Oxidative Catabolism of (+)-Pinoresinol Is Initiated by an Unusual Flavocytochrome Encoded by Translationally Coupled Genes within a Cluster of (+)-Pinoresinol-Coinduced Genes in Pseudomonas sp. Strain SG-MS2

(+)-Pinoresinol is an important plant defense compound, a major food lignan for humans and some other animals, and the model compound used to study degradation of the β-β′ linkages in lignin. We report a gene cluster, in one strain each of Pseudomonas and Burkholderia, that is involved in the oxidative catabolism of (+)-pinoresinol. The flavoprotein component of the α-hydroxylase which heads the pathway belongs to the 4-phenol oxidizing (4PO) subgroup of the vanillyl alcohol oxidase/p-cresol methyl hydroxylase (VAO/PCMH) enzyme family but constitutes a novel combination of cofactor and electron acceptor properties for the family. It is translationally coupled with a cytochrome gene whose product is also required for activity. The work casts new light on the biology of (+)-pinoresinol and its transformation to other bioactive molecules. Potential applications of the findings include new options for deconstructing lignin into useful chemicals and the generation of new phytoestrogenic enterolactones from lignans.

ABSTRACT

Burkholderia sp. strain SG-MS1 and Pseudomonas sp. strain SG-MS2 have previously been found to mineralize (+)-pinoresinol through a common catabolic pathway. Here, we used comparative genomics, proteomics, protein semipurification, and heterologous expression to identify a flavoprotein from the vanillyl alcohol oxidase/p-cresol methyl hydroxylase (VAO/PCMH) enzyme family in SG-MS2 that carries out the initial hydroxylation of (+)-pinoresinol at the benzylic carbon. The cognate gene is translationally coupled with a downstream cytochrome gene, and the cytochrome is required for activity. The flavoprotein has a unique combination of cofactor binding and cytochrome requirements for the VAO/PCMH family. The heterologously expressed enzyme has a K_m of 1.17 μM for (+)-pinoresinol. The enzyme is overexpressed in strain SG-MS2 upon exposure to (+)-pinoresinol, along with 45 other proteins, 22 of which were found to be encoded by genes in an approximately 35.1-kb cluster also containing the flavoprotein and cytochrome genes. Homologs of 18 of these 22 genes, plus the flavoprotein and cytochrome genes, were also found in a 38.7-kb cluster in SG-MS1. The amino acid identities of four of the other proteins within the SG-MS2 cluster suggest they catalyze conversion of hydroxylated pinoresinol to protocatechuate and 2-methoxyhydroquinone. Nine other proteins upregulated in SG-MS2 on exposure to (+)-pinoresinol appear to be homologs of proteins known to comprise the protocatechuate and 2-methoxyhydroquinone catabolic pathways, but only three of the cognate genes lie within the cluster containing the flavoprotein and cytochrome genes.

IMPORTANCE (+)-Pinoresinol is an important plant defense compound, a major food lignan for humans and some other animals, and the model compound used to study degradation of the β-β′ linkages in lignin. We report a gene cluster, in one strain each of Pseudomonas and Burkholderia, that is involved in the oxidative catabolism of (+)-pinoresinol. The flavoprotein component of the α-hydroxylase which heads the pathway belongs to the 4-phenol oxidizing (4PO) subgroup of the vanillyl alcohol oxidase/p-cresol methyl hydroxylase (VAO/PCMH) enzyme family but constitutes a novel combination of cofactor and electron acceptor properties for the family. It is translationally coupled with a cytochrome gene whose product is also required for activity. The work casts new light on the biology of (+)-pinoresinol and its transformation to other bioactive molecules. Potential applications of the findings include new options for deconstructing lignin into useful chemicals and the generation of new phytoestrogenic enterolactones from lignans.

INTRODUCTION

Pinoresinol [(+)-pinoresinol] is a plant defense molecule with both anthelmintic and antifungal properties and is a model dimer representing the key β-β′ (resinol) linkage in the complex polymeric structure of lignin (1–4). Plants produce it intracellularly by enantioselective oxidative dimerization of two coniferyl alcohol monomers with the aid of a dirigent protein (5, 6). Animals consume pinoresinol in various seeds, fruits, and vegetables, and humans also consume it in tea, coffee, and red wine (7–11). Reductive pathways for the degradation of the compound/linkage in decaying or ingested plant material have been characterized to some degree, but oxidative catabolism is much less well understood (1, 2).

Pinoresinol can be reduced by soil and gut bacteria in the presence of alternative carbon sources by a mechanism that is also shared with plants (12). This involves transformation to secoisolariciresinol via lariciresinol in two reductive steps by well-characterized pinoresinol-lariciresinol reductases (9). In plants, these metabolites are further transformed into other bioactive polyphenolic compounds (7–11). In mammals, gut microflora further transform secoisolariciresinol to enterolactone and enterodiols, which are collectively called mammalian phytoestrogens because of their weak estrogenic activities (7, 8, 11).

We have recently isolated two pinoresinol-mineralizing bacterial strains, Burkholderia sp. strain SG-MS1 and Pseudomonas sp. strain SG-MS2, which use an inducible pathway initiated by an oxidative step involving benzylic hydroxylation, generating a hemiketal via a quinone methide intermediate, which then is hydrated at the benzylic carbon by water (13). The hemiketal, which stays in equilibrium with the corresponding keto alcohol, undergoes an aryl-alkyl cleavage to generate a lactone and 2-methoxyhydroquinone (13). While the fate of 2-methoxyhydroquinone was not investigated further, it was assumed to be assimilated by ring cleavage (13). The lactone was shown to be further metabolized via two routes, namely, lactone ring cleavage and benzylic hydroxylation via a quinone methide intermediate (13). It was also shown that both routes of lactone metabolism lead to vanillin and vanillic acid, which were eventually assimilated by strain SG-MS2 (13).

Here, we investigate the genetics and biochemistry of the pinoresinol pathway in strain SG-MS2. Comparative genomics and proteomics combined with partial enzyme purification and heterologous expression are used to identify and characterize an unusual flavoprotein that, in complex with a cytochrome encoded by a downstream, translationally coupled gene, catalyzes the initial hydroxylation step. The genes encoding these two proteins are found to lie in a cluster of 34 genes, 22 of which are upregulated on pinoresinol exposure. The annotation of sequence-related genes also suggests some of the genes in this cluster encode enzymes catalyzing subsequent steps in a pinoresinol catabolic pathway which produces 2-methoxyhydroquinone and protocatechuate via vanillin and vanillic acid. Some other upregulated genes that are scattered across the genome are close homologs of genes known to be involved in vanillic acid assimilation in other strains.

RESULTS

Candidate pinoresinol hydroxylases from comparative genomics.

As a first step toward identifying gene-enzyme systems that might be associated with pinoresinol degradation in the two strains, all bidirectional best BLASTp hits of the proteins inferred from the annotated assemblies of the SG-MS1 and SG-MS2 genomes (see Materials and Methods) that showed >70% identity between them were identified. The SG-MS2 sequences among these hits then were screened against the annotated genome of Pseudomonas putida KT2440 (14–16), which cannot degrade pinoresinol under conditions identical to those under which SG-MS1 and SG-MS2 can (data not shown). Only seven proteins that were conserved between SG-MS1 and SG-MS2 had <26% amino acid identities with any proteins in KT2440 (Table 1). Four of the seven were annotated as hypothetical proteins in SG-MS2, and three of these were similarly annotated in SG-MS1, the fourth being annotated as p-cresol methyl hydroxylase in that strain. The other three of the seven were annotated as glyoxylase, l-rhamnonate dehydratase, and cupin in both strains. Four of these seven proteins, the glyoxylase and three hypothetical proteins, were found in a genomic cluster in both SG-MS1 and SG-MS2 (described below). Two of these hypothetical proteins were translationally coupled, with ATGA stop-start codons in both SG-MS1 and SG-MS2, noting also a putative ribosome binding site (RBS; AGGAGG) 10 bp upstream of this start codon for the cytochrome gene (Fig. 1).

TABLE 1.

Comparative genomics showing the genes/proteins present in both pinoresinol-degrading strains but absent from the non-pinoresinol-degrading strain KT2550^a

Pseudomonas sp. strain SG-MS2 (GenBank accession no. MOMO00000000)			Burkholderia sp. strain SG-MS1 (GenBank accession no. NPIH00000000)
GenBank locus tag	Length (aa)	Annotated function	Hitb	GenBank locus tag	% Identity	Annotated function
BLX42_10690	558	Hypothetical protein	bi	CIC12_20375	74.86	p-Cresol methylhydroxylase
BLX42_10695	113	Hypothetical protein	bi	CIC12_20380	76.54	Hypothetical protein
BLX42_10720	319	Glyoxalase	bi	CIC12_20295	77.35	Glyoxalase
BLX42_10760	529	Hypothetical protein	bi	CIC12_20345	72.32	Hypothetical protein
BLX42_12780	400	l-Rhamnonate dehydratase	bi	CIC12_01470	80.96	l-Rhamnonate dehydratase
BLX42_17155	169	Cupin	bi	CIC12_10925	75.44	Cupin
BLX42_21760	177	Hypothetical protein	bi	CIC12_15005	75.72	Hypothetical protein

^aGenes/proteins present, >70% amino acid identity; genes/proteins absent, <26% amino acid identity.

^bbi, bidirectional.

FIG 1.

Nucleotide-level details of the translationally coupled flavoprotein and cytochrome genes and locations of other potential regulatory and structural elements in SG-MS2, plus PCR primers to generate the three-gene construct pCR2.1-FCA.

The protein encoded by the upstream gene (locus tags CIC12_20375 and BLX42_10690 in SG-MS1 and SG-MS2, respectively; 74% identity to each other) of this pair had BLASTp best hits with functionally characterized flavoproteins, namely, the fungal enzyme vanillyl alcohol oxidase (VAO; NCBI accession no. P56216; 36% and 35% identities, respectively) and the bacterial enzyme p-cresol methyl hydroxylase (PCMH; accession no. P09788; 32% and 33% identities, respectively). Both of these proteins lie in the 4-phenol oxidizing (4PO) subgroup of the recently described VAO/PCMH enzyme family (17). All five genetically and biochemically characterized members of this subgroup characterized to date, VAO, PCMH, eugenol hydroxylase (EUGH), eugenol oxidase (EUGO), and chondrochloren decarboxylase (CCD), catalyze the oxidation of the α-carbon of para-substituted phenols, a reaction similar to that of the α-carbon hydroxylation of (+)-pinoresinol reported earlier for SG-MS1 and SG-MS2 (13, 18). (Another 4PO flavocytochome [4-ethylphenol methylenehydroxylase from Pseudomonas putida JD1] has been functionally characterized, but the cognate gene has not yet been reported [19].)

The best BLASTp match of the hypothetical protein encoded by the downstream gene (locus tags CIC12_20380 and BLX42_10695 in SG-MS1 and SG-MS2, respectively; 61% identity to each other) was a cytochrome subunit attached to PCMH (GenBank accession no. P09787.2; 30% and 24% identities to SG-MS1 and SG-MS2, respectively). Intriguingly, while the protein encoded by the upstream gene has higher identities to the oxidase VAO than to the hydroxylating dehydrogenase PCMH (as also indicated in the phylogenetic analysis shown in Fig. 4 below), the protein encoded by the downstream gene has similarities to the cytochrome found in PCMH but not in VAO. The cytochrome gene associated with PCMH (pchC) in the Pseudomonas sp. strains NCIMB 9866 (GenBank accession no. KC762703.1) and NCIMB 9869 (accession no. U96339.1) was also found to be closely linked to the PCMH gene (pchF) but in that case was not downstream of it and translationally coupled but upstream of it and separated by a 684-bp gene (pchX) encoding a hypothetical protein (20).

FIG 4.

Phylogenetic tree of p-phenol-oxidizing enzymes in the VAO/PCMH flavoprotein family showing four distinct clades, which are also differentiated based on their FAD binding and electron acceptor properties. Footnote 1, the SG-MS1 protein CIC12_20345 is a suspected (+)-pinoresinol α-hydroxylase with 74.9% identity to the functionally characterized enzyme BLX42_10609 from SG-MS2. Footnote 2, the FAD binding is predicted on the basis of the sequence alignment used to generate the phylogenetic tree provided in Fig. S4. Footnote 3, the in vivo electron acceptor of clade 3 is not characterized. In vitro, however, phenazine methosulfate (PMS) can accept electrons from BLX42_10690 as well as from flavoproteins in clades 1 and 2 and the cytochrome subunit of functionally active flavocytochrome of clade 1.

Semipurification of pinoresinol hydroxylase activity.

Pinoresinol α-hydroxylase activity was semipurified from pinoresinol-induced cells of strain SG-MS2 via ammonium sulfate precipitation and two ion exchange columns. An 80.6-fold enrichment of this activity was achieved (Table 2), but this still did not yield a homogeneous preparation on a Coomassie blue-stained SDS-PAGE gel (see Fig. S1A in the supplemental material). This preparation, which contained too little protein to quantify, was further concentrated using a 50-kDa-molecular-weight-cutoff centrifugal filter, and product formation still could be observed by liquid chromatography–time-of-flight mass spectrometry (LC-TOF/MS) in a pinoresinol degradation assay of this material. Coomassie blue staining of it on SDS-PAGE revealed twelve visible proteins (Fig. S1B).

TABLE 2.

Purification of pinoresinol α-hydroxylase from Pseudomonas sp. strain SG-MS2

Purification step	Total protein (mg)	Sp act (μM min−1 mg−1)	Total activity (μM min−1)	Fold purification	Yield (%)
Cell free crude enzyme	11,847	0.35	4,146	1	100
Ammonium sulfate precipitate (40–60%)	417	1.93	805	5.51	19.4
Anion exchange column (pH 7)	3.6	22.5	80.9	64.2	1.95
Cation exchange column (pH 6)	1.2	28.2	33.8	80.6	0.80

These twelve bands were individually eluted and digested with trypsin, and the resulting peptides were sequenced. Details of the identified proteins are listed in Table S1. Notably, the hypothetical protein (BLX42_10690) encoded by the upstream member of the translationally coupled genes described above was present in one of the sequenced bands (protein band number 5 in Fig. S1B). Twenty-six unique peptides were detected covering 54% of this 52.4-kDa protein (data not shown). However, no peptide corresponding to the downstream putative cytochrome subunit was detected in any of the sequenced bands. This could mean either that cytochrome is not a required subunit for the flavoprotein (as per the flavoprotein VAO) or that it is required (as per the flavocytochrome PCMH) but was not sufficiently abundant in the sample to be visible in the SDS-PAGE.

Substrate catalysis in VAO and PCMH is a result of two half reactions, and the first half reaction in both involves the reduction of FAD by their respective substrates, which themselves are oxidized into a p-quinone methide intermediate (17). In the second half-reaction, FAD is reoxidized and the enzyme-bound p-quinone methide intermediate is either hydrated or undergoes rearrangement, depending on the structure of the substrate (17, 21). However, in the second half-reaction the FAD is reoxidized by molecular oxygen in VAO but not in PCMH, where instead the electrons are transferred to a cytochrome subunit. The cytochrome is in turn reoxidized by transferring electrons to a small copper-containing redox protein, azurin, which is colocated in the periplasm with the flavocytochrome in P. putida and is well known to operate as an electron shuttle via oxidation-reduction between Cu(I) and Cu(II) (17, 22). Artificial electron acceptors such as phenazine methosulfate (PMS) also can reoxidize VAO and the cytochrome but not flavin subunit of PCMH (17, 23, 24). Notably, the gene immediately downstream of the translationally coupled pair in SG-MS2 (locus tag BLX42_10700) was annotated as azurin (Fig. 1). No homologue of azurin was found anywhere in the SG-MS1 genome (see below). PCMH, along with azurin, previously was shown to be present in the periplasm in two p-cresol-degrading Pseudomonas strains (17, 22), but we have no empirical data on their location in SG-MS2.

Heterologous expression of candidate genes.

Given the evidence above implicating the flavoprotein (BLX42_10690), either alone or together with the cytochrome (BLX42_10695), in the hydroxylation of pinoresinol in strain SG-MS2, we tested it for this function directly. Three constructs were tested in Escherichia coli BL21(DE3) cells: a pCR2.1 plasmid containing the flavoprotein, cytochrome, and azurin genes under the control of their native promoters (pCR2.1FCA); a pETMCSIII (25) plasmid containing the flavoprotein and cytochrome genes in their native orientations but with a His tag sequence inserted at the 5′ end of the flavoprotein gene and under the control of the T7 promoter (pETMCSIII-FC1); and one identical to the second construct but lacking the cytochrome gene (pETMCSIII-Flavo). Figure 2A shows Coomassie blue-stained SDS-PAGE gels of extracts of cells containing the three constructs.

FIG 2.

Expression profiles of pinoresinol hydroxylase in E. coli BL21(DE3). (A) Expression of E. coli BL21(DE3) alone and as a host to plasmids containing the flavoprotein, cytochrome, and azurin genes (pCR2.1-FCA), the flavoprotein and cytochrome genes (pETMCSIII-FC1), and just the flavoprotein gene (pETMCSIII-flavo). (B) Cell lysate and His-tagged purified protein fractions from pETMCSIII-flavo (flavo) and pETMCSIII-FC1 (FC1). (C) Gel shown in panel B after excitation at 488 nm and emission at 520 nm.

Extracts of cells containing the three-gene plasmid (pCR2.1-FCA) produced no obvious bands corresponding to any of the respective gene products on SDS-PAGE gels stained with Coomassie blue but could transform pinoresinol to its hydroxylated product if provided with oxidized PMS as an electron acceptor (Fig. S2). Extracts of both induced and uninduced cells containing the plasmid with His-tagged flavoprotein and cytochrome genes produced obvious bands at mobilities corresponding to the expected molecular weights of the two proteins and also transformed pinoresinol to its hydroxylated product in the presence of PMS (data not shown). Extracts of induced and uninduced cells containing the plasmid with just the flavoprotein gene produced the expected SDS-PAGE band but did not transform the pinoresinol (data not shown).

The identity of the hydroxylated pinoresinol (retention time, 3.90 min) produced from the first two constructs was confirmed using LC-TOF/MS, which showed an m/z of 373.136, in precise agreement with that of the purified hydroxylated pinoresinol previously obtained in pinoresinol degradation assays with strain SG-MS2 (13). The identity of the latter had in turn been confirmed using LC-TOF/MS, gas chromatography-mass spectrometry (GC-MS), and nuclear magnetic resonance (NMR) (13).

The flavoprotein-cytochrome complex produced by the cells with the flavoprotein-cytochrome two-gene construct (pETMCSIII-FC1) was purified using affinity chromatography (Fig. 2B) and showed benzylic oxidation of pinoresinol in LC-TOF/MS in the presence of PMS (data not shown). Peptide sequencing of the Coomassie blue-stained SDS-PAGE bands from gel slices confirmed the identity of both proteins, with 84% and 40% peptide coverage, respectively (Fig. S3).

These results indicate that the pinoresinol-oxidizing enzyme in SG-MS2 is a flavocytochrome complex produced by the translationally coupled BLX42_10690 and BLX42_10695 genes.

Properties of the purified flavocytochrome.

Three of the four members of the 4PO subgroup of the VAO/PCMH family which were previously characterized empirically for cofactor interactions have been shown to have FAD covalently attached to a loop close to the C terminus of the protein (17). The exception is CCD, which does not seem to have it covalently bound (17). Crystal structures have confirmed that the linkage in the other three occurs through an 8α-N³-histidyl-FAD bond in the case of VAO and EUGO and an 8α-O-tyrosyl-FAD bond in the case of PCMH (17). Three lines of evidence indicate our flavocytochrome also has a covalently attached FAD.

First, laser scanning (excitation at 488 nm, emission at 520 nm) of an unstained SDS-PAGE gel of a boiled aliquot (in the presence of SDS) of the His-tagged enzyme purified from E. coli showed fluorescence where the flavoprotein was present (Fig. 2C), indicating flavin was still covalently attached to the protein despite the boiling (26). Second, the visible spectrum of the oxidized enzyme showed an absorption maximum at 410 nm and a shoulder at 450 nm, which is indicative of the presence of both a cytochrome and flavin, respectively (Fig. 3A). The 410-nm absorption maximum was shifted to 414 nm, and the 450-nm shoulder was bleached upon reduction of the enzyme by 0.6 mM pinoresinol (Fig. 3A), with peaks at 550 and 520 nm, characteristic of c-type cytochromes, appearing instead (Fig. 3A). These spectral shifts of the purified enzyme are similar to the 4PO flavocytochromes 4-ethylphenol methylenehydroxylase and PCMH (19, 27). Third, structure-based sequence alignments show our flavoprotein has an H411 that is equivalent to H422 of VAO and H390 of EUGO, which are known to form 8α-N³-histidyl-FAD linkages (Fig. S4).

FIG 3.

Spectroscopic and kinetic properties of heterologously expressed, His-tagged purified pinoresinol-hydroxylase flavocytochrome. (A) Spectral changes on titration with pinoresinol. (B) Michaelis-Menten kinetics.

We conclude that the flavoprotein discovered here indeed has a covalently attached flavin and further suggest that it is attached to H411 through an 8α-N³-histidyl-FAD linkage. Phylogenetic analysis (Fig. 4) based on the alignments described above show that our flavoprotein sits in a lineage, along with VAO and EUGO, that all share the 8α-N³-histidyl-FAD linkage, but it is the only one in that lineage (possibly apart from its presumptive orthologue in SG-MS1) known to require cytochrome for activity. As such, it represents a new subclade with a distinctive combination of covalently 8α-N³-histidyl-linked FAD and cytochrome requirement for the subgroup.

The Michaelis-Menten constant (K_m) of the His-tagged purified enzyme for pinoresinol was determined based on a previously established spectrophotometric assay for PCMH (24). Enzyme activity was determined as a function of the decrease in the optical density at 600 nm (OD₆₀₀) at different pinoresinol concentrations, with PMS and 2,6-dichloroindophenol (DCIP) provided as electron acceptors. Reduction of DCIP (as a result of successive electron transfers from pinoresinol to the enzyme, PMS, and DCIP) was monitored spectrophotometrically at 600 nm. The enzyme showed a K_m of 1.17 μM (Fig. 3B), suggesting pinoresinol is a natural substrate for it.

The pinoresinol hydroxylase genes are clustered in SG-MS2 with coinduced genes that could encode subsequent catabolic steps.

Comparative proteomics was carried out on SG-MS2 cells, which were either uninduced or induced with pinoresinol (0.5 mM or 0 mM in minimal salt medium [MSM] for 3 h) using sequential window acquisition of all theoretical mass spectra (SWATH) analysis (28, 29). A total of 2,165 proteins, representing 43.3% of the theoretical SG-MS2 proteome, were detected, and 78 of these showed a >2-fold difference in abundance between the two types of cells with a false discovery rate (FDR)-adjusted P value of <0.05. Forty-six of these were upregulated (by 2.01- to 32.65-fold) and 32 downregulated (by 2.03- to 4.36-fold) in the induced cells. Details of all upregulated and downregulated proteins are provided in Table S2.

The 32 downregulated genes were not clustered on the SG-MS2 chromosome. PGAAP annotation showed five of them encoded putative ABC transporters (BLX42_11110, BLX42_01595, BLX_14825, BLX42_15085, and BLX42_20280) and one encoded a two-component system (BLX42_14530) for chemotaxis signal transduction, which suggests some negative regulatory responses to the pinoresinol.

In contrast, 22 of the 46 upregulated (2.8- to 32.7-fold) genes were located in a 35.1-kb cluster of 34 genes (Fig. 5). Fifteen of these 22 were annotated by PGAAP to encode hypothetical proteins, one of which we have shown above is the flavoprotein component of the pinoresinol hydroxylase. The other seven were annotated to encode an RND transporter (BLX42_10645), azurin (BLX42_10700), maleylacetate reductase (BLX42_10715), glyoxylase (BLX42_10720), oxidoreductase (BLX42_10730), gamma carboxymuconolactone decarboxylase (BLX42_10735), and glutathione-independent formaldehyde dehydrogenase (BLX42_10785). Three of the overexpressed proteins in the cluster (the flavoprotein, glyoxylase, and hypothetical protein BLX42_10760) also were among the seven proteins found to be in both SG-MS1 and SG-MS2 but not the non-pinoresinol-degrading strain KT2440 in the comparative genomic analysis described above (Table 1).

FIG 5.

Pinoresinol-induced genes in the 35.1-kb gene cluster in Pseudomonas sp. strain SG-MS2 and their homologs in a 38.7-kb cluster in another pinoresinol-degrading strain, Burkholderia sp. strain SG-MS1. Five potential operons, which are also present in strain SG-MS1 in the same or inverted orientation, are also shown for that strain. The accession number prefixes BLX42_ and CIC12_ for SG-MS2 and SG-MS1 genes, respectively, are omitted for clarity. Genes with demonstrated or predicted roles in the upper (U) or lower (L) pinoresinol catabolic pathways are indicated per the nomenclature used in Fig. 6. Protein-level fold change expression (FC) and percent identity between homologous genes in the two bacterial strains are also shown.

Eighteen of the 22 pinoresinol-induced, overexpressed proteins in the 35.1-kb cluster were encoded by genes whose intergenic distances and orientations in strain SG-MS2 (and largely conserved positions in SG-MS1; see below) suggest they are arranged within five putative operons (here termed “operons”) (Fig. 5). All the genes in operons 1, 4, and 5 (three, four, and five genes, respectively) were overexpressed on pinoresinol exposure. Two of the four genes in operon 2, encoding the flavoprotein and the cytochrome (the latter was not found to be overexpressed as a protein), were among the seven associated with pinoresinol catabolism in the comparative genomic analysis described above. This operon also included the overexpressed gene encoding the putative electron acceptor for the flavocytochrome azurin (BLX42_10700). In total, the five operons harbored 21 of the 34 genes in the cluster.

In addition to the flavoprotein, cytochrome, and azurin genes, BLASTp best-hit analysis showed that a further eight of the genes in the 35.1-kb cluster could be involved in subsequent steps in the degradation of pinoresinol. Figure 6 (see also Table S2 and Text S1) illustrates how these genes could encode enzymes in the route for catabolism of pinoresinol deduced by Shettigar et al. (13). This route comprises an upper and a lower pathway. The upper pathway transforms pinoresinol to 2-methoxyhydroquinone and protocatechuate via vanillin/vanillic acid, with nonaromatic side products formaldehyde and 4-hydroxy-2-(hydroxymethyl)butanoic acid, and the lower pathway further metabolizes these products, including converting 2-methoxyhydroquinone and protocatechuate to tricarboxylic acid (TCA) cycle intermediates (13) (Text S1).

FIG 6.

Pinoresinol catabolic pathway intermediates and cognate gene-enzyme systems. The enzyme names shown in blue either have significant amino acid identities with functionally characterized enzymes or are functionally validated (PinA and PinB; flavocytochrome) in this study. Enzymes shown in red are predicted to catalyze the reactions indicated. All genes/enzymes (except esterase and PcaG in the lower pathway) shown here are overexpressed in response to pinoresinol exposure. The upper (U) and lower (L) pathway steps are also indicated.

Three of the six enzyme functions in the upper pathway, in addition to the α-hydroxylation of pinoresinol, could be encoded by four genes within the cluster. These are alkyl-aryl cleavage of the hydroxylated pinoresinol, which could be catalyzed by an ipso-hydroxylase (BLX42_10775, annotated as a hypothetical protein, 3-fold upregulated, located in operon 4), three lactone ring cleavages by either a serine hydrolase (BLX42_10795, annotated as a hypothetical protein, 11-fold upregulated, operon 5) or lactonase (BLX42_10805, annotated as a hypothetical protein, 8-fold upregulated, operon 5), and the C-3–C-4 bond cleavage of 4-(4′-hydroxy-3′-methoxyphenyl)-2,3-bis(hydroxymethyl)-4-oxobutanoic acid (metabolite 3) (Fig. 6) by a dienoate hydrolase (BLX42_10710, annotated as a hypothetical protein, 14-fold upregulated). The two upper pathway functions not encoded by the cluster are conversion of vanillin to vanillic acid and then protocatechuate, and the second of these could be encoded by copies of the VanA and VanB genes adjacent to the cluster (Fig. 5 and 6).

Five of the 10 lower pathway steps identified in Fig. 6 also could be encoded by the cluster, and four of these are upregulated. The latter are a putative hydroquinone dioxygenase (BLX42_10720, annotated as glyoxylase, 15-fold upregulated, operon 3), maleylacetate reductase (BLX42_10715, annotated as a hypothetical protein, 10-fold upregulated, operon 3), 4-carboxymuconolactone decarboxylase (BLX42_10735, annotated as gamma carboxymuconolactone decarboxylase), and formaldehyde dehydrogenase (BLX42_10785, annotated as a hypothetical protein but with 95% amino acid identity with a functionally characterized formaldehyde dehydrogenase from another P. putida strain, 8-fold upregulated).

Key components of the SG-MS2 cluster are also clustered in SG-MS1.

Twenty-five of the 34 genes in the 35.1-kb SG-MS2 cluster described above, including 20 of the 21 in the five operons described above, have homologs in SG-MS1 that are also clustered across about 38.7 kb in that strain (Fig. 5). Eight of the remaining nine (all annotated as hypothetical proteins) have no homologs anywhere in the SG-MS1 genome, the exception being BLX_10780, which has a homologue (CIC12_19750) in SG-MS1 (annotated as glyoxylase in both genomes). The 25 genes that are common to the two clusters include 20 that were upregulated on pinoresinol exposure in SG-MS2. Percent amino acid identities between the respective members of the 25 homologue pairs range from 26.7 to 78.5%. The flavoprotein and cytochrome homologue pairs were in the upper end of this range, showing 74.8% and 76.5% amino acid identities, respectively, in the two strains.

Three of the five other candidate genes for the upper pathway in the SG-MS2 cluster (i.e., the candidate ipso-hydroxylase BLX42_10775, serine hydrolase BLX42_10795, and lactonase BLX42_10805) have homologs in the SG-MS1 cluster, with over 55% amino acid identities within the various pairs. All four of the upregulated lower pathway genes are present in this cluster. Maleylacetate reductase (BLX42_10715), 2-methoxyhydroquinone dioxygenase (BLX42_10720), and 4-carboxymuconolactone decarboxylase (BLX42_10735) are all in operon 3 and have homologs (>64% amino acid identity) in the SG-MS1 cluster. The fourth lower pathway gene, formaldehyde dehydrogenase (BLX42_10785), is not in any of the five operons but has an SG-MS1 homolog (CIC12_20260; 78% identity).

The structures of the five operons are largely conserved between the two strains, with just one rearrangement difference (operon 4) and the absence of the azurin gene in strain SG-MS1. Two of the nine genes in the SG-MS2 cluster that have no homologs in the other cluster in fact have no homologs anywhere else in the SG-MS1 genome. These are azurin, as noted, and one hypothetical gene (BLX42_10710; 44.5% identity with 2-hydroxy-6-oxo-6-phenylhexa-2-4-dienoate hydrolase). Ten genes in the SG-MS1 cluster have no homologs in the SG-MS2 cluster (Fig. 5).

Genes associated with vanillin degradation are also induced by pinoresinol exposure in SG-MS2.

Vanillin and vanillic acid were previously shown to be downstream catabolic intermediates of pinoresinol mineralization in strain SG-MS2 (13). It was further shown that both compounds are mineralized when provided as sole sources of carbon and energy, and vanillic acid transiently appears as a catabolic intermediate during the growth of the strain on vanillin (13). The comparative proteomics experiment discussed above concerning upper pinoresinol catabolic pathway enzymes also provided concrete evidence for the upregulation of genes with products homologous to enzymes known to be involved in vanillin catabolism in SG-MS2 cells exposed to pinoresinol (vanillin to vanillic acid and then protocatechuate) in the upper pathway noted above plus subsequent lower pathway genes (Fig. 6).

Specifically, 13 of the 46 upregulated proteins in strain SG-MS2 have high amino acid identities (35.3 to 98.8%) with functionally characterized enzymes involved in the catabolism of vanillin and 2-methoxyhydroquinone in other bacteria (Table S2 and Text S1). Four of these were implicated in the conversion of vanillic acid to protocatechuate in the upper pathway, namely, two copies of vanillate O-demethylase, subunit 1 (VanA; BLX42_10625, adjacent to the cluster, and BLX42_19765) and two copies of vanillate O-demethylase, subunit 2 (VanB; BLX42_10630, adjacent to the cluster, and BLX42_19770). The other nine were implicated in the lower pathway, namely, 3-carboxy-cis,cis-muconate cycloisomerase (PcaB; BLX42_20880), 4-carboxymuconolactone decarboxylase (PcaC; BLX42_10735), 3-oxoadipate enol-lactonase (PcaD; BLX42_20875), β-ketoadipyl coenzyme A (CoA) thiolase (PcaF; BLX42_20890), protocatechuate 3,4-dioxygenase, beta subunit (PcaH; BLX42_15620), 3-oxoadipate CoA-transferase, subunit A (PcaI; BLX42_21970), 3-oxoadipate CoA-transferase, subunit B (PcaJ; BLX42_21975), 2-methoxyhydroquinone dioxygenase (MHQDi; BLX42_10720), and maleylacetate reductase (MAR; BLX42_10720), and were overexpressed in the pinoresinol-exposed cells.

All 13 of these enzymes have homologs with 60 to 99% amino acid identity in SG-MS1. As noted, three of them, the 2-methoxyhydroquinone dioxygenase, maleylacetate reductase, and 4-carboxymuconolactone decarboxylase, are in the cluster and in fact are in the same operon (operon 3 in SG-MS2; Fig. 5) in both clusters. In fact, all five genes of operon 3 have homologues in the same order and orientation in SG-MS1, with 49 to 77% identity to respective pairs (Fig. 5). The other eight are scattered elsewhere in the genome (Table S2). Interestingly, 11 of these proteins (there are only single copies of VanA and VanB) were also found to be upregulated and scattered across the genome in a previous proteome analysis of the response of Pseudomonas putida KT2440 to vanillin (30).

A similar cluster is present in Pseudomonas sp. strain MYb187.

A discontinuous MegaBLAST search using the 35.1-kb SG-MS2 cluster to probe Pseudomonadales (taxid 72274) whole-genome shotgun contig sequences identified a sequence in Pseudomonas sp. strain MYb187 (MYb187_contig000021; GenBank accession no. PCOA01000021.1) with 99.99% (35,121/35,127 nucleotides) identity to the cluster. The order and orientation of the genes in this contig are exactly the same as those in SG-MS2. Two of the six nucleotide differences are in noncoding regions, and the other four are physicochemically conserved missense mutations in genes BLX42_10735 (S→T), BLX42_10750 (S→T), BLX42_10705 (N→D), and BLX42_10810 (S→A). Three of these four genes, BLX42_10735, BLX42_10705, and BLX42_10810, are induced on exposure of strain SG-MS2 to pinoresinol, but only BLX42_10795 (esterase/lactonase) is tentatively assigned to the pinoresinol catabolic pathway. Notably, strain MYb187 was isolated from a nematode from composting plant matter/soil in Roxel, Germany, in 2008 (31, 32), while strain SG-MS2 was isolated directly from a wheat straw/manure compost from Stirling, South Australia, in 2013 (13).

No significant identities were found in a similar discontinuous MegaBLAST search using the SG-MS1 cluster as a probe.

DISCUSSION

The pinoresinol α-hydroxylase we have characterized from Pseudomonas sp. strain SG-MS2 is a new member of the 4PO subgroup of the VAO/PCMH (vanillyl alcohol oxidase/p-cresol methyl hydroxylase) family, which catalyzes hydroxylation of the α-carbon of pinoresinol. It initiates a chain of catabolic reactions and leads to assimilation of the compound as a source of carbon and energy in this strain. Pinoresinol shares a phenol group at the 4 position with the substrates (vanillyl alcohol, p-cresol, eugenol, and chondrochloren) of other enzymes in the 4PO subgroup. The catalytic reaction of these enzymes begins with the abstraction of a hydride from the α-carbon of the enzyme-substrate complex, generating a para-quinone methide intermediate that subsequently, depending on the substrate, rearranges or reacts with water to give the reaction product (17, 33, 34).

The evidence provided here indicates that the pinoresinol α-hydroxylase in SG-MS2 has a covalently linked FAD cofactor, which is consistent with the cofactor properties of the other characterized 4PO subgroup enzymes, except for CCD (17, 35–38). Previously characterized oxidases with the 8α-N³-histidyl-FAD linkage in this subgroup (VAO and EUGO) transfer electrons to oxygen, while the dehydrogenases (flavocytochromes) in the subgroup with the 8α-O-tyrosyl-FAD (PCMH and predicted in EUGH) transfer them to cytochrome c (35, 39–41). However, our evidence indicates pinoresinol α-hydroxylase (a flavocytochrome) has an 8α-N³-histidyl-FAD linkage and transfers electrons to cytochrome c. Notably, it also appears that azurin is used as the terminal electron acceptor for the cytochrome in SG-MS2, whereas no azurin gene is found in either the cluster or anywhere else in the SG-MS1 genome. This raises the possibility that the flavocytochrome in SG-MS1 uses molecular oxygen as the terminal electron acceptor and acts as an oxygenase rather than a dehydrogenase in this strain. This is discussed further below.

We found that many genes with annotations consistent with a function in subsequent steps in pinoresinol catabolism were overexpressed upon exposure of SG-MS2, and several of them were tightly linked in a 35.1-kb cluster in this strain. This suggests that the pathway and cluster have evolved to coordinate the catabolism of pinoresinol and avoid metabolic bottlenecks. However, we also note that the cluster contains several other genes that are not induced by pinoresinol and have no annotations suggesting a role in its catabolism. Perhaps the cluster encodes the machinery for catabolizing multiple structurally similar substrates, with some enzymes shared between pathways and others unique to a particular pathway. Other compounds that could be substrates include pinoresinol derivatives (pinoresinol di-β-d-glucoside, sesamin, eudesmine, paulownin, matairesinol, secoisolariciresinol, syringaresinol, arctigenin, 7-hydroxymatairesinol, isolariciresinol, and lariciresinol) and substrates of other 4PO subgroup enzymes [p-cresol, eugenol, 4-hydroxybenzyl alcohols, 4-hydroxybenzylamines, and 4-(methoxymethyl)phenols].

Most of the upregulated genes implicated in the upper pinoresinol catabolic pathway that lie within the 35.1-kb cluster in SG-MS2 have homologs in an equivalent 38.7-kb cluster in SG-MS1. The arrangement of many of the genes in putative operons is also largely conserved between the two strains, notwithstanding some differences in the relative positions of some putative operons. Homologs of almost all the candidate downstream pathway genes identified in SG-MS2 also were recovered in the SG-MS1 genome, but generally they were not located within the cluster in either strain. We conclude that overall the genetic and biochemical basis of both the upper and lower pinoresinol catabolic pathways are largely conserved in the two strains.

However, we also note that there is no evidence of horizontal gene transfer between the strains for any of the candidate genes for either of the pathways; amino acid identities between homologs are never greater than 78%. Moreover, several genes in each of the two clusters are not induced by pinoresinol and not found in the other cluster. There is also the key difference noted above concerning azurin and the possibility that the flavoprotein in the latter strain must use a different electron acceptor and may be an oxygenase. These differences suggest that the two clusters have been evolving essentially independently for some time and are adapted to metabolize overlapping but slightly different suites of secondary compounds and xenobiotics.

Given the differences between the SG-MS1 and SG-MS2 clusters, which were collected from the same Australian composting plant/soil sample (13), it is notable that a cluster so much more similar to SG-MS2 than SG-MS1 was recovered from Caenorhabditis elegans in composting soil from Europe. This is an even tighter conservation than that seen in the lin cluster responsible for catabolism of hexachlorocyclohexane isomers in sphingomonads from different continents (42, 43). This suggests that there is quite strong selection to conserve the function of the SG-MS2 cluster.

In conclusion, we suggest that the SG-MS1 and SG-MS2 clusters now are fruitful models for studying the evolution of bacterial pathways for the catabolism of complex secondary compounds. Given that pinoresinol is an important plant defense compound, a major food lignan for humans and some other animals, and the model compound used to study degradation of the β-β′ linkages in lignin, we further suggest that the functions encoded by the clusters deserve further attention for potential applications in crop protection as well as human and animal nutrition and the production of useful chemistries from lignocellulosic waste.

MATERIALS AND METHODS

Pinoresinol, bacterial strains, and growth conditions.

(+)-Pinoresinol (>98% purity) was purchased from Arbonova Sales (Turku, Finland). The pinoresinol-degrading Burkholderia sp. strain SG-MS1 and Pseudomonas sp. strain SG-MS2 were previously isolated in our laboratory (13). Growth media and culture conditions for these strains were as described previously (13). P. putida KT2440 was sourced from DSMZ and cultivated under identical conditions. E. coli strains were grown at 28°C in the presence of 100 μg·ml⁻¹ either Luria-Bertani (LB) medium or LB agar. Liquid cultures were shaken at 200 rpm.

Genome sequencing and annotation.

Genomic DNA from overnight-grown cultures of SG-MS1 and SG-MS2 in LB medium was prepared using a QIAamp DNA minikit (Qiagen, Venlo, Netherlands) by following the manufacturer’s protocol. Sequencing libraries were prepared from 500-bp DNA fragments and sequenced using Illumina NextSeq technology at Macrogen (Seoul, South Korea). Genome assembly was performed with the ABySS de novo assembler (version 2.0.2), setting a k-mer size of 63 (44). Sequencing of the genomes of strains SG-MS1 and SG-MS2 yielded 433,355 and 258,981 sequence reads, respectively, that passed the quality filters for use in genome assemblies. The 7.3-Mb assembly obtained for strain SG-MS1 comprised 276 contigs greater than 500 bp in length, giving an N₅₀ of nearly 47 kb. The 5.6-Mb assembly for SG-MS2 comprised 278 contigs greater than 500 bp in length, with an N₅₀ of 33 kb. Annotation of these genomes with the NCBI Prokaryotic Genomes Automatic Annotation Pipeline (PGAAP) (45) identified 6,247 and 4,988 protein-coding genes in SG-MS1 and SG-MS2, respectively. Automated annotation of these genomes was performed with the NCBI Prokaryotic Genomes Automatic Annotation Pipeline (PGAAP) (45).

Comparative genomics.

PGAAP-annotated proteins from SG-MS2 were identified that had over >70% identities with annotated proteins in SG-MS1 but <30% identities with any protein in strain KT2440. These hits were matched to annotated proteins of the genomes of strains SG-MS1, SG-MS2, and KT2440.

Protein purification.

A 30-liter culture of strain SG-MS2 was grown in LB in 1-liter batches in separate 2-liter flasks overnight (OD₆₀₀ of 0.6 to 0.8). Cell were harvested by centrifugation at 4,000 × g for 10 min (at room temperature) and washed by resuspension into carbon-free minimal salt medium (MSM). Pinoresinol-degrading genes were induced by resuspending the washed cells in three separate batches of 1 liter of MSM supplemented with 55.8 μM pinoresinol as a sole carbon source with shaking at 200 rpm and 28°C for 3.5 h. Cell lysis was achieved by passing the cells three times through a homogenizer operating at 21,000 lb/in² at 4°C. Cell debris was removed by spinning down the crude cell lysate at 25,000 × g at 4°C. Protein purification was initiated by salting out with 2 M ammonium sulfate at 4°C. The precipitated proteins were removed by centrifugation at 10,000 × g, and the collected supernatant was dialyzed twice against 50 mM potassium phosphate buffer (pH 7.0) through a 50-kDa Amicon Ultra-15 centrifugal filter unit (Merck, Germany). The dialyzed sample then was loaded onto a Fractogel DEAE column (EMD Millipore, MA, USA), and the proteins were eluted with a 0 to 1 M NaCl gradient in 50 mM potassium phosphate buffer (pH 7.0). Active fractions were pooled and passed through 50-kDa spin filters to concentrate them and change the buffer to 50 mM potassium phosphate (pH 6.0). The concentrated fractions then were loaded onto a HiTrap SP HP cation exchange column (GE Healthcare Life Sciences, Sweden) and the proteins eluted with a pH gradient from pH 6.0 to pH 7.0 in 50 mM potassium phosphate buffer. Active fractions again were pooled and concentrated in 50 mM potassium phosphate buffer (pH 8.0) with 50-kDa spin filters (Amicon).

The enzyme was assayed during purification using a previously described protocol (13).

Protein concentration, gel electrophoresis, and fluorescence detection of enzyme-bound FAD.

Protein concentration was measured using Direct Detect (Merck Millipore), and 4 to 12% SDS-PAGE (Bolt Bis-Tris plus gel; Invitrogen, USA) was run using the manufacturer’s protocols. The fluorescence of enzyme-bound FAD was determined as described previously (26). In brief, the unstained SDS-PAGE gel was scanned in a PhosphorImager (Typhoon; Amersham Biosciences, UK) with excitation at 488 nm and emission detection at 520 nm. After fluorescence detection, the gel was stained with Coomassie blue using AcquaStain protein gel stain (Bulldog Bio, USA).

Peptide sequencing and protein identification.

Twelve protein bands resulting from SDS-PAGE of the semipurified protein were excised from the gel and cut into 1-mm cubes, using separate scalpel blades for each band. In-gel digestion with trypsin and elution were performed as described previously (46). Details of the peptide sequencing protocol are provided as Text S2 in the supplemental material. The peptide sequences obtained were matched to a custom database generated from all the proteins predicted from the PGAAP-annotated SG-MS2 genome.

Proteins were identified using the Sequest algorithm in Proteome Discoverer v2.1 (ThermoFisher, USA). Carbamidomethyl was selected as the alkylating agent, and trypsin was selected as the digestion enzyme. Dynamic modifications for oxidation on histidine, methionine, tryptophan, and deamidated asparagine and glutamine were selected, with a maximum of three modifications on each of these amino acids. The tandem mass spectrometry data were searched against a database containing all proteins from the PGAAP-annotated genome of SG-MS2. The database search results were manually curated to yield the protein identifications using a 1% global false discovery rate (FDR), as determined by the built-in FDR tool within Proteome Discoverer.

Comparative proteomics.

For sample preparation, a single colony of Pseudomonas sp. strain SG-MS2 was used to inoculate 32.5 ml of LB, and the culture was shaken at 180 rpm at 28°C until its OD₆₀₀ reached 0.6. Cells were harvested by centrifugation at 4,000 × g for 10 min, washed with 50 ml of carbon-free MSM, and recollected by spinning down at 4,000 × g for 10 min. Centrifugation and washing were done at 4°C. The collected cells were finally resuspended in MSM, and the OD₆₀₀ of the resuspended cells was adjusted to 1.05. Six aliquots of 600 μl each were taken from the cell suspension and put into different 50-ml Falcon tubes as the basis for three replicates each of pinoresinol-induced and uninduced cells. The cells in three of the aliquots were induced with pinoresinol for 3 h by incubation in MSM containing 125 μM pinoresinol. Uninduced cells were prepared in a similar way from the other three aliquots, except that the 4.4 ml of pinoresinol was replaced by 4.4 ml MSM. All samples were kept at 28°C in an incubator and constantly shaken at 180 rpm during the incubation, as described above.

Following incubation, cells were collected by spinning at 4,000 × g for 10 min at 4°C and resuspended in 50 mM triethylammonium bicarbonate buffer (pH 9.5) supplemented with 1%, wt/vol, sodium deoxycholate. Cell lysis was achieved by incubating the samples in boiling water for 5 min. Benzonuclease (50 U) was added to the samples at room temperature (22°C) and incubated for 30 min. The protein concentration of these samples was measured the following day by Direct Detect (Merck Millipore, USA), and 90 μg of protein was taken from these samples and the volume adjusted to 500 μl with the buffer described above. Dithiothreitol (final concentration, 10 mM) and iodoacetamide (final concentration, 20 mM) were added sequentially and incubated for 10 min and 5 min, respectively. Finally, the tryptic digest of these samples was prepared by adding 1.8 μg of mass spectrometry-grade trypsin (Promega, WI, USA) and incubating at 37°C for 16 h. The trypsin was inactivated by adding 1%, vol/vol, formic acid, which also resulted in the precipitation of deoxycholate from the solution. Samples were spun down at 18,000 × g for 5 min at 37°C, and the supernatant was collected and dried using a Genevac (performed sequentially at 30°C for 3 h, 35°C for 1 h, and 40°C for 1 h). The dried sample was finally resuspended in 140 μl of loading buffer (2% acetonitrile and 0.1% formic acid). Digested samples (3.3 μl) from each sample group (pinoresinol induced and uninduced; each in triplicate) were pooled to run for one-dimensional (1D) IDA (information-dependent acquisition) analysis as described below. A pool was prepared from 39 μl of each sample (total of ∼150 μg) to perform high-pH reverse-phase fractionation on a high-performance liquid chromatography (HPLC) column. A total of 12 fractions were pooled from collected fractions (0 to 82 min), dried, and resuspended in 20 μl of loading buffer, and 10 μl of each fraction was transferred to vials for 2D IDA analysis. For SWATH analysis, 10 μl from each digested sample was transferred to HPLC vials, and each sample was run twice.

Protein samples described above were provided to the Australian Proteome Analysis Facility (APAF) for quantitative comparative proteomics analysis. Details of the protocols for 1D IDA and data-independent acquisition (SWATH), along with data processing and SWATH quantitation, are provided in Text S3. Results provided to us by the APAF are available as Text S2 in the supplemental material.

Gene cloning.

Three expression constructs were generated, the first containing the flavoprotein, cytochrome, and azurin genes plus 550 bp upstream and 51 bp downstream (pCR2.1-FCA), the second containing N-terminally 6-His-tagged flavoprotein and cytochrome (pETMCSIII-FC1), and the third containing just the N-terminally 6-His-tagged flavoprotein (pETMCSIII-Flavo).

For generating the three-gene construct (pCR2.1-FCA), a 3,064-bp DNA fragment containing the flavoprotein, cytochrome, and azurin genes from the genomic SG-MS2 DNA was PCR amplified using forward (5′-ATCCAGCATTCAGCTGAAAGCCATCCTTTCTGACCGC-3′) and reverse (5′-ACTACGATTGGGCCGTCGATGACCGCAAGCTGGAAG-3′) primers. This PCR fragment also contained 550 bp upstream of the flavoprotein and 51 bp downstream of the azurin gene. PCR was conducted using Phusion polymerase (NEB, MA, USA) and the primers described above under the conditions recommended by the manufacturer. The amplicon was resolved and purified from a 0.6% agarose gel. 3′ A-overhangs were added to the amplified DNA using one unit of Taq polymerase and the manufacturer’s protocol (ThermoFisher, MA, USA). The resultant overhang DNA fragment was cloned into pCR2.1-TOPO vector (ThermoFisher) using the manufacturer’s protocol. The plasmid was termed pCR2.1-FCA. The cloned gene then was transformed into BL21(DE3) electrocompetent cells as described earlier.

The translationally coupled native flavoprotein and cytochrome genes in their native orientation were synthesized as a DNA string (FC1) by GeneArt (Invitrogen, USA). FC1 was designed to carry NdeI (5′) and EcoRI (3′) restriction enzyme sites, along with buffering nucleotides to facilitate enzyme digestion. It was cloned into the NdeI-EcoRI sites of the pETMCSIII T7 promoter vector for overexpression of 6× His-tagged (N terminus) flavocytochrome to generate the pETMCSII-FC1 plasmid, which was transformed into BL21(DE3) electrocompetent cells (NEB).

The flavoprotein (N-terminal 6× His-tagged) gene was PCR amplified from 20 ng FC1 DNA string with forward (5′-CATATGCTACCCCCCGGTATCG-3′) and reverse (5′-GAATTCGTCAGCTTCATACGGAGTCGCCTC-3′) primers, using PfuUltra DNA polymerase (Agilent Technologies, MA, USA) under the manufacturer’s recommended conditions. The 1,690-bp DNA product was A-tailed using Platinum Taq DNA polymerase (Life Technologies, Australia) and ligated into pGEM-T-Easy vector (Promega, Australia) using a LigaFast rapid DNA ligation system (Promega). Standard transformation into chemically competent DH5α (Life Technologies) cells produced a pGEM-Flavo clone containing the RE-flanked flavoprotein gene. The flavoprotein gene then was excised using NdeI and EcoRI (NEB) and subsequently subcloned into pETMCSIII to generate pETMCSIII-flavo, which was transformed and expressed in BL21(DE3) cells (NEB).

Heterologous gene expression and purification.

pCR2.1-FCA, pETMCSIII-FC1, and pETMCSIII-flavo constructs were expressed in BL21(DE3) cells. A colony from a fresh streak of frozen glycerol stocks was used to inoculate a 10-ml starter culture in LB supplemented with 100 μg·ml⁻¹ ampicillin (Sigma-Aldrich, USA). The culture was grown for 18 h at 37°C and 200 rpm, and then 5 ml was used to seed 500 ml LB supplemented with 100 μg·ml⁻¹ ampicillin in a 2-liter fluted culture flask. The culture was grown at 28°C and 200 rpm overnight, resulting in a final OD₆₀₀ of around 3. Cells were harvested by centrifugation at 5,000 × g for 15 min at 4°C and washed once in 20 mM Tris (pH 7.5). Cell pellets were stored at –80°C if required. Pellets were thawed and gently resuspended in 100 ml of His tag buffer (20 mM sodium phosphate, 0.5 M NaCl, pH 7.4). His-tagged protein was purified as described previously (47). Samples of the resulting fractions were run on SDS-PAGE gels and analyzed for activity with pinoresinol by LC quadrupole-TOF/MS (LC Q-TOF/MS) as described previously (13).

Enzyme assay.

The protocol for assaying pinoresinol-hydroxylating activity from the wild-type strain was based on LC-MS/TOF as described previously (13). E. coli clones of the three constructs described above were individually grown for 16 h at 28°C in LB containing 100 μg ml⁻¹ ampicillin to generate a seed culture. This seed culture was inoculated (1%, vol/vol) into 30 ml LB and grown until an OD₆₀₀ of 0.8 at 28°C in the presence of 100 μg ml⁻¹ ampicillin. The cells were centrifuged at 3,500 × g for 10 min at 4°C, and the cell pellet was washed by resuspension in 100 mM NaCl containing 50 mM potassium phosphate buffer (pH 7.2) and recentrifuged as described above. The resulting cell pellet was lysed in phosphate buffer that also contained 1-fold BugBuster (Merck Millipore, USA), 558 μM pinoresinol (substrate), and 3 mg ml⁻¹ fully oxidized PMS (electron acceptor). After 30 min of incubation, the reaction was stopped by adding formic acid (0.1% final volume) and then an equal volume of acetonitrile. This was filtered through a 0.22-μm syringe filter (Merck Millipore) and qualitatively analyzed for transformation of pinoresinol to its hydroxylated product by LC Q-TOF/MS as described previously (13). E. coli BL21(DE3) cells were taken as a negative control in these experiments.

An enzyme assay to determine the Michaelis-Menten constant (K_m) of heterologously purified His-tagged pinoresinol α-hydroxylase (flavocytochrome) for pinoresinol was performed spectrophotometrically, as described previously for PCMH (24). Specifically, the assay was performed in clear flat-bottom 96-well plates in a final volume of 150 μl (equivalent to 0.5-cm path length) over 2 min at 25°C with 3-s recordings. Final concentrations of fully oxidized DCIP, PMS, and the purified enzyme were 100 μM, 150 μM, and 0.5 μl (6.25 μg), respectively. Enzyme activity was determined as a function of the decrease in OD₆₀₀ at different pinoresinol concentrations (0 to 40 μM), with PMS as an electron acceptor and 2,6-dichloroindophenol (DCIP) as the terminal acceptor. Reduction of DCIP was monitored spectrophotometrically at 600 nm.

Phylogenetic analysis.

To perform phylogenetic analysis, structures of functionally characterized eugenol oxidases (PDB entry 5FXP), vanillyl alcohol oxidase (1VAO), and p-cresol methylhydroxylase (1WVF) first were aligned using the PROMALS3D multiple-sequence and structure alignment server using default settings (48). Protein sequences of functionally characterized eugenol oxidases (GenBank accession no. AAM21269 and CAB64355), chondrochloren decarboxylase (accession no. CAQ43085), (+)-pinoresinol α-hydroxylase (BLX42_10690), and CIC12_20345 from strain SG-MS1 were added to the structure alignment using the MAFFT webserver (version 7.310) under default settings (49, 50). This alignment (Fig. S4) was provided to the IQ-TREE webserver (51) with the substitution model set to auto (52) and specifying FreeRate heterogeneity to generate a maximum likelihood tree with 1,000 ultrafast bootstraps (53).

Data availability.

The whole-genome shotgun projects for strains SG-MS1 and SG-MS2 have been deposited at DDBJ/ENA/GenBank under the accession numbers NPIH00000000 and MOMO00000000, respectively. The versions described in this paper for strains SG-MS1 and SG-MS2 are versions NPIH01000000 and MOMO00000000, respectively.