A New Group of Aromatic Prenyltransferases in Fungi, Catalyzing a 2,7-Dihydroxynaphthalene 3-Dimethylallyl-transferase Reaction

Abstract

Five fungal genomes from the Ascomycota (sac fungi) were found to contain a gene with sequence similarity to a recently discovered small group of bacterial prenyltransferases that catalyze the C-prenylation of aromatic substrates in secondary metabolism. The genes from Aspergillus terreus NIH2624, Botryotinia fuckeliana B05.10 and Sclerotinia sclerotiorum 1980 were expressed in Escherichia coli, and the resulting His₈-tagged proteins were purified and investigated biochemically. Their substrate specificity was found to be different from that of any other prenyltransferase investigated previously. Using 2,7-dihydroxynaphthalene (2,7-DHN) and dimethylallyl diphosphate as substrates, they catalyzed a regiospecific Friedel-Crafts alkylation of 2,7-DHN at position 3. Using the enzyme of A. terreus, the K_m values for 2,7-DHN and dimethylallyl diphosphate were determined as 324 ± 25 μm and 325 ± 35 μm, respectively, and k_cat as 0.026 ± 0.001 s⁻¹. A significantly lower level of prenylation activity was found using dihydrophenazine-1-carboxylic acid as aromatic substrate, and only traces of products were detected with aspulvinone E, flaviolin, or 4-hydroxybenzoic acid. No product was formed with l-tryptophan, l-tyrosine, or 4-hydroxyphenylpyruvate. The genes for these fungal prenyltransferases are not located within recognizable secondary metabolic gene clusters. Their physiological function is yet unknown.

Introduction

Recently, a new family of prenyltransferases has been discovered in bacteria of the genus Streptomyces (1–4). They are involved in the biosynthesis of secondary metabolites, e.g. antibiotics. The members of this family are characterized by a new protein fold, termed PT barrel (1). The PT barrel consists of five repetitive ααββ elements, with the β strands forming a central barrel. In contrast to the TIM barrel, the PT barrel consists of 10 antiparallel β strands and contains the active site of the enzyme in its spacious lumen. The members of this enzyme family catalyze Friedel-Crafts alkylations of aromatic substrates, i.e. the formation of carbon-carbon-bonds between C-1 or C-3 of the isoprenoid substrate and an aromatic carbon of the acceptor substrate. In contrast to the membrane-bound aromatic prenyltransferases of ubiquinone, menaquinone, and plastoquinone biosynthesis (5), the enzymes containing the PT barrel are soluble proteins and do not contain (N/D)DxxD motifs for binding of the isoprenoid substrate.

In fungi, the prenylation of aromatic substrates leads to a large and important class of secondary metabolites, the prenylated indole alkaloids (6). Also, the fungal indole prenyltransferases are soluble enzymes without (N/D)DxxD motifs (7). Very recently, we solved the first structure of a fungal indole prenyltransferase, i.e. of the dimethylallyltryptophan synthase. This enzyme also showed the PT barrel structure (8). For prenyltransferases characterized by the PT barrel fold, the name ABBA prenyltransferases has been suggested (3). From the currently available data, the ABBA prenyltransferases can be divided into two groups. In the dimethylallyltryptophan synthase/LtxC group, PSI-BLAST searches reveal more than a hundred entries in the database with sequence similarity to dimethylallyltryptophan synthase, mostly in fungal genomes but also some bacterial enzymes, e.g. LtxC (9) and CymD (10). Typically, the members of this group catalyze the prenylation of indole moieties (7). In the CloQ/NphB group, PSI-BLAST searches reveal 12 entries within the bacterial genus Streptomyces (and five fungal genes discussed below) that show sequence similarity to CloQ of clorobiocin biosynthesis (2) and to NphB of naphterpin biosynthesis (1) (see Fig. 1). The currently known members of this group catalyze the prenylation of phenols and phenazine derivatives (4). Sequence similarity between these two groups cannot be detected with PSI-BLAST searches but can be detected with more powerful bioinformatic techniques like HHpred (11), indicating a distant evolutionary relationship. Structure predictions (12, 13) suggest that all members of the two groups exhibit the PT barrel fold.

FIGURE 1.

Reactions catalyzed by the bacterial enzymes CloQ, NphB, and SCO7190 and by the fungal aspulvinone dimethylallyltransferase. The genuine aromatic substrate of NphB in the biosynthesis of naphterpin is unknown; 2,7-dihydroxynaphthalene is one of the artificial substrates accepted in vitro. GPP, geranyl diphosphate.

In the last years, genome-sequencing projects of fungi of the subphylum Pezizomycotina (the largest group of the phylum Ascomycota, i.e. sac fungi) revealed in five different species, including A. terreus, a gene with no sequence similarity to any other fungal gene but with obvious similarity to the bacterial ABBA prenyltransferases of the CloQ/NphB group (3). No data have been published on their possible function. However, 30 years ago, a soluble aromatic prenyltransferase was identified in A. terreus and purified to apparent homogeneity (14, 15). It was functionally identified as aspulvinone dimethylallyltransferase, catalyzing a reaction similar to that of the bacterial ABBA prenyltransferase CloQ (Fig. 1). Yet, the gene coding for aspulvinone dimethylallyltransferase has not been identified. In the present study, we expressed the A. terreus gene with sequence similarity to CloQ as well as two orthologs from other fungal species, purified the resulting proteins, and investigated them biochemically to determine whether they indeed are aromatic prenyltransferases and whether they may represent the aspulvinone dimethylallyltransferase.

EXPERIMENTAL PROCEDURES

Chemicals

DMAPP² and geranyl diphosphate were synthesized according to Woodside et al. (16). Flaviolin was prepared as described by Gross et al. (17). Dihydrophenazine-1-carboxylic acid was generated as described by Saleh et al. (18). Aspulvinone E was synthesized according to Bernier and Brückner (19). IPTG, Tris, NaCl, glycerol, dithiothreitol, MgCl₂, formic acid, sodium dodecyl sulfate, polyacrylamide, and EDTA were bought at Carl Roth, Karlsruhe, Germany. 1,6-Dihydroxynaphthalene, 4-hydroxyphenyl-pyruvate, 4-hydroxybenzoic acid, TAPS, methanol, Tween 20, and imidazole were bought at Sigma Aldrich, Steinheim, Germany. 2,7-Dihydroxynaphthalene and 1,3-dihydroxynaphthalene were bought at Acros Organics. Merck delivered dipotassium hydrogen phosphate, potassium dihydrogen phosphate, β-mercaptoethanol, sodium ascorbate, methanol D₄, l-tryptophan, and l-tyrosine. Lysozyme was bought at Boehringer Ingelheim, Heidelberg, Germany.

Protein Expression and Purification

The nucleotide sequences of ptf_Ss, ptf_Bf, and the truncated ptf_Pm-S (lacking the coding sequence for the first 57 amino acids of XP_002143864) were optimized for expression in Escherichia coli with the Gene Designer Tool and synthesized commercially by DNA2.0 (Basel, Switzerland). All three genes were excised from vector pJ201 (DNA2.0) with NcoI and EcoRI and ligated into vector pHis8 (20) using the same restriction sites.

Likewise, the sequence of ptf_At was optimized with the Gene Designer Tool and was synthesized including the His₈ tag and the linker region as found in pHis8 (20). It was delivered in the expression vector pJExpress411 (DNA2.0) and used without further cloning. The His₈ tag and linker region were thereby identical in all four constructs.

35 ml of an overnight culture in Luria-Bertani medium (50 μg ml⁻¹ kanamycin, 25 μg ml⁻¹ chloramphenicol) of E. coli Rosetta 2 (DE3) pLysS cells harboring one of the plasmids were used to inoculate 1 liter of terrific broth (21) (50 μg ml⁻¹ kanamycin, 25 μg ml⁻¹ chloramphenicol). The cultures were grown at 37 °C and 250 rpm to an A₆₀₀ of 0.6. Temperature was lowered to 20 °C, and IPTG was added to a final concentration of 0.5 mm. Synthesis of recombinant protein was allowed to proceed for 6 h. Cells were harvested by centrifugation for 20 min at 2700 × g at 4 °C, and the pellet was stored at −20 °C. Cells were thawed and resuspended in lysis buffer (50 mm Tris-HCl, pH 8.0, 500 mm NaCl, 10% glycerol, 10 mm β-mercaptoethanol, 20 mm imidazol, 1% Tween 20, 0.5 mg ml⁻¹ lysozyme, 0.5 mm phenylmethylsulfonyl fluoride) using a ratio of 25 ml lysis buffer for 10 g of pellet. After stirring for 30 min at 4 °C, cells were ruptured with a Branson sonifier. Debris and membranes were removed by centrifugation at 38,720 × g for 45 min. The supernatant was applied to a nickel-nitrilotriacetic acid-agarose resin column (GE Healthcare) according to the manufacturer's instructions, using a linear gradient of 0–60% 250 mm imidazole (in 50 mm Tris-HCl, pH 8.0, 500 mm NaCl, 10% glycerol, 10 mm β-mercaptoethanol) in 60 min for elution. The eluate was passed over PD-10 columns (Amersham Biosciences, Freiburg, Germany) and equilibrated and eluted with 25 mm Tris-HCl, pH 8.0, 100 mm NaCl, 15% glycerol, 2 mm dithiothreitol. The proteins were stored in this buffer at −80 °C.

For further purification, Ptf_At was loaded on an anion exchange column (Resource Q, 30 × 6.4 mm, 15 μm; GE Healthcare) according to the manufacturer's instructions and eluted with a linear gradient of 0–50% 1 m NaCl in 25 mm Tris-HCl, pH 8.0. The eluate was again passed over PD-10 columns as described above.

Assay for Prenyltransferase Activity

The reaction mixture (100 μl) contained 100 mm sodium-TAPS pH 8.7, 2 mm aromatic substrate, 2 mm isoprenoid substrate, 10 mm MgCl₂, 10 mm sodium ascorbate, and either 7.2 μg Ptf_At or 15 μg Ptf_Ss, or 50 μg Ptf_Bf. After incubation for 10 min at 30 °C, the reaction was stopped with 100 μl of ethyl acetate/formic acid (40:1). After vortexing and centrifugation, 75 μl of the organic layer was transferred to an Eppendorf tube. The solvent was evaporated, and the residue was dissolved in 50 μl methanol and analyzed by HPLC, using an Eclipse XDB-C18 column (150 × 4.6 mm, 5 μm, Agilent, Waldbronn, Germany) at a flow rate of 1 ml min⁻¹. Water (solvent A) and methanol (solvent B) each containing 1% (v/v) formic acid were used as eluents. A linear gradient was ran from 20–100% solvent B in 30 min. Products were detected with a photo diode array detector. Assays containing amino acids as aromatic substrates were stopped after incubation with 10 μl formic acid. The protein was removed by centrifugation at 21,460 × g for 10 min. The supernatant was directly analyzed by HPLC as described above.

¹H NMR Data of the Enzymatic Product 3-Dimethylallyl-2,7-dihydroxynaphthalene

The NMR spectrum was recorded at 250 MHz in a Bruker AC-250 spectrometer (Bruker, Rheinstetten, Germany), with CD₃OD as solvent. δ values are given in ppm. The solvent signal (3.3 ppm) was used as reference: δ, 1.73 (br s, 3H, H-4′), 1.76 (br s, 3H, H-5′), 3.37 (d, J = 7.32 Hz, 2H, H-1′), 5.39 (m, 1H, H-2′), 6.78 (dd, J = 8.67 Hz, 2.4 Hz, 1H, H-6), 6.84 (d, J = 2.4 Hz, 1H, H-8), 6.86 (s, 1H, H-1), 7.34 (s, 1H, H-4), and 7.48 (d, J = 8.85 Hz, 1H, H-5).

Sequence Analysis

Database searches were performed with BLASTP (http://www.ncbi.nlm.nih.gov/). Secondary structure predictions were performed with GenTHREADER (12, 13), and sequences were aligned with ClustalW2 (22) and visualized with ESPript (23).

Calculation of Kinetic Constants

K_m and k_cat were calculated using GraphPad Prism software, version 5.01 for Windows, GraphPad Software, Inc., La Jolla, CA.

RESULTS

Sequence Analysis

A gene with sequence similarity to the CloQ/NphB group of ABBA prenyltransferases is found in the genomes of only five fungal species in the database: A. terreus NIH2624, Botryotinia fuckeliana B05.10, Sclerotinia sclerotiorum 1980, Penicillium marneffei ATCC18224, and Microsporum canis CBS113480. In this study, these genes will be designated as ptf_At, ptf_Bf, ptf_Ss, ptf_Pm, and ptf_Mc, respectively. In the genome sequences of the respective organisms, they have been designated as ATEG_00821, BC1G_01295, SS1G_09465, PMAA_022320, and MCYG_05060, respectively. An alignment of the amino acid sequence of their predicted gene products with the sequences of the bacterial enzymes CloQ (2), NovQ (24), and NphB (1) is shown in Fig. 2. This reveals obvious similarities between the sequences as well as conservation of several of the amino acids found to interact with the pyrophosphate moiety of the isoprenoid substrate in the x-ray structure of NphB (1).

FIGURE 2.

A, sequence comparison of the bacterial prenyltransferases NphB, CloQ, and NovQ with the fungal proteins Ptf_At, Ptf_Bf, Ptf_Ss, Ptf_Pm, and Ptf_Mc. For Ptf_At, both the current sequence in the database (Ptf_At-Or) and the corrected sequence (Ptf_At) are displayed. The amino acids of NphB involved in the binding of the pyrophosphate moiety (▴, Lys¹¹⁹, Asn¹⁷³, Tyr²¹⁶, Arg²²⁸, and Lys²⁸⁴; over H₂O ■, Tyr²⁸²) and in the coordination of the Mg²⁺ ion (●, Asp⁶²) are marked. Likewise, the position of the lysine residue of CloQ and NovQ, suggested to functionally replace the positive charge of the Mg²⁺ ion, is marked (★). The sequence was created with Clustal W2 (22) and visualized by ESPript 2.2 (23). The sequence shows the secondary structure elements of NphB: α, α helices; η, 3₁₀ helices; β, β strands; TT, strict β turns. Strict sequence identity is shown by a black box with a white character, and similarity is shown by bold characters in a black frame. B, sequence identities (%) of the predicted gene products of the fungal genes ptf_At, ptf_Bf, ptf_Ss, ptf_Pm-S, and ptf_Mc with the bacterial ABBA prenyltransferases NphB, CloQ, and NovQ. For ptf_Pm, an N-terminally truncated version (ptf_Pm-S) was used for this calculation (see text).

However, in the case of the A. terreus protein (GenBank^TM accession number XP_001210907, Ptf_At-Or in Fig. 2), similarity to the other enzymes stops abruptly downstream of the amino acid corresponding to position 253 in NphB (Fig. 2). Closer inspection of the nucleotide sequence showed that this position marks the beginning of the first of two predicted introns in the current annotation of the gene ptf_At (ATEG_00821). As the bioinformatics prediction of introns is prone to errors, we generated a translation of the genomic sequence without introns (sequence Ptf_At in Fig. 2), and this showed close similarity to the other fungal and bacterial enzymes. We therefore suggest that the gene ptf_At does not contain introns, just like ptf_Bf, ptf_Ss, and ptf_Pm. This suggestion was later on supported by the demonstration of enzymatic activity of the protein generated from the genomic sequence of ptf_At in E. coli (see below). The gene ptf_Mc is annotated to contain a small intron near its 3′ end; bioinformatic analysis does not allow to decide whether this annotation is correct.

The N terminus of the predicted gene product of ptf_Pm was >50 amino acids longer than that of the four other fungal proteins. For the current study, we generated a truncated version of this gene (termed ptf_Pm-S), lacking the coding sequence for the first 57 amino acids and thereby having a similar size as the other proteins in this study. The active site of the protein, located inside the β barrel, should not be affected by this truncation.

The sequence identity in between the five predicted fungal proteins ranges from 27 to 81% (Fig. 2B). As noted previously (3), they are closely related to the 4-hydroxyphenylpyruvate-3-dimethylallyltransferases CloQ and NovQ (2, 24), but only distantly to NphB (1) and the other nine bacterial ABBA prenyltransferases of the CloQ/NphB group in the database.

NphB is magnesium-dependent in its catalytic activity, but CloQ and NovQ are not. Correspondingly, the aspartate residue required for the coordination of Mg²⁺ in NphB (Asp⁶², Fig. 2) is not conserved in CloQ and NovQ, and neither is this residue conserved in the five predicted fungal proteins. Kuzuyama et al. (1) suggested that the positive charge of Mg²⁺ in NphB is functionally replaced by a positively charged lysine residue in CloQ and NovQ and by an arginine residue in the related prenyltransferase SCO7190. As shown in Fig. 2, the five fungal proteins also contain arginine or lysine residues in this position, predicting that they may represent Mg²⁺-independent prenyltransferases. This was later confirmed by biochemical investigations (see below). Structure prediction using GenTHREADER (12, 13) showed that all five fungal enzymes are likely to have the PT barrel fold.

Protein Expression and Purification

The nucleotide sequences of ptf_At, ptf_Bf, and ptf_Ss as well as of the truncated ptf_Pm-S were optimized for expression in E. coli and synthesized commercially (see “Experimental Procedures”). The expression plasmids were introduced into E. coli. After induction with IPTG, the cells were harvested, and the His₈-tagged proteins were purified by Ni²⁺ affinity chromatography (Fig. 3A–C). Ptf_At was further purified by ion exchange chromatography to apparent homogeneity (Fig. 3A). Although a part of the proteins remained insoluble, Ptf_At, Ptf_Bf, and Ptf_Ss could be obtained in yields of 21, 11, and 25 mg per liter culture, respectively. In contrast, Ptf_Pm-S was completely insoluble and could not be purified. The sequence of ptf_Mc has only been deposited in the database when the current study was already in progress. Biochemical investigations were therefore carried out initially with Ptf_At and later on also with Ptf_Bf and Ptf_Ss.

FIGURE 3.

Purification of the fungal prenyltransferases Ptf_At (A), Ptf_Bf (B), and Ptf_Ss (C) after expression in the form of His₈-tagged fusion proteins. Lane 1, molecular mass standards; lane 2, total protein before IPTG induction; lane 3, total protein after IPTG induction; lane 4, soluble protein after IPTG induction; lane 5, insoluble protein after IPTG induction; lane 6, protein after Ni²⁺ affinity chromatography; lane 7, molecular mass standards; lane 8, protein after ion exchange chromatography. The calculated masses are 36.6 kDa for Ptf_At, 36.7 kDa for Ptf_Ss, and 36.5 kDa for Ptf_Bf. The 12% polyacrylamide gel was stained with Coomassie Brilliant Blue R-250.

Demonstration of C-prenyltransferase Activity of Ptf_At

2,7-Dihydroxynaphthalene (2,7-DHN) is one of the aromatic substrates accepted by the bacterial ABBA prenyltransferases NphB and SCO7190 (Fig. 1) (25). When purified Ptf_At was incubated with DMAPP and 2,7-DHN, the enzyme- and time-dependent formation of a prenylated product was readily observed in HPLC-UV (Fig. 4B). HPLC-MS analysis in the positive mode showed the molecular ion of the product at m/z = 229 [M+H]⁺, corresponding to a monoprenylated derivative of 2,7-DHN. Subsequently, the assay was repeated on a large scale, and the product was isolated by preparative HPLC and investigated by ¹H NMR and ¹³C NMR (supplemental Figs. S1 and S2). In the ¹H NMR spectrum of the substrate 2,7-DHN, the signals of the aromatic protons can be detected at 6.91 ppm (H-1), 6.83 ppm (H-3), and 7.54 ppm (H-4). The signals of the protons at position 3 and 4 show coupling with 8.82 Hz, while the more distant protons at position 1 and 3 couple with 2.37 Hz. This results in a double doublet signal for H-3. In the symmetrical 2,7-DHN molecule, the signals of H-5, H-6, and H-8 are identical to those of H-4, H-3, and H-1, respectively.

FIGURE 4.

A, the aromatic prenylation reaction catalyzed by Ptf_At. B, HPLC analysis of prenyltransferase assays with Ptf_At. Detection: UV, 285 nm.

The ¹H NMR spectrum of the enzymatic product clearly showed the signals of a dimethylallyl moiety (see supplemental Fig. S1). The chemical shift of the signal of H-1′ of the dimethylallyl moiety (3.37 ppm) showed that it is attached to an aromatic carbon, not to an oxygen atom. The NMR signals of the aromatic protons H-5, H-6, and H-8 were essentially unchanged in comparison with the substrate. However, the signal of H-3 had disappeared, and correspondingly, the signals of H-1 and H-4 now appeared as singlets rather than doublets. This proved that substitution had occurred at C-3 of the aromatic nucleus (Fig. 4A). Therefore, Ptf_At catalyzes a C-prenylation of the aromatic substrate, i.e. a Friedel-Crafts alkylation, similar as reported for other ABBA prenyltransferases (3, 4). The regiospecificity of the prenylation by Ptf_At (Fig. 4A) is in contrast to the prenylation of 2,7-DHN by the bacterial prenyltransferases NphB and SCO7190 (Fig. 1), both of which lead to substitution at C-1 of 2,7-DHN (25).

Testing of Aspulvinone E as Substrate of Ptf_At

Previous biochemical investigations of aspulvinone dimethylallyltransferase from A. terreus showed that aspulvinone E (Fig. 1) is readily accepted as substrate of this enzyme, leading predominantly to a diprenylated product (14, 15). For the present study, aspulvinone E was synthesized according to a published procedure (19), and the identity of the compound was confirmed by ¹H and ¹³C NMR. When aspulvinone E was incubated with Ptf_At and DMAPP, no product formation could be observed in HPLC-UV analysis. The more sensitive HPLC-MS analysis, however, showed a low formation of a monoprenylated product (data not shown). The reaction velocity of the prenylation of aspulvinone E by Ptf_At was at least 60 times lower than the prenylation of 2,7-DHN, and much lower than described previously for the aspulvinone dimethylallyltransferase purified from cell-free extracts of A. terreus (14, 15). Therefore, aspulvinone E is not the preferred substrate of Ptf_At.

Biochemical Properties of Ptf_At

The biochemical properties of the enzyme were investigated using 2,7-DHN and DMAPP as substrates. In the assay described in the “Experimental Procedures,” the formation of the prenylated product showed linear dependence on the amount of Ptf_At (up to 10 μg) and on reaction time (up to 30 min). During prolonged incubation, a precipitation of protein was sometimes observed. Inclusion of NaCl into the incubation mixture was not useful to resolve this problem because it reduced the reaction velocity of the prenylation (50% reduction by 500 mm NaCl). Therefore, NaCl was not included into the assays, but incubation time was reduced to 10 min. In sharp contrast to the bacterial ABBA prenyltransferase NphB (1), to the membrane-bound aromatic prenyltransferase UbiA (26), and to the transprenyltransferases such as FPP synthase (27), Ptf_At showed catalytic activity in the complete absence of Mg²⁺ or any other divalent cation. Addition of 10 mm EDTA did not influence the reaction velocity. Therefore, binding of the isoprenoid substrate does not occur in form of a magnesium complex, as shown for FPP synthase (27). Nevertheless, addition of MgCl₂ to the incubation mixture increased the reaction velocity. The optimal concentration of MgCl₂ was found to be 10 mm, leading to a 2.4-fold increase of product formation. Similar observations have been reported both for fungal and for bacterial aromatic prenyltransferases of the ABBA family, e.g. for dimethylallyltryptophan synthase (28) and for CloQ (2). The pH optimum of Ptf_At was determined as 8.7, with half-maximal reaction velocities at pH 8.1 and 9.4. The enzyme could be stored at −80 °C, with <5% loss of activity within two months.

Using a constant concentration of 2,7-DHN (2 mm) and varying concentrations of DMAPP, a typical hyperbolic curve of product formation over substrate concentration was obtained, indicating that the reaction followed Michaelis-Menten kinetics (Fig. 5A). Nonlinear regression analysis resulted in a K_m value for DMAPP of 325 ± 35 μm and a k_cat of 0.026 ± 0.001 s⁻¹. Using different concentrations of 2,7-DHN (up to 1.6 mm) in the presence of 2 mm DMAPP, likewise a hyperbolic curve was observed, resulting in a K_m value for 2,7-DHN of 324 ± 25 μm (Fig. 5B). Using these values the catalytic efficiency (k_cat/K_m) of the conversion of the non-genuine substrate 2,7-DHN by Ptf_At was calculated as 80 m⁻¹ s⁻¹. This is >50 times higher than the value of 1.4 m⁻¹ s⁻¹ reported for the conversion of 2,7-DHN by NphB (25) but significantly lower than the values of 5280, 12400 or 46250 m⁻¹ s⁻¹ calculated for the prenylation of the genuine substrates of the ABBA prenyltransferases CloQ (2), PpzP (18), or dimethylallyltryptophan synthase (8). When 2,7-DHN was used in concentrations higher than 2 mm, substrate inhibition was observed. 20 mm 2,7-DHN inhibited product formation almost completely. It was not possible to calculate a K_i value, as no typical substrate inhibition curve was obtained. This may indicate multiple interactions of the phenolic substrate 2,7-DHN with the protein at high concentrations.

FIGURE 5.

Determination of K_m values of Ptf_At for DMAPP and 2,7-dihydroxynaphthalene. In A, 2,7-diydroxynaphthalene was kept constant at 2 mm. In B, DMAPP was kept constant at 2 mm. K_m and k_cat values were determined by nonlinear regression, using GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA).

Substrate Specificity of Ptf_At

A range of aromatic compounds, previously found to be accepted by various bacterial or fungal ABBA prenyltransferases, were tested as substrates of Ptf_At, each in a concentration of 2 mm. Of these, only 5,10-dihydrophenazine 1-carboxylic acid yielded a product formation that was detectable by HPLC-UV analysis. HPLC-MS analysis showed that it was a monoprenylated product with the same properties as the product formed by PpzP (18) in a positive control assay. Yet, reaction velocity with this substrate was nearly 50 times lower than with 2,7-DHN (Table 1). Incubation of Ptf_At with flaviolin (2,5,7-trihydroxy-1,4-naphthoquinone) yielded a very low enzymatic formation of a monoprenylated product, detectable only in HPLC-MS. For other aromatic substances, including l-tryptophan, l-tyrosine, 4-hydroxyphenylpyruvate, 4-hydroxybenzoic acid, 1,3-dihydroxynaphthalene, and 1,6-dihydroxy-naphthalene, neither HPLC-UV nor HPLC-MS analysis showed any enzymatic product formation, whereas positive control assays with dimethylallyltryptophan synthase (28), CloQ (2), Fnq26 (29), and NphB (1) readily showed the formation of prenylated products from these aromatic substrates. This proves that the specificity of Ptf_At for its aromatic substrate is different from that of any other prenyltransferase investigated so far.

TABLE 1.

Formation of prenylated products by the fungal prenyltransferases Ptf_At, Ptf_Ss, and Ptf_Bf from different aromatic substrates and DMAPP. n.d., no product detectable in HPLC-UV and HPLC-MS analysis. See text for further explanations.

Aromatic substrate	PtfAt		PtfSs		PtfBf
	pmol product × s−1 × mg protein−1	%	pmol product × s−1 × mg protein−1	%	pmol product × s−1 × mg protein−1	%
2,7-Dihydroxy-naphthalene	570	100	280	100	76	100
5,10-Dihydro-phenazine-1-carboxylic acid	12	2.1	7.5	2.7	7.7	10.1
Flaviolin	<10	<1.7	<5	<1.8	<1.5	<2
Aspulvinon E	<10	<1.7	<5	<1.8	n.d.	n.d.
p-Hydroxy-benzoic acid	n.d.	n.d.	<9	<3.2	11	14.5

Ptf_At was found to be specific for DMAPP. When DMAPP was replaced by geranyl diphosphate, no formation of a prenylated product from 2,7-DHN was observed. As expected, also isopentenyl diphosphate gave no product formation.

Biochemical Investigation of Ptf_Ss and Ptf_Bf

The enzymes from S. sclerotiorum and B. fuckeliana, Ptf_Ss and Ptf_Bf, showed very similar biochemical properties as Ptf_At from A. terreus. Upon incubation with 2,7-DHN and DMAPP, both enzymes gave a single product that was identical to that formed by Ptf_At in HPLC and mass spectrometry. Both enzymes were active in the absence of Mg²⁺, but addition of 10 mm MgCl₂ increased reaction velocity by a factor of 2.6 for Ptf_Ss and 2.2 for Ptf_Bf.

The K_m values for 2,7-DHN were determined as 590 ± 136 μm for Ptf_Ss and 996 ± 177 μm for Ptf_Bf. For DMAPP, the respective values were 1232 ± 75 μm and 1010 ± 125 μm. The catalytic efficiency (k_cat/K_m) for the prenylation of 2,7-DHN resulted as 22 and 5 m⁻¹ s⁻¹, still higher than the value of 1.4 m⁻¹ s⁻¹ reported for the prenylation of 2,7-DHN by NphB (25) but somewhat lower than the value observed of Ptf_At.

Like Ptf_At, the enzymes from S. sclerotiorum and B. fuckeliana were specific for DMAPP and did not accept geranyl diphosphate or isopentenyl diphosphate. Also their specificity for aromatic substrates resembled that of Ptf_At (Table 1); other than 2,7-DHN, only 5,10-dihydrophenazine-1-carboxylic acid gave an enzymatic product formation that was detectable in HPLC-UV analysis. With flaviolin, product formation was detectable only in HPLC-MS. Product formation with aspulvinone E could be detected in HPLC-MS for Ptf_Ss but remained below detection limit for Ptf_Bf. In all cases, the prenylation products formed by Ptf_Ss and Ptf_Bf were identical to those formed by Ptf_At in their chromatographic and mass spectrometric properties.

In contrast to Ptf_At, both Ptf_Ss and Ptf_Bf gave a monoprenylated product with 4-hydroxybenzoic acid, but due to the low amount of product, its structure could not be solved. Just as for Ptf_At, no product formation was observed when Ptf_Ss or Ptf_Bf were incubated with l-tryptophan, l-tyrosine, 4-hydroxyphenylpyruvate, 1,3-dihydroxynaphthalene, or 1,6-dihydroxy-naphthalene as aromatic substrates. Therefore, the properties of Ptf_Ss and Ptf_Bf are similar to those of Ptf_At but at variance to those reported for any other previously investigated prenyltransferase.

DISCUSSION

The present study proves that the ptf genes of A. terreus, S. sclerotiorum, and B. fuckeliana code for functional prenyltransferases that attach isoprenoid moieties to carbon atoms of aromatic substrates in an enzyme-catalyzed Friedel-Crafts reaction. Their substrate specificity is clearly different from that of any other known prenyltransferase. In the prenylation of the artificial substrate 2,7-DHN they show a different regiospecificity than the previously examined bacterial enzymes NphB and SCO7190 (1, 25). Therefore, these fungal enzymes represent an interesting addition to the presently available aromatic prenyltransferases, which can be used for the chemoenzymatic synthesis of bioactive molecules (25, 30–32).

The amino acid sequences of Ptf_At, Ptf_Ss, and Ptf_Bf, as well as those of the gene products of ptf_Pm and ptf_Mc, are closely related to the bacterial enzymes NovQ and CloQ, which catalyze the 3-prenylation of 4-hydroxyphenylpyruvate in the biosynthesis of clorobiocin and novobiocin (2, 24, 33). The protein encoded by ptf_Mc has therefore been annotated as “NovQ” (GenBank^TM accession number EEQ32241). However, in clear contrast to NovQ and CloQ, none of the three fungal proteins investigated in this study accepted 4-hydroxyphenylpyruvate as substrate.

The physiological function of these enzymes in the respective fungal organism remains unknown. Our study disproves the hypothesis that they may be responsible for the prenylation reaction in the biosynthesis of aspulvinone H and J (Fig. 1) (14, 15). An inspection of the genes located in the vicinity of the ptf genes in the genomes of the fungi A. terreus, B. fuckeliana, S. sclerotiorum, P. marneffei, and M. canis gives no indication of a secondary metabolic gene cluster to which the ptf genes may belong (see supplemental Tables S1–S5). This fact, i.e. that the ptf genes are not part of a recognizable secondary metabolic gene cluster, is at variance with the previously investigated fungal genes for indole prenyltransferases of the dimethylallyltryptophan synthase/LtxC group (7), but reminiscent of the bacterial genes SCO7190 and SCO7467 in Streptomyces coelicolor A3(2), the best examined bacterium of the genus Streptomyces (34). The predicted gene products of both genes show obvious similarity to aromatic prenyltransferases with a PT barrel fold, and the enzymatic activity of the protein encoded by SCO7190 has already been demonstrated in vitro (1, 25). Yet, a physiological function is not known for either of these two bacterial genes, and no primary or secondary metabolite resulting from an aromatic prenylation reaction possibly catalyzed by the enzymes encoded by these genes has been identified in S. coelicolor. Therefore, the present study clearly proved that Ptf_At, Ptf_Ss, and Ptf_Bf exhibit a specific, new prenyltransferase activity, but it remains unknown whether these enzymes have a physiological function in their respective fungal organisms.