Combinatorial biosynthesis of flavones and flavonols in Escherichia coli

(2S)-Flavanones (naringenin and pinocembrin) are key intermediates in the flavonoid biosynthetic pathway in plants. Recombinant Escherichia coli cells containing four genes for a phenylalanine ammonia-lyase, cinnamate/coumarate:CoA ligase, chalcone synthase, and chalcone isomerase, in addition to the acetyl-CoA carboxylase, have been established for efficient production of (2S)-naringenin from tyrosine and (2S)-pinocembrin from phenylalanine. Further introduction of the flavone synthase I gene from Petroselinum crispum under the control of the T7 promoter and the synthetic ribosome-binding sequence in pACYCDuet-1 caused the E. coli cells to produce flavones: apigenin (13 mg/l) from tyrosine and chrysin (9.4 mg/l) from phenylalanine. Introduction into the E. coli cells of the flavanone 3β-hydroxylase and flavonol synthase genes from the plant Citrus species led to production of flavonols: kaempferol (15.1 mg/l) from tyrosine and galangin (1.1 mg/l) from phenylalanine. The combinatorial biosynthesis of the flavones and flavonols in E. coli is promising for the construction of a library of various flavonoid compounds and un-natural flavonoids in bacteria.     

Heterologous production of flavanones in<Emphasis Type="Italic"> Escherichia coli</Emphasis>: potential for combinatorial biosynthesis of flavonoids in bacteria

Chalcones, the central precursor of flavonoids, are synthesized exclusively in plants from tyrosine and phenylalanine via the sequential reaction of phenylalanine ammonia-lyase (PAL), cinnamate-4-hydroxylase (C4H), 4-coumarate:coenzyme A ligase (4CL) and chalcone synthase (CHS). Chalcones are converted into the corresponding flavanones by the action of chalcone isomerase (CHI), or non-enzymatically under alkaline conditions. PAL from the yeast Rhodotorula rubra, 4CL from an actinomycete Streptomyces coelicolor A3(2), and CHS from a licorice plant Glycyrrhiza echinata, assembled as artificial gene clusters in different organizations, were used for fermentation production of flavanones in Escherichia coli. Because the bacterial 4CL enzyme attaches CoA to both cinnamic acid and 4-coumaric acid, the designed biosynthetic pathway bypassed the C4H step. E. coli carrying one of the designed gene clusters produced about 450 μg naringenin/l from tyrosine and 750 μg pinocembrin/l from phenylalanine. The successful production of plant-specific flavanones in bacteria demonstrates the usefulness of combinatorial biosynthesis approaches not only for the production of various compounds of plant and animal origin but also for the construction of libraries of "unnatural" natural compounds. Dedicated to Professor Sir David Hopwood.     

Efficient production of (2S)-flavanones by Escherichia coli containing an artificial biosynthetic gene cluster

For the fermentative production of plant-specific flavanones (naringenin, pinocembrin) by Escherichia coli, a plasmid was constructed which carried an artificial biosynthetic gene cluster, including PAL encoding a phenylalanine ammonia-lyase from a yeast, ScCCL encoding a cinnamate/coumarate:CoA ligase from the actinomycete Streptomyces coelicolor A3(2), CHS encoding a chalcone synthase from a licorice plant and CHI encoding a chalcone isomerase from the Pueraria plant. The recombinant E. coli cells produced (2S)-naringenin from tyrosine and (2S)-pinocembrin from phenylalanine. When the two subunit genes of acetyl-CoA carboxylase from Corynebacterium glutamicum were expressed under the control of the T7 promoter and the ribosome-binding sequence in the recombinant E. coli cells, the flavanone yields were greatly increased, probably because enhanced expression of acetyl-CoA carboxylase increased a pool of malonyl-CoA that was available for flavanone synthesis. Under cultural conditions where E. coli at a cell density of 50 g/l was incubated in the presence of 3 mM tyrosine or phenylalanine, the yields of naringenin and pinocembrin reached about 60 mg/l. The fermentative production of flavanones in E. coli is the first step in the construction of a library of flavonoid compounds and un-natural flavonoids in bacteria.     

One-pot synthesis of genistein from tyrosine by coincubation of genetically engineered <Emphasis Type="Italic">Escherichia coli</Emphasis> and <Emphasis Type="Italic">Saccharomyces cerevisiae</Emphasis> cells

For production of genistein from N-acetylcysteamine-attached p-coumarate (p-coumaroyl-NAC) supplemented to the medium, a chalcone synthase (CHS) gene from Glycyrrhiza echinata, a chalcone isomerase (CHI) gene from Pueraria lobata, and an isoflavone synthase (IFS) gene from G. echinata were placed under the control of the galactose-inducible GAL promoters in pESC vector and were introduced in Saccharomyces cerevisiae. When the recombinant yeast cells (0.5 g wet weight) were used as “enzyme bags” and incubated at 30°C for 48 h in 100 ml of the buffer containing galactose and 1 mM (265 mg/l) p-coumaroyl-NAC, ca. 340 μg genistein/l was produced. Another system consisting of two enzyme bags was also generated for the purpose of production of genistein from tyrosine. One enzyme bag was an Escherichia coli cell containing a phenylalanine ammonia-lyase gene from a yeast, a 4-coumarate/cinnamate:CoA ligase gene from the actinomycete Streptomyces coelicolor A3(2), the CHS gene, and the CHI gene, in addition to the acetyl-CoA carboxylase gene from Corynebacterium glutamicum, all of which were under the control of the isopropyl-β-d-thiogalactopyranoside-inducible T7 promoter, and thus producing (S)-naringenin from tyrosine. The other enzyme bag was a S. cerevisiae cell containing the IFS gene. Coincubation of the E. coli cells (0.5 g wet weight) and S. cerevisiae cells (0.5 g wet weight) at 26°C for 60 h in 20 ml of the buffer containing 3 mM (543 mg/l) tyrosine as the starting substrate yielded ca. 6 mg genistein/l.     

Investigation of Two Distinct Flavone Synthases for Plant-Specific Flavone Biosynthesis in Saccharomyces cerevisiae

Flavones are plant secondary metabolites that have wide pharmaceutical and nutraceutical applications. We previously constructed a recombinant flavanone pathway by expressing in Saccharomyces cerevisiae a four-step recombinant pathway that consists of cinnamate-4 hydroxylase, 4-coumaroyl:coenzyme A ligase, chalcone synthase, and chalcone isomerase. In the present work, the biosynthesis of flavones by two distinct flavone synthases was evaluated by introducing a soluble flavone synthase I (FSI) and a membrane-bound flavone synthase II (FSII) into the flavanone-producing recombinant yeast strain. The resulting recombinant strains were able to convert various phenylpropanoid acid precursors into the flavone molecules chrysin, apigenin, and luteolin, and the intermediate flavanones pinocembrin, naringenin, and eriodictyol accumulated in the medium. Improvement of flavone biosynthesis was achieved by overexpressing the yeast P450 reductase CPR1 in the FSII-expressing recombinant strain and by using acetate rather than glucose or raffinose as the carbon source. Overall, the FSI-expressing recombinant strain produced 50% more apigenin and six times less naringenin than the FSII-expressing recombinant strain when p-coumaric acid was used as a precursor phenylpropanoid acid. Further experiments indicated that unlike luteolin, the 5,7,4′-trihydroxyflavone apigenin inhibits flavanone biosynthesis in vivo in a nonlinear, dose-dependent manner.     

Polymyxin B as inhibitor of LPS contamination of Schistosoma mansoni recombinant proteins in human cytokine analysis

Background  

Recombinant proteins expressed in Escherichia coli vectors are generally contaminated with endotoxin. In this study, we evaluated the ability of Polymyxin B to neutralize the effect of LPS present as contaminant on Schistosoma mansoni recombinant proteins produced in E. coli in inducing TNF-α and IL-10. Peripheral blood mononuclear cells from individuals chronically infected with S. mansoni were stimulated in vitro with recombinant Sm22.6, Sm14 and P24 antigens (10 μg/mL) in the presence of Polymyxin B (10 μg/mL).     

Optimization of glucose feeding approaches for enhanced glucosamine and N-acetylglucosamine production by an engineered Escherichia coli

In this work, a recombinant Escherichia coli was constructed by overexpressing glucosamine (GlcN) synthase and GlcN-6-P N-acetyltransferase for highly efficient production of GlcN and N-acetylglucosamine (GlcNAc). For further enhancement of GlcN and GlcNAc production, the effects of different glucose feeding strategies including constant-rate feeding, interval feeding, and exponential feeding on GlcN and GlcNAc production were investigated. The results indicated that exponential feeding resulted in relatively high cell growth rate and low acetate formation rate, while constant feeding contributed to the highest specific GlcN and GlcNAc production rate. Based on this, a multistage glucose supply approach was proposed to enhance GlcN and GlcNAc production. In the first stage (0–2 h), batch culture with initial glucose concentration of 27 g/l was conducted, whereas the second culture stage (2–10 h) was performed with exponential feeding at μ _set = 0.20 h⁻¹, followed by feeding concentrated glucose (300 g/l) at constant rate of 32 ml/h in the third stage (10–16 h). With this time-variant glucose feeding strategy, the total GlcN and GlcNAc yield reached 69.66 g/l, which was enhanced by 1.59-fold in comparison with that of batch culture with the same total glucose concentration. The time-dependent glucose feeding approach developed here may be useful for production of other fine chemicals by recombinant E. coli.     

Characterization of dihydroflavonol 4-reductases for recombinant plant pigment biosynthesis applications

Anthocyanins are colorful plant pigments with promising applications as pharmaceuticals and colorants. In order to engineer efficient pigment biosynthesis in Escherichia coli, the activities of various dihydroflavonol 4-reductases (DFRs) were characterized for the three primary dihydroflavonol substrates. The biochemical assays demonstrated variable DFR activities for dihydroflavonol with one B-ring hydroxyl group, the precursor of pelargonidin derivatives. In contrast, dihydroflavonols with two and three B-ring hydroxylation were metabolized with comparable efficiency. Furthermore, the catalysis of DFR for the secondary substrates, flavanones, also depended on the number of B-ring hydroxyl groups. Engineering the expression of the DFR clones together with plant-specific 4-coumaroyl:CoA ligase, chalcone synthase, chalcone isomerase, and flavanone 3-hydroxylase in E. coli resulted in the synthesis of pelargonidin at various levels, from p-coumaric acids. The identification of a robust DFR from this study can also be used for engineering recombinant synthesis of other bioactive flavonoids, such as flavan-3-ols.     

Study of the Action of Se and Cu on the Growth Metabolism of <Emphasis Type="Italic">Escherichia coli</Emphasis> by Microcalorimetry

The biological effect of Se and Cu²⁺ on Escherichia coli (E. coli) growth was studied by using a 3114/3236 TAM Air Isothermal Calorimeter, ampoule method, at 37°C. From the thermogenesis curves, the thermokinetic equations were established under different conditions. The kinetics showed that a low concentration of Se (1–10 μg/mL) promoted the growth of E. coli, and a high concentration of Se (>10 μg/mL) inhibited the growth, but the Cu²⁺ was always inhibiting the growth of E. coli. Moreover, there was an antagonistic or positive synergistic effect of Se and Cu²⁺ on E. coli in the different culture medium when Se was 1–10 μg/ml and Cu²⁺ was 1–20 μg/ml. There was a negative synergistic effect of Se and Cu²⁺ on E. coli when Se was higher than 10 μg/ml and Cu²⁺ was higher than 20 μg/ml. The antagonistic or synergistic effect between Se and Cu²⁺ on E. coli was related to the formation of Cu–Se complexes under the different experimental conditions chosen.     

Enhanced <Emphasis Type="SmallCaps">l</Emphasis>-phenylalanine production by recombinant <Emphasis Type="Italic">Escherichia coli</Emphasis> BR-42 (pAP-B03) resistant to bacteriophage BP-1 via a two-stage feeding approach

The l-phenylalanine (l-Phe) production by Escherichia coli WSH-Z06 (pAP-B03) was frequently prevented by bacteriophage BP-1 infestation. To cope with the bacteriophage BP-1 problem for an improved l-Phe production, one bacteriophage BP-1-resistant mutant, E. coli BR-42, was obtained from 416 mutant colonies of E. coli WSH-Z06 after N-methyl-N’-nitro-N-nitrosoguanidine (NTG) mutagenesis by selection for resistance to bacteriophage BP-1. The recombinant E. coli BR-42-carrying plasmid pAP-B03 had a high capacity in l-Phe production and a remarkable tolerance to 1 × 10¹⁰ pfu (plaque-forming unit)/ml bacteriophage stock. For an enhanced l-Phe production by E. coli BR-42 (pAP-B03), the effects of different feeding strategies including pH–stat, constant rate feeding, linear decreasing rate feeding, and exponential feeding on l-Phe production were investigated; and a two-stage feeding strategy, namely exponential feeding at μ _set = 0.18 h⁻¹ in the first 20 h and a following linear varying rate feeding with F = (−0.55 × t + 18.6) ml/h, was developed to improve l-Phe production. With this two-stage feeding approach, a maximum l-Phe titer of 57.63 g/l with a high l-Phe productivity (1.15 g/l/h) was achieved, which was 15% higher than the highest level (50 g/l) reported so far according to our knowledge. The recombinant E. coli BR-42 (pAP-B03) is a potential l-Phe over-producer in substantial prevention of bacteriophage BP-1 infestation compared to its parent strain WSH-Z06 (pAP-B03).     

Enhanced production of ɛ-caprolactone by overexpression of NADPH-regenerating glucose 6-phosphate dehydrogenase in recombinant <Emphasis Type="Italic">Escherichia coli</Emphasis> harboring cyclohexanone monooxygenase gene

Whole-cell conversion of cyclohexanone to ɛ-caprolactone was attempted by recombinant Escherichia coli BL21(DE3) expressing cyclohexanone monooxygenase (CHMO) of Acinetobacter calcoaceticus NCIMB 9871. High concentrations of cyclohexanone and ɛ-caprolactone reduced CHMO-mediated bioconversion of cyclohexanone to ɛ-caprolactone in the resting recombinant E. coli cells. Metabolically active cells were employed by adopting a fed-batch culture to improve the production of ɛ-caprolactone from cyclohexanone. A glucose-limited fed-batch Baeyer–Villiger oxidation where a cyclohexanone level was maintained less than 6 g/l resulted in a maximum ɛ-caprolactone concentration of 11.0 g/l. The maximum ɛ-caprolactone concentration was improved further to 15.3 g/l by coexpression of glucose-6-phosphate dehydrogenase, an NADPH-generating enzyme encoded by the zwf gene which corresponded to a 39% enhancement in ɛ-caprolactone concentration compared with the control experiment performed under the same conditions.     

Cis-2′, 3′-Dihydrodiol production on flavone B-Ring by Biphenyl Dioxygenase from Pseudomonas pseudoalcaligenes KF707 expressed in Escherichia coli

Escherichia coli JM109 strains expressing either toluene dioxygenase from Pseudomonas putida F1 or biphenyl dioxygenase from Pseudomonas pseudoalcaligenes KF707 were examined for their ability to catalyze flavones. Biphenyl dioxygenase produced metabolites from flavone and 5,7-dihydroxyflavone which were not found in the control experiments. The absorption maxima of UV-visible spectra for the metabolites from flavone and 5,7-dihydroxyflavone were found at 337 and 348 nm respectively by using a photodiode array detector in the HPLC. Liquid chromatography/mass spectroscopy (LC/MS) showed molecular weights 256 and 288 for the metabolites, respectively. The metabolite of flavone, which was isolated and purified from the bacterial culture, was further subjected to analysis by ¹H and ¹³C nuclear magnetic resonance (NMR) spectroscopy. Based on the LC/MS and NMR results, biphenyl dioxygenase inserted oxygen at C2′ and C3′ on the B-ring of flavone, resulting in the formation of flavone cis-2′, 3′-dihydrodiol (2-[3,4-dihydroxy-1.5-cyclohexadienyl]-4H-chromen-4-one). Since this product is not found in Chemical Abstracts, this compound is considered a novel one. In addition, biotransformation of flavones by biphenyl dioxygenase suggested a potential role of bacterial dioxygenase to synthesize novel compounds from plant secondary metabolites. This revised version was published online in June 2006 with corrections to the Cover Date.     

<Emphasis Type="Italic">Escherichia coli</Emphasis> Heat Stable (STa) Enterotoxin and the Upper Small Intestine: Lack of Evidence in Vivo for Net Fluid Secretion

Heat stable (STa) enterotoxin from E. coli reduced fluid absorption in vivo in the perfused jejunum of the anaesthetized rat in Krebs-phosphate buffer containing lactate and glucose (nutrient buffer), in glucose saline and in glucose free saline. Bicarbonate ion enhanced fluid absorption of 98 ± 7 (6) μl/cm/h was very significantly (P < 0.0001) reduced by STa to 19 ± 4 (6) μl/cm/h, but net secretion was not found. When impermeant MES substituted for bicarbonate ion, net fluid absorption of 29 ± 3 (6) μl/cm/h was less (P < 0.01) than the values for phosphate buffer and bicarbonate buffer. With STa in MES buffer, fluid absorption of 3 ± 2 (6) μl/cm/h was less than (P < 0.001) that in the absence of STa and not significantly different from zero net fluid absorption. E. coli STa did not cause net fluid secretion in vivo under any of the above circumstances. Neither bumetanide nor NPPB when co-perfused with STa restored the rate of fluid absorption. In experiments with zero sodium ion-containing perfusates, STa further reduced fluid absorption modestly by 20 μl/cm/h. Perfusion of ethyl-isopropyl-amiloride (EIPA) with STa in zero sodium ion buffers prevented the small increment in fluid entry into the lumen caused by STa, indicating that the STa effect was attributable to residual sodium ion and fluid uptake that zero sodium-ion perfusates did not eradicate. These experiments, using a technique that directly measures mass transport of fluid into and out of the in vivo proximal jejunum, do not support the concept that E. coli STa acts by stimulating a secretory response.     

Metabolic engineering of carotenoid accumulation in Escherichia coli by modulation of the isoprenoid precursor pool with expression of deoxyxylulose phosphate synthase

The recently discovered non-mevalonate pathway to isoprenoids, which uses glycolytic intermediates, has been modulated by overexpression of Escherichia coli d-1-deoxyxylulose 5-phosphate synthase (DXS) to increase deoxyxylulose 5-phosphate and, consequently, increase the isoprenoid precursor pool in E. coli. Carotenoids are a large class of biologically important compounds synthesized from isoprenoid precursors and of interest for metabolic engineering. However, carotenoids are not ordinarily present in E. coli. Co-overexpression of E. coli dxs with Erwinia uredovora gene clusters encoding carotenoid biosynthetic enzymes led to an increased accumulation of the carotenoids lycopene or zeaxanthin over controls not expressing DXS. Thus, rate-controlling enzymes encoded by the carotenogenic gene clusters are responsive to an increase in isoprenoid precursor pools. Levels of accumulated carotenoids were increased up to 10.8 times the levels of controls not overexpressing DXS. Lycopene accumulated to a level as high as 1333 μg/g dw and zeaxanthin accumulated to a level as high as 592 μg/g dw, when pigments were extracted from colonies. Zeaxanthin-producing colonies grew about twice as fast as lycopene-producing colonies throughout a time course of 11 days. Metabolic engineering of carbon flow from simple glucose metabolites to representatives of the largest class of natural products was demonstrated in this model system. Received: 6 August 1999 / Received revision: 25 October 1999 / Accepted: 5 November 1999     

Efficient secretory production of alkaline phosphatase by high cell density culture of recombinant Escherichia coli using the Bacillus sp. endoxylanase signal sequence

New secretion vectors containing the Bacillus sp. endoxylanase signal sequence were constructed for the secretory production of recombinant proteins in Escherichia coli. The E. coli alkaline phosphatase structural gene fused to the endoxylanase signal sequence was expressed from the trc promoter in various E. coli strains by induction with IPTG. Among those tested, E. coli HB101 showed the highest efficiency of secretion (up to 25.3% of total proteins). When cells were induced with 1 mM IPTG, most of the secreted alkaline phosphatase formed inclusion bodies in the periplasm. However, alkaline phosphatase could be produced as a soluble form without reduction of expression level by inducing with less (0.01 mM) IPTG, and greater than 90% of alkaline phosphatase could be recovered from the periplasm by the simple osmotic shock method. Fed-batch cultures were carried out to examine the possibility of secretory protein production at high cell density. Up to 5.2 g/l soluble alkaline phosphatase could be produced in the periplasm by the pH-stat fed-batch cultivation of E. coli HB101 harboring pTrcS1PhoA. These results demonstrate the possibility of efficient secretory production of recombinant proteins in E. coli by high cell density cultivation. Received: 8 September 1999 / Received revision: 3 January 2000 / Accepted 4 January 2000     

Purification and characterization of novel antifungal peptide,mouse beta defensin-1, in Escherichia coli

Mouse beta defensin-1 (mBD-1) is a cationic peptide with broad antimicrobial activity. The mBD-1 gene was cloned and fused with TrxA to construct pET32-mBD1, which was transformed into E. coli BL21 (DE3). The optimal expression conditions of fusion protein TrxA–mBD1 were: cultivation at 32°C in 2 × YT medium, induction with 0.2 mM isopropylthio-d-galactoside (IPTG), and post-induction expression for 8 h. The fusion protein was highly soluble (90.0%) and accounted for 65% of the total soluble protein; and its volumetric productivity reached 0.67 g/l, i.e., 0.14 g/l of recombinant mBD-1. At 5 μM, purified recombinant mBD-1 killed 50% of Candida albicans. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

The use of EDTA or polymyxin with lysozyme for the recovery of intracellular products fromEscherichia. coli

Summary Escherichia coli cells taken from exponential and late stationary (or decline) phases of culture were very susceptible to lysis by EDTA/lysozyme. Log phase cells were most susceptible to lysis by polymyxin/lysozyme. Treatment ofE. coli with EDTA and lysozyme compared favourably with sonication as a method for release of intracellular protein. Concentration ranges for optimal lysis were 100–800 μg/ml for EDTA and 25–50 μg/ml for lysozyme.     

Production and Characterization of a Single-Chain Fv Antibody–Alkaline Phosphatase Fusion Protein Specific for Clenbuterol

The production and characterization of an anti-clenbuterol single-chain Fv antibody (CBLscFv)–bacterial alkaline phosphatase (AP) fusion protein are described. The CBLscFv and the phoA gene of Escherichia coli strain K12 chromosomal DNA were cloned by PCR and sequentially inserted into the expression vector pBV220 to express the CBLscFv–AP fusion protein in E. coli strain BL21(DE3)pLysS. SDS–PAGE and western blot analyses revealed that the fusion protein showed a molecular weight of 73 kDa and bound with the antibacterial AP monoclonal antibody. Determination of enzymatic activity indicated that k _cat and K _m values of the fusion protein were 113.60 s⁻¹ and 29.82 μM, respectively. Competitive direct enzyme-linked immunosorbent assay based on the obtained fusion protein indicated that the average concentration required for 50% inhibition of binding (IC₅₀) and the limit of detection for CBL were 4.74 ± 0.003 (n = 3) and 0.54 ± 0.004 (n = 3) μg/l, respectively, and the linear response range extended from 1.13 to 69.68 μg/l. Cross-reactivity studies showed that the fusion protein did not cross-react with CBL analogs. The present findings indicate that the production of the CBLscFv–AP fusion protein in E. coli strain BL21(DE3)pLysS is feasible and suggest that it could be further used to develop a one-step ELISA for the specific detection of CBL.     

Investigation of two distinct flavone synthases for plant-specific flavone biosynthesis in Saccharomyces cerevisiae

Flavones are plant secondary metabolites that have wide pharmaceutical and nutraceutical applications. We previously constructed a recombinant flavanone pathway by expressing in Saccharomyces cerevisiae a four-step recombinant pathway that consists of cinnamate-4 hydroxylase, 4-coumaroyl:coenzyme A ligase, chalcone synthase, and chalcone isomerase. In the present work, the biosynthesis of flavones by two distinct flavone synthases was evaluated by introducing a soluble flavone synthase I (FSI) and a membrane-bound flavone synthase II (FSII) into the flavanone-producing recombinant yeast strain. The resulting recombinant strains were able to convert various phenylpropanoid acid precursors into the flavone molecules chrysin, apigenin, and luteolin, and the intermediate flavanones pinocembrin, naringenin, and eriodictyol accumulated in the medium. Improvement of flavone biosynthesis was achieved by overexpressing the yeast P450 reductase CPR1 in the FSII-expressing recombinant strain and by using acetate rather than glucose or raffinose as the carbon source. Overall, the FSI-expressing recombinant strain produced 50% more apigenin and six times less naringenin than the FSII-expressing recombinant strain when p-coumaric acid was used as a precursor phenylpropanoid acid. Further experiments indicated that unlike luteolin, the 5,7,4'-trihydroxyflavone apigenin inhibits flavanone biosynthesis in vivo in a nonlinear, dose-dependent manner.