• xanthones;
  • phloroglucinols;
  • benzoic acids;
  • flavonoids;
  • phylogeny;
  • secondary metabolism


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References

Benzophenone derivatives, such as polyprenylated benzoylphloroglucinols and xanthones, are biologically active secondary metabolites. The formation of their C13 skeleton is catalyzed by benzophenone synthase (BPS; EC that has been cloned from cell cultures of Hypericum androsaemum. BPS is a novel member of the superfamily of plant polyketide synthases (PKSs), also termed type III PKSs, with 53–63% amino acid sequence identity. Heterologously expressed BPS was a homodimer with a subunit molecular mass of 42.8 kDa. Its preferred starter substrate was benzoyl-CoA that was stepwise condensed with three malonyl-CoAs to give 2,4,6-trihydroxybenzophenone. BPS did not accept activated cinnamic acids as starter molecules. In contrast, recombinant chalcone synthase (CHS; EC from the same cell cultures preferentially used 4-coumaroyl-CoA and also converted CoA esters of benzoic acids. The enzyme shared 60.1% amino acid sequence identity with BPS. In a phylogenetic tree, the two PKSs occurred in different clusters. One cluster was formed by CHSs including the one from H. androsaemum. BPS grouped together with the PKSs that functionally differ from CHS. Site-directed mutagenesis of amino acids shaping the initiation/elongation cavity of CHS yielded a triple mutant (L263M/F265Y/S338G) that preferred benzoyl-CoA over 4-coumaroyl-CoA.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References

Benzophenone derivatives are phenolic natural products with interesting pharmacologic properties. A series of polyprenylated benzophenones, such as guttiferone F (Figure 1), exhibit anti-HIV and antimicrobial activities (Cuesta Rubio et al., 1999; Fuller et al., 1999). Their occurrence is confined to Hypericaceae (Guttiferae) that have proved to be a valuable source of leads to HIV-inhibitory natural products. Structurally more complex benzophenone derivatives are sampsoniones that have been isolated from the Chinese medicinal plant Hypericum sampsonii (Hu and Sim, 2000). They are characterized by unique caged tetracyclic skeletons, and they exhibit cytotoxic activity. Besides polyprenylation, benzophenones can undergo intramolecular cyclization to give xanthones (Figure 1). 2,3′,4,6-Tetrahydroxybenzophenone is converted to 1,3,5- and 1,3,7-trihydroxyxanthones by regioselective oxidative phenol couplings that are catalyzed by cytochrome P450 enzymes (Peters et al., 1998). 1,3,5-Trihydroxyxanthone is the precursor of the antitumor agent psorospermin that occurs in the African plant Psorospermum febrifugum and is active against drug-resistant human leukemia lines and AIDS-related lymphoma (Habib et al., 1987; Kwok and Hurley, 1998). This DNA-alkylating topoisomerase II poison serves as a lead for the design of new drugs for cancer chemotherapy. 1,3,7-Trihydroxyxanthone is the precursor of rubraxanthone that has been isolated from Garcinia mangostana and even surpasses the antibiotic vancomycin in its antibacterial activity against methicillin-resistant strains of Staphylococcus aureus (Iinuma et al., 1996a). Further conversions that benzophenones can undergo are C- and O-glycosylation and dimerization by C–C coupling (Ferrari et al., 2000; Iinuma et al., 1996b; Kitanov and Nedialkov, 2001).


Figure 1. Examples of pharmacologically active benzophenone derivatives and their biogenic relationships.

Download figure to PowerPoint

A crucial reaction in the biosynthesis of benzophenones and xanthones is the formation of the C13 skeleton that is catalyzed by benzophenone synthase (BPS; EC In cell cultures of Hypericum androsaemum, this enzyme stepwise condenses one molecule of benzoyl-CoA with three molecules of malonyl-CoA to give a tetraketide intermediate that is cyclized by intramolecular Claisen condensation into 2,4,6-trihydroxybenzophenone (Figure 2; Schmidt and Beerhues, 1997). The subsequent 3′-hydroxylation is catalyzed by a cytochrome P450 monooxygenase. In cell cultures of Centaurium erythraea, the preferred starter substrate for benzophenone synthase is 3-hydroxybenzoyl-CoA, yielding immediately 2,3′,4,6-tetrahydroxybenzophenone, the substrate for the regioselective cyclizations (Beerhues, 1996).


Figure 2. Branching of the general phenylpropanoid pathway and the benzoic acid biosynthetic route that lead to the formation of chalcone and benzophenone derivatives, respectively. PAL: phenylalanine ammonia-lyase.

Download figure to PowerPoint

Chalcones are the precursors of flavonoids that serve multiple functions in plants. They are flower pigments attracting pollinators, protectants against damaging UV irradiation, phytoalexins against invading pathogens, and signal molecules in plant–microbe interactions (Brouillard, 1988; Caldwell et al., 1983; Dixon and Paiva, 1995; Verma, 1992). The biosynthesis of chalcones is catalyzed by chalcone synthase (CHS; EC that uses 4-coumaroyl-CoA as the starter substrate and three malonyl-CoAs as the extender molecules (Schröder, 1999a). The resulting 2′,4,4′,6′-tetrahydroxychalcone undergoes either enzyme-catalyzed stereospecific or spontaneous random isomerization to the corresponding flavanone naringenin (Jez et al., 2000a).

Here, we report cloning and functional expression of BPS. The molecular and kinetic properties of this novel plant polyketide synthase (PKS) were compared with those of recombinant CHS from the same cell cultures. Furthermore, we constructed a phylogenetic tree and exchanged active site residues between the two enzymes by site-directed mutagenesis.

Results and discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References

Isolation of BPS and CHS cDNAs

Our previous enzymologic studies of xanthone biosynthesis in H. androsaemum were carried out with 5-day-old cell cultures that contained high benzophenone synthase activity (Schmidt and Beerhues, 1997). A cDNA library was constructed using mRNA from 3-day-old cell cultures and screened at low stringency with CHS cDNAs of Arabidopsis thaliana and Petroselinum crispum. This procedure, however, failed to identify any PKS clones, and a polymerase chain reaction (PCR)-based approach was used. Degenerate primers were designed on the basis of three conserved regions of plant PKSs (Figure 3). PCR using the primer pair 1 and 2 and the cDNA library as a template gave a 446 bp fragment whose deduced amino acid sequence exhibited approximately 90% identity with some previously cloned CHSs. The 5′ and 3′ ends were obtained by rapid amplification of cDNA ends (RACE). The 1402 bp full-length cDNA (clone 1) contained a 1173 bp open-reading frame (ORF) encoding a 42.7 kDa protein that consisted of 390 amino acids and had a pI of 6.55. Upstream of the translation start site, there was a stop codon in frame with the ORF, ruling out the usage of a prior start codon.


Figure 3. Alignment of the deduced amino acid sequences of BPS and CHS. Dots indicate amino acid identity, dashes represent gaps, and the asterisk marks the catalytic cysteine. Arrows indicate the positions of the three degenerate primers used for cloning, and boxes mark active site residues that were exchanged for each other by site-directed mutagenesis.

Download figure to PowerPoint

Another experiment used mRNA from 4-day-old cell cultures and the primer pairs 1 and 3. Reverse transcriptase-mediated PCR (RT-PCR) resulted in the amplification of a 651 bp fragment whose deduced amino acid sequence had about 65% identity with some previously cloned PKSs. 5′ and 3′ RACE gave a 1398 bp full-length cDNA (clone 2) that exhibited 63.9% nucleotide sequence identity with clone 1. The 1188 bp ORF encoded a protein that consisted of 395 amino acids and had a molecular mass of 42.8 kDa and a pI of 5.78. Again, a stop codon occurred upstream of the translation start site. The protein shared 60.1% identity with the PKS encoded by clone 1 (Figure 3) and 53–63% identity with other members of the superfamily, with the highest homology being observed with Arabis alpina CHS.

Characterization of recombinant BPS and CHS

The two PKSs were overexpressed in Escherichia coli as glutathione-S-transferase (GST) fusion proteins that were isolated by affinity chromatography and were cleaved directly in front of the PKS start methionine using factor Xa. The mature enzyme encoded by clone 2 preferred benzoyl-CoA as a starter substrate (Table 1). The enzymatic product was identified as 2,4,6-trihydroxybenzophenone by HPLC-MS and co-chromatography with a sample of authentic reference compound. Thus, cDNA 2 encoded BPS. The pH and temperature optima were 6.5–7.0 and 35°C, respectively. The enzyme assays contained 50 µm dithiothreitol (DTT) for maximum BPS activity. Without DTT, the enzyme activity was reduced by approximately 40%. Side products were not detected. The kcat and kcat/Km values of BPS (Table 2) were higher than those reported for other PKSs (Jez et al., 2000b, 2002). Besides benzoyl-CoA, 3-hydroxybenzoyl-CoA and N-methylanthraniloyl-CoA were accepted as starter substrates, but the relative enzyme activities were low (Table 1). The respective products 2,3′,4,6-tetrahydroxybenzophenone and 1,3-dihydroxy-N-methylacridone were identified by co-chromatography (HPLC) with samples of authentic reference substances. BPS did not accept CoA esters of cinnamic acids as the starter molecules. Also, acetyl-CoA was not converted. The Mr of BPS on a calibrated gel filtration column was 85 000, indicating that the enzyme is active as a homodimer. This is also true for the other plant PKSs so far studied (Schröder, 1999a).

Table 1.  Substrate specificities of recombinant BPS and CHS from H. androsaemum cell cultures
SubstrateEnzyme activity (% of maximum each)
Table 2.  Steady-state kinetic parameters for BPS, CHS and the CHS mutant M9 a
 kcat (min−1)Kmm)kcat/Km (m−1 sec−1)kcat (min−1)Kmm)kcat/Km (m−1 sec−1)Kmm)
  • a

    Data are means of two independent experiments.


The enzyme encoded by clone 1 exhibited the highest activity with 4-coumaroyl-CoA and somewhat lower activity with cinnamoyl-CoA as the starter substrates (Table 1). The enzymatic products were identified as naringenin and pinocembrin, respectively, by HPLC-MS and co-chromatography with samples of authentic reference substances. Thus, cDNA 1 encoded CHS. The catalytic efficiency of this PKS was approximately three times lower than that of BPS (Table 2). Optimum assay conditions were pH 7.0, 35°C and 10 µm DTT. Without DTT, CHS activity was reduced by approximately 30%. Bis-noryangonin and 4-coumaroyltriacetic acid lactone occurred as side products; their amounts were 8 and 18%, respectively, of the yield of naringenin. These derailment products result from spontaneous lactonization of the free triketide and tetraketide intermediates after hydrolysis of these polyketides from the active site cysteine by solvent water (Suh et al., 2000). Benzoyl-CoA and 3-hydroxybenzoyl-CoA also served as starter molecules for CHS, leading to the formation of the above benzophenones.

BPS and CHS use the same mechanism of reaction, i.e. multiple decarboxylative condensations with malonyl-CoA followed by cyclization and aromatization of the intermediate tetraketide, but they use different starter substrates. Benzoyl-CoA and 4-coumaroyl-CoA arise biosynthetically from cinnamic acid, itself supplied by the phenylalanine ammonia-lyase (PAL)-catalyzed oxidative deamination of phenylalanine (Figure 2). Cinnamic acid is an important branch point of secondary metabolism in H. androsaemum cell cultures. It is channeled into either the general phenylpropanoid pathway or the benzoic acid biosynthetic route. Recently, the biosynthesis of benzoyl-CoA from cinnamic acid in H. androsaemum cell cultures has been found to be CoA dependent and non-β-oxidative (Abd El-Mawla and Beerhues, 2002). Contrary to this pathway via cinnamic acid, 3-hydroxybenzoic acid in cell cultures of Centaurium erythraea originates directly from an intermediate of the shikimate pathway (Abd El-Mawla et al., 2001).

Benzoyl-CoA, the starter molecule for BPS, is a minor substrate for other plant PKSs. An exception is 2-pyrone synthase (2-PS) from the ornamental plant Gerbera hybrida that, however, performs only two condensations with malonyl-CoA to give the triketide 6-phenyl-4-hydroxy-2-pyrone (Eckermann et al., 1998). As the plant fails to accumulate phenylpyrone derivatives, the physiologic starter substrate for 2-PS appears to be acetyl-CoA and not benzoyl-CoA, resulting in the formation of a methylpyrone. Acetyl-CoA is a poor starter substrate for other PKSs (Schröder, 1999b), including BPS (Table 1). An enzyme that catalyzes only one condensation reaction with malonyl-CoA is benzalacetone synthase (BAS) that is involved in phenylbutanone biosynthesis (Abe et al., 2001).

Phylogenetic analysis

To construct a phylogenetic tree, we selected one CHS each from different families of ferns, gymnosperms, monocots, and dicots and included all types of PKS that functionally differ from CHS (Figure 4). BPS and CHS from H. androsaemum cell cultures occurred in different clusters. One cluster was formed by CHSs including the one from H. androsaemum. In a second cluster, the functionally different CHS-like proteins including BPS grouped together. This cluster also contained Petunia hybrida CHS B that was found to be highly divergent from the remaining Petunia sequences (Durbin et al., 1995). Other pairs of PKSs that were cloned from one species also divided between these two clusters, indicating an ancient duplication and diversification of the ancestral gene. An exception were the enzymes from ferns and gymnosperms that formed individual clusters containing both CHSs and CHS-like enzymes. Stilbene synthase has been reported previously to have evolved independently from CHS several times during evolution (Tropf et al., 1994).


Figure 4. A neighbor-joining tree depicting the relationships of plant PKSs including CHSs from different families and all types of PKS that functionally differ from CHS. Multiple sequence alignment was performed using ClustalX (version 1.81). Numbers at the forks indicate how often the group to the right appeared among 1000 bootstrap replicates. A bacterial type III PKS served as an outgroup. BPS, benzophenone synthase; CHS, chalcone synthase; STS, stilbene synthase; ACS, acridone synthase; BBS, bibenzyl synthase; 2PS, 2-pyrone synthase; CTS, 4-coumaroyl triacetic acid lactone synthase; VAS, valerophenone synthase; BAS, benzalacetone synthase.

Download figure to PowerPoint

We used a type III PKS from Streptomyces griseus as an outgroup. Bacterial CHS-like proteins possess 20–30% amino acid sequence identity with plant enzymes, and they appear to be of common ancestry (Moore and Hopke, 2001). The PKS from S. griseus catalyzes the formation of 1,3,6,8-tetrahydroxynaphthalene from five molecules of malonyl-CoA that serves here both as starter and as extender (Funa et al., 1999). The balhimycin biosynthetic gene cluster of Amycolatopsis mediterranei contains a PKS gene that participates in the biosynthesis of 3,5-dihydroxyphenylacetic acid (Pfeifer et al., 2001). Furthermore, acetylphloroglucinol in Pseudomonas fluorescens has been proposed to result from the condensation of acetoacetyl-CoA with two malonyl-CoAs (Bangera and Thomashow, 1999).

Site-directed mutagenesis

The crystal structures of alfalfa CHS2 and Gerbera hybrida 2-PS that have recently been determined (Ferrer et al., 1999; Jez et al., 2000b) provide a framework for understanding the substrate and product specificities of plant PKSs and facilitate rational engineering of new enzyme activities. Three catalytic residues are highly conserved in both plant and bacterial type III PKSs, including BPS. Cys 164 (numbering in alfalfa CHS2; Figure 3) serves as the nucleophile in the loading reaction and as the attachment site of the polyketide in the elongation reactions. The thiolate anion is stabilized by an ionic interaction with His 303 as an imidazolium cation (Jez and Noel, 2000). His 303 and Asn 336 catalyze the decarboxylation of malonyl-CoA and stabilize the transition state during the condensation steps. The crystal structures, in addition, defined a number of conserved amino acids lining the initiation/elongation cavity that binds the starter molecule and accommodates the growing polyketide chain. Some of these positions vary between BPS and CHS and may account for starter molecule selectivity. Ala260, Met267, Tyr269, and Gly342 of BPS replace Gly256, Leu263, Phe265, and Ser338, respectively, of CHS (Figure 3). These amino acids were exchanged between the two enzymes by site-directed mutagenesis. The resulting proteins carrying either single or multiple mutations were heterologously expressed and functionally analyzed using benzoyl-CoA and 4-coumaroyl-CoA as the starter substrates. The BPS mutants either resembled functionally the wild-type enzyme (G342S, A260G/M267L/Y269F) or lacked the ability for benzophenone and chalcone formation (A260G, A260G/G342S, M267L/Y269F/G342S, A260G/M267L/Y269L/G342S).

In contrast, CHS was converted into an enzyme that exhibited higher activity with benzoyl-CoA than with 4-coumaroyl-CoA (Table 3). This transformation, however, was not achieved by single exchanges of amino acids determining substrate specificity (M1, M2) and polyketide size (M3, M4). An appreciable shift of the starter molecule selectivity towards benzoyl-CoA was observed with the double mutant M8 (F265Y/S338G). Introduction of a third mutation (L263M) further reduced enzyme activity strongly with 4-coumaroyl-CoA, which was mainly caused by a decrease in the kcat value (Table 2). The resulting triple mutant M9 (L263M/F265Y/S338G) preferentially accepted benzoyl-CoA. Structural knowledge of BPS is required to reveal in detail the steric differences between the active site cavities of BPS and CHS. Substitution of the catalytic cysteine (C164 in CHS and C167 in BPS; Figure 3) for alanine inactivated both enzymes, as expected from previous findings (Lanz et al., 1991) and the above crystal structures.

Table 3.  Relative activities (%) of CHS mutants with 4-coumaroyl-CoA and benzoyl-CoA as starter substrates a
  • a

    The activity of wild-type CHS with 4-coumaroyl-CoA was set as 100%.

WTWild type10022.34.5 : 1
M1L263M103.225.74.0 : 1
M2F265Y67.49.47.2 : 1
M3G256A62.815.74.0 : 1
M4S338G101.723.24.4 : 1
M5L263M/F265Y53.914.43.7 : 1
M6G256A/S338G46.815.53.0 : 1
M7L263M/S338G105.327.33.9 : 1
M8F265Y/S338G65.729.62.2 : 1
M9L263M/F265Y/S338G16.431.81.0 : 1.9
M10G256A/L263M/F265Y33.72.712.5 : 1
M11G256A/L263M/F265Y/S338G16.47.62.2 : 1

In alfalfa CHS2, three amino acid substitutions reduced the size of the initiation/elongation cavity, and the resulting triple mutant was functionally identical to 2-PS (Jez et al., 2000b). A single point mutation (F215S) yielded an enzyme that preferentially converted N-methylanthraniloyl-CoA and generated N-methylanthraniloyl triacetic acid lactone, a novel alkaloid (Jez et al., 2002). N-methylanthraniloyl-CoA is not a substrate for wild-type CHS but the preferred starter for acridone synthase (ACS). This enzyme has been converted into CHS by triple mutation (Lukačin et al., 2001). However, the opposite, i.e. generation of ACS from CHS that lacks ACS side activity, appeared to be more difficult, which agrees with our observations of site-directed mutagenesis of BPS that does not exhibit any CHS side activity.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References


Benzoyl-CoA, malonyl-CoA, and acetyl-CoA were purchased from Sigma (Taufkirchen, Germany). 2-Hydroxybenzoyl-CoA, 3-hydroxybenzoyl-CoA, 4-hydroxybenzoyl-CoA, cinnamoyl-CoA, 2-coumaroyl-CoA, and 3-coumaroyl-CoA were synthesized as described by Abd El-Mawla and Beerhues (2002) and Beerhues (1996). 4-Coumaroyl-CoA, caffeoyl-CoA, and feruloyl-CoA were kindly provided by Professor R. Wiermann (Universität Münster, Germany). N-Methylanthraniloyl-CoA and 1,3-dihydroxy-N-methylacridone were kind gifts from Professor U. Matern (Universität Marburg, Germany). Bis-noryangonin was kindly provided by Professor J. Schröder (Universität Freiburg, Germany). 2,4,6-Trihydroxybenzophenone and naringenin were obtained from ICN (Meckenheim, Germany) and Sigma, respectively. 2,3′,4,6-Tetrahydroxybenzophenone was synthesized as described by Peters et al. (1998).

Cell cultures

Cell cultures of H. androsaemum were grown as described by Peters et al. (1998).

cDNA library construction and screening

We used the QuickPrep Micro mRNA Purification Kit (Amersham, Freiburg, Germany) to extract poly(A)-rich RNA from 3-day-old cell cultures (0.2 g) and the TimeSaver cDNA Synthesis Kit (Amersham) to prepare a cDNA library from 5 µg of mRNA. cDNAs were ligated into λExCell NotI/EcoRI/CIP (Amersham) and were packaged using the Gigapack III Gold Packaging Extract (Stratagene, Amsterdam, the Netherlands). The library was screened at low stringency with radioactively labeled CHS cDNAs of P. crispum and A. thaliana using established protocols (Sambrook et al., 1989). The probes were kindly provided by Professor K. Hahlbrock (Max-Planck-Institut für Züchtungsforschung, Köln, Germany).

Isolation of PKS cDNAs and DNA sequencing

Core cDNA fragments were amplified by PCR and RT-PCR using degenerate primers derived from conserved regions of plant PKSs: The forward primer was 5′-ATG ATG TA(CT) CA(AG) CA(AG) GGN TG-3′. The reverse primers were 5′-CCN CCN GG(AG) TGN GC(AG) ATC C-3′ and 5′-AAN CC(AG) AAN A(AG)N CAN CCC C-3′. N represents a mixture of all four nucleotides. PCR was carried out using Taq DNA polymerase (Peqlab, Erlangen, Germany). After denaturation at 95°C (2 min), 30 cycles were performed at 95°C (30 sec), 42°C (30 sec), and 72°C (1 min). The final extension was at 72°C for 15 min. The template for PCR was either the cDNA library or reverse-transcribed mRNA prepared by using the QuickPrep Micro mRNA Purification Kit (Amersham) and reverse transcriptase (Invitrogen, Karlsruhe, Germany). On the basis of the amplified core fragments, gene-specific primers were designed and RACE was performed using the protocol of the SMART RACE cDNA Amplification Kit (Clontech, Heidelberg, Germany). When the cDNA library was the template, the gene-specific primers were paired with the M13 reverse and universal primers located at the λ arms upstream and downstream, respectively, of the cDNA inserts. The PCR program included a 5 min period at 98°C, 30 cycles at 95°C (30 sec), 55°C (30 sec) and 72°C (2 min), and a final 15 min extension at 72°C. Nucleotide sequences were determined from both strands using the BigDye Terminator Cycle Sequencing Kit (Applied Biosystems, Darmstadt, Germany) and the ABI PRISM 377 Genetic Analyzer. Sequence analysis employed the program 1989–99 dnastar (Applied Biosystems).

Expression of PKSs in Escherichia coli

The ORFs of the cDNAs were re-amplified by PCR using Platinum Pfx DNA Polymerase (Invitrogen) and the following primers: 5′-ATG GCC CCG GCG ATG GAG TAC TCA AC-3′ (BPS, forward), 5′-GGG GTA CCT CAC TGG AGA ATT GGG ACA CTC TGG-3′ (BPS, reverse), 5′ATG GTG ACC GTG GAA GAA GTC AGG-3′ (CHS, forward), and 5′-GGG GTA CCC TAG TTA ATG GCG ACA CTG TGA AGG AC-3′ (CHS, reverse). The PCR mixture (50 µl) contained 1× Pfx amplification buffer, 0.2 mm dNTPs (Amersham), 1 mm MgSO4, 15 pmol of each primer, and 2 units of Pfx DNA polymerase. PCR cycling included a 2 min period at 94°C, 25 cycles at 94°C (30 sec) and 68°C (3 min), and a final 10 min extension at 68°C. The amplified ORFs were ligated into an expression vector derived from the pGEX vector series (Goerlach and Schmid, 1996). The restriction sites used were StuI (forward) and KpnI (reverse). Recombinant plasmids were introduced into E. coli BL21-CodonPlus (DE3)-RIL (Stratagene) for overexpression. The bacteria were grown at 200 r.p.m. and 37°C in LB medium (500 ml) containing ampicillin (100 µg ml−1) and chloramphenicol (30 µg ml−1). At an OD600 of 0.6–0.8, IPTG (1 mm) was added and the incubation temperature was reduced to 25°C. After 3–4 h of incubation, the cells were harvested by centrifugation and were re-suspended in 8 ml of 0.1 m potassium phosphate buffer, pH 7.5, containing 1 mm DTT. Sonication of the cells was carried out on ice for 5 min at 50% pulses using a (Branson Sonifier B15, Heinemann, Schwäbisch Gmünd, Germany). After centrifugation at 10 000 g and 4°C for 10 min, an aliquot of the supernatant was subjected to SDS–PAGE and the residual portion applied to an affinity column.

Purification and factor Xa cleavage of fusion proteins

PKSs were heterologously expressed as GST fusion proteins with molecular masses of 69 kDa and were purified on a GSTrap affinity column (1 ml; Amersham). Unbound proteins were removed, and 10 units of factor Xa per milligram of fusion protein was injected into the column. During the incubation at 4°C for 18 h, cleavage occurred directly in front of the PKS start methionine, thereby releasing the mature enzymes without any overhang at the N-terminus. The proteins were eluted with 0.1 mm potassium phosphate buffer, pH 7.0, and the efficiency of the cleavage was assayed by SDS–PAGE. The PKSs gave single bands at Mr of 43 kDa. The final washing step with reduced glutathione (10 mm) released GST only; fusion proteins were no longer detectable.

Enzyme assays and product analysis

Enzyme assays (250 µl) contained 15 µm starter CoA ester, 56 µm malonyl-CoA, 0.1 m potassium phosphate buffer, pH 7.0, 50 µm DTT (BPS) or 10 µm DTT (CHS), and approximately 2 µg protein. After incubation at 35°C for 30 min, the reaction mixture was extracted twice with 250 µl ethyl acetate and was centrifuged at 10 000 g for 10 min. For detection of side products, acetic acid was added prior to extraction to give a final concentration of 5%. The combined organic phases were dried under vacuum, and the residue was dissolved in 50 µl of methanol. Analysis of the enzymatic products was performed by HPLC on a Symmetry C18 5 µm column (4.6 mm × 150 mm; Waters, Eschburn, Germany). The eluents were water (A) and methanol (B) at a flow rate of 1 ml min−1. The following gradients were used: 30% B for 3 min, 30–60% B for 15 min, and 60% B for 2 min (analysis of benzophenones); 30% B for 3 min, 30–80% B for 17 min, and 80% B for 3 min (analysis of flavanones). The detection wavelengths were 289 nm (flavanones, benzoyltriacetic acid lactone), 306 nm (benzophenones), 325 nm (phenylpyrone), 330 nm (4-coumaroyltriacetic acid lactone), and 365 nm (bis-noryangonin). Standard solutions of reference compounds served for quantification.

Alternatively, enzyme assays contained, in addition, [2-14C]malonyl-CoA (0.93 kBq). The extracted products were separated by TLC (silica gel 60 F254; chloroform:methanol:water = 65 : 25 : 4). Detection and quantification were carried out using the TLC radioscanner Rita (Raytest, Germany).

Kinetic data

The Km and kcat values were determined from Michaelis–Menten plots using five different substrate concentrations covering the range of 0.2–3 Km. The concentration of the second substrate was saturating. The incubation time was restricted to 10 min.

Relative molecular mass determination by gel filtration

The recombinant proteins were chromatographed on an FPLC column Superdex 200 HR10/30 (Amersham). The mobile phase was 10 mm potassium phosphate buffer, pH 7.0, the flow rate was 0.5 ml min−1, and the fraction size was 0.2 ml. The calibration proteins were obtained from Roche (Mannheim, Germany)(Combithek 18 000–380 000: chymotrypsinogen A, ovalbumin, bovine serum albumin, aldolase, and catalase).

Site-directed mutagenesis

Point mutations were introduced using the QuikChange XL Site-Directed Mutagenesis Kit (Stratagene) and verified by DNA sequencing. The PKS mutants were expressed in E. coli, as described above. Aliquots of the bacterial crude extracts were subjected to SDS–PAGE, which established that the enzyme mutants were expressed at similar levels and their subunit molecular masses were correct. Purification of enzyme mutants was performed using affinity chromatography, as described above.

Phylogenetic tree construction

Amino acid sequences of type III PKSs were aligned with ClustalX version 1.81 (Thompson et al., 1997) using default parameters. Including gaps, a 469-amino-acid length was aligned. The aligned sequences were first subjected to bootstrapping using the program seqboot in the phylip package (Felsenstein, 1993). Genetic distances within 1000 bootstrap replicates were calculated with the phylip program protdist using a Dayhoff's PAM 001 matrix. The distance matrices were then analyzed with the phylip program neighbor using neighbor-joining algorithm. The multiple data from the above calculation were analyzed with the phylip program consense to obtain the bootstrap values reflecting the consistency of the tree branch pattern. A type III PKS from S. griseus served as an outgroup. The final tree was viewed using TreeView Win32 (Page, 1996).


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References

We thank Professor Ulrich Engelhardt and Dr Gerold Jerz for carrying out HPLC-MS; Professor Daming Zhang for help with the phylogenetic analysis; Professors Ulrich Matern, Joachim Schröder, and Rolf Wiermann for kindly providing CoA esters and reference compounds; and Dr Dietrich Ober for valuable advice. We gratefully acknowledge the gifts of the modified pGEX vector and the heterologous CHS cDNAs from the laboratories of Professors Nikolaus Amrhein and Klaus Hahlbrock, respectively. This work was supported by a postdoctoral fellowship from the Biosciences Special Program of the Deutscher Akademischer Austauschdienst (to B.L.) and a grant from the Deutsche Forschungsgemeinschaft (to L.B.).


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  • Abd El-Mawla, A.M.A. and Beerhues, L. (2002) Benzoic acid biosynthesis in cell cultures of Hypericum androsaemum. Planta, 214, 727733.
  • Abd El-Mawla, A.M.A., Schmidt, W. and Beerhues, L. (2001) Cinnamic acid is a precursor of benzoic acids in cell cultures of Hypericum androsaemum L. but not in cell cultures of Centaurium erythraea RAFN. Planta, 212, 288293.
  • Abe, I., Takahashi, Y., Morita, H. and Noguchi, H. (2001) Benzalacetone synthase. A novel polyketide synthase that plays a crucial role in the biosynthesis of phenylbutanones in Rheum palmatum. Eur. J. Biochem. 268, 33543359.
  • Bangera, M.G. and Thomashow, L.S. (1999) Identification and characterization of a gene cluster for synthesis of the polyketide antibiotic 2,4-diacetylphloroglucinol from Pseudomonas fluorescens Q2–87. J. Bacteriol. 181, 31553163.
  • Beerhues, L. (1996) Benzophenone synthase from cultured cells of Centaurium erythraea. FEBS Lett. 383, 264266.
  • Brouillard, R. (1988) Flavonoids and flower colour. In The Flavonoids (Harborne, J.B., ed.). London: Chapman & Hall, pp. 525538.
  • Caldwell, M.M., Robberecht, R. and Flint, S.D. (1983) Internal filters: prospects for UV-acclimation in higher plants. Physiol. Plant. 58, 445450.
  • Cuesta Rubio, O., Cuellar Cuellar, A., Rojas, N., Velez Castro, H., Rastrelli, L. and Aquino, R. (1999) A polyisoprenylated benzophenone from Cuban propolis. J. Nat. Prod. 62, 10131015.
  • Dixon, R.A. and Paiva, N.L. (1995) Stress-induced phenylpropanoid metabolism. Plant Cell, 7, 10851097.
  • Durbin, M.L., Learn, G.H., Huttley, G.A. and Clegg, M.T. (1995) Evolution of the chalcone synthase gene family in the genus Ipomoea. Proc. Natl. Acad. Sci. USA, 92, 33383342.
  • Eckermann, S., Schröder, G., Schmidt, J. et al. (1998) New pathway to polyketides in plants. Nature, 396, 387390.
  • Felsenstein, J. (1993) Phylip (Phylogeny Inference Package) Version 3.5c. Distributed by the Author. University of Washington, Seattle: Department of Genetics.
  • Ferrer, J.L., Jez, J.M., Bowman, M.E., Dixon, R.A. and Noel, J.P. (1999) Structure of chalcone synthase and the molecular basis of plant polyketide biosynthesis. Nat. Struct. Biol. 6, 775784.
  • Ferrari, J., Terreaux, C., Sahpaz, S., Msonthi, J.D., Wolfender, J.L. and Hostettmann, K. (2000) Benzophenone glycosides from Gnidia involucrata. Phytochemistry, 54, 883889.
  • Fuller, R.W., Blunt, J.W., Boswell, J.L., Cardellina, J.H., II and Boyd, M.R. (1999) Guttiferone F, the first prenylated benzophenone from Allanblackia stuhlmannii. J. Nat. Prod. 62, 130132.
  • Funa, N., Ohnishi, Y., Fujii, I., Shibuya, M., Ebizuka, Y. and Horinouchi, S. (1999) A new pathway for polyketide synthesis in microorganisms. Nature, 400, 897899.
  • Goerlach, J. and Schmid, J. (1996) Introducing StuI sites improves vectors for the expression of fusion proteins with factor Xa cleavage sites. Gene, 170, 145146.
  • Habib, A.M., Ho, D.K., Masuda, S., McCloud, T., Reddy, K.S., Aboushoer, M., McKenzie, A., Byrn, S.R., Chang, C.J. and Cassady, J.M. (1987) Structure and stereochemistry of psorospermin and related cytotoxic dihydrofuranoxanthones from Psorospermum febrifugum. J. Org. Chem. 52, 412418.
  • Hu, L.H. and Sim, K.Y. (2000) Sampsoniones A-M, a unique family of caged polyprenylated benzoylphloroglucinol derivatives, from Hypericum sampsonii. Tetrahedron, 56, 13791386.
  • Iinuma, M., Tosa, H., Tanaka, T., Asai, F., Kobayashi, Y., Shimano, R. and Miyauchi, K.I. (1996a) Antibacterial activity of xanthones from Guttiferaeous plants against methicillin-resistant Staphylococcus aureus. J. Pharm. Pharmacol. 48, 861865.
  • Iinuma, M., Tosa, H., Ito, T., Tanaka, T. and Riswan, S. (1996b) Three new benzophenone-xanthone dimers from the root of Garcinia dulcis. Chem. Pharm. Bull. 44, 17441747.
  • Jez, J.M. and Noel, J.P. (2000) Mechanism of chalcone synthase. pKa of the catalytic cysteine and the role of the conserved histidine in a plant polyketide synthase. J. Biol. Chem. 275, 3964039646.
  • Jez, J.M., Bowman, M.E., Dixon, R. and Noel, J.P. (2000a) Structure and mechanism of the evolutionarily unique plant enzyme chalcone isomerase. Nat. Struct. Biol. 7, 786791.
  • Jez, J.M., Austin, M.B., Ferrer, J.L., Bowman, M.E., Schröder, J. and Noel, J.P. (2000b) Structural control of polyketide formation in plant-specific polyketide synthases. Chem. Biol. 7, 919930.
  • Jez, J.M., Bowman, M.E. and Noel, J.P. (2002) Expanding the biosynthetic repertoire of plant type III polyketide synthases by altering starter molecule specificity. Proc. Natl. Acad. Sci. USA, 99, 53195324.
  • Kitanov, G.M. and Nedialkov, P.T. (2001) Benzophenone O-glucoside, a biogenic precursor of 1,3,7-trioxygenated xanthones in Hypericum annulatum. Phytochemistry, 57, 12371243.
  • Kwok, Y. and Hurley, L.H. (1998) Topoisomerase II site-directed alkylation of DNA by psorospermin and its effect on topoisomerase II-mediated DNA cleavage. J. Biol. Chem. 273, 3302033026.
  • Lanz, T., Tropf, S., Marner, F.J., Schröder, J. and Schröder, G. (1991) The role of cysteines in polyketide synthases: site-directed mutagenesis of resveratrol and chalcone synthases, two enzymes in different plant-specific pathways. J. Biol. Chem. 266, 99719976.
  • Lukačin, R., Schreiner, S. and Matern, U. (2001) Transformation of acridone synthase to chalcone synthase. FEBS Lett. 508, 413417.
  • Moore, B.S. and Hopke, J.N. (2001) Discovery of a new bacterial polyketide biosynthetic pathway. ChemBioChem. 2, 3538.
  • Page, R.D.M. (1996) treeview: an application to display phylogenetic trees on personal computers. Comput. Appl. Biosci. 12, 357358.
  • Peters, S., Schmidt, W. and Beerhues, L. (1998) Regioselective oxidative phenol couplings of 2,3′,4,6-tetrahydroxybenzophenone in cell cultures of Centaurium erythraea RAFN and Hypericum androsaemum L. Planta, 204, 6469.
  • Pfeifer, V., Nicholson, G.J., Ries, J., Recktenwald, J., Schefer, A.B., Shawky, R.M., Schröder, J., Wohlleben, W. and Pelzer, S. (2001) A polyketide synthase in glycopeptide biosynthesis – the biosynthesis of the non-proteinogenic amino acid (S)-3,5-dihydroxyphenylglycine. J. Biol. Chem. 276, 3837038377.
  • Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor: Cold Spring. Harbor Laboratory Press.
  • Schmidt, W. and Beerhues, L. (1997) Alternative pathways of xanthone biosynthesis in cell cultures of Hypericum androsaemum. FEBS Lett. 420, 143146.
  • Schröder, J. (1999a) The chalcone/stilbene synthase-type family of condensing enzymes. In Comprehensive Natural Products Chemistry, Vol. 1 (Sankawa, U., ed.). Amsterdam: Elsevier Science, pp. 749771.
  • Schröder, J. (1999b) Probing plant polyketide biosynthesis. Nat. Struct. Biol. 6, 714716.
  • Suh, D.Y., Fukuma, K., Kagami, J., Yamazaki, Y., Shibuya, M., Ebizuka, Y. and Sankawa, U. (2000) Identification of amino acid residues important in the cyclization reactions of chalcone and stilbene synthases. Biochem. J. 350, 229235.
  • Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. and Higgins, D.G. (1997) The ClustalX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucl. Acids Res. 24, 48764882.
  • Tropf, S., Lanz, T., Rensing, S.A., Schröder, J. and Schröder, G. (1994) Evidence that stilbene synthases have developed from chalcone synthases several times in the course of evolution. J. Mol. Evol. 38, 610618.
  • Verma, D.P.S. (1992) Signals in root nodule organogenesis and endocytosis of Rhizobium. Plant Cell, 4, 373382.

Accession numbers: AF352395 (BPS from Hypericum androsaemum), AF315345 (CHS from Hypericum androsaemum).