Sugar O-methylation shields algal polysaccharides against microbial hydrolytic enzymes. Here, we describe cytochrome P450 monooxygenases from marine bacteria that, together with appropriate redox-partner proteins, catalyze the oxidative demethylation of 6-O-methyl-d-galactose, which is an abundant monosaccharide of the algal polysaccharides agarose and porphyran. This previously unknown biological function extends the group of carbohydrate-active enzymes to include the class of cytochrome P450 monooxygenases.
Oceanic and terrestrial ecosystems contribute equally to the world’s net primary production, each accounting for about 52 petagrams of fixed carbon per year1. Polysaccharides constitute more than 50% of algal dry matter2 and have a high potential in biotechnological processes, where they are used as carbon source for microbial fermentation to produce second-generation biofuels and other products2,3,4. The depolymerization of polysaccharides is catalyzed by carbohydrate-active enzymes (CAZymes) and enables the microbes to utilize the monomeric sugars in their cellular metabolism. In algal polysaccharides5, as well as in hemicellulose6, O-methylation is a common modification that inhibits hydrolysis of polysaccharides by preventing productive binding by glycoside hydrolases7. Compared to esters and sulfates, methyl ethers are difficult to remove chemically, which is why they are rarely used as protecting groups in organic syntheses. The idea that O-methylations serve as protective adaptations, which restrain microbial catabolism in biotechnological applications and in the wild, is supported by the recent report of accumulation of O-methylated sugars in sea water5. How microbes digest and catabolize methoxy sugars remains unknown. We focused our analysis on the pathways involved in the degradation of agarose and porphyran, agar polysaccharides produced by red algae. Porphyran is an O-methylated polysaccharide that consists mainly of alternating 3-linked β-d-galactose and 4-linked α-l-galactose-6-sulfate or 3,6-anhydro-α-l-galactose (the latter is more abundant in agarose)8. 6-O-Methyl-d-galactose (G6Me) can represent up to 28% of porphyran8 and is also a constituent of agarose and carrageenan9. Given the stability of methyl ethers, it is likely that G6Me needs to be demethylated before it can enter the usual glycolytic pathway in the cellular metabolism.
Among the putative candidates for O-demethylations are the cytochrome P450 monooxygenases, which are known to catalyze a wide range of oxidation reactions such as hydroxylation and epoxidation, making them valuable enzymes for biocatalysis10. One example of their many applications is the selective deprotection of chemically permethylated sugars by mutants of CYP102A111, which shows that sugar demethylation is in principle feasible with P450 monooxygenases. In nature, these enzymes’ broad range of biological functions includes the acquisition of specific alternative carbon sources for microorganisms12. Oxidative enzymes are known to be involved in the degradation of complex carbohydrates, with laccases, manganese peroxidases and the recently described lytic polysaccharide monooxygenases (LPMOs) as the common terrestrial enzymes13,14. These enzymes, among other redox enzymes related to polysaccharide degradation, are grouped in the class of auxiliary activities (AA) in the CAZy database15. These enzyme families have recently received substantial interest as they significantly improve the bioconversion of recalcitrant polysaccharides, such as that of cellulose into glucose16, leading to the discovery of new enzyme functions17. However, so far no evidence has been provided that P450 monooxygenases also are involved in the degradation of carbohydrates.
In many marine bacteria that decompose complex polysaccharides, genes encoding the respective degradation enzymes are clustered in the genome in polysaccharide-utilization loci (PULs). Enzymes encoded by a given PUL cooperate in the deconstruction of a specific complex polysaccharide. Querying the genomes from marine bacteria with PULs involved in agar (porphyran and agarose) degradation, we found that genes for putative cytochrome P450 monooxygenases are conserved in close proximity to putative CAZyme genes. Furthermore, these genes are co-localized with genes encoding for putative ferredoxins and ferredoxin reductases, which might deliver the electrons from NAD(P)H to the monooxygenase. We reasoned that the presence of P450 monooxygenases close to or within PULs dedicated to the degradation of porphyran and agarose, for example, in Formosa agariphila KMM 3901T18 (Fig. 1a) suggests that they may be involved in the turnover of the O-methylated sugars that are part of these polysaccharides. This argument is further supported by the phylogenetic tree of the P450 proteins, which reveals that they are conserved at the protein sequence level and that they are present in Bacteroidetes and Gammaproteobacteria that are agarolytic (Supplementary Fig. 1). Notably, we did not detect close homologs of the P450s in microbial genomes lacking agar degrading enzymes. Moreover, a systematic analysis of bacterial genomes reveals that the protein triplet of ferredoxin, ferredoxin reductase and P450 co-occurs in PULs with agarases from Gammaproteobacteria, Bacteroidetes and Planctomycetes (Fig. 1b), all together suggesting that these enzymes contribute to agar degradation.
To test this hypothesis, we selected two putative P450 monooxygenases located in different branches of the phylogenetic tree. The enzyme from F. agariphila (P450FoAg, WP_038530297.1) was assigned as CYP236A20 and the one from Zobellia galactanivorans (P450ZoGa, WP_013995999.1) as CYP236A2, showing that they belong to the same, so-far uncharacterized P450 subfamily (D. Nelson, personal communication). Both enzymes have 72% amino acid sequence identity and were predicted to lack transmembrane helices. They were expressed in Escherichia coli and purified (Supplementary Fig. 2). The carbon monoxide difference spectrum confirmed the presence of the heme group and the identity of these enzymes as P450 monooxygenases (Supplementary Fig. 3a,b). Furthermore, the native ferredoxin (FoX, WP_038530300.1) and the ferredoxin reductase (FoR, WP_038530304.1) of F. agariphila located next to the P450FoAg gene were cloned, expressed and purified (Supplementary Fig. 2) to restore the electron transport chain in vitro. Both purified proteins showed the typical color and expected spectrum resulting from the flavin cofactor in FoR and the 2Fe-2S cluster in FoX (Supplementary Fig. 3c,d). The preferred redox cofactor of FoR was identified as NADH by monitoring the reduction of K3FeCN6 with NADH or NADPH (Supplementary Fig. 4). In vitro biocatalysis reactions indeed confirmed that the agar-derived methoxy sugar G6Me is a substrate for both marine P450 monooxygenases and that NADH is significantly oxidized only when G6Me is present (Supplementary Fig. 5). The first reaction product was confirmed by GC/MS analysis to be d-galactose (Fig. 2a), whereas the second was shown via the Purpald assay to be formaldehyde (Fig. 2b). These results prove that the function of these enzymes is the oxidative demethylation of G6Me (Fig. 2c). The enzyme activity was dependent on the complete electron transport chain from NADH via FoR and FoX to the P450 monooxygenase (Fig. 2b). The results also show that the redox system of P450FoAg is fully compatible with P450ZoGa (Fig. 2b).
After G6Me was confirmed to be a substrate for P450FoAg and P450ZoGa, the binding affinities of G6Me to the active center of the monooxygenases were determined by the type I spectral shift. Both enzymes showed the expected spectral changes of the heme center upon substrate binding (Fig. 2d, Supplementary Fig. 6), revealing a dissociation constant, Kd, of 0.68 ± 0.04 mM for P450FoAg and 1.80 ± 0.05 mM for P450ZoGa. To determine the kinetic parameters of P450FoAg and P450ZoGa, the concentration-dependent activity rates were measured. To ascertain that the reaction rate was mostly limited by the concentration of the monooxygenase, the redox proteins FoR and FoX had to be used in 25-fold and 250-fold molar excess, respectively (Supplementary Fig. 7). This requirement for relatively high excess of the redox proteins in this analysis is probably due to the hindered protein–protein interaction caused by the presence of His-tags on all proteins. The initial activities at varying G6Me concentration are displayed in Fig. 2e. Fitting of the data to Michaelis–Menten kinetics revealed a KM of 0.44 ± 0.01 mM and kcat of 23.1 ± 0.3 s−1 for P450FoAg and a KM of 0.97 ± 0.10 mM and kcat of 19.8 ± 1.2 s−1 for P450ZoGa.
Next, we asked whether the redox proteins of P450FoAg are interchangeable with those of other P450 monooxygenase systems. To address this, the ferredoxin reductase PdR, the ferredoxin PdX and the P450 monooxygenase P450cam (CYP101A1) from the camphor hydroxylation system of Pseudomonas putida12 and the cytochrome P450-diflavin reductase (CPR) from Candida apicola19 were expressed and purified. P450FoAg and P450cam were used with their native substrates in reactions with combinations of the ferredoxin reductases with the ferredoxins and additionally with CPR. Product measurements showed that both monooxygenases reach full conversion only with their native redox partners. Exchanging either the ferredoxin or the ferredoxin reductase abolishes the conversion completely or reduces it to < 7% in the case of exchanged reductases (Fig. 2f). This experiment shows that the optimal electron transfer between ferredoxin reductases and ferredoxins as well as between ferredoxins and P450 monooxygenase is a specific reaction that requires a coevolved set of proteins to function. Problems in electron transfer are likely caused by inadequate protein–protein interactions, indicating that the use of appropriate redox proteins will be crucial for functional applications of these enzymes. Furthermore, we tested whether both monooxygenases were active with H2O2 instead of the redox proteins, but could detect neither d-galactose nor formaldehyde in these reactions. To test whether these marine P450 monooxygenases are exclusively specific for the demethylation of G6Me, a substrate screening was carried out using mainly physiological P450 monooxygenase substrates, including a variety of fatty acids, fatty acid esters, lactones, long-chain alcohols, alkenes, steroids, terpenes, lignin monomers and other methylated carbohydrates (see Supplementary Table 2). Beside HPLC and GC analysis to detect substrate conversion, all reactions were subjected to analysis for formaldehyde formation. No conversion could be detected for any of the additional tested compounds. This demonstrates the specificity of the marine P450 monooxygenases CYP236A20 and CYP236A2 for the demethylation of G6Me.
Our findings thus suggest a new pathway for the complete degradation of O-methylated polysaccharides by marine bacteria. This discovery could also have implications for biotechnological processes in which microbial hosts are equipped with enzymes enabling the processing of algal polysaccharides as alternative carbon and energy sources3, 4. The transfer of the P450 monooxygenase system described here to a host will lead to higher product yields when porphyran or agarose are used as a carbon source for bioethanol production20. In summary, we identified two unique P450 monooxygenases from marine bacteria, which were shown to catalyze the demethylation of the methoxy sugar G6Me derived from the marine polysaccharides porphyran and agarose, a demethylation that is expected to be the key step for the further metabolism of G6Me. Our phylogenetic analysis of these two P450 protein sequences indicates that similar degradation pathways are highly conserved among marine bacteria that degrade agars and other algal polysaccharides. This is, to the best of our knowledge, the first example of a sugar demethylation with a cytochrome P450 monooxygenase in a carbohydrate degradation context. Our discovery contributes substantially to the understanding of the degradation of marine polysaccharides. Related proteins might also be involved in the conversion of terrestrial carbohydrates. Beside the recently discovered LPMOs, these novel P450 monooxygenases represent a second group of monooxygenases involved in carbohydrate degradation.
6-O-Methyl-d-galactose (≥98%), 2-O-methyl-d-glucose (≥98%, TLC), 4-O-methyl-d-glucuronic acid (one spot on TLC), d-pinitol (99.1%, HPLC), and 5-O-methyl-myo-inositol (elementary analysis and NMR confirmed) were purchased from Carbosynth, UK. Methyl 3,6-anhydro-α-d-galactopyranoside (>95%, 13C NMR) was purchased from Dextra, UK and methyl α-d-xylopyranoside (98%) was purchased from abcr. All other chemicals were purchased at the highest purity from Sigma-Aldrich, Carl Roth, Alfa Aesar or Acros.
The expression vectors for the camphor hydroxylation system from P. putida (pET28a-camA, pET28a-camB, pET28a-camC-C334A) and the reductase from C. apicola (pET28-CPRΔ22;His6) were provided by V. Urlacher (Heinrich-Heine-Universität Düsseldorf)19, 21, 22. The plasmids pET22 FoAg and pET22 ZoGa containing the genes for P450FoAg and P450ZoGa with C-terminal His-tags were synthesized by GenScript with codon optimization for E. coli. FastCloning23 was used to generate the expression vectors pET28a-FoR and pET28a-FoX for the ferredoxin reductase FoR with N-terminal His-tag and the ferredoxin FoX with C-terminal His-tag. The vector was amplified with the primer pair pET28-FC-fwd (5′-GCGGCCGCACTCGAGCA-3′) and pET28-FC-rev (5′-GCGCGGCAGCCAT ATG-3′). The gene for FoR was amplified from genomic DNA with FoR-fwd (5′-CACAGCAGCGGCCTGGTGCCGCGCGGCAGCCATATGTTACAGGATTCTAAAAACAAAATC-3′) and FoR-rev (5′-CAGTGGTGGT GGTGGTGGTGCTCGAGTGCGGCCGCTTAATCTTTTAGAAAGCTCGTCG-3′). FoX was first cloned with N-terminal His-tag using FoX-fwd (5′-TCATCACAGCAGCGGCCTGGTGCCGCGCGGCAGCCATATGGCTAAAATAATTTTTGTAACAAAG-3′) and FoX-rev (5′-TCATCACAGCAGCGGCCTGGTGCCGCGCGGCAGCCATAT GGCTAAAATAATTTTTGTAACAAAG-3′), and then amplified from the plasmid to delete the N-terminal His-tag using NHis-FoX-fwd (5′-CTTTAAGAAGGAGATATACCATGGCTAAAATAATTTTTG-3′) and NHis-FoX-rev (5′-CAAAAATTATTTTAGCCATGGTATATCTCCTTCTTAAAG-3′). To include the C-terminal His-tag from the vector, site-directed mutagenesis was performed using QC-FoX-fwd (5′-CCTTTTAAAA GTTGCTCAATCAGCGGCCGCACTCGAG-3′) and QC-FoX-rev (5′-CTCGAGTGCGGCCGCTGATTGAGCAACTTTTAAAAGG-3′).
TB medium with 50 µg/mL kanamycin was inoculated from an overnight culture and incubated at 37 °C until the culture reached an OD600 of 0.8 unless stated otherwise. The cultures were then cooled to 20 °C and induced with 1 mM IPTG for 24 h. For the cultivation of the P450FoAg- and P450ZoGa-overexpressing strains, 100 µg/mL ampicillin was used. The culture for CPR was induced with 0.25 mM IPTG, while the culture for FoX was induced at an OD600 of 1.4 with 0.3 mM IPTG. Cultures expressing P450 monooxygenases were supplemented with 0.5 mM δ-aminolevulinic acid and 0.3 mM FeSO4 upon induction, while cultures for ferredoxins were supplemented with 0.3 mM FeSO4. Cells were harvested by centrifugation (15 min, 4,500 g, 4 °C).
For protein purification, cell pellets were resuspended in buffer A (50 mM Na2HPO4/NaH2PO4 pH 7.4, 500 mM NaCl, 20 mM imidazole). Cells were lysed by sonication (4 × 2 min 0.2 cycle, 50% power) and cell debris were removed by centrifugation (20 min, 8,500 g, 4 °C). 2 mL of the column material Roti®garose-His/Ni beads were equilibrated with buffer A and incubated with the cell lysate for 1 h on ice in batch format and then transferred to the column. The column was washed with buffer A and bound proteins were eluted with buffer B (50 mM Na2HPO4/NaH2PO4 pH 7.4, 500 mM NaCl, 300 mM imidazole). The protein solution was desalted with Sephadex G-25 PD 10 columns (GE Healthcare) and buffer C (50 mM Na2HPO4/NaH2PO4, pH 7.4) according to the manufacture’s protocol. All buffers for the purification of FoX additionally contained 10 mM β-mercaptoethanol. For the purification of P450cam, 10% (v/v) glycerol and 1 mM phenylmethylsulfonyl fluoride (PMSF) were added to buffer A or buffer B. P450cam was stored in buffer C containing 100 µM ( + )-camphor. For FoR, 100 mM NaCl were added to the storage buffer C. The protein concentrations of PdR, PdX, FoR and FoX were determined spectrophotometrically using the extinction coefficients (ε378nm = 9.7 mM−1 cm−1, ε454nm = 10.0 mM−1 cm−1 and ε480nm = 8.5 mM−1 cm−1 for the ferredoxin reductases; ε415nm = 11.1 mM−1 cm−1 and ε455nm = 10.4 mM−1 cm−1 for the ferredoxins24; ε380nm = 16.1 mM−1 cm−1 and ε453nm = 17.9 mM−1 cm−1 for CPR19). Protein purity was checked by SDS-PAGE.
The concentrations of P450 monooxygenases were determined via UV/Vis spectroscopy as described previously25 with the extinction coefficient ε450 = 91 mM−1 cm−1. Purified P450 monooxygenase was diluted with buffer C and 2 µM Safranin T was added. Sodium dithionite was added to 2 mL of the solution until the pink color was absent. 1 mL of this solution was transferred into a cuvette and the baseline spectrum was recorded. The remaining protein solution was saturated with 60 bubbles of carbon monoxide and used to measure the difference spectrum.
Determination of K d
Spin-state shift upon substrate binding with P450FoAg was analyzed spectrophotometrically at room temperature under aerobic conditions26. The dissociation constant, Kd, for 6-O-methyl-d-galactose (G6Me) was determined from type I spectra by titrating a P450FoAg or P450ZoGa solution in buffer C with the same enzyme solution containing G6Me to avoid baseline shifts. The absorbance difference between the maximum and minimum of the difference spectra (389 nm and 424 nm for P450FoAg; 391 nm and 425 nm for P450ZoGa) were plotted against the G6Me concentration and Kd was determined by curve fitting with OriginPro 9.0. All measurements were performed in triplicate.
The activity of the purified reductases FoR, PdR and CPR was measured via the reduction of ferricyanide. The reaction was investigated in buffer C. 25 nM of the reductase and 1 mM K3Fe(CN)6 were mixed in a cuvette and the reaction was started with the addition of 3 mM NADH or NADPH. The decrease of the absorption was measured at 420 nm. Negative controls were performed without reductases. All measurements were performed in triplicate.
Substrate screening with in vitro biotransformation
Reactions with purified enzymes were performed in glass vials with a total volume of 500 µL. The enzyme concentrations were 0.5 µM FoR, 2.5 µM FoX and 0.5 µM P450FoAg. Reactions were performed with 1 mM substrate and 3 mM NADH. All potential substrates except the sugars were dissolved in ethanol before addition to the reaction leading to a final ethanol concentration of 2.0% (v/v). For G6Me, the activity was also verified with 2.0% (v/v) ethanol. Up to three substrates from one substance class were pooled in one reaction. Samples were shaken at room temperature for 24 h before they were analyzed with GC, GC/MS or HPLC depending on the used potential substrate.
Formaldehyde was quantified using the Purpald assay11. 200 µL of a sample were mixed with 50 µL 0.16 M Purpald in 2 M NaOH. After 40 min incubation time at room temperature the absorbance at 550 nm was measured and the concentration of formaldehyde was calculated using a standard curve.
NADH oxidation with G6Me
The NADH oxidation after G6Me addition was measured with 0.02 µM P450FoAg or P450ZoGa, 0.1 µM FoR, 0.25 µM FoX and 0.15 mM NADH in buffer C at room temperature at 340 nm. The background NADH oxidation was recorded for the first 3 min after which G6Me was added to 2.5 mM. To verify that NADH oxidation is dependent on the presence of the correct substrate, a control with the addition of d-galactose instead of G6Me was performed.
Determination of the protein ratio for the kinetic measurement
To determine the ratio of P450 to both of the redox enzymes, the NADH consumption was monitored with varying concentrations of FoR and FoX. The starting ratio was 0.01 µM P450FoAg, 2.5 µM FoX and 0.5 µM FoR with 0.5 mM NADH in a total volume of 200 µL buffer C. First, the background oxidation of NADH was monitored without substrate followed by the addition of 3 mM G6Me. The concentrations of FoR and FoX were reduced by 50% until the reduction of the redox partners reduced the overall NADH consumption after G6Me addition. Due to practical reasons, the concentration of FoX could not be increased higher than 2.5 µM.
The kinetic parameters of P450FoAg or P450ZoGa were determined based on the consumption of NADH at room temperature, which was spectrophotometrically measured at 340 nm in a final volume of 200 µL. The P450 monooxygenases were used at a concentration of 0.01 µM, while the ferredoxin was used at 2.5 µM and the ferredoxin reductase at 0.25 µM. β-Mercaptoethanol originating from the FoX solution was present at 0.41 mM in the final reaction. The G6Me concentration was varied between 50 and 1,500 µM. Reactions were started by the addition of NADH to a final concentration of 0.5 mM and activities were calculated from the initial slope of NADH consumption. Curve fitting was performed with OriginPro 9.0. All measurements were performed in triplicate.
Compatibility of the redox systems
The compatibility of the redox system with P450 monooxygenases were investigated by in vitro biotransformations with a total volume of 200 µL. The following combinations of the redox systems were examined: FoR/FoX, PdR/PdX, FoR/PdX, PdR/FoX and CPR. The final enzyme concentrations were 0.5 µM for reductases, 1.25 µM for ferredoxins and 0.1 µM for P450 monooxygenases. G6Me was used at a concentration of 1 mM as substrate for P450FoAg and 1 mM ( + )-camphor as substrate for P450cam. The reactions were started by the addition of 1.2 mM NADH when using FoR or PdR and 1.2 mM NADPH when using CPR and incubated at room temperature for 4 h. Samples with G6Me were analyzed with GC/MS, while samples with (+)-camphor were analyzed by GC. The conversion was determined from the peak areas of substrates and products. All measurements were performed in triplicate.
Test for the activation by H2O2
P450FoAg and P450ZoGa were investigated for their ability to utilize H2O2 for their activation. These reactions were performed as described in the previous paragraph with G6Me as substrate. The redox partners and NADH were substituted by a total of 1 mM H2O2. The measurements were performed in triplicate.
Gas chromatography (GC) analysis
Terpenes were analyzed by GC. Samples were extracted with 250 µL or 200 µL dichloromethane in the substrate screening or the test for compatibility of the redox systems respectively. The samples of the substrate screening were analyzed on the column Lipodex G (Macherey & Nagel, 50 m × 0.25 mm). 1 µL was injected at 200 °C. The column temperature stayed 13 min at 75 °C before it was increased to 88 °C with 1 °C/min. After 10 min at 88 °C the temperature was raised to 120 °C with 2 °C/min and then to 200 °C with 10 °C/min. FID was used for detection.
The samples of the biocatalysis reactions with P450cam were analyzed with the column Hydrodex-β-3P (Macherey & Nagel, 25 m × 0.25 mm). The injection temperature was 220 °C and the start temperature of the column was 60 °C. After 10 min, the temperature was increased to 180 °C where it remained for another 10 min. FID was used for detection.
Fatty acids, alkenes, steroids, long chain alcohols and lactones were analyzed with GC/MS. Samples were acidified with 10 µL 10% (w/v) HCl and extracted three times with equal volume of diethyl ether. The organic phases were collected and dried using anhydrous Na2SO4. After solvent evaporation the residue was dissolved in 60 µL N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) with 1% (v/v) trimethylchlorosilane (TMCS) and incubated for 30 min at 80 °C before GC/MS analysis. 1 µL of a sample was injected at 250 °C. The temperature stayed at 80 °C for 5 min and was then raised to 250 °C at a rate of 10 °C/min and after this to 340 °C at a rate of 20 °C/min where it remained for 6 min.
Monosaccharides were derivatized as trimethylsilyl oximes27. Proteins in the samples were precipitated with an equal volume of methanol and separated by centrifugation (15 min at 13,000 g, 4 °C). The supernatant was evaporated to dryness and the remaining substances were dissolved in 50 µL pyridine solution with 2.5% (w/v) hydroxylamine hydrochloride. Insoluble parts were separated from the sample by centrifugation. 40 µL of the supernatant was transferred to a GC vial and the formation of oximes was performed for 30 min at 75 °C. The samples were cooled down to room temperature and equal volumes of BSTFA with 1% (v/v) TCMS were added. The derivatization was performed for 30 min at 80 °C before the samples were analyzed by GC/MS. The method used started with a temperature of 80 °C for 10 min. The temperature was raised to 250 °C at a rate of 10 °C/min where it remained for 6 min. A BPX-5 column (SGE Analytik, 25 m × 0.25 mm) was used in all cases.
Phylogenetic tree calculation was carried out with MEGA728. The data set consisted of 100 sequences with identity over 52% and query coverage over 98% to the Formosa agariphila P450 amino acid sequence (WP_038530297.1). The data set was assembled with blastp (protein-protein BLAST)29 against the non-redundant protein sequence database hosted at NCBI using the F. agariphila protein sequence (WP_038530297.1) as query. The sequence of the cytochrome P450 CYP109E1 from Bacillus megaterium (PDBid 5L90) was included to serve as an outgroup. The protein sequence alignment was calculated with MUSCLE30 with default parameters. The maximum likelihood tree was calculated with the most suitable amino acid evolutionary model (LG + G + I), which was tested in MEGA7. If a site was not present in at least 80% of the sequences, for example due to a gap in some of the sequences, then this amino acid site was not used for the calculation of the phylogenetic tree. The phylogeny was tested with 100 bootstrap replications. The complete tree, with bootstrap values, all 100 sequences and the NCBI Reference Sequence identifiers in addition to the species names is presented in Supplementary Fig. 1.
For the agar PUL analysis, public bacterial genomes were downloaded from the NCBI site. For each genome, protein sequences were extracted and labeled with the corresponding locus tag. An in-house pipeline that will be published in detail elsewhere was used to find gene clusters containing agarases or porphyranases belonging to family GH86 as well as agarases of family GH50 and neoagarobiose hydrolases of family GH117 that is a keystone enzyme family indicative of agar metabolism31. Clusters with agarases, which also contained P450 proteins, were extracted for further analysis. P450, ferredoxin and ferredoxin reductase were identified through their corresponding pfam domains (P450, Fer2 and Reductase_C)32. Agarases from families GH50, GH86 and GH117 were identified using the hidden Markov models of CAZyme domains from the dbCAN database33. Blastp was used to find the identities between the different P450 protein sequences (cut-off 1e-10, 50% identity). Cluster information and sequence identity between P450 was displayed in Circos34.
The prediction of transmembrane helices was carried out with the TMHMM Server v. 2.0 using the default settings35.
Life Sciences Reporting Summary
Further information on experimental design is available in the Life Sciences Reporting Summary.
The sequences of the newly characterized P450 monooxygenases and the redox enzymes are available in the NCBI reference sequence database (P450FoAg, WP_038530297.1; P450ZoGa, WP_013995999.1; FoX, WP_038530300.1; FoR, WP_038530304.1). The data that support the findings of this study are available from the corresponding authors upon reasonable request.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
A correction to this article is available online at https://doi.org/10.1038/s41589-018-0020-9.
We thank the German Research Foundation (DFG) for funding through the Research Unit FOR2406. J.-H.H. acknowledges funding by the Emmy Noether Program of the DFG, grant number HE 7217/1-1. We are also grateful to V. Urlacher (Düsseldorf, Germany) for providing the genes encoding P450cam, PdX, PdR and CPR. We thank D. Nelson (Memphis, USA) for assigning the P450s to a subfamily in the P450 database.
About this article
Biochemical characterization of an ulvan lyase from the marine flavobacterium Formosa agariphila KMM 3901T
Applied Microbiology and Biotechnology (2018)