Skip to main content
Acta Crystallographica Section F: Structural Biology Communications logoLink to Acta Crystallographica Section F: Structural Biology Communications
. 2020 Nov 25;76(Pt 12):597–603. doi: 10.1107/S2053230X20014818

The crystal structure of benzophenone synthase from Garcinia mangostana L. pericarps reveals the basis for substrate specificity and catalysis

Chomphunuch Songsiriritthigul a,b,*, Natsajee Nualkaew c, James Ketudat-Cairns b,d, Chun-Jung Chen e
PMCID: PMC7716263  PMID: 33263571

The X-ray crystal structure of benzophenone synthase from G. mangostana L. reveals its active-site cavity and key residues catalyzing the conversion of its preferred substrate benzoyl-CoA to produce 2,4,6-trihydroxybenzophenone as the major product.

Keywords: benzophenone synthase; Garcinia mangostana L.; 2,4,6-trihydroxybenzophenone; polyketide synthases; cyclization reaction

Abstract

Benzophenone synthase (BPS) catalyzes the production of 2,4,6-trihydroxy­benzophenone via the condensation of benzoyl-CoA and three units of malonyl-CoA. The biosynthetic pathway proceeds with the formation of the prenylated xanthone α-mangostin from 2,4,6-trihydroxybenzophenone. Structural elucidation was performed to gain a better understanding of the structural basis of the function of Garcinia mangostana L. (mangosteen) BPS (GmBPS). The structure reveals the common core consisting of a five-layer αβαβα fold as found in other type III polyketide synthase enzymes. The three residues Met264, Tyr266 and Gly339 are proposed to have a significant impact on the substrate-binding specificity of the active site. Crystallographic and docking studies indicate why benzoyl-CoA is preferred over 4-coumaroyl-CoA as the substrate for GmBPS. Met264 and Tyr266 in GmBPS are properly oriented for accommodation of the 2,4,6-trihydroxybenzophenone product but not of naringenin. Gly339 offers a minimal steric hindrance to accommodate the extended substrate. Moreover, the structural arrangement of Thr133 provides the elongation activity and consequently facilitates extension of the polyketide chain. In addition to its impact on the substrate selectivity, Ala257 expands the horizontal cavity and might serve to facilitate the initiation/cyclization reaction. The detailed structure of GmBPS explains its catalytic function, facilitating further structure-based engineering to alter its substrate specificity and obtain the desired products.

1. Introduction  

The type III polyketide synthase (PKS) superfamily are also referred to as chalcone synthase-like enzymes (Austin & Noel, 2003). Type III PKS enzymes were discovered in plants, but were subsequently also found in some bacteria (Austin & Noel, 2003). Type III PKS enzymes have functional diversity related to the variety of their substrates, the number of malonyl-CoA molecules that are condensed and the cyclization reaction, which can be one of three types: Claisen and aldol cyclizations and lactonization (Abe & Morita, 2010).

Based on structural information, PKSs could be engineered to accept various types of substrate, such as diverse thioester substrates, nitrogen-containing substrates and an aminoacyl-CoA substrate unit (Abe & Morita, 2010). Site-directed mutagenesis of the active site could alter the ability to accept different substrate molecules, the number of malonyl-CoA units that are condensed and the mechanism of the cyclization reaction (Abe, 2008, 2012). Site-directed mutagenesis of the active-site residues based on structural information from crystallography could generate diverse polyketides in order to obtain novel compounds or to engineer unnatural novel polyketide reactions (Abe, 2008, 2012; Morita et al., 2019).

Garcinia mangostana L. is a tree of the Clusiaceae (Guttiferae) family, the fruit of which (mangosteen) is widely known as the ‘queen of tropical fruits’. α-Mangostin, a natural prenylated xanthone isolated from the pericarp of mango­steen, possesses antibacterial, antioxidant, antitumor and anti­carcinogenic activities (Pedraza-Chaverri et al., 2008).

Benzophenone synthase (BPS; EC 2.3.1.151) is a member of the type III PKS superfamily. It catalyzes the condensation of benzoyl-CoA and three units of malonyl-CoA, which proceeds via a Claisen condensation and produces 2,4,6-trihydroxy­benzophenone (phlorobenzophenone), which is the first intermediate in the xanthone biosynthetic pathway. In addition to BPS, biphenyl synthase (BIS) is also a benzoic acid-specific type III PKS. BIS also prefers benzoyl-CoA as a substrate, similar to BPS, but condenses it with three molecules of malonyl-CoA via an aldol condensation, thus producing 3,5-dihydroxybiphenyl as a product (Stewart et al., 2017).

G. mangostana BPS (GmBPS) has previously been cloned from the pericarp of G. mangostana fruit, expressed, purified and characterized (Nualkaew et al., 2012). It can utilise various substrates, including aliphatic, branched and aromatic thio­esters, and condense these with 1–3 units of malonyl-CoA (Nualkaew et al., 2012).

Site-directed mutagenesis of GmBPS, which was previously performed based on a homology model derived from chalcone synthase from Medicago sativa (MsCHS), could not achieve the production of naringenin from 4-coumaroyl-CoA as a substrate, despite this proceeding via the same cyclization as the native reaction (Nualkaew et al., 2012). This chalcone synthase-catalyzed reaction is a critical step in the metabolic engineering of microorganisms to produce flavonoids and other phenylpropanoids. Enabling the GmBPS enzyme to produce naringenin, from which other flavonoids can be produced, is a critical benchmark in diversifying its use in speciality chemical production with lower cost compared with chemical synthesis.

To precisely define the crucial amino-acid residues for the substrate-specificity, initiation, elongation and condensation mechanisms of the enzyme, the crystal structure of GmBPS was determined. Elucidation of the structure–function relationship provides clues to substrate binding and the rational design of mutagenesis for further application of this enzyme.

2. Materials and methods  

2.1. Determination of particle size and homogeneity by dynamic light scattering  

Expression and immobilized metal-affinity chromatography (IMAC) purification of GmBPS have been described previously (Nualkaew et al., 2012). Fractions containing BPS from IMAC were pooled, desalted, concentrated and exchanged into buffer A (20 mM Tris–HCl pH 8.0) in Amicon Ultra-15 centrifugal filter units (10 kDa molecular-weight cutoff; Millipore, Billerica, Massachusetts, USA). The concentrated BPS was applied onto a 1 ml Resource Q anion-exchange column (GE Healthcare, Bucking­hamshire, UK) equilibrated in buffer A. The column was washed and eluted using a linear gradient of 0–1 M NaCl in buffer A over 60 min at a flow rate of 4.0 ml min−1. The protein concentration was determined by the Bradford assay (Bradford, 1976).

A solution of 2.3 mg ml−1 purified protein in 20 mM Tris–HCl pH 8.0 containing 1 mM EDTA was analyzed by dynamic light scattering on a DynaPro NanoStar (Wyatt, California, USA) with an acquisition time of 30 s and ten rounds of determination at a controlled temperature of 25°C.

2.2. Crystallization  

Crystallization conditions for purified GmBPS solubilized in 20 mM Tris–HCl pH 7.0, 1 mM EDTA were screened using the microbatch method at 291 K. Diffraction-quality crystals of GmBPS were obtained from condition 12 of the PEG/Ion Screen (Hampton Research, Aliso Viejo, California, USA) consisting of 20%(w/v) PEG 3350, 0.2 M ammonium iodide pH 6.2 after one week of incubation. These crystals of GmBPS exhibited a rectangular morphology with approximate dimensions of 400 × 20 × 10 µm (Supplementary Fig. S1) and were used in the X-ray diffraction experiments.

2.3. Data collection, structure determination and refinement  

The optimized crystals were briefly transferred into reservoir solution containing 20%(v/v) glycerol as a cryoprotectant and were flash-cooled in liquid nitrogen. X-ray diffraction was performed on the BL13C1 beamline at the National Synchrotron Radiation Research Center (NSRRC), Hsinchu, Taiwan. Diffraction data were collected with a 0.975 Å wavelength X-ray beam using a Quantum 315r CCD area detector (ADSC, Poway, California, USA), with the crystal mounted on a Huber single-axis goniometer (Huber Diffraktion­stechnik, Rimsting, Germany). The data set was indexed, integrated and scaled with the HKL-2000 package (Otwinowski & Minor, 1997).

The structure was solved by molecular replacement with MOLREP (Vagin & Teplyakov, 2010) in the CCP4 suite (Winn et al., 2011) using the structure of MsCHS (PDB entry 1cgk, 59% sequence identity; Ferrer et al., 1999) as the search model and was rebuilt with WinCoot (Emsley et al., 2010). Refinement of the GmBPS structure was performed with REFMAC5 (Murshudov et al., 2011) with an initial R work of 45.06% and R free of 44.48%, and water molecules were included at a later stage of refinement with ARP/wARP (Perrakis et al., 1999). The refined structure with an R work of 16.97% and an R free of 21.52% at 2.30 Å resolution was validated with MolProbity (Chen et al., 2010). Most residues (97.62%) were in the favoured regions and no residues were found in outlier regions of the Ramachandran plot. The statistics for data collection and structure refinement of GmBPS are summarized in Table 1.

Table 1. Crystallographic data-collection and refinement statistics.

Values in parentheses are for the outer shell.

Data collection
 Space group P212121
a, b, c (Å) 70.81, 98.33, 112.02
 Mosaicity (°) 0.232–0.659
 Resolution range (Å) 30.0–2.30 (2.38–2.30)
 Total No. of reflections 66619 (6660)
 No. of unique reflections 35272 (3469)
 Completeness (%) 99.5 (99.7)
 Multiplicity 2.4 (2.4)
 〈I/σ(I)〉 9.67 (3.14)
R merge (%) 15.9 (49.4)
 CC1/2 (0.82)
Refinement
R work (%) 16.97
R free (%) 21.52
 No. of protein atoms 5859
 No. of ligand atoms 44
 No. of water atoms 317
 Average B factors (Å2)
  Protein atoms 17.23
  Ligand atoms 42.10
  Water atoms 19.48
 R.m.s. deviations
  Bond lengths (Å) 0.0067
  Angles (°) 1.418
 Ramachandran plot
  Most favoured (%) 97.62
  Allowed (%) 100

2.4. Amino-acid sequence alignment  

A sequence alignment of GmBPS (PDB entry 7cbf) with Hypericum androsaemum BPS (PDB entry 5uco; Stewart et al., 2017), M. sativa CHS (PDB entry 1cgk), M. domestica BIS (PDB entry 5w8q; Stewart et al., 2017), Arachis hypogaea stilbene synthase (PDB entry 1z1e; Shomura et al., 2005) and G. hybrida 2-pyrone synthase (PDB entry 1ee0; Jez et al., 2000) was obtained with the ClustalW multiple sequence-alignment software (Thompson et al., 1994). Subsequently, ESPript (Robert & Gouet, 2014) was used to display the secondary structure of our structure (PDB entry 7cbf).

2.5. Docking study  

After the completion of model building and refinement, our refined apo structure was superimposed with selected PKS structures using SUPERPOSE from the CCP4 suite. The modelling of naringenin bound to our structure was obtained from superimposition of our GmBPS structure with the holo MsCHS structure with bound naringenin (PDB entry 1cgk). Structure data files for the docked ligands were downloaded from PubChem (https://pubchem.ncbi.nlm.nih.gov) and were further formatted into structural coordinates by eLBOW (Moriarty et al., 2009).

The coordinates of 2,4,6-trihydroxybenzophenone were docked into the active site of GmBPS using AutoDock Vina (Trott & Olson, 2010). The selected conformation of phloro­benzophenone had a binding free energy of −8.1 kcal mol−1. Moreover, the selected conformation of 3,5-dihydroxy­biphenyl was docked into the structure of biphenyl synthase (MdBIS3) from Malus × domestica, yielding a binding free energy of −9.2 kcal mol−1. The structure of benzoyl-CoA docked into the structure of GmBPS had a binding free energy of −7.1 kcal mol−1. All structures were created and displayed with PyMOL (DeLano, 2002).

3. Results and discussion  

3.1. Overall structure of GmBPS  

GmBPS containing a C-terminal 6×His tag forms a homodimer with a total molecular mass of ∼88 kDa in solution (Nualkaew et al., 2012); this was confirmed by dynamic light scattering, which revealed monodisperse GmBPS with a polydispersity of 18% and a hydrodynamic diameter of 10.02 nm. The two monomeric structures in the GmBPS dimer are similar, with a root-mean-square deviation (r.m.s.d.) of 0.77 Å on the Cα atoms of 377 residues. The GmBPS homodimer structure comprises a total of 760 amino-acid residues with 317 water molecules, two molecules of imidazole, ten iodide ions and four molecules of glycerol (Fig. 1). The final model gave a good fit to the electron-density map, with an average atomic temperature factor (B factor) of 17.23 Å2 over two molecules.

Figure 1.

Figure 1

The overall homodimeric structure of GmBPS (PDB entry 7cbf). The monomers are coloured salmon and grey in the ribbon diagram. The N-­terminus and C-terminus of each monomer are indicated. The labelled amino-acid residues are shown as stick models with C atoms in salmon and grey, S atoms in yellow, N atoms in blue and O atoms in red.

Each molecule comprises the common core of the five-layer αβαβα fold that is generally found in class III PKS enzymes (Austin & Noel, 2003). The GmBPS structure reveals the conserved catalytic triad, Cys165/His304/Asn337 (Supplementary Fig. S2), positioned at the top of an internal active-site cavity, whereas two conserved aromatic residues, Phe216 and Tyr266 (Supplementary Fig. S2), are located at the bottom (Fig. 1).

Similar to the structure of MsCHS (PDB entry 1cgk), Met138 protrudes from one monomer into the other GmBPS monomer (Fig. 1). Moreover, a cis-peptide bond is formed between Met138 and Pro139 to position the side chain of Met138 as a knob on the monomer surface that forms a wall of the active-site cavity of the other monomer. The residues in the dimeric interface, from Thr132 to Ala141 at the other end of the loop, separate the active sites of each monomer.

3.2. Structural analysis of GmBPS structure  

3.2.1. Residues in the active-site cavity  

The structures of GmBPS and 2-pyrone synthase (2-PS) from G. hybrida (PDB entry 1ee0) superimposed with an r.m.s.d. of 0.75 Å between 376 equivalent Cα atoms, indicating a high similarity of the overall folds of these two proteins. Superposition of GmBPS with MsCHS (PDB entry 1cgk) gives a higher r.m.s.d. of 1.16 Å between 387 equivalent Cα atoms.

A naringenin molecule bound in the MsCHS complex structure superimposed on the GmBPS structure is shown in Figs. 2(a) and 2(b). Fig. 2(b) also shows phlorobenzophenone bound in the GmBPS structure by molecular docking. In MsCHS, the catalytic residue Cys164 acts as the nucleophile and subsequently forms a hydrogen bond to His303 during the formation of thiolate, while Asn336 is proposed to play a role in condensation (Ferrer et al., 1999). These catalytic triad residues are fully conserved and structurally analogous in all type III PKS enzymes (Austin & Noel, 2003).

Figure 2.

Figure 2

Structural comparison of GmBPS with other PKSs. The superposed structures of GmBPS (PDB entry 7cbf, salmon), MsCHS (PDB entry 1cgk, purple) and 2-PS (PDB entry 1ee0, lime green). The labelled amino-acid residues are shown as stick models with C atoms in the colours noted above, S atoms in yellow, N atoms in blue and O atoms in red. The residues are labelled according to GmBPS. The naringenin modelled in the structure of GmBPS by superposition of the MsCHS complex with GmBPS is shown as a ball-and-stick model with C atoms in purple and O atoms in red. (a) Catalytic triad residues (Cys165/His304/Asn337) and the gatekeeper (Phe216). (b) Buried active-site cavities. Only Cys165 in GmBPS is shown for clarity. The modelled 2,4,6-trihydroxybenzophenone in the structure of GmBPS is shown as a ball-and-stick model with C atoms in salmon and O atoms in red.

Phe216 in GmBPS is structurally conserved with the homologous gatekeeper residues Phe215 in MsCHS and Phe220 in 2-PS (Fig. 2 a), which act in substrate binding (Ferrer et al., 1999). In Fig. 2(b) and Supplementary Fig. S3, the other gatekeeper is Tyr266 in GmBPS (Phe265 in MsCHS and Phe270 in 2-PS), which is located at the interface between the CoA-binding tunnel and the buried active-site cavity (Jez et al., 2002). These two gatekeepers are important for the substrate preference (Ferrer et al., 1999; Jez et al., 2002).

Fig. 2(b) reveals that Met264, Tyr266 and Gly339 in GmBPS are aligned at the active site, rendering its substrate preference for benzoyl-CoA. The side chain of Met264 in GmBPS projects towards the cavity and is bulkier than the homologous Leu263 in MsCHS, thus favouring the smaller substrate benzoyl-CoA. Therefore, GmBPS does not produce naringenin, since Met264 sterically interferes with the flavanone moiety, as illustrated in Fig. 2(b). This conjecture is supported by kinetic studies of H. androsaemum CHS mutants, as its triple mutant (L263M/F265Y/S338G) also preferred benzoyl-CoA to 4-coumaroyl-CoA (Liu et al., 2003). Mutant residues which may affect substrate selectivity were identified by sequence analysis, but this alteration in activity of H. androsaemum CHS was not achieved by any of the single mutants L263M, F265Y or S338G.

GmBPS accepted 4-coumaroyl-CoA and allowed poly­ketide elongation to form minor products of the triketide lactone type, but not naringenin (Nualkaew et al., 2012). Our result clearly confirmed that naringenin cannot properly fit into the active site of GmBPS (Fig. 2 b).

In Fig. 2(b), Tyr266 in GmBPS may function as a mobile steric gate in selecting the substrate molecule, as does Phe265 in MsCHS (Ferrer et al., 1999), but their orientations are significantly different. The orientation of Tyr266 at the bottom of the active-site cavity facilitates the accommodation of the two phenyl rings of benzophenone (Fig. 2 b).

In Fig. 2(b), the substitution of the homologous Ile343 in 2-PS with Gly339 in GmBPS significantly expands the active-site cavity near the catalytic residue Cys165, thus accounting for its preference for the aromatic moiety of benzoyl-CoA over the acetyl moiety of acetyl-CoA. Gly339 in GmBPS particularly expands the space for benzoyl-CoA and the hydroxyl group of C2 in benzophenone (Fig. 2 b), and therefore the Gly339 residue is important for acceptance of the substrate and chain elongation. Supplementary Fig. S2 demonstrates that the homologous Ser338 in MsCHS is conserved in most plant type III PKSs (Austin & Noel, 2003). Substitution of Gly339 with serine allowed the GmBPS G339S mutant to maintain its activity against benzoyl-CoA, producing benzophenone, whereas the bulkier valine substitution in the G339V mutant showed no enzymatic activity with either benzoyl-CoA or 4-coumaroyl-CoA (Nualkaew et al., 2012).

Fig. 2(b) supports the finding that the larger steric bulk of the Leu202, Leu261 and Ile343 residues of 2-PS, which are homologous to Thr198 (GmBPS)/Thr197 (MsCHS), Ala257 (GmBPS)/Gly256 (MsCHS) and Gly339 (GmBPS)/Ser338 (MsCHS), respectively, occlude benzoyl-CoA binding, similar to 4-coumaroyl-CoA binding (Lim et al., 2016). The side chains of Leu202, Leu261 and Ile343 in 2-PS project into the active-site entrance compared with those of the homologous Thr198, Ala257 and Gly339 in GmBPS (Fig. 2 b), thus narrowing the buried pocket for substrate accommodation. These structural observations agree with the previous kinetic studies demonstrating that the active-site residues corresponding to Thr198, Ala257 and Gly339 in GmBPS are crucial for the substrate and product specificities of type III PKSs (Austin & Noel, 2003; Abe, 2008).

3.2.2. The stabilized Tyr266 is involved in the preference for benzoyl-CoA  

The molecule of benzoyl-CoA docked into GmBPS reveals that the side chain of Tyr266 (homologous to Tyr260 in MdBIS3 and Phe265 in MsCHS) aligns with the bottom of the incoming benzoyl-CoA (Supplementary Fig. S3), which is similar to Tyr269 in H. androsaemum BPS (HaBPS; Stewart et al., 2017). The benzyl ring of the benzoyl-CoA substrate docked to GmBPS stacks against Phe216 (Supplementary Fig. S3). Moreover, the backbone carbonyl of Leu268 in GmBPS forms a hydrogen bond to the adenosine moiety of the docked benzoyl-CoA, with a predicted distance of 2.6 Å (Supplementary Fig. S3).

MdBIS3 shares 59% sequence identity with GmBPS (Supplementary Fig. S2). The structural superimposition of GmBPS with MdBIS3 (PDB entry 5w8q) gives an r.m.s.d. of 1.25 Å between 377 equivalent Cα atoms. The side-chain O atom of the gatekeeper residue, Tyr260 in MdBIS3, is stabilized by a water-mediated hydrogen bond to the carbonyl group of Phe192 and the amide group of Gln207, affecting the specificity of the enzyme towards benzoyl-CoA. Wat660 is well defined, with an average atomic temperature factor (B factor) of 8.36 Å2 (Fig. 3 a). The corresponding Tyr266 residue in GmBPS is apparently stabilized by a similar water-mediated hydrogen bond to the backbone carbonyl O atom of Thr198 and the side-chain amide of Gln213 via Wat125, which has a B factor of 16.03 Å2 (Fig. 3 b). In the previously determined structure of HaBPS (PDB entry 5uco), the presence of such a water molecule was inconclusive owing to the low resolution of the X-ray diffraction data (Stewart et al., 2017), so hydrogen bonding between Thr200 and Gln215 was instead suggested (Fig. 3 c). This water-mediated interaction is common in PKS enzymes that prefer benzoyl-CoA (Stewart et al., 2017). As noted earlier, F265Y was one of the three important mutations (L263M/F265Y/S338G) in H. androsaemum CHS that altered its specificity from 4-coumaroyl-CoA to benzoyl-CoA (Liu et al., 2003).

Figure 3.

Figure 3

A water-mediated hydrogen bond which stabilizes Tyr266 is involved in the preference for benzoyl-CoA. (a) MdBIS3 (PDB entry 5w8q, yellow). (b) GmBPS (this work; PDB entry 7cbf, salmon). (c) HaBPS (PDB entry 5uco, purple). Black dashed lines represent hydrogen bonds formed in the crystal structures; their distances are given in Å.

3.2.3. Regions which contribute to the initiation/elongation cavity  

As shown in Fig. 4(a), Thr132, Thr197, Ile254, Gly256, Leu263 and Ser338 contribute to the initiation/elongation cavity in the structure of the CHS–naringenin complex. The structurally analogous residues are Thr133 (GmBPS)/Ala127 (MdBIS3), Thr198 (GmBPS)/Phe192 (MdBIS3), Ile255 (GmBPS)/Val249 (MdBIS3), Ala257 (GmBPS)/Ala251 (MdBIS3), Met264 (GmBPS)/Phe258 (MdBIS3) and Gly339 (GmBPS)/Gly335 (MdBIS3), as shown in Figs. 4(b) and 4(c), respectively. The products of GmBPS and MdBIS3 are modelled into their active-site cavities (Figs. 4 b and 4 c).

Figure 4.

Figure 4

Amino-acid residues contributing to the initiation/elongation cavity. (a) MsCHS (PDB entry 1cgk, purple). (b) GmBPS (this work; PDB entry 7cbf, salmon). (c) MdBIS3 (PDB entry 5w8q, yellow). Black dashed lines represent hydrogen bonds; their distances are given in Å.

Similar to the corresponding residue in the HaBPS structure, the side-chain O atom of Thr133 forms hydrogen bonds to the backbone amides of both Gln163 and Gly164 with distances of 3.3 and 2.9 Å, respectively (Supplementary Fig. S4). This hydrogen-bond formation apparently facilitates the orientation of the side chain of Thr133 supporting the accommodation of the tetraketide in wild-type GmBPS. The corresponding peanut stilbene synthase residue, Thr132, is also important for its aldol condensation specificity, which is electronically mediated by the interaction of the hydroxyl side chain with a water molecule close to the catalytic residue Cys164 (Shomura et al., 2005).

The hydrogen-bonding network between Glu193 and Thr133 is apparently mediated by Wat60 (B factor of 4.88 Å2) in the GmBPS structure. This is quite similar to MsCHS, in which the hydrogen-bond network between Glu192 and Thr132 has also been identified to enable catalysis of a C6 to C1 Claisen condensation.

Mutation of Thr133 to leucine created greater steric hindrance and significantly diminished the entrance to the active site, resulting in the major production of triketide lactone and the production of a small amount of benzo­phenone (Nualkaew et al., 2012). An altered product speci­ficity in the T133L mutant of GmBPS (Nualkaew et al., 2012) is presumably owing to disruption of the hydrogen-bond network involving its side chain, thus generating a narrower pocket, as previously described for HaBPS (Klundt et al., 2009; Beerhues & Liu, 2009). Mutation of T135L in HaBPS significantly changed its substrate and product specificities, whereas T135S and T135F mutations reduced the activity, and other point mutations of Thr135 were catalytically inactive (Klundt et al., 2009; Beerhues & Liu, 2009).

It is proposed that Thr198 in GmBPS affects both the shape and the volume of the initiation and elongation cavity of the active site compared with the active-site cavity of MdBIS3, which is significantly narrower owing to the replacement of Thr198 (GmBPS)/Thr197 (MsCHS) with Phe192 in MdBIS3. In contrast, substitution of Ile254 (MsCHS)/Ile255 (GmBPS) with Val249 in MdBIS3 may not affect the cavity volume, since the I254M mutant of MsCHS produced a product profile similar to that of wild-type MsCHS (Jez et al., 2000).

The replacement of Gly256 in MsCHS with the bulkier residues Ala257 (GmBPS) and Ala251 (MdBIS3) presumably shifts the substrate preference from 4-coumaroyl-CoA to the smaller substrate benzoyl-CoA in GmBPS and MdBIS3. This correlates with the previous report that Leu261 in 2-PS significantly reduced the cavity size for initiation/elongation compared with Gly256 in MsCHS (Jez et al., 2000). Therefore, the chain-elongation and cyclization specificity are determined by the volume of the active-site cavity (Klundt et al., 2009). The orientation of Met264 in GmBPS (Fig. 4 b) differs from that of the homologous residue Phe258 in MdBIS3 (Fig. 4 c). The orientation of Met264 in GmBPS also expands the cavity for its product, whereas Phe258 in MdBIS3 stabilizes the position of Phe192, thus reducing the bottom of the active-site cavity to a smaller size compared with that in GmBPS. On the contrary, Gly339 (GmBPS) and Gly335 (MdBIS3) (Figs. 4 b and 4 c), which replace Ser338 (MsCHS) (Fig. 4 a), expand the active site near Cys165 (GmBPS) and Cys159 (MdBIS3), respectively, thus generating pockets for specific binding to benzoyl-CoA and extending the initiation/elongation cavity of the active site.

3.2.4. Residues that are important for the cyclization mechanisms  

The mutations W281G in the ArsB and G284W in the ArsC type III PKSs from Azotobacter affected the cavity sizes of the enzymes, thus altering their cyclization specificities (Satou et al., 2013). Based on sequence alignment, both residues are homologous to Ala257 in GmBPS, Ala260 in HaBPS and Gly256 in MsCHS. The A257G mutant of GmBPS accepted both 4-coumaroyl-CoA and benzoyl-CoA, since the substitution with glycine did not restrict the binding pocket horizontally (Nualkaew et al., 2012). In addition to its role in the initiation/cyclization pocket, Ala257 in GmBPS is also proposed to be an important residue for the substrate preference. G256A and G256V mutants of MsCHS increased the production of coumaroyltriacetic acid lactone, whereas G256L and G256F mutants produced bisnoryangonin and methylpyrone (Klundt et al., 2009). The A260V mutant of HaBPS increased the proportion of phenylpyrone product compared with the wild-type enzyme by derailing further condensation, whereas the A260L mutant was catalytically inactive (Klundt et al., 2009).

4. Conclusions  

The substrate preference of the GmBPS enzyme for benzoyl-CoA is profoundly related to its active-site architecture. The collective contribution of Met264, Tyr266 and Gly339 facilitates specific binding to benzoyl-CoA, especially with support from Tyr266, which is stabilized by a bound water molecule. Moreover, Gly339 in GmBPS relatively expands the active site near the catalytic residue, providing the pocket for specific binding to benzoyl-CoA. Thr133 participates in elongation, whereas Ala257 is involved in initiation and cyclization. Thr198, which is structurally homologous to Phe192 in MdBIS3, supports the bottom of the pocket in GmBPS to give the phlorobenzophenone product. In contrast, the larger Phe192 in MdBIS3 protrudes towards the active site, resulting in the smaller product 3,5-dihydroxybiphenyl. Moreover, the phenyl side chain of Phe216 in GmBPS is proposed to make a π–π stacking interaction with the phenyl ring of benzoyl-CoA.

Although GmBPS has a binding preference for benzoyl-CoA, an alteration of substrate preference and product specificity would be achieved by substitution of the important residues elaborated from our reported structure of GmBPS. A triple mutation (M264L/Y266F/G339S) of the GmBPS enzyme may generate chalcone synthase activity, leading to a preference for 4-coumaroyl-CoA over benzoyl-CoA substrate. Moreover, a single mutation of Ala257 in GmBPS to glycine may support the 4-coumaroyl binding pocket of GmBPS in a similar fashion as in MsCHS, providing the possibility of accommodating various substrates. Thus, structural understanding of GmBPS will allow its engineering to alter its substrate preferences and lead to the production of novel products.

Supplementary Material

PDB reference: benzophenone synthase, 7cbf

Supplementary Figures. DOI: 10.1107/S2053230X20014818/nw5104sup1.pdf

f-76-00597-sup1.pdf (744.6KB, pdf)

Acknowledgments

We would like to thank all of the staff of beamline 13C1 at NSRRC for technical assistance during data collection.

Funding Statement

This work was funded by Khon Kaen University grant . Synchrotron Light Research Institute grant .

References

  1. Abe, I. (2008). Chem. Pharm. Bull. 56, 1505–1514. [DOI] [PubMed]
  2. Abe, I. (2012). Curr. Opin. Chem. Biol. 16, 179–185. [DOI] [PubMed]
  3. Abe, I. & Morita, H. (2010). Nat. Prod. Rep. 27, 809–838. [DOI] [PubMed]
  4. Austin, M. B. & Noel, J. P. (2003). Nat. Prod. Rep. 20, 79–110. [DOI] [PubMed]
  5. Beerhues, L. & Liu, B. (2009). Phytochemistry, 70, 1719–1727. [DOI] [PubMed]
  6. Bradford, M. M. (1976). Anal. Biochem. 72, 248–254. [DOI] [PubMed]
  7. Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S. & Richardson, D. C. (2010). Acta Cryst. D66, 12–21. [DOI] [PMC free article] [PubMed]
  8. DeLano, W. L. (2002). PyMOL. http://www.pymol.org.
  9. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). Acta Cryst. D66, 486–501. [DOI] [PMC free article] [PubMed]
  10. Ferrer, J.-L., Jez, J. M., Bowman, M. E., Dixon, R. A. & Noel, J. P. (1999). Nat. Struct. Biol. 6, 775–784. [DOI] [PubMed]
  11. Jez, J. M., Austin, M. B., Ferrer, J.-L., Bowman, M. E., Schröder, J. & Noel, J. P. (2000). Chem. Biol. 7, 919–930. [DOI] [PubMed]
  12. Jez, J. M., Bowman, M. E. & Noel, J. P. (2002). Proc. Natl Acad. Sci. USA, 99, 5319–5324. [DOI] [PMC free article] [PubMed]
  13. Klundt, T., Bocola, M., Lütge, M., Beuerle, T., Liu, B. & Beerhues, L. (2009). J. Biol. Chem. 284, 30957–30964. [DOI] [PMC free article] [PubMed]
  14. Lim, Y. P., Go, M. K. & Yew, W. S. (2016). Molecules, 21, 806.
  15. Liu, B., Falkenstein-Paul, H., Schmidt, W. & Beerhues, L. (2003). Plant J. 34, 847–855. [DOI] [PubMed]
  16. Moriarty, N. W., Grosse-Kunstleve, R. W. & Adams, P. D. (2009). Acta Cryst. D65, 1074–1080. [DOI] [PMC free article] [PubMed]
  17. Morita, H., Wong, C. P. & Abe, I. (2019). J. Biol. Chem. 294, 15121–15136. [DOI] [PMC free article] [PubMed]
  18. Murshudov, G. N., Skubák, P., Lebedev, A. A., Pannu, N. S., Steiner, R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011). Acta Cryst. D67, 355–367. [DOI] [PMC free article] [PubMed]
  19. Nualkaew, N., Morita, H., Shimokawa, Y., Kinjo, K., Kushiro, T., De-Eknamkul, W., Ebizuka, Y. & Abe, I. (2012). Phytochemistry, 77, 60–69. [DOI] [PubMed]
  20. Otwinowski, Z. & Minor, W. (1997). Methods Enzymol. 276, 307–326. [DOI] [PubMed]
  21. Pedraza-Chaverri, J., Cárdenas-Rodríguez, N., Orozco-Ibarra, M. & Pérez-Rojas, J. M. (2008). Food Chem. Toxicol. 46, 3227–3239. [DOI] [PubMed]
  22. Perrakis, A., Morris, R. & Lamzin, V. S. (1999). Nat. Struct. Biol. 6, 458–463. [DOI] [PubMed]
  23. Robert, X. & Gouet, P. (2014). Nucleic Acids Res. 42, W320–W324. [DOI] [PMC free article] [PubMed]
  24. Satou, R., Miyanaga, A., Ozawa, H., Funa, N., Katsuyama, Y., Miyazono, K., Tanokura, M., Ohnishi, Y. & Horinouchi, S. (2013). J. Biol. Chem. 288, 34146–34157. [DOI] [PMC free article] [PubMed]
  25. Shomura, Y., Torayama, I., Suh, D.-Y., Xiang, T., Kita, A., Sankawa, U. & Miki, K. (2005). Proteins, 60, 803–806. [DOI] [PubMed]
  26. Stewart, C., Woods, K., Macias, G., Allan, A. C., Hellens, R. P. & Noel, J. P. (2017). Acta Cryst. D73, 1007–1019. [DOI] [PMC free article] [PubMed]
  27. Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). Nucleic Acids Res. 22, 4673–4680. [DOI] [PMC free article] [PubMed]
  28. Trott, O. & Olson, A. J. (2010). J. Comput. Chem. 31, 455–461. [DOI] [PMC free article] [PubMed]
  29. Vagin, A. & Teplyakov, A. (2010). Acta Cryst. D66, 22–25. [DOI] [PubMed]
  30. Winn, M. D., Ballard, C. C., Cowtan, K. D., Dodson, E. J., Emsley, P., Evans, P. R., Keegan, R. M., Krissinel, E. B., Leslie, A. G. W., McCoy, A., McNicholas, S. J., Murshudov, G. N., Pannu, N. S., Potterton, E. A., Powell, H. R., Read, R. J., Vagin, A. & Wilson, K. S. (2011). Acta Cryst. D67, 235–242. [DOI] [PMC free article] [PubMed]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

PDB reference: benzophenone synthase, 7cbf

Supplementary Figures. DOI: 10.1107/S2053230X20014818/nw5104sup1.pdf

f-76-00597-sup1.pdf (744.6KB, pdf)

Articles from Acta Crystallographica. Section F, Structural Biology Communications are provided here courtesy of International Union of Crystallography

RESOURCES