HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 281, Issue 42, 32036-32047, October 20, 2006

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 281/42/32036    most recent M606500200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Zha, W. Articles by Zhao, H. Search for Related Content PubMed PubMed Citation Articles by Zha, W. Articles by Zhao, H. Social Bookmarking                      What's this?

Characterization of the Substrate Specificity of PhlD, a Type III Polyketide Synthase from Pseudomonas fluorescens^*

From the Departments of Chemical and Biomolecular Engineering, and Chemistry, Center for Biophysics and Computational Biology, Institute for Genomic Biology, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801

Received for publication, July 10, 2006 , and in revised form, August 23, 2006.

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
PhlD, a type III polyketide synthase from Pseudomonas fluorescens, catalyzes the synthesis of phloroglucinol from three molecules of malonyl-CoA. Kinetic analysis by direct measurement of the appearance of the CoASH product (k_cat = 24 ± 4 min^-1 and K_m = 13 ± 1 µM) gave a k_cat value more than an order of magnitude higher than that of any other known type III polyketide synthase. PhlD exhibits broad substrate specificity, accepting C₄-C₁₂ aliphatic acyl-CoAs and phenylacetyl-CoA as the starters to form C6-polyoxoalkylated -pyrones from sequential condensation with malonyl-CoA. Interestingly, when primed with long chain acyl-CoAs, PhlD catalyzed extra polyketide elongation to form up to heptaketide products. A homology structural model of PhlD showed the presence of a buried tunnel extending out from the active site to assist the binding of long chain acyl-CoAs. To probe the structural basis for the unusual ability of PhlD to accept long chain acyl-CoAs, both site-directed mutagenesis and saturation mutagenesis were carried out on key residues lining the tunnel. Three mutations, M21I, H24V, and L59M, were found to significantly reduce the reactivity of PhlD with lauroyl-CoA while still retaining its physiological activity to synthesize phloroglucinol. Our homology modeling and mutational studies indicated that even subtle changes in the tunnel volume could affect the ability of PhlD to accept long chain acyl-CoAs. This suggested novel strategies for combinatorial biosynthesis of unnatural pharmaceutically important polyketides.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Polyketide synthases (PKSs)³ from bacteria, fungi, and plants produce an amazing array of polyketides that have important biological and pharmacological activities (1-6). These PKSs are divided into three classes. Type I PKSs, the so-called modular PKSs, are multifunctional enzymes and catalyze the synthesis of polyketide in a modular fashion (7-11). Type II PKSs, on the other hand, are multienzyme complexes with each domain located on a single protein (12, 13). Type III PKSs, the simplest PKSs, are homodimeric enzymes that iteratively carry out the polyketide synthesis at one single active site (14-16). This third type of PKSs shares a common mechanism for assembly of the polyketide scaffold: loading of the starter substrate (derived from an acyl-CoA) onto the active site cysteine, repetitive decarboxylative condensation reaction of the extension substrate (usually malonyl-CoA) to elongate the polyketide chain, and the final cyclization of polyketide product that is subsequently released from the PKS (5, 16). Although simple, type III PKSs produce a wide variety of polyketides. The diversity of the products arises from the choices of different starter units (aliphatic or aromatic acyl-CoAs) and extender units (malonyl or modified malonyl-CoAs), the number of elongation cycles to control the size of the polyketides, and the final cyclization pattern of the linear polyketides (Claisen condensation, aldol condensation, or heterocyclic lactonization).

Type III PKSs are traditionally isolated from higher plants, where they play important roles in the defense system against microorganisms, but recently a handful of type III PKSs have been discovered in prokaryotes. The first one is RppA from Streptomyces griseus, which catalyzes the condensation of five malonyl-CoA molecules to synthesize 1,3,6,8-tetrahydroxynaphthalene (THN), a key intermediate in the formation of melanin in Streptomyces (17, 18). Other members of the bacterial type III PKSs include DpgA from Amycolatopsis orientalis, which produces 3,5-dihydroxyphenylacetyl-CoA from four molecules of malonyl-CoA (19, 20); PKS10 and PKS18 from Mycobacterium tuberculosis, which are capable of accepting long chain acyl-CoAs (C₁₂-C₂₀) to produce -pyrones (21); and PhlD from Pseudomonas fluorescens (22). PhlD was first discovered in the gene cluster responsible for the biosynthesis of 2,4-diacetylphloroglucinol in a variety of P. fluorescens strains (23, 24). However, its catalytic function was not defined until very recently (22). In a proposed mechanism (22), PhlD first condenses three molecules of malonyl-CoA to form a triketide. The cyclization of the triketide is then initiated by a decarboxylation reaction of the carboxyl group derived from the starter malonyl-CoA, leading to a C6 carbanion that attacks C1 in a Claisen condensation (Fig. 1A). A similar mechanism was also seen with RppA, where the Claisen condensation of the C10 carbanion to C1 cyclizes the linear pentaketide into THN (Fig. 1B) (18, 25). Although the plant type III PKSs typically share a high level of sequence identity (70-95%), those of bacterial origin are more dissimilar. RppA and PhlD share about 40% sequence identity, the highest among all the bacterial type III PKSs, whereas the rest share only 20%.

View larger version (16K): [in this window] [in a new window]   FIGURE 1. Proposed reaction mechanisms of PhlD (A) and THNS (or RppA; B).

Here we report the biochemical, mutational, and structural characterization of the substrate specificity of PhlD. Enzyme kinetics was determined by direct measurement of the free CoASH produced by the reaction. The thermal stability of PhlD was also investigated. A starter substrate profiling employing various acyl-CoAs revealed an unusual substrate specificity of PhlD. A homology structural model accompanied by mutational analysis was further used to probe the structural basis for this unusual property. A buried tunnel, similar to that seen in RppA, was identified adjacent to the active site in the modeled PhlD structure and was proposed to aid the acceptance of long chain acyl-CoAs. Surprisingly, it was shown that subtle changes in the tunnel volume can drastically affect the substrate specificity of PhlD.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Materials—The pET-26b(+) expression vector and BL21(DE3) Escherichia coli strain were obtained from Novagen (Madison, WI). Phusion DNA polymerase, T4 DNA ligase, and restriction endonucleases were purchased from New England Biolabs (Beverly, MA). QIAprep spin plasmid miniprep kit, QIAquick gel extraction kit, QIAquick PCR purification kit, and Ni-NTA-agarose resins were obtained from Qiagen. Oligonucleotide primers were obtained from Integrated DNA Technologies (Coralville, IA). [2-¹⁴C]Malonyl-CoA (1.74 GBq/mmol) was purchased from PerkinElmer Life Sciences. Nonlabeled acyl-CoAs and all of the other reagents were obtained from Sigma-Aldrich.

Cloning, Expression, and Purification—The phlD gene was PCR-amplified from the genomic DNA of P. fluorescens strain Pf-5 with the specific forward primer (PhlDfor) and reverse primer (PhlDrev) designed according to the putative phlD sequence of that strain (GenBank^TM accession number CP000076 [GenBank] ): PhlDfor, GGATCGGTCATATGTCTACACTTTGCCTTCCACAC (the NdeI restriction site is underlined and the ATG start codon is italicized); PhlDrev: TCATTACTCGAGGGCGGTCCACTCGCCCAC (the XhoI restriction site is underlined). The PCR amplification was carried out using Phusion DNA polymerase (New England Biolabs). The amplified product was then cloned into the expression vector pET-26b(+) using the restriction enzymes NdeI and XhoI. The stop codon was not included in the reverse primer to add a His₆ tag to the C terminus of PhlD. After confirming the accuracy of the cloned phlD gene by DNA sequencing, the plasmid pET26-PhlD was transformed into E. coli BL21 (DE3) for protein overexpression. The cells harboring the plasmid were initially cultured in LB medium at 37 °C to an A₆₀₀ of 0.6. Then 0.25 mM isopropyl--D-thiogalactoside inducer was added, and the cells were further grown at 20 °C overnight to achieve maximal protein expression. The cells were harvested by centrifugation (3,000 x g, 4 °C, 20 min), resuspended in lysis buffer (50 mM sodium phosphate buffer, 300 mM NaCl, 10 mM imidazole, pH 8.0), supplemented with protease inhibitor mixture (Novagen), and stored at -70 °C when needed. PhlD protein with a C-terminal His₆ tag was purified to near homogeneity by one-step affinity chromatography using Ni-NTA resins. Briefly, the cells were lysed by two passes through the French press at 16,000 p.s.i. After removal of the cell debris by centrifugation (5,000 x g, 4 °C, 20 min), the cleared lysate was incubated with the Ni-NTA resins at 4 °C for 1 h. The mixture was then loaded onto a polypropylene column, and the resins were washed extensively (50 mM sodium phosphate buffer, 300 mM NaCl, 20 mM imidazole, pH 8.0). PhlD protein was eluted with elution buffer (50 mM sodium phosphate buffer, 300 mM NaCl, 100 mM imidazole, pH 8.0). Finally, the purified PhlD protein was buffer-exchanged and concentrated to 2-3 mg/ml using dry polyethyleneglycol 20,000 sprinkled around the dialysis bag (10,000 MWCO).

Determination of Enzyme Kinetics and Stability—The kinetic parameters of PhlD for its substrate malonyl-CoA were determined based on the method used by Tseng et al. (20), with some modifications. The experiments were carried out in triplicate using a range of concentrations of malonyl-CoA (5-100 µM). For each reaction, substrate was incubated with PhlD (4-16 µg/ml) in buffer (50 mM Tris-HCl, pH 7.5, 500 µl) at 25 °C using a refrigerated water bath. Buffer and malonyl-CoA were preincubated at 25 °C for 20 min before addition of enzyme. At 10 s, 1 min, and 2 min, 100-µl aliquots of the reaction were quenched by the addition of 20 µl of 4% trifluoroacetic acid (TFA) (final concentration 0.667% (TFA)) and kept at room temperature until analysis by HPLC. The reactions were analyzed using an Agilent 1100 Series HPLC and ZORBAX SB-C18 reverse-phase column (3.0 x 150 mm, 3.5 µm) (Agilent Technologies, Palo Alto, CA) monitoring at 258 nm (0.5 ml/min; 25 °C; 0-1 min, 3% B; 1-15 min, 3-15% B; where A = H₂O, 0.1% TFA and B = acetonitrile, 0.1% TFA). Peaks at 7.5, 8.9, and 10 min correspond to authentic CoASH, malonyl-CoA, and acetyl-CoA, respectively. The initial rates of CoASH production determined from these reactions were fitted to the Michaelis-Menten equation using nonlinear least squares regression analysis in Microcal Origin 5.0 (Microcal Software, Northampton, MA) to calculate k_cat and K_m.

To determine thermal inactivation, samples of PhlD enzyme (1 mg/ml concentration) were incubated at 25, 30, and 37 °C in a MJ Research PTC-200 thermocycler (±0.3 °C). The aliquots were taken over a range of times and cooled on ice, and the enzymatic activity was subsequently determined using the two-enzyme system utilized by Achkar et al. (22), with some slight adjustments. Each reaction contained 50 µg of PhlD, 150 milli-units of -ketoglutarate dehydrogenase enzyme, 2 mM -ketoglutarate, 0.3 mM -NAD⁺, and 160 µM malonyl-CoA in potassium phosphate buffer (50 mM, pH 7.0). The initial reaction rate was determined at 25 °C by following the appearance of NADH at 340 nm ( = 6220 M^-1 cm^-1) in a Varian Cary 100 Bio UV-visible spectrophotometer (Palo Alto, CA). Typically inactivation was monitored until activity was diminished by at least 70%.

Substrate Specificity Characterization—The standard reaction was carried out in 50 mM Tris-HCl buffer, pH 7.5, containing 1 mM Tris-(2-carboxyethyl)-phosphine hydrochloride, 200 µM starter acyl-CoA, 20 µM malonyl-CoA (including 1.7 µM [2-¹⁴C]malonyl-CoA), and 60 µg of purified wild type or mutant PhlD protein in a total volume of 600 µl. The reaction was incubated at 30 °C for 30 min and quenched with 20 µl of 6 M HCl. The products were extracted twice with 600 µl of ethyl acetate, and the combined extracts were dried under N₂ flush. Finally, the residue material was dissolved in 50 µl of methanol for HPLC analysis. A ZORBAX SB-C18 reverse-phase column (3.0 x 150 mm, 3.5 µm; Agilent) was run at room temperature on a SpectraSYSTEM LC system (Thermo, Waltham, MA) with a flow rate of 0.5 ml/min and UV absorbance detection at 280 nm (UV₂₈₀). Gradient elution was carried out with H₂O (solvent A) and acetonitrile (solvent B), both containing 0.1% TFA. The mobile phase linear gradient was set for reactions A, B, and C (Fig. 2): 5% B to 40% B in 30 min; for reactions D, E, F, and H: 20% B to 80% B in 30 min; and for reaction G: 40% B to 100% B in 30 min. The elution was collected every minute, and the radioactivity was measured by a LS 6500 Multi-Purpose Scintillation Counter (Beckman Coulter, Fullerton, CA), which was plotted on the same HPLC profile as UV₂₈₀. For direct detection of the products under UV₂₈₀, the substrates in the reaction were increased to 0.5 mM for the starter acyl-CoA and 0.2 mM for malonyl-CoA without the addition of radiolabeled malonyl-CoA.

Product Characterization—The polyketide products were characterized by liquid chromatography-electrospray ionization mass spectrometry (LC/ESI-MS) on a LCQ DECAXP ion trap (Thermo, Waltham, MA). Reactions for LC/ESI-MS were scaled up 10-fold compared with those prepared for UV detection, and only nonradioactive malonyl-CoA was used.

Homology Modeling—The crystal structure of 1,3,6,8-tetrahydroxynaphthalene synthase (THNS) from S. coelicolor was downloaded from the Protein Data Bank (accession code 1U0M [PDB] ). The dimer coordinates of the structure were prepared by PyMol (pymol.sourceforge.net/). The amino acid sequence of Pf-5 PhlD was then aligned with THNS using the sequence alignment tool provided by Insight II software (Insight II, version 2000; Accelrys Inc., San Diego, CA), followed by manual inspection and correction to make sure that the conserved residues were well aligned. This sequence alignment was subjected to the automatic model building module of Insight II, MODELER, using default parameters with moderate refinement for the loop regions to build the homology structure model of PhlD. Of nine models created, the best model was selected based on visual inspection, such as the proper orientation of the key active site residues, the score from the Profile3D estimation, and the values of the and angles calculated by ProStat. The reaction intermediate, lauroyl-primed triketide, was generated and manually docked into the active site cavity of the created model using Molecular Operating Environment (Chemical Computing Group Inc., Montreal, Canada), according to the configuration of the polyethyleneglycol molecule co-crystallized with THNS. Using forcefield MMFF94s (33), the whole structure was subjected to energy minimization to relieve steric and torsional artifacts from the modeling and docking processes.

Site-directed Mutagenesis—Site-directed mutagenesis was carried out using the megaprimer PCR method (34), which consists of two sequential PCRs, both employing the wild type phlD gene as the template. Briefly, the first PCR was carried out with the mutagenesis primer containing the desired mutation with 20-21 additional nucleotides on either side (5' and 3') and one of the flanking primers, PhlDfor or PhlDrev, previously used for amplification. The first PCR product was then used as the megaprimer in the second PCR in conjunction with the other flanking primer to amplify the full-length product. The mutant genes were subcloned into pET-26b(+) using the restriction enzymes NdeI and XhoI and sequenced to confirm that only the desired mutations were incorporated.

View larger version (38K): [in this window] [in a new window]   FIGURE 2. HPLC analysis of the products synthesized by PhlD from various acyl-CoA starter substrates. The reactions for radiolabeling HPLC contain 20 µM malonyl-CoA including 1.7 µM [2-14C]malonyl-CoA and 200 µM acyl-CoA starters. A, malonyl-CoA; B, acetoacetyl-CoA (2a); C, butyryl-CoA (2b); D, hexanoyl-CoA (2c); E, octanoyl-CoA (2d); F, decanoyl-CoA (2e); G, lauroyl-CoA (2f); H, phenylacetyl-CoA (2g). For substrate numbering, see also Fig. 4. The reactions for UV280 HPLC contain 200 µM nonradioactive malonyl-CoA and 500 µM of each acyl-CoA starters. The UV280 HPLC for A was prepared by injection of pure phloroglucinol. The solid lines represent the UV absorbance at 280 nm (milli-absorbance unit (mAU), left y axis), and the solid lines with filled circles represent radioactivity measurement (disintegrations/min (dpm), right y axis).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Kinetics and Stability of PhlD—The recent characterization of native PhlD by Achkar et al. (22) reported significantly different kinetic values (k_cat = 10 min^-1, K_m = 5.6 µM, k_cat/K_m = 1787 mM^-1 min^-1) than other chalcone synthase superfamily enzymes, for example the related bacterial enzymes RppA from S. griseus (k_cat = 0.77 min^-1, K_m = 0.93 µM, k_cat/K_m = 828 mM^-1 min^-1) (27) and DpgA from A. orientalis (kcat = 0.81 min^-1, K_m = 15 µM, k_cat/K_m = 54 mM^-1 min^-1) (20), which condense and cyclize five and four malonyl-CoAs, respectively, whereas PhlD utilizes three. We determined the kinetic parameters of PhlD to be k_cat = 24 ± 4 min^-1, K_m = 13 ± 1 µM, and k_cat/K_m = 1883 mM^-1 min^-1. Interestingly, these results do corroborate the finding that, in terms of k_cat value, PhlD is highly unusual among the chalcone synthase superfamily of type III PKS enzymes. Additionally, the inactivation of PhlD over time was analyzed at 25, 30, and 37 °C. Loss of activity did not follow the typical first order kinetics, so results are not reported as the half-life of thermal inactivation. The times required to reduce the activity of the PhlD enzyme by 50% at 25, 30, and 37 °C were found to be 128, 74, and 7.2 min, respectively, showing that PhlD is relatively unstable even at lower temperatures.

Starter Substrate Specificity of PhlD—The starter acyl-CoA specificity of PhlD was investigated using [2-¹⁴C]malonyl-CoA mixed with various acyl-CoAs. The reactivity of PhlD toward each acyl-CoA was then deduced from the amount of [2-¹⁴C]malonyl-CoA incorporation as the extension unit into the final polyketide products. Efficient acceptance of unnatural acyl-CoAs could only be seen when they were in at least 2-fold excess over malonyl-CoA (data not shown). This indicates a clear preference of PhlD to select malonyl-CoA over other acyl-CoAs as the starter substrates, similar to what was seen with RppA that also uses malonyl-CoA as the sole substrate (27). When nonradiolabeled malonyl-CoA was used in place of unnatural acyl-CoA in the reaction mixture, only one radioactive product was observed, which co-eluted with authentic phloroglucinol (Fig. 2A). This confirmed the function of PhlD as a phloroglucinol synthase from solely the malonyl-CoA substrate.

To further probe the substrate specificity of PhlD, we first tested aliphatic acyl-CoAs with different chain length. PhlD did not react very well with short chain acyl-CoAs. It showed barely detectable activity toward acetyl-CoA. When butyryl-CoA (2b) was used as the starter substrate, beside phloroglucinol (1), a small amount of product was also detected by both the UV absorbance at 280 nm and the radioactivity measurement (Fig. 2C). The product (3b) was identified by LC/ESI-MS to be a triketide (M_r = 154) from the butyryl-primed reaction. The amount of the triketide and phloroglucinol were calculated from their radioactivity as a ratio of 3.1:100. Interestingly, PhlD showed a greater ability to accept long chain aliphatic acyl-CoAs. With hexanoyl-CoA (2c) as the starter, PhlD produced a triketide (3c, M_r = 182) and a tetraketide (4c, M_r = 224) as determined by LC/ESI-MS (Fig. 2D). The product amount ratio was 100 (phloroglucinol):17.2 (triketide):0.9 (tetraketide). When octanoyl-CoA (2d) was incubated with PhlD as the starter, in addition to the triketide (3d, M_r = 210) and tetraketide (4d, M_r = 252) products as observed with the hexanoyl-CoA reaction, two more products were identified (Fig. 2E). Product 6d had a M_r of 336 by LC/ESI-MS, which corresponds to a hexaketide resulted from five condensation cycles of the octanoyl moiety with malonyl-CoA. Product 7d was a heptaketide (M_r = 378) from six cycles of condensation. The product amount from octanoyl-CoA reaction had a ratio of 100 (phloroglucinol):12.9 (triketide):3.5 (tetraketide):10 (hexaketide):1 (heptaketide). Decanoyl-CoA (2e) was accepted by PhlD with even greater efficiency, leading to the formation of a triketide (3e, M_r = 238), tetraketide (4e, M_r = 280), pentaketide (5e, M_r = 322), and hexaketide (6e, M_r = 364) (Fig. 2F). The product amount from the decanoyl-CoA starter was significantly increased compared with that from hexanoyl and octanoyl-CoA, with the ratio of 100 (phloroglucinol):15.1 (triketide):42.1 (tetraketide):8.9 (hexaketide):4.6 (heptaketide). PhlD could still react with lauroyl-CoA (2f) to produce a triketide (3f, M_r = 266) and a tetraketide (4f, M_r = 308) (Fig. 2G), whose amounts were 6.2 and 30.8, respectively, compared with the amount of phloroglucinol that was set to 100. However, when the aliphatic chain in an acyl-CoA was further elongated, such as in myristoyl-CoA and palmitoyl-CoA, no more activity was detected with PhlD. Therefore, PhlD could accept aliphatic acyl-CoA with carbon chain length of up to 12.

Five additional acyl-CoAs were also investigated for their reactivity with PhlD. Acetoacetyl-CoA (2a), whose structure mimics an intermediate in the polyketide elongation pathway, was not used very well by PhlD, yielding only a triketide (3a, M_r = 126; Fig. 2B) in addition to phloroglucinol (1). Under the assay conditions, PhlD poorly accepted acyl-CoAs with a branched position, such as isobutyryl-CoA and could not accept acyl-CoAs with a rigid position, such as crotonoyl-CoA. Regarding acyl-CoAs with aromatic side chains, benzoyl-CoA was not accepted by PhlD, but phenylacetyl-CoA (2g) could be incorporated by PhlD with moderate efficiency, generating a triketide (3g, M_r = 202) and a tetraketide (4g, M_r = 244) (Fig. 2H). The product amount was distributed as 100 (phloroglucinol):1.9 (triketide):5.3 (tetraketide). Therefore, PhlD exhibits very broad substrate specificity and produces a wide range of products (see Fig. 4). The amount of phloroglucinol produced in the malonyl-CoA reaction was set as the 100% standard for calculation of the relative incorporation efficiency of other acyl-CoAs by PhlD as the starter substrates, and the data are summarized in Table 1. The remarkable ability of PhlD to react with long chain and bulky acyl-CoA substrates and to carry out extensive polyketide elongation is not reflected by its physiological activity, which only produces a triketide from the condensation of three molecules of malonyl-CoA. These results could not be plausibly interpreted by the prevailing theory for type III PKS catalysis, which states that the size of the polyketide product, influenced by the choice of the starter molecule and the number of extension steps, is controlled by the volume of the active site cavity (16, 29, 30).

Homology Modeling—To elucidate the structural basis for the unusual ability of PhlD to accept long chain acyl-CoAs, we built a homology structural model for PhlD based on the only crystal structure that showed high sequence identity with PhlD: THNS from S. coelicolor (Protein Data Bank accession code 1U0M [PDB] ; 43.4% identity). When the modeled PhlD structure was superimposed onto the structure of THNS, all of the residues lining the active site cavity were superimposed well; in particular, the catalytic triads, Cys-His-Asn (located at position 138, 272, and 305, respectively in PhlD), overlapped excellently.

View larger version (25K): [in this window] [in a new window]   FIGURE 3. Selected MS/MS spectra of the products from unnatural acyl-CoAs. A, 4-hydroxyl-6-(2',4'-dioxotridecyl)-2-pyrone (5e), precursor ion at m/z = 321. B, 4-hydroxyl-6-(2',4',6'-trioxotridecyl)-2-pyrone (6d), precursor ion at m/z = 335. C, 4-hydroxyl-6-(2',4',6'-trioxopentadecyl)-2-pyrone (6e), precursor ion at m/z = 363. D, 4-hydroxyl-6-(2',4',6',8'-tetraoxopentadecyl)-2-pyrone (7d), precursor ion at m/z = 377.

Mutational Analysis—To investigate the role of the above six different residues between PhlD and RppA on substrate binding individually, each of the six residues in PhlD was changed, respectively, into the corresponding residue in RppA by site-directed mutagenesis. The ability of these mutants to accept long chain acyl-CoAs was probed using lauroyl-CoA, the substrate used by PhlD but not by RppA. None of the six mutants had the expected effect on the acceptance of long chain acyl-CoAs: L59T did not produce any soluble protein; M21T and A60L still retained the wild type activity to react with lauroyl-CoA; H24L, although it showed reduced reactivity toward lauroyl-CoA, had lost the ability to produce phloroglucinol to the same extent as the wild type PhlD; and A185D was completely inactive. Unlike the successful examples from plant type III PKSs, where one PKS could be converted into a functional equivalent of another PKS by just mutating a few active site residues (29-31), the low sequence homology between bacterial type III PKSs, such as PhlD and RppA, made such rational design difficult. Note that the function of residues lining the tunnels of PhlD and RppA could be affected by their interacting residues farther away. However, given their important location, these six selected sites might still be important in determining the better ability of PhlD to accept long chain acyl-CoAs than that of RppA.

View larger version (31K): [in this window] [in a new window]   FIGURE 4. Summary of PhlD reactions with various acyl-CoA starter substrates.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
The closest homolog of PhlD in the type III PKS family is RppA. Notably, PhlD from P. fluorescens Pf-5 and RppA from S. griseus have 45% identity in the amino acid sequence, the highest among all known bacterial type III PKSs. PhlD and RppA share several unique features in catalysis. They use solely malonyl-CoA as their physiological substrates with the difference being that PhlD condenses three molecules of malonyl-CoA, whereas RppA condenses five. Another unique feature of PhlD and RppA is that their polyketide products are cyclized through a Claisen condensation of the end carbanion group (generated via the decarboxylation of the starter malonyl-CoA derived carboxyl group) to C1 (linked to the thiol group of the active site cysteine). RppA was shown to have broad starter substrate specificity (27). This study demonstrated that PhlD can also utilize a wide variety of acyl-CoAs as its starter substrate. To fairly compare the reactivity of PhlD toward unnatural acyl-CoAs with that of RppA, we used the previously reported reaction conditions for S. griseus RppA with some slight modification (27). Our data revealed a distinct substrate specificity profile for PhlD.

View larger version (13K): [in this window] [in a new window]   FIGURE 5. Comparison between the mutants and wild type PhlD of the ability to react with lauroyl-CoA () and the retained activity to synthesize phloroglucinol (). The lauroyl-CoA reaction was carried out as described for the direct detection of products under UV280, containing 500 µM lauroyl-CoA, 200 µM malonyl-CoA, and 0.2 mg/ml of each protein. The products after ethyl acetate extraction were analyzed on HPLC, and their relative amounts were compared by the integrated peak area. The reaction for phloroglucinol synthesis contained only 200 µM malonyl-CoA and 0.2 mg/ml of each protein. The relative amount of phloroglucinol produced was estimated by the cinnamaldehyde colorimetric assay monitored at 466 nm. WT, wild type.

To understand the structural basis of the unusual ability of PhlD to accept long chain acyl-CoAs, we built a homology model for PhlD. This modeled structure of PhlD revealed a similar buried tunnel as found in RppA and PKS18, which was believed to accommodate the binding of long chain acyl-CoAs (25, 32). The comparison of the residues lining the tunnel between PhlD and S. griseus RppA identified six sites that are occupied by different residues. These sites were hypothesized to contribute to the different binding capabilities of long chain acyl-CoAs between PhlD and RppA. However, direct replacement of the residues in PhlD by the corresponding ones in RppA affected not only PhlD binding of long chain acyl-CoAs but also its physiological activity to synthesize phloroglucinol. This might not be surprising considering the low sequence homology between PhlD and RppA. The function of the residues lining the tunnel could not be accurately predicted because they might be affected by the surrounding interacting residues on the second shell or even the third shell of the tunnel. Therefore, saturation mutagenesis was carried out on the six sites of PhlD individually to more comprehensively explore their effects on the acceptance of long chain acyl-CoAs. Interestingly, the six saturation mutagenesis libraries yielded only 30-60% active fraction that produced phloroglucinol. Although located toward the bottom of the buried tunnel and far away from the actual active site cavity, these six sites still impinged on the physiological activity of PhlD to a great extent.

Three mutations identified from the library screening, which significantly reduced the reactivity of PhlD with lauroyl-CoA but not its physiological activity to synthesize phloroglucinol, were mapped onto the structural model of PhlD. At site 21, the original methionine residue and the new mutation isoleucine both have hydrophobic side chains of similar size. Nevertheless, the branched side chain in isoleucine increased the contacting surface on one side of the tunnel and consequently pushed this side slightly into the tunnel, causing a small reduction in the volume (Fig. 6C) compared with the wild type enzyme (Fig. 6B). Indeed, using the cavity volume calculation function provided by Molecular Operating Environment, we found that the change of Met to Ile decreased the volume of the tunnel slightly from 1227 to 1199 ų. The new narrowed tunnel might not accommodate the large aliphatic moiety of lauroyl-CoA as well as the wild type PhlD, resulting in its loss of ability to accept this acyl-CoA substrate by more than half. At site 24, the imidazole plane of histidine in the wild type PhlD did not directly point into the tunnel because of the presence of surrounding bulky residues such as Tyr¹⁷² (Fig. 6A, shown as a stick). Valine, on the contrary, contains a small side chain that can be positioned toward the tunnel. As a result, the substitution of His with Val also caused a reduction in the tunnel volume by 78 ų. The bottom of the tunnel in H24V was raised up high enough to hinder the binding of lauroyl-CoA, giving only of the reactivity compared with that of the wild type enzyme (Fig. 6, D and E). The mutation of Leu⁵⁹ to Met made the most obvious change on the tunnel volume as the long side chain of methionine extended into the tunnel, pushing the side of the tunnel inwards (Fig. 6, F and G). Although a reduction of 94 ų in the tunnel volume was obtained with L59M, the most significant of all the three identified mutations, this mutation only rendered 50% decrease in the ability to accept lauroyl-CoA.

View larger version (46K): [in this window] [in a new window]   FIGURE 6. Modeling and mutational studies of the tunnel for binding of long chain acyl-CoAs. A, schematic representation of the key residues lining the tunnel and the active site cavity. The lauroyl-primed triketide intermediate, covalently linked to the active site cysteine (Cys138), was modeled in the PhlD active site cavity and was shown with an electron density cloud around it (generated by Molecular Operating Environment). B-G, molecular surface at the tunnel. The residues lining the front side of the tunnel in A were removed for a clear view of the tunnel surface. B and F, wild type PhlD (WT) showing Met21 and Leu59; C, mutant M21I; D, WT showing His24; E, mutant H24V; G, mutant L59M. B, C, F, and G were viewed from the bottom of the tunnel, whereas D and E were viewed from the right side of the tunnel as shown in A. Green, blue, and red represent hydrophobic, hydrophilic, and exposed (due the removal of the front residues) surfaces, respectively.

Acyl-CoAs with shorter chains were incorporated by PhlD much less efficiently than by RppA. Butyryl and hexanoyl-CoA as the starter substrates gave only 0.2 and 8% incorporation rates, respectively (Table 1). The tunnel in PhlD seemed to be capable of accommodating larger acyl chain moieties than that in RppA. On the other hand, acyl-CoAs with short chains, such as butyryl and hexanoyl-CoA might not access, bind to, and thus stabilize the tunnel in PhlD as well as the RppA tunnel. Interestingly, the reactivity of PhlD toward acetoacetyl-CoA was only 1.6% (Table 1), less than of that of RppA if the amount of the triketide product, 6-methyl-4-hydroxyl-pyrone (3a), relative to the malonyl-CoA product in the reaction was compared (6.6 triketide:100 phloroglucinal for PhlD versus 188 triketide:100 THN for RppA) (Fig. 2B) (27). The acetoacetyl moiety, after being loaded onto the thiol group of the active site cysteine, imitated the structure of the diketide in both PhlD and RppA polyketide elongation pathways. According to the reaction mechanism proposed for the phloroglucinol synthesis by PhlD, a carboxyl group or its decarboxylated product, carbanion group, was necessary to be retained at the end position in the linear polyketide intermediate to carry out the final Claisen condensation for ring closure (Fig. 1A). However, the end position in the acetoacetyl starter-derived polyketide was occupied by a methyl group. The environment of the PhlD active site cavity that favors the presence of a negatively charged end group must hinder the binding of the acetoacetyl moiety. On the other hand, RppA produces a much longer polyketide intermediate, and thus its active site cavity might be more flexible to accommodate the end methyl group in the acetoacetyl-primed reaction. PhlD was unable to accept the acyl-CoAs with a branched position, exemplified by isobutyryl-CoA, whereas RppA could still react with such substrates to a certain extent. The comparison between the active site cavities of the two proteins revealed two different residues near the catalytic triad: Val¹³⁹ and Phe³³⁸ in PhlD versus Ala¹³⁹ and Ile³³⁷ in RppA (Fig. 6A). The two bulkier substitutions in PhlD might result in a steric repulsion to the side group of isobutyryl-CoA. Neither PhlD nor RppA reacted with crotonoyl-CoA or benzoyl-CoA, presumably because of the unfavorable electronic or steric environment conferred by the active site residues to such substrates with a rigid position. However, phenylacetyl-CoA could be incorporated by PhlD with moderate efficiency (3.3% incorporation rate; Table 1) but not by RppA. Because the more extended phenyl group in the phenylacetyl moiety was very likely to reach to the buried tunnel, the different reactivity seen in this case might imply a bigger and more flexible tunnel in PhlD than in RppA.

Interestingly, when long chain acyl-CoAs, such as octanoyl-CoA or decanoyl-CoA, were incubated with PhlD, extension steps beyond tetraketide were observed, whereas this did not occur with acyl-CoAs containing shorter chains. The same trend was also observed in RppA, where only the longest acceptable acyl-CoA, octanoyl-CoA, was able to produce polyketides with more than three extension steps (27). The crystal structure of S. coelicolor THNS (RppA) showed that the three helices that formed the buried polyethyleneglycol binding tunnel underwent dynamic motions, which assisted the active site cavity to selectively increase its volume in response to long chain acyl-CoA starters to form polyketide products with further extension than tetraketide (25). Given the relatively high sequence identity between THNS and PhlD, the same mechanism might be responsible for the ability of PhlD to catalyze multiple extension steps when bound with long chain acyl-CoA starters.

The substrate specificity study of RppA from S. griseus reported the formation of phloroglucinols from tetraketides when hexanoyl-CoA and octanoyl-CoA were used as the starters (27). However in our product analysis by LC/ESI-MS, we did not observe the formation of any phloroglucinol type products from the unnatural acyl-CoAs by PhlD within the detection limits of our UV absorbance and radioactivity based HPLC. It is clear that PhlD produces -pyrones as its major products from the unnatural acyl-CoA starters. Even for RppA, only trace amounts of phloroglucinols were detected from hexanoyl and octanoyl-CoA, which were barely distinguishable from the noise on the HPLC chromatograms (27). The formation of -pyrones by PhlD is likely to follow the mechanism by which 2-pyrone synthase produces 6-methyl-4-hydroxyl-pyrone (29). The electron distribution and geometry of the PhlD active site cavity stimulate the formation of the C5 oxyanion in the unnatural acyl-CoA starter-initiated linear polyketides and make it adopt a configuration favorable for the lactonizational attack of C5 oxyanion to C1, leading to -pyrones.

In summary, PhlD exhibits relatively broad substrate specificity compared with other members of the type III PKS family. In addition to its ability to produce phloroglucinol from malonyl CoA, it accepts C₄-C₁₂ aliphatic acyl-CoAs and phenylacetyl-CoA as the starter substrates to form from tri- to heptaketide -pyrones. The diverse products produced by PhlD from unnatural acyl-CoA starters greatly expand the existing reservoir of polyketides. Homology modeling of PhlD indicated the presence of a buried tunnel extending out from the active site cavity to accommodate the binding of long chain acyl-CoAs. Subsequent saturation mutagenesis guided by the structural knowledge from modeling on the tunnel successfully altered the substrate specificity of PhlD, which demonstrated that even subtle changes in the tunnel volume could affect the ability of PhlD to accept long chain acyl-CoAs. The powerful combination of random mutagenesis and rational design can be generally applied to tailor other properties of polyketide synthases beside the substrate specificity. This provides novel strategies for combinatorial biosynthesis of unnatural pharmaceutically important polyketides.

	FOOTNOTES

 
^* This work was supported by an Office of Naval Research Grant N00014-02-1-0725. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The on-line version of this article (available at http://www.jbc.org) contains supplemental Table S1 and Fig. S1.

¹ Supported by the United States National Science Foundation Graduate Research Fellowship Program.

² To whom correspondence should be addressed: Depts. of Chemical and Biomolecular Engineering and Chemistry, Center for Biophysics and Computational Biology, Institute for Genomic Biology, University of Illinois at Urbana-Champaign, 600 S. Mathews Ave., Urbana, IL 61801. Tel.: 217-333-2631; Fax: 217-333-5052; E-mail: zhao5{at}uiuc.edu.

³ The abbreviations used are: PKS, polyketide synthase; THN, 1,3,6,8-tetrahydroxynaphthalene; THNS, THN synthase; Ni-NTA, nickel-nitrilotriacetic acid; HPLC, high pressure liquid chromatography; LC, liquid chromatography; ESI, electrospray ionization; MS, mass spectrometry; TFA, trifluoroacetic acid.

	ACKNOWLEDGMENTS

 
We thank Mike Nickels for help in radioactive HPLC analysis.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 

1. Edwards, M. L., Stemreick, D. M., and Sunkara, P. S. (1990) J. Med. Chem. 33, 1948-1954[CrossRef][Medline] [Order article via Infotrieve]
2. Hopwood, D. A. (1997) Chem. Rev. 97, 2465-2497[CrossRef][Medline] [Order article via Infotrieve]
3. Jang, M. (1997) Science 275, 218-220[Abstract/Free Full Text]
4. Katz, L., and Donadia, S. (1993) Annu. Rev. Microbiol. 47, 875-912[CrossRef][Medline] [Order article via Infotrieve]
5. Keating, T. A., and Walsh, C. T. (1999) Curr. Opin. Chem. Biol. 3, 598-606[CrossRef][Medline] [Order article via Infotrieve]
6. Khosla, C., Gokhale, R. S., Jacobsen, J. R., and Cane, D. E. (1999) Annu. Rev. Biochem. 68, 219-253[CrossRef][Medline] [Order article via Infotrieve]
7. Donadia, S., Staver, M. J., McAlpine, J. B., Swanson, S. J., and Katz, L. (1991) Science 252, 675-679[Abstract/Free Full Text]
8. Shen, B. (2000) Biosynthesis: Aromatic Polyketides, Isoprenoids, Alkaloids 209, 1-51
9. Shen, B., and Hutchinson, C. R. (1993) Science 262, 1535-1540[Abstract/Free Full Text]
10. Rawlings, B. J. (2001) Nat. Prod. Rep. 18, 190-227[CrossRef][Medline] [Order article via Infotrieve]
11. Rawlings, B. J. (2001) Nat. Prod. Rep. 18, 231-281[CrossRef][Medline] [Order article via Infotrieve]
12. Staunton, J., and Weissman, K. J. (2001) Nat. Prod. Rep. 18, 380-416[CrossRef][Medline] [Order article via Infotrieve]
13. Rawlings, B. J. (1999) Nat. Prod. Rep. 16, 425-484[CrossRef][Medline] [Order article via Infotrieve]
14. Tropf, S., Kaercher, B., Schroeder, J., and Schroeder, G. (1995) J. Biol. Chem. 270, 7922-7928[Abstract/Free Full Text]
15. Moore, B. S., and Hopke, J. N. (2001) Chembiochem. 2, 35-38[CrossRef][Medline] [Order article via Infotrieve]
16. Austin, M. B., and Noel, J. P. (2003) Nat. Prod. Rep. 20, 79-110[CrossRef][Medline] [Order article via Infotrieve]
17. Ueda, K., Kim, K. M., Beppu, T., and Horinouchi, S. (1995) J. Antibiot. (Tokyo) 48, 638-646[Medline] [Order article via Infotrieve]
18. Funa, N., Ohnishi, Y., Fujii, I., Shibuya, M., Ebizuka, Y., and Horinouchi, S. (1999) Nature 400, 897-899[CrossRef][Medline] [Order article via Infotrieve]
19. Chen, H., Tseng, C. C., Hubbard, B. K., and Walsh, C. T. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 14901-14906[Abstract/Free Full Text]
20. Tseng, C. C., McLoughlin, S. M., Kelleher, N. L., and Walsh, C. T. (2004) Biochemistry 43, 970-980[Medline] [Order article via Infotrieve]
21. Saxena, P., Yadav, G., Mohanty, D., and Gokhale, R. S. (2003) J. Biol. Chem. 278, 44780-44790[Abstract/Free Full Text]
22. Achkar, J., Xian, M., Zhao, H. M., and Frost, J. W. (2005) J. Am. Chem. Soc. 127, 5332-5333[CrossRef][Medline] [Order article via Infotrieve]
23. Bangera, M. G., and Thomashow, L. S. (1996) Mol. Plant Microbe. Interact. 9, 83-90[Medline] [Order article via Infotrieve]
24. Bangera, M. G., and Thomashow, L. S. (1999) J. Bacteriol. 181, 3155-3163[Abstract/Free Full Text]
25. Austin, M. B., Izumikawa, M., Bowman, M. E., Udwary, D. W., Ferrer, J. L., Moore, B. S., and Noel, J. P. (2004) J. Biol. Chem. 279, 45162-45174[Abstract/Free Full Text]
26. Schuz, R., Heller, W., and Hahlbrock, K. (1983) J. Biol. Chem. 258, 6730-6734[Abstract/Free Full Text]
27. Funa, N., Ohnishi, Y., Ebizuka, Y., and Horinouchi, S. (2002) J. Biol. Chem. 277, 4628-4635[Abstract/Free Full Text]
28. Ferrer, J. L., Jez, J. M., Bowman, M. E., Dixon, R. A., and Noel, J. P. (1999) Nat. Struct. Biol. 6, 775-784[CrossRef][Medline] [Order article via Infotrieve]
29. Jez, J. M., Austin, M. B., Ferrer, J. L., Bowman, M. E., Schroder, J., and Noel, J. P. (2000) Chem. Biol. 7, 919-930[CrossRef][Medline] [Order article via Infotrieve]
30. Abe, I., Oguro, S., Utsumi, Y., Sano, Y., and Noguchi, H. (2005) J. Am. Chem. Soc. 127, 12709-12716[CrossRef][Medline] [Order article via Infotrieve]
31. Austin, M. B., Bowman, M. E., Ferrer, J. L., Schroder, J., and Noel, J. P. (2004) Chem. Biol. 11, 1179-1194[CrossRef][Medline] [Order article via Infotrieve]
32. Sankaranarayanan, R., Saxena, P., Marathe, U. B., Gokhale, R. S., Shanmugam, V. M., and Rukmini, R. (2004) Nat. Struct. Mol. Biol. 11, 894-900[CrossRef][Medline] [Order article via Infotrieve]
33. Halgren, T. A. (1999) J. Comput. Chem. 20, 720-729[CrossRef]
34. Sarkar, G., and Sommer, S. S. (1990) BioTechniques 8, 404-407[Medline] [Order article via Infotrieve]
35. Georgescu, R., Bandara, G., and Sun, L. (2003) Methods Mol. Biol. 231, 75-83[Medline] [Order article via Infotrieve]
36. Chockalingam, K., Chen, Z. L., Katzenellenbogen, J. A., and Zhao, H. M. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 5691-5696[Abstract/Free Full Text]
37. Lee, J., Ang, E. L., and Zhao, H. M. (2006) J. Bacteriol. 188, 6179-6183[Abstract/Free Full Text]

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				H. Zhou, J. Zhan, K. Watanabe, X. Xie, and Y. Tang A polyketide macrolactone synthase from the filamentous fungus Gibberella zeae PNAS, April 29, 2008; 105(17): 6249 - 6254. [Abstract] [Full Text] [PDF]

				S. Li, S. Gruschow, J. S. Dordick, and D. H. Sherman Molecular Analysis of the Role of Tyrosine 224 in the Active Site of Streptomyces coelicolor RppA, a Bacterial Type III Polyketide Synthase J. Biol. Chem., April 27, 2007; 282(17): 12765 - 12772. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 281/42/32036    most recent M606500200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Zha, W. Articles by Zhao, H. Search for Related Content PubMed PubMed Citation Articles by Zha, W. Articles by Zhao, H. Social Bookmarking                      What's this?