The O-Methyltransferase SrsB Catalyzes the Decarboxylative Methylation of Alkylresorcylic Acid during Phenolic Lipid Biosynthesis by Streptomyces griseus

Abstract

Streptomyces griseus contains the srs operon, which is required for phenolic lipid biosynthesis. The operon consists of srsA, srsB, and srsC, which encode a type III polyketide synthase, an O-methyltransferase, and a flavoprotein hydroxylase, respectively. We previously reported that the recombinant SrsA protein synthesized 3-(13′-methyltetradecyl)-4-methylresorcinol, using iso-C₁₆ fatty acyl-coenzyme A (CoA) as a starter substrate and malonyl-CoA and methylmalonyl-CoA as extender substrates. An in vitro SrsA reaction using [¹³C₃]malonyl-CoA confirmed that the order of extender substrate condensation was methylmalonyl-CoA, followed by two extensions with malonyl-CoA. Furthermore, SrsA was revealed to produce an alkylresorcylic acid as its direct product rather than an alkylresorcinol. The functional SrsB protein was produced in the membrane fraction in Streptomyces lividans and used for the in vitro SrsB reaction. When the SrsA reaction was coupled, SrsB produced alkylresorcinol methyl ether in the presence of S-adenosyl-l-methionine (SAM). SrsB was incapable of catalyzing the O-methylation of alkylresorcinol, indicating that alkylresorcylic acid was the substrate of SrsB and that SrsB catalyzed the conversion of alkylresorcylic acid to alkylresorcinol methyl ether, namely, by both the O-methylation of the hydroxyl group (C-6) and the decarboxylation of the neighboring carboxyl group (C-1). O-methylated alkylresorcylic acid was not detected in the in vitro SrsAB reaction, although it was presumably stable, indicating that O-methylation did not precede decarboxylation. We therefore postulated that O-methylation was coupled with decarboxylation and proposed that SrsB catalyzed the feasible SAM-dependent decarboxylative methylation of alkylresorcylic acid. To the best of our knowledge, this is the first report of a methyltransferase that catalyzes decarboxylative methylation.

INTRODUCTION

Type III polyketide synthases (PKSs) are simple homodimers of ketosynthases that condense an extender substrate(s) onto a starter substrate iteratively by decarboxylative Claisen condensation reactions (1). They also catalyze the cyclization of the resulting β-ketoacyl molecule to give aromatic polyketides (1). Type III PKSs are distributed in diverse organisms, including plants, fungi, and bacteria, and are responsible for the syntheses of various biologically and pharmaceutically important compounds (1). In 1999, we revealed the function of the first bacterial type III PKS, the 1,3,6,8-tetrahydroxynaphthalene synthase RppA, from the Gram-positive, filamentous bacterium Streptomyces griseus (5). The complete genome sequence of S. griseus subsequently revealed another type III PKS gene (srsA) (12). We previously characterized the function of the srs operon, which consists of srsA, srsB, and srsC (7). This operon is responsible for the production of methylated phenolic lipids, which confer resistance to β-lactam antibiotics on the producer strain (7). We proposed a biosynthetic pathway for the methylated phenolic lipids, as shown in Fig. 1A (7). First, the type III PKS SrsA synthesizes 3-alkyl-4-methylresorcinols. The O-methyltransferase SrsB then catalyzes the O-methylation of the 3-alkyl-4-methylresorcinols to yield 3-alkyl-4-methylresorcinol-1-methyl ethers. Finally, the hydroxylase SrsC hydroxylates the alkyl-methylresorcinol-methyl ethers, resulting in 6-alkyl-2-methoxy-5-methylhydroquinones, which undergo spontaneous oxidation to give 6-alkyl-2-methoxy-5-methylquinones. However, detailed in vitro analyses are required to confirm the proposed enzymatic reactions, especially those reactions catalyzed by SrsA and SrsB, for the reasons given below.

Fig 1.

The Srs biosynthetic pathway. (A) Alkylquinone formation by Srs proteins from 14-methylpentadecanoyl-CoA (iC₁₆-CoA), malonyl-CoA, and methylmalonyl-CoA, as described in our previous study. (B) Possible pathways for alkylresorcinol methyl ether formation by SrsA and SrsB. A dashed arrow indicates an unfeasible reaction; presumably, alkylresorcylic acid methyl ether is stable and is not readily decarboxylated. The pathway proposed by this study is highlighted with thick arrows.

The type III PKS gene, srsA, was found to direct the production of alkylresorcinols using intrinsic branched-chain fatty acids as starter units (for example, iso-C₁₅, iso-C₁₆, iso-C₁₇, and anteiso-C₁₇ fatty acids) when expressed in Streptomyces lividans (7). The recombinant SrsA protein also produced 3-(13′-methyltetradecyl)-4-methylresorcinol (compound 1), using 14-methylpentadecanoyl-coenzyme A (CoA) (iso-C₁₆ fatty acyl-CoA and iC₁₆-CoA) as the starter substrate and two molecules of malonyl-CoA and one molecule of methylmalonyl-CoA as extender substrates (Fig. 1A). Although alkylpyrones were also detected as trace by-products in both in vivo and in vitro experiments (7), we do not now consider alkylpyrone formation to be a significant biological or enzymatic function of SrsA; alkylpyrone formation by SrsA is not discussed further in this study. The structure of compound 1 allows two possible extender substrate condensation patterns to be proposed (Fig. 2). However, no experimental evidence on the order of the condensation of the two molecules of malonyl-CoA and the one molecule of methylmalonyl-CoA has been obtained to date. Furthermore, it remains to be elucidated whether SrsA produces alkylresorcinol as a direct product. Alkylresorcylic acids are readily converted to alkylresorcinols by facile nonenzymatic decarboxylation, and it is therefore possible that the direct product of the SrsA reaction is alkylresorcylic acid (Fig. 1B).

Fig 2.

Two possible extender substrate condensation patterns for alkylresorcinol production by SrsA. The first pattern uses one methylmalonyl-CoA followed by two malonyl-CoAs (top). The second pattern uses two molecules of malonyl-CoA followed by one molecule of methylmalonyl-CoA (bottom). These patterns can be distinguished by studying the incorporation of [¹³C₃]malonyl-CoA, since three or four ¹³C carbon atoms are incorporated into the alkylresorcinol product in the case of the former and latter patterns, respectively. The ¹³C carbon atoms are indicated by black circles.

The second gene of the srs operon, srsB, encodes a 177-amino-acid protein that includes an isoprenylcysteine carboxyl methyltransferase (ICMT) domain. ICMT is an endoplasmic reticulum membrane-localized enzyme that catalyzes the O-methylation of the α-carboxyl group of the newly exposed cysteine residue of prenylated proteins (14). ICMT lacks homology with other methyltransferases. Protein localization predictions from the primary amino acid sequence (SOSUI analysis [http://bp.nuap.nagoya-u.ac.jp/sosui/]) indicated that SrsB is also a membrane-bound protein possessing two transmembrane helices (R67 to W89 and E131 to A153). When srsAB was expressed in S. lividans, the alkylresorcinols produced by SrsA appeared to be converted to alkylresorcinol methyl ethers, indicating that SrsB is an O-methyltransferase that acts on alkylresorcinols. However, the reaction catalyzed by SrsB has not been characterized in vitro. If the product of the SrsA-catalyzed reaction were an alkylresorcylic acid, the genuine substrate of SrsB may be alkylresorcylic acid rather than alkylresorcinol (Fig. 1B).

The purpose of this study was to carry out a detailed in vitro characterization of the reactions catalyzed by SrsA and SrsB. We revealed that the type III PKS SrsA produces an alkylresorcylic acid, while the membrane-bound O-methyltransferase SrsB catalyzes an unprecedented decarboxylative methylation of the alkylresorcylic acid to produce an alkylresorcinol methyl ether. A possible mechanism for the reaction catalyzed by SrsB is discussed.

MATERIALS AND METHODS

Materials.

Escherichia coli strain JM109, plasmid pUC19, restriction enzymes, and other DNA-modifying enzymes used for DNA manipulation were purchased from Takara Bio (Otsu, Japan). S. griseus IFO13350 was obtained from the Institute of Fermentation, Osaka, Japan (IFO). S. lividans TK21 and plasmid pIJ6021 (13) were obtained from the John Innes Centre (Norwich, United Kingdom). Plasmid pHSA81 was obtained from Michihiko Kobayashi (University of Tsukuba, Ibaraki, Japan). [2-¹⁴C]malonyl-CoA and dl-2-[methyl-¹⁴C]methylmalonyl-CoA were purchased from American Radiolabeled Chemicals (St. Louis, MO).

[¹³C₃]malonyl-CoA was purchased from Isotec (Miamisburg, OH). Nonlabeled CoA esters and S-adenosyl-l-methionine (SAM) chloride were purchased from Sigma. 14-Methylpentadecanoyl-CoA (iC₁₆-CoA) was prepared according to a method described previously by Blecher (3). The S. griseus chromosomal DNA was used as a template for PCRs. The absence of undesired alterations during PCR was confirmed by nucleotide sequencing.

Construction of pIJ6021-srsB.

A 0.6-kb DNA fragment containing the srsB coding sequence was amplified by PCR with primer I (5′-GCGGAATTCCATATGACCTCGACCTGGTACACC-3′ [the EcoRI site is underlined, the NdeI site is shown in italic type, and the start codon of srsB is shown in boldface type]) and primer II (5′-GCGAAGCTTCTCGAGGCCGACGACCAGGACGTCGA-3′ [the HindIII site is shown in italic type]) and cloned between the EcoRI and HindIII sites of pUC19, resulting in pUC19-srsB. The NdeI-HindIII fragment excised from pUC19-srsB was cloned between the NdeI and HindIII sites of pIJ6021, resulting in pIJ6021-srsB.

Construction of pHSA81-srsBhis.

A 0.6-kb DNA fragment containing the srsB coding sequence was amplified by PCR with primer III (5′-GCGGGTACCCATATGACCTCGACCTGGTACACCC-3′ [the KpnI site is underlined, the NdeI site is shown in italic type, and the start codon of srsB is shown in boldface type]) and primer IV (5′-GCGAAGCTTTCAATGATGATGATGATGATGCGCGGGCGCGTCCGCGGCCGGGAG-3′ [the His tag-coding sequence is underlined, the HindIII site is shown in italic type, and the stop codon of srsB is shown in boldface type]) and cloned between the KpnI and HindIII sites of pUC19, resulting in pUC19-srsBhis. The NdeI-HindIII fragment excised from pUC19-srsBhis was cloned between the NdeI and HindIII sites of pHSA81, resulting in pHSA81-srsBhis.

Production of SrsB.

S. lividans TK21 harboring pIJ6021-srsB was used to inoculate yeast extract-malt extract (YEME) medium (1% glucose, 34% sucrose, 0.3% yeast extract, 0.3% malt extract, 0.5% Bacto peptone, 0.1% MgCl₂ · 6H₂O [pH 7.0] [100 ml]) containing 5 μg/ml kanamycin and incubated at 30°C. After 24 h, 10 μg/ml thiostrepton was added to induce the tip promoter, and the incubation of the culture was continued for a further 48 h. Cells were harvested by centrifugation and resuspended in buffer A (50 mM Tris-HCl, 10% glycerol [pH 8.0] [20 ml]) and then disrupted by sonication. The cell-free crude extract was prepared by the removal of the cell debris by centrifugation at 10,000 × g for 20 min. The precipitate was resuspended in buffer A (20 ml) and used as the insoluble fraction. The crude extract was centrifuged at 150,000 × g for 1 h at 4°C. The supernatant and precipitate (resuspended in buffer A [20 ml]) were used as cytosol and membrane fractions, respectively.

Production and purification of C-terminally His₆-tagged SrsB.

S. lividans TK21 harboring pHSA81-srsBhis was used to inoculate YEME medium (200 ml) containing 5 μg/ml thiostrepton and incubated at 30°C for 72 h. Cells were harvested by centrifugation and resuspended in buffer B (50 mM Tris-HCl, 150 mM NaCl, 5 mM imidazole, 10% glycerol [pH 8.0] [20 ml]) and then disrupted by sonication. The crude extract was prepared by the removal of the cell debris by centrifugation at 10,000 × g for 20 min. The crude extract was centrifuged at 150,000 × g for 1 h at 4°C to pellet the membrane. SrsB-containing membranes were suspended in buffer B containing 0.1% n-dodecyl-β-d-maltopyranoside (DDM) (20 ml) at 4°C for 30 min. The solution was subsequently centrifuged at 300,000 × g for 1 h at 4°C to remove the membrane. The supernatant containing solubilized SrsB was applied onto a nickel-nitrilotriacetic acid (Ni-NTA) superflow column (Qiagen), washed 10 times with buffer B containing 0.1% DDM and 20 mM imidazole, and then eluted with buffer B containing 0.1% DDM and 250 mM imidazole. Excess imidazole was removed by Superdex 200 gel filtration chromatography in buffer C (50 mM Tris-HCl, 150 mM NaCl, 10% glycerol, 0.1% DDM [pH 8.0]). The purity of recombinant SrsB was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Protein concentrations were measured by using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Wilmington, DE).

Analysis of the order of extender substrate condensation in the SrsA reaction.

The C-terminally His-tagged SrsA protein, which had the sequence SrsA (Met-1 to Trp-350)-DPNSSSVDKLAAALE-His₆, was prepared as described previously (7). The reaction mixture consisted of 50 mM Tris-HCl (pH 8.0), 100 μM iC₁₆-CoA, 100 μM [¹³C₃]malonyl-CoA, 100 μM methylmalonyl-CoA, and 5 μM SrsA, in a total volume of 500 μl. Reaction mixtures were incubated at 30°C for 2 h. The products were extracted with 500 μl of ethyl acetate. The organic layer was evaporated, and the residue was dissolved in 20 μl of methanol. The sample (10 μl) was used for reverse-phase liquid chromatography–atmospheric pressure chemical ionization mass spectrometric (LC-APCIMS) analysis using the Esquire HCT system (Bruker Daltonics) equipped with a reversed-phase Pegasil-B C₄ column (4.6 by 250 mm; Senshu Scientific, Tokyo, Japan). The compounds were eluted isocratically with 50% CH₃CN and 0.1% acetic acid in water, at a flow rate of 1.0 ml/min.

Analysis of alkylresorcylic acids produced by SrsA.

For the analysis of alkylresorcylic acids, the reaction mixture consisted of 50 mM Tris-HCl (pH 7.0), 100 μM iC₁₆-CoA, 100 μM malonyl-CoA, 100 μM methylmalonyl-CoA, and 5 μM SrsA, in a total volume of 500 μl. Reaction mixtures were incubated at 30°C for 5, 20, and 40 min. The products were extracted with 200 μl of ethyl acetate. Because alkylresorcylic acids are readily decarboxylated by overevaporation, the organic layer was evaporated to approximately 10 μl by visual evaluation, methanol was added (50 μl), and the mixture was evaporated again to 15 μl. During the evaporation process, the test tubes containing the organic layers were kept cool, at approximately 4°C. The samples were analyzed by LC-APCIMS analysis with a reversed-phase Pegasil-B C₄ column (4.6 by 250 mm). The compounds were eluted with a linear acetonitrile-water gradient (both containing 0.1% acetic acid and 50 to 100% acetonitrile for 60 min at a flow rate of 1.0 ml/min). To study the time course of the production of compounds 1 and 3 in the SrsA reaction, 5 μM SrsA was incubated with 100 μM iC₁₆-CoA, 50 μM malonyl-CoA, 20 μM [¹⁴C]malonyl-CoA, 50 μM methylmalonyl-CoA, and 20 μM [¹⁴C]methylmalonyl-CoA, in a total volume of 100 μl (50 mM Tris-HCl, pH 7.0) at 30°C for 5, 10, 20, 30, 45, and 60 min. The products were extracted with 100 μl of ethyl acetate and dissolved in methanol (20 μl) as described above for the preparation of alkylresorcylic acid samples. The ¹⁴C-labeled products were separated by thin-layer chromatography (TLC) using a silica gel 60 WF254 TLC plate (Merck, Darmstadt, Germany) (benzene-acetone-acetic acid ratio of 85:15:1, vol/vol/vol) and detected by autoradiography using a BAS-MS imaging plate (Fuji Film, Tokyo, Japan).

In vitro SrsB reaction.

For LC-APCIMS analysis, the reaction mixture consisted of 50 mM Tris-HCl (pH 7.0), 100 μM iC₁₆-CoA, 100 μM malonyl-CoA, 100 μM methylmalonyl-CoA, 5 μM SrsA, 100 μM S-adenosyl-l-methionine chloride, and the SrsB-containing crude extract (5 μl), in a total volume of 500 μl. The standard reaction mixture for TLC analysis contained 50 mM Tris-HCl (pH 7.0), 100 μM iC₁₆-CoA, 50 μM malonyl-CoA, 20 μM [¹⁴C]malonyl-CoA, 50 μM methylmalonyl-CoA, 20 μM [¹⁴C]methylmalonyl-CoA, 5 μM SrsA, 100 μM S-adenosyl-l-methionine chloride, and 0.5 μM purified SrsB or the SrsB-containing fraction (1 μl of the crude cell extract, cytosolic fraction, or membrane fraction), in a total volume of 100 μl. When purified SrsB was used, DDM (0.1%) was added to the reaction mixture. The reaction mixture was incubated at 30°C for the designated time periods and then extracted with ethyl acetate. The organic layer was concentrated by evaporation and dissolved in 20 μl methanol for TLC and LC-APCIMS analyses. For in vitro analysis with compound 1, the reaction mixture consisted of 50 mM Tris-HCl (pH 7.0), 50 μM compound 1, 100 μM S-adenosyl-l-methionine chloride, and the SrsB-containing crude extract (5 μl), in a total volume of 500 μl. The reaction mixture was incubated at 30°C for 2 h and extracted with 500 μl of ethyl acetate. The organic layer was concentrated by evaporation and dissolved in 20 μl methanol for high-performance liquid chromatography (HPLC) analysis. The conditions for analytical HPLC were as follows: a Pegasil-B C₄ column (4.6 by 250 mm) eluted isocratically with 80% CH₃CN and 0.1% trifluoroacetic acid in water at a flow rate of 1.0 ml/min.

RESULTS

SrsA uses methylmalonyl-CoA as the first extender substrate, followed by two molecules of malonyl-CoA as the second and third extender substrates.

We previously reported that the recombinant SrsA protein synthesized 3-(13′-methyltetradecyl)-4-methylresorcinol (compound 1), using iC₁₆-CoA as a starter substrate and two molecules of malonyl-CoA and one molecule of methylmalonyl-CoA as extender substrates (Fig. 1A) (7). To analyze the order of extender substrate condensation, we examined an in vitro SrsA reaction using [¹³C₃]malonyl-CoA, in which all three carbon atoms of the malonate moiety were labeled with ¹³C. The structure of compound 1 allows two possible extender substrate condensation patterns to be proposed. The first pattern involves the condensation of one methylmalonyl-CoA extender unit, followed by two malonyl-CoAs (Fig. 2, top). In this case, three ¹³C carbon atoms would be expected to be incorporated into compound 1 when SrsA is incubated with [¹³C₃]malonyl-CoA, methylmalonyl-CoA, and iC₁₆-CoA. The second pattern involves the condensation of two molecules of malonyl-CoA followed by one molecule of methylmalonyl-CoA (Fig. 2, bottom). In contrast to the first pattern, four ¹³C carbon atoms would be expected to be incorporated into compound 1 in this case. LC-APCIMS analysis of compound 1 produced by the reaction mixture containing [¹³C₃]malonyl-CoA revealed a molecular ion peak of [M + H]⁺ at m/z 338.3 (see Fig. S1 in the supplemental material), corresponding to the incorporation of three ¹³C carbon atoms ([M + H]⁺ at m/z 335.3 was observed for compound 1 produced in the presence of unlabeled malonyl-CoA, as shown in Fig. S1 in the supplemental material). From this result, we can conclude that the first pattern corresponds to the natural biosynthetic pathway, as previously proposed.

SrsA produces alkylresorcylic acid as its direct product.

We analyzed the products of the SrsA-catalyzed reaction in detail. When SrsA was incubated with iC₁₆-CoA, malonyl-CoA, and methylmalonyl-CoA for 40 min at 30°C, alkylresorcinol 1 was detected (Fig. 3), in accordance with our previous result (7). However, trace quantities of another product (compound 3) were also detected in the reaction. When the incubation period was shortened to 5 min, compound 3 was detected as the major product, and only trace quantities of compound 1 were observed (Fig. 3). To analyze the time course of the production of compounds 1 and 3 in the SrsA reaction more precisely, SrsA was incubated with iC₁₆-CoA, malonyl-CoA, [¹⁴C]malonyl-CoA, methylmalonyl-CoA, and [¹⁴C]methylmalonyl-CoA at 30°C for 5, 10, 20, 30, 45, and 60 min, and the reaction mixtures were analyzed by radio-TLC (Fig. 4A). Compound 3 was the major product observed after a 5-min incubation, but the amount present decreased as the incubation period was prolonged, with none being detected after incubation times of 45 and 60 min. In contrast, compound 1 was the minor product after 5 min of incubation but was detected in increasingly high proportions after prolonged incubation. This result indicated that compound 3 was the initial product of the reaction and was then converted to compound 1. In the LC-APCIMS analysis, an [M − H]⁻ molecular ion peak at m/z 377.3 was observed for compound 3 (data not shown), indicating that the molecular mass of compound 3 is 378.3 Da, 44 Da higher than that of compound 1 (corresponding to the molecular mass of a molecule of CO₂). From these results, together with the condensation order of the extender substrates described above (Fig. 2, top), we concluded that compound 3 was 2-(13′-methyltetradecyl)-3-methylresorcylic acid (molecular weight [MW] of 378.3) (see Fig. 1B for the structure). We also concluded that this alkylresorcylic acid was the direct product of SrsA, rather than the alkylresorcinol 1 originally proposed.

Fig 3.

Analysis of in vitro SrsA reaction products by HPLC. Shown are HPLC chromatograms of the products generated by SrsA with iC₁₆-CoA, malonyl-CoA, and methylmalonyl-CoA after incubation for 5, 20, and 40 min and authentic alkylresorcinol (compound 1). mAU, milli-absorbance unit.

Fig 4.

Analysis of in vitro SrsA and SrsB reaction products by radio-TLC. Assays were conducted at 30°C for 5, 10, 20, 30, 45, and 60 min. (A) SrsA was incubated with iC₁₆-CoA, malonyl-CoA, [¹⁴C]malonyl-CoA, methylmalonyl-CoA, and [¹⁴C]methylmalonyl-CoA. (B) SrsB was incubated with SrsA, iC₁₆-CoA, malonyl-CoA, [¹⁴C]malonyl-CoA, methylmalonyl-CoA, and [¹⁴C]methylmalonyl-CoA in the absence of SAM. (C) SrsB was incubated with SrsA, iC₁₆-CoA, malonyl-CoA, [¹⁴C]malonyl-CoA, methylmalonyl-CoA, and [¹⁴C]methylmalonyl-CoA in the presence of SAM.

In vitro analysis of the SrsB reaction using the crude extract of S. lividans cells expressing srsB.

To allow in vitro studies of the SrsB-catalyzed reaction, we attempted to produce SrsB in a conventional Streptomyces host, S. lividans. For this purpose, the srsB gene was placed under the control of the strong, thiostrepton-inducible tipA promoter on the high-copy-number plasmid pIJ6021, resulting in pIJ6021-srsB. The plasmid was introduced into S. lividans, and the production of SrsB was analyzed by SDS-PAGE (Fig. 5A). The recombinant SrsB protein (molecular mass of approximately 19 kDa) was detected in the crude cell extract of S. lividans harboring pIJ6021-srsB and was absent in a control strain (S. lividans harboring the empty vector pIJ6021) (Fig. 5A, soluble fraction, lane S). Note that the soluble fraction contained both cytoplasmic and membrane proteins. We used the cell-free crude extract of S. lividans harboring pIJ6021-srsB for studies of the in vitro SrsB-catalyzed reaction.

Fig 5.

Localization analysis of SrsB. (A) SDS-PAGE analysis of subcellular fractions of S. lividans harboring pIJ6021-srsB or the empty vector pIJ6021 as a negative control. S and I denote the soluble and insoluble fractions, respectively, obtained by centrifugation (10,000 × g for 20 min) of the sonicated cells. Note that the soluble fraction is referred to as the “cell-free crude extract” in the text. C and M denote the cytosolic and membrane fractions, respectively, obtained by the centrifugation (150,000 × g for 1 h) of the soluble fraction. A molecular mass marker was applied to lane M. (B) SDS-PAGE analysis of purified His₆-tagged SrsB from S. lividans harboring pHSA81-srsBhis. Lane 1, membrane fraction; lane 2, solubilized proteins after 0.1% DDM treatment of the membrane fraction; lane 3, membrane fraction after treatment with 0.1% DDM; lane 4, His₆-tagged SrsB purified by Ni-NTA chromatography; lane M, molecular mass marker. (C) Radio-TLC analysis. Crude extract, cytosolic, and membrane fractions prepared from S. lividans harboring pIJ6021-srsB and purified SrsB prepared from S. lividans harboring pHSA81-srsBhis were assayed for methyltransferase activity (at 30°C for 20 min). A crude extract prepared from S. lividans harboring pIJ6021 was used as a negative control (“control crude”).

Since the direct product of SrsA had been revealed to be alkylresorcylic acid, as described above, we postulated that the genuine substrate of SrsB is not alkylresorcinol but alkylresorcylic acid. However, alkylresorcylic acid cannot be readily prepared for use as a substrate in in vitro experiments, due to its facile decarboxylation. We therefore coupled the SrsB reaction with the alkylresorcylic acid-producing SrsA reaction. The crude cell extract containing SrsB was incubated for 5, 10, 20, 30, 45, and 60 min with SrsA, malonyl-CoA, [¹⁴C]malonyl-CoA, methylmalonyl-CoA, [¹⁴C]methylmalonyl-CoA, and iC₁₆-CoA in the presence of the probable methyl group donor, S-adenosyl-l-methionine (SAM), and the reaction mixtures were analyzed by radio-TLC (Fig. 4C). As expected, a new product (compound 2) was detected, even after only 5 min of incubation (Fig. 4C). When SAM was absent from the reaction mixture, compound 2 was not observed (Fig. 4B), indicating that compound 2 is the product of the methylation of alkylresorcylic acid 3 or alkylresorcinol 1. We next examined the reaction using nonlabeled substrates. When the reaction mixture generated after 5 min of incubation was analyzed by LC-APCIMS, the production of compound 2 was again observed (Fig. 6B). Compound 2 was not detected when either the SrsB-containing crude extract or SAM was absent from the reaction mixture (Fig. 6A). Alkylresorcinol 1 and its O-methylated form (the hydroxyl group at C-1 of compound 1 is O-methylated), which had been previously purified from S. lividans strains expressing srsA and srsAB, respectively (7), were used as authentic samples to determine the structure of compound 2 (Fig. 6C). This allowed compound 2 to be identified as the alkylresorcinol methyl ether. Note that in the in vitro SrsAB reaction (Fig. 4C and 6B), we did not detect any other products that could give evidence for the formation of O-methylated alkylresorcylic acid.

Fig 6.

Analysis of in vitro SrsA and SrsAB reaction products by HPLC. (A) HPLC chromatogram of the products generated by SrsA with iC₁₆-CoA, malonyl-CoA, and methylmalonyl-CoA for 5 min. (B) HPLC chromatograms of the products generated by SrsB with SrsA, iC₁₆-CoA, malonyl-CoA, methylmalonyl-CoA, and SAM for 5 min. (C) HPLC chromatogram of authentic alkylresorcinol (compound 1) and alkylresorcinol methyl ether (compound 2).

Next, we examined whether SrsB could use alkylresorcinol 1 as a substrate. The SrsB-containing crude extract was incubated at 30°C for 2 h with alkylresorcinol 1 in the presence of SAM, and the reaction mixture was analyzed by HPLC. No conversion of alkylresorcinol 1 was detected (see Fig. S2 in the supplemental material). This result strongly indicated that the genuine substrate of SrsB in the in vitro SrsAB reaction was alkylresorcylic acid 3 rather than alkylresorcinol 1. The conversion of alkylresorcylic acid 3 to alkylresorcinol methyl ether 2 requires not only the O-methylation of the hydroxyl group at C-6 of compound 3 but also the decarboxylation of compound 3 (Fig. 1B). Because alkylresorcylic acid 3 is readily converted to alkylresorcinols by facile nonenzymatic decarboxylation (Fig. 3 and 4), we could not purify alkylresorcylic acid 3 for a further in vitro analysis of SrsB. The reaction catalyzed by SrsB is discussed further in Discussion.

Localization of SrsB in the membrane.

The prediction of the localization of SrsB using the SOSUI program indicated that SrsB is likely to be a membrane-bound protein. To confirm this, we separated the SrsB-containing crude extract into cytoplasmic and membrane fractions by high-speed centrifugation (150,000 × g for 1 h at 4°C). The recombinant SrsB protein was detected in the membrane fraction by SDS-PAGE (Fig. 5A). SrsB activity was also detected in the membrane fraction but not in the cytoplasmic fraction, as determined by radio-TLC analysis (Fig. 5C). From these results, we concluded that SrsB is localized in the cell membrane. Alkylresorcylic acid, the substrate of SrsB, is likely to be localized in the cell membrane, so the localization of SrsB in the membrane is reasonable.

In vitro analysis using purified SrsB.

As described above, SrsB activity was detected in the membrane fraction of S. lividans expressing srsB. We next attempted to purify the recombinant SrsB protein from the membrane fraction, and C-terminally His₆-tagged SrsB was produced in S. lividans. The His₆-tagged SrsB protein was detected in the membrane fraction by SDS-PAGE, as expected (Fig. 5B, lane 1). The SrsB-containing membrane was suspended in a buffer containing 0.1% DDM and incubated at 4°C for 30 min to solubilize the His₆-tagged SrsB protein from the membrane. The solution was subsequently centrifuged at 300,000 × g for 1 h at 4°C to remove the membrane (Fig. 5B, lanes 2 and 3). The solubilized His₆-tagged SrsB was purified from the supernatant by Ni-NTA superflow chromatography to give a single major protein band of 20 kDa on SDS-PAGE gels (Fig. 5B, lane 4). Purified His₆-tagged SrsB showed clear SrsB activity, as determined by radio-TLC analysis (Fig. 5C). This result indicates that no additional enzymes are required for the conversion of alkylresorcylic acid 3 to alkylresorcinol methyl ether 2 by SrsB.

DISCUSSION

Previously, we reported that the srs operon directs the biosynthesis of methylated phenolic lipids in S. griseus. In the present study, we carried out a detailed in vitro characterization of the reactions catalyzed by SrsA and SrsB and revealed the functions of these proteins.

By using [¹³C₃]malonyl-CoA in an in vitro SrsA-catalyzed reaction, we confirmed that the type III PKS SrsA uses methylmalonyl-CoA as the first extender substrate, followed by two molecules of malonyl-CoA as the second and third extender substrates. Furthermore, we revealed that SrsA is an alkylresorcylic acid synthase, with the direct product of SrsA being alkylresorcylic acid rather than alkylresorcinol. Very recently, we have shown that a type III PKS (FtpA) from Myxococcus xanthus catalyzes the same reaction as that catalyzed by SrsA (8). FtpA shows a relatively high level of amino acid sequence similarity (52% identity) with SrsA. Alkylresorcylic acid synthesis by SrsA and FtpA is unique due to their strict regulation of the extender unit condensation order. Little is known about the mechanism of regulation of the extender unit condensation order in type III PKSs.

Our studies of microbial type III PKSs involved in phenolic lipid biosynthesis have revealed that these PKSs can be classified into four groups: alkylpyrone synthases, alkylresorcinol synthases, alkylresorcylic acid synthases, and alkylresorcylic acid/alkylresorcinol synthases. ArsC from Azotobacter vinelandii (6) and BpsA from Bacillus subtilis (11) are examples of alkylpyrone synthases, while ArsB from A. vinelandii (6) is an alkylresorcinol synthase. 2′-Oxoalkylresorcylic acid synthase (ORAS) from Neurospora crassa (4), SrsA, and FtpA are alkylresorcylic acid synthases. AgqA from Actinoplanes missouriensis (2) produces both alkylresorcylic acid and alkylresorcinol as the direct products. As the alkylresorcylic acids produced by alkylresorcylic acid synthases are rapidly converted to alkylresorcinols nonenzymatically, the biological significance of alkylresorcylic acid production in some of these systems has been unknown to date. However, in the present study, we have confirmed the production of alkylresorcylic acid by SrsA during phenolic lipid biosynthesis in S. griseus. Alkylresorcinol does not serve as a substrate of the tailoring enzyme SrsB, and alkylresorcylic acid production by SrsA is therefore essential for the successful production of the final polyketide product.

We successfully produced a functional SrsB protein in S. lividans, which enabled us to develop an in vitro SrsB reaction system in which the SrsB reaction was coupled with the alkylresorcylic acid-producing SrsA reaction. Note that the SrsB-catalyzed decarboxylative methylation of alkylresorcylic acid 3 may be a rate-limiting step in this system, at least under the conditions used; therefore, a considerable amount of alkylresorcinol 1 was nonenzymatically produced from alkylresorcylic acid 3 (Fig. 4 and 5). We confirmed that SrsB is localized in the membrane. His₆-tagged SrsB could be purified after solubilization from the membrane by using detergent. The purified His₆-tagged SrsB was found to catalyze the conversion of alkylresorcylic acid 3 to alkylresorcinol methyl ether 2. This requires not only the O-methylation of the hydroxyl group at C-6 but also the decarboxylation of compound 3 (Fig. 1B). SrsB was found to be incapable of methylating alkylresorcinol 1, indicating that decarboxylation does not precede O-methylation. On the other hand, O-methylated alkylresorcylic acid was not detected in the in vitro SrsAB reaction, while it is presumed that the decarboxylation of O-methylated alkylresorcylic acid does not occur readily. From these results, we postulate that O-methylation is coupled with decarboxylation in this case and that SrsB therefore catalyzes the decarboxylative methylation of alkylresorcylic acid. Here we propose a mechanism for the reaction catalyzed by SrsB (Fig. 7A), in which decarboxylation is inevitable during the course of O-methylation. Alkylresorcylic acid 3 is first tautomerized to compound 3′. A methyl group is then transferred from SAM onto the oxygen atom at C-6 of compound 3′ by SrsB, and this step is accompanied by the decarboxylation of compound 3′ to produce alkylresorcinol methyl ether 2. The nonenzymatic decarboxylation of alkylresorcylic acid is readily explained by a similar reaction mechanism (Fig. 7B). The C-6-specific O-methylation of compound 3 by SrsB (no product corresponding to O-methylation at the hydroxyl group at C-4 of compound 3 has been observed to date) also supports the proposed reaction mechanism. It should be noted that SrsB did not catalyze the decarboxylation of alkylresorcylic acid 3 to produce alkylresorcinol 1 in the absence of SAM (compare Fig. 4B and A). Thus, SAM is essential not only for O-methylation but also for decarboxylation by SrsB. This also supports the proposed reaction mechanism.

Fig 7.

Reaction proposed to be catalyzed by SrsB. (A) Proposed biosynthetic pathway for the production of alkylresorcinol methyl ether by SrsA and SrsB. First, SrsA produces alkylresorcylic acid 3, which is tautomerized to compound 3′. The methyl group of SAM is then transferred onto the oxygen atom at C-6 of compound 3′ by SrsB, accompanied by the decarboxylation of compound 3′, to produce alkylresorcinol methyl ether 2. (B) Mechanism of nonenzymatic decarboxylation of alkylresorcylic acid to produce alkylresorcinol. Decarboxylation occurs via the tautomerism of alkylresorcylic acid.

Some O-methyltransferases, including catechol-O-methyltransferases, chalcone O-methyltransferases, and isoflavone O-methyltransferases, catalyze the transfer of a methyl group from SAM to a hydroxyl group (10, 15). However, the reaction catalyzed by SrsB is different from the reactions catalyzed by these O-methyltransferases, as decarboxylation accompanies O-methylation. SrsB belongs to the ICMT family and is not homologous with these O-methyltransferases. Furthermore, SrsB, as with other ICMT family members, does not contain any conserved consensus sequences described for known soluble methyltransferases (9). The molecular mechanism of SrsB catalysis remains to be elucidated. Operons analogous to srsAB and srsABC are widely distributed among Gram-positive and Gram-negative bacteria (7), and some of the SrsB homologs encoded by these operons may catalyze the decarboxylative methylation of alkylresorcylic acids, similarly to SrsB. For example, FtpB (54% amino acid sequence identity with SrsB) from M. xanthus was postulated to have the same function as that of SrsB as a result of our recent in vivo analysis (8). However, some SrsB homologs encoded by srsAB- and srsABC-like operons catalyze O-methylation without concomitant decarboxylation. For example, BpsB (39% amino acid sequence identity with SrsB) from B. subtilis was demonstrated previously to catalyze the O-methylation of alkylpyrones produced by the type III PKS BpsA in vivo (11). Characterizations of the reactions catalyzed by other SrsB homologs, comparisons of their amino acid sequences, and mutational analysis would provide insight into the molecular mechanism of SrsB catalysis.

In conclusion, we have revealed that the type III PKS SrsA produces an alkylresorcylic acid and that the membrane-bound O-methyltransferase SrsB then catalyzes the decarboxylative methylation of alkylresorcylic acid to produce an alkylresorcinol methyl ether. We have proposed a plausible mechanism for the reaction catalyzed by SrsB (Fig. 7A). Although the molecular mechanism of SrsB catalysis remains to be elucidated, this study is significant, as to the best of our knowledge, it represents the first in vitro demonstration of an enzyme that catalyzes decarboxylative methylation. SrsB is the first bacterial ICMT family methyltransferase to be characterized in an in vitro assay.