak at 450 nm; however, under the present culture conditions, a peak for cytochrome P420, the denatured form of CYP93C20, was also observed.The reduced-CO difference spectrum of microsomes from yeast expressing the full-length CYP93C20 also exhibited a major peak at 450 nm; however, after the microsomes had been kept in the spectrophotometer for 3 min, or if the enzyme remained at roomtemperature for 5-10 min prior to reduction, a peak at 420 nm would be observed. Very similar culture conditions have been used in our laboratory to express another plant P450 enzyme, allene oxide synthase, in E. coli with no appreciable degradation to cytochrome P420. Our observations suggest that CYP93C20 is inherently unstable when expressed either as the full-length protein in yeast or as an engineered form in E. coli, and may require additional factors and more careful handling to maintain it in the active P450 form during expression and purification.Purification of recombinant engineered CYP93C20The engineered CYP93C20 pCW ori+ 2-34 10ifsH expressed in XL1-blue or BL21(DE3) cells was purified using two chromatography steps. Chromatography on Ni2+-NTA column appeared to be an efficient purification step and led to a significant enrichment of the P450. Subsequent gel filtration chromatography resulted in nearly homogeneous CYP93C20, which appeared as a single band of SDS-PAGE analysis (Fig. 5).The gel filtration profile using a HiLoad 16/60 Superdex 200 preparative column showed that engineered CYP93C20 was a monomer, with a molecular weight value matching that calculated based on its primary sequence. The red-colored fractions were pooled and used for crystallization of CYP93C20. In some circumstances after gel filtration, a small proportion of soluble aggregated protein was also observed (Fig. 6). The polymer had the same purity, but with very low heme content and was not used for crystallization.Expression of the construct pCW ori+ 2-34 10ifsH in E. coli XL1-blue or BL21(DE3) cells and subsequent purification by Ni2+-NTA affinity chromatography yielded ~25 mg CYP93C20 per liter of culture medium as assayed by the Bradford method [30].Enzymatic activity of recombinant engineered CYP93C20To determine whether the recombinant engineered CYP93C20 retains full catalytic activity, enzymatic activity assays were performed using liquiritigenin as substrate, in the presence of NADPH and NADPH-P450 reductase (either as microsomes from yeast expressing the Arabidopsis reductase, or commercial human P450 oxidoreductase). Reaction products were separated by HPLC and their identities confirmed by comparison of retention times and UV spectra to those of authentic standards.When E. coli expressed recombinant CYP93C20 was incubated with yeast microsomes containing plant P450 reductase, liquiritigenin, and NADPH, a major product, daidzein, was observed, along with a smaller amount of 2,7,4’-trihydroxyisoflavanone (Fig. 7b). This indicated that the engineered CYP93C20 converted liquiritigenin to 2,7,4’-trihydroxyisoflavanone, and was subsequently dehydrated non-enzymatically to daidzein. Similar amounts of daidzein were produced under these conditions when yeast-expressed full-length CYP93C20 was incubated with liquiritigenin and NADPH (Fig. 7a, b). Daidzein was also formed when the soluble engineered CYP93C20 was incubated with liquiritigenin, NADPH and human NADPH-P450 oxidoreductase, but at levels less than 10% of those produced when using the Arabidopsis reductase (Fig. 7c). Thus, the solubleengineered CYP93C20 maintained full catalytic activity when the plant P450 reductase was supplied as the electron donor from NADPH.ConclusionsOverall, our work has demonstrated that it is possible to optimize construct design and expression system to generate large amounts of soluble, active IFS. These approaches should be applicable to other plant P450s. The success in CYP93C20 expression, purification will facilitate future structural studies to understand how IFS carries out the novel aryl-ring migration for conversion of flavanone to 2-hydroxyisoflavanone, and to reveal the molecular basis of the catalysis and exquisite substrate specificity of IFS and other plant biosynthetic P450s.AcknowledgementsWe thank Dr. Z. Lei for assistance with the MALDI-TOF-MS peptide fingerprint analysis, J. Lin for initial work on cloning and expression, Dr. B. Deavours for help with HPLC analysis, Dr. L. Pedersen (NIEHS, NIH) for kindly providing pCWori+ vector, Dr. F. W. Dahlquist (Institute of Molecular Biology, University of Oregon) for permission to use pCWori+, Drs. R. A. Dixon and C.C. Yin for critical reading of the manuscript. This work was supported by grant 30970572 from the National Natural Science Foundation of China.Figure LegendsFig. 1. Alignment of amino acid sequences of CYP93C20 from Medicago truncatula with rabbit P450s 2C5 and 2B4. Box A indicates the proline-rich region (Kemper, 2004). Box B indicates the highly conserved “I” helix suggested to be involved in O2 binding by P450s, and box C highlights the heme binding motif (Graham-Lorence, 1996). Conserved residues are highlighted. Sequence analysis was performed using ClustalW search on DeCypher Protein Sequences.Fig. 2. SDS-PAGE analysis of engineered CYP93C20 expressed in different E. coli competent cells. BLs, XLs, JMs, DHs, RSs indicated the supernatant of BL21(DE3), XL1-blue, JM109, DH5 , Rosetta (DE3), respectively; BLp, XLp, JMp, DHp, RSp indicated the protein of each cell which was purified with Ni2+-NTA affinity column. Bands 1 and 2 were excised for protein identification by peptide mass fingerprinting.Fig. 3. Peptide mass spectrums obtained using MALDI-TOF-MS. Mass spectral peaks are labeled with monoisotopic mass-to-charge ratio (m/z) values used for database searching. IFS (A) and chaperone Hsp70 (B) were successfully identified using the NCBInr database.Fig. 4. UV-visible absorption spectrum (A) and CO difference spectrum (B) of supernatant of engineered CYP93C20 expressed in E. coli XL1-blue cells.Fig. 5. SDS-PAGE analysis of the purification of CYP93C20 expressed in XL1 blue cells. Lanes 1-3, fractions eluted from Ni2+-NTA with 250 mM imidazole; Lane 4, cell supernatant; Lane 5, fraction from gel filtration chromatography; Lane 7, Pre-stained protein standardsFig. 6. Gel filtration chromatography of CYP93C20 on a HiLoad 16/60 Superdex 200 column. Elution was monitored at 280 nm at a flow rate of 1.0 ml/min.Fig. 7. HPLC elution profiles of enzyme assay mixtures. a, yeast microsomes from pYeDP60-IFS 51865 strain incubated with liquiritigenin and NADPH; b, CYP93C20 incubated with liquiritigenin, NADPH and yeast microsome from pYeDP60 control strain (expressing Arabidopsis P450 reductase); c, CYP93C20 incubated with liquiritigenin, NADPH and human NADPH-P450 oxidoreductase; d, Liquiritigenin standard. L, liquiririgenin; D, Daidzein; H, 2,7,4’-trihydroxyisoflavanone.。