Identification of a Phylogenetically Divergent Vanillate O-Demethylase from Rhodococcus ruber R1 Supporting Growth on Meta-Methoxylated Aromatic Acids

Rieske-type two-component vanillate O-demethylases (VanODs) catalyze conversion of the lignin-derived monomer vanillate into protocatechuate in several bacterial species. Currently, VanODs have received attention because of the demand of effective lignin valorization technologies, since these enzymes own the potential to catalyze methoxy group demethylation of distinct lignin monomers. In this work, we identified a phylogenetically divergent VanOD from Rhodococcus ruber R1, only distantly related to previously described homologues and whose presence, along with a 3-hydroxybenzoate/gentisate pathway, correlated with the ability to grow on other meta-methoxylated aromatics, such as 3-methoxybenzoate and 5-methoxysalicylate. The complementation of catabolic abilities by heterologous expression in a host strain unable to grow on vanillate, and subsequent resting cell assays, suggest that the vanAB genes of R1 strain encode a proficient VanOD acting on different vanillate-like substrates; and also revealed that a methoxy group in the meta position and a carboxylic acid moiety in the aromatic ring are key for substrate recognition. Phylogenetic analysis of the oxygenase subunit of bacterial VanODs revealed three divergent groups constituted by homologues found in Proteobacteria (Type I), Actinobacteria (Type II), or Proteobacteria/Actinobacteria (Type III) in which the R1 VanOD is placed. These results suggest that VanOD from R1 strain, and its type III homologues, expand the range of methoxylated aromatics used as substrates by bacteria.


Construction of a Plasmid Expressing vanAB Genes and Growth Tests of Strain Derivatives
To obtain pBS1-vanAB plasmid (Table 1), which contain the vanAB genes under the control of an L-arabinose-inducible promoter, a restriction enzymes approach was used. In brief, PCR product comprising vanAB genes (locus tags: E2561_01225-E2561_01230), was obtained using oligos FW_vanAB_R1_EcoRI (5 -TGACGAATTCGAAGGAACGACATGACC GATC-3 ) and RV_vanAB_R1_XbaI (5 -GTACTCTAGATGTATCCGATGACCAGGCC-3 ) including underlined restriction sites for EcoRI and XbaI enzymes. The amplified DNA fragment was purified and double digested to be ligated into EcoRI/XbaI restriction sites Microorganisms 2023, 11,78 3 of 14 of pBS1 [30], forming pBS1-vanAB plasmid, that was electroporated into E. coli Mach1. Transformed cells were selected in LB medium supplemented with gentamycin 30 µg mL −1 ; and selected transformants were checked by PCR for proper insertion of the vanAB genes. The full-length gene construct was again checked by Sanger sequencing for errors, and the pBS1-derived plasmid was transferred into strain JMP134 for phenotypic analysis. To evaluate growth proficiency, derivatives of JMP134 strain carrying vanAB-expressing plasmid were grown overnight on LB medium, and then inoculated at 0.2% in cultures containing 5 mM VA, PCA, 3-MB, 3-HB, 5-MS or gentisate as the sole carbon and energy source. For expression of vanAB genes driven by the heterologous P BAD promoter, these derivatives were exposed to 1 mM L-arabinose in addition to growth substrates. The cultures were incubated in a 96-well microplate (Thermo Fisher Scientific, Rochester, NY, USA) at 30 • C and the OD 600 was recorded in a Synergy HTX Multi-Mode plate reader (BioTek, Winooski, VT, USA).

Resting Cell Assays
Resting cells of strain JMP134 derivatives were grown on 5 mM VA or 3-HB plus 1 mM arabinose where appropriate. These cells were washed twice with 1 volume of phosphate buffer (14 g/L Na 2 HPO 4 ·12H 2 O, 2 g/L KH 2 PO 4 ), 5X concentrated, and subsequently incubated with 1 mM of each compound to be assayed where appropriate. Samples were obtained at different times, filtered (0.22 µm), and stored at −20 • C.
Concentrations are represented as percentages of the initial substrate concentration. Three biological replicates were performed for growth measurements. Error bars indicate SEM.

Heterologous Expression and Resting Cell Assays Suggest a Key Role of VanOD from Rhodococcus ruber R1 in meta-Methoxylated Aromatic Acids Degradation
Our original observation about the ability of R. ruber R1 to grow on VA as a sole carbon and energy source ( Figure 1A), in the absence of a canonical VanOD as the one described in R. jostii RHA1 [11], prompted us to analyze its growth profile on other meta-methoxylated aromatic substrates, such as 3-methoxybenzoate (3-MB) and 5-methoxysalicylate (5-MS). Interestingly, 3-MB and 5-MS also supported cell proliferation (see Figure 1B,C for a de-tailed growth curve of R1 cells), suggesting O-demethylation of these substrates into 3-hydroxybenzoate (3-HB) and gentisate respectively, as depicted in Figure 1D. For 3-MB consumption, the O-demethylation activity was additionally suggested by the transient accumulation of 3-HB ( Figure 1B). In the case of 5-MS growth, the absence of gentisate in the supernatant of R1 cell cultures would be correlated with a lower rate of substrate consumption, in comparison to VA and 3-MB consumption curves, as shown by Figure 1A-C; avoiding the accumulation of intermediates. Further catabolism of 3-HB and gentisate is correlated with the presence of genes encoding 3-hydroxybenzoate 6-hydroxylase (locus tag: E2561_07550) and gentisate 1,2-dioxygenase (locus tag: E2561_07565) enzymes, comprising the catabolic route for 3-HB via gentisate in strain R1, which are closely related to the enzymes described for R. jostii RHA1 [38,39]. The presence of O-demethylation activities for VA, 3MB, and 5-MS in R. ruber R1 raise the possibility that VanOD encoded in the genome of this strain would be responsible for all of them.
In order to gain comprehension about the whole function of the divergent VanOD from R. ruber R1, a plasmid construct containing the vanAB genes of this strain was introduced into C. pinatubonensis JMP134, a well-known aromatics-degrader bacterium unable to grow on VA, 3-MB, and 5-MS, but that harbors PCA, 3-HB, and gentisate degradation routes ( Figure 2B,D,F) [40], allowing complementation of the catabolic abilities. The expression of the vanAB genes was controlled by the L-arabinose-inducible P BAD promoter, which was chosen since L-arabinose is non-toxic and is not a carbon source for C. pinatubonensis JMP134, permitting reliable growth tests in this strain [30,41]. Remarkably, the presence of vanAB genes was sufficient to allow L-arabinose-depending growth on VA, 3-MB, and 5-MS of JMP134 strain (Figure 2A,C,E), strongly suggesting that VanOD of R1 strain has Odemethylation activity toward the three meta-methoxylated aromatic substrates. It should be noted that, in the absence of L-arabinose as an inducer, no growth was observed (data not shown), and that the presence of the empty pBS1 vector has no effect on cell proliferation of JMP134 strain on these substrates (Figure 2A,C,E). Moreover, resting cell assays of JMP134 (pBS1-vanAB) cells previously grown on VA showed a sharp decrease in the concentration of VA, and a slower consumption rate for 3-MB and 5-MS, also detecting the occurrence of 3-HB in 3-MB-incubated cells ( Figure 3A-C); which provides further support for VA/3-MB/5-MS O-demethylase activity encoded by R1 vanAB genes. This inference was additionally supported by detecting a small accumulation of formaldehyde in parallel to substrates consumption ( Figure 3A-C), which is the by-product of O-demethylation by VanOD [26].
The presence of functional groups in the potential substrates of the divergent VanOD of R1 strain was the next interesting issue to be determined. Nishimura et al. [19] reported that a carboxylic acid on the benzene ring in conjunction with a hydroxyl group in para-orientation, as occurs in VA or syringate molecules, is required for efficient methoxy oxidation in meta-position of the VanAB substrates in Streptomyces sp. NL15-2K, which is homologous to VanAB from RHA1 (70% aa identity for oxygenase subunit). Recently, the properties of VanOD from Pseudomonas sp. HR199 were extensively examined, confirming that the presence of a carboxylic acid moiety is essential, and that catalysis occurs selectively at the meta-position relative to the -COOH group in the aromatic ring, although exposing specific differences in substrate recognition in comparison to VanAB from Streptomyces sp. NL15-2K [19,28]. To confirm that previous observations also apply to VanOD from R. ruber R1, resting cell assays considering additional potential substrates were performed in C. pinatubonensis JMP134 carrying the plasmid that contains the vanAB genes from R1 strain. The doubly meta-methoxylated syringate that carries a -OH group in the para-position relative to the carboxylic acid was a proper substrate for VanOD of the R1 strain; being 3-O-methylgallate, the partially demethoxylated analog, identified as the only conversion product of its catalysis ( Figure 3D). Meanwhile, 3-O-methylgallate apparently was not recognized as a substrate by the VanOD of R1 strain ( Figure 3E), similar to what was described for the VanAB from HR199 strain [28], but unlike VanAB of NL15-2K strain which is able to generate a mixture of 3-O-methylgallate and gallate, the fully demethoxylated analog, in the presence of syringate [19]. These results were supported by introduction of vanAB genes of R1 strain into Pseudomonas putida KT2440, which is unable to grow on syringate or 3-O-methylgallate but contains a functional gallate degradation pathway [42,43], being the product of two consecutive O-demethylations over syringate comprising 3-O-methylgallate as intermediate, as mentioned before. The P. putida KT2440 (pBS1-vanAB) strain was unable to grow on syringate as a sole carbon and energy source (data not shown), suggesting that inefficient O-demethylation of 3-O-methylgallate by VanOD from R1 could be the reason for this phenotype.    Additional compounds including differences in the key positions of functional groups for the recognition of substrates by this enzyme, such as 2-methoxybenzoate and 4-methoxybenzoate (methoxy group in orthoor para-position in relation to -COOH group), isovanillate (methoxy group in para-position in relation to -COOH group with an adjacent -OH group in meta-position), homovanillate (VA analog with a -CH 2 COOH replacing -COOH group), 3-methoxyphenylacetate (3 MB analog with a -CH 2 COOH replacing -COOH group), 3-methoxysalicylate (methoxy group in meta-position in relation to -COOH group with an adjacent -OH group in ortho-position), and guaiacol (methoxy group with an adjacent -OH group but lacking -COOH group) were not degraded by resting cells of C. pinatubonensis JMP134 carrying the vanAB genes of R1 strain (see Figure 3F-H, for isovanillate, guaiacol, and 3-methoxysalicylate as representative examples).
In summary, the results of growth tests and resting cell assays suggest that divergent VanOD from strain R1 not only recognizes VA, but also 3-MB, 5-MS, and syringate as proper substrates to a lesser extent ( Figure 3A-D). According to this, vanAB genes could be key not only on VA degradation, but also on the potential catabolism of 3-MB and 5-MS in additional Rhodococcus species that carry this divergent version of VanOD.

Rhodococcus Strains Carrying VanAB Homologues Closely Related to VanOD of R. ruber R1 Strain Are Able to Grow on VA, 3-MB, and 5-MS as a Sole Carbon and Energy Sources
To explore the phenotypic differences of selected Rhodococcus species carrying divergent VanOD related to VanAB from R. jostii RHA1 strain or VanAB from R. ruber R1 strain, we analyzed their growth profile on several meta-methoxylated substrates structurally related to VA, such as 3-MB, 5-MS, syringate, 3-methoxysalicylate, 3-methoxyphenylacetate, and homovanillate, in addition to some of the putative products of O-demethylation such as PCA, 3-HB, gentisate, 2,3-dihydroxybenzoate and 3-hydroxyphenylacetate (Figure 4).
Results showed that Rhodococcus strains harboring vanAB-like genes similar to R. ruber R1 as R. ruber DSM 43338 T , R. ruber Chol-4 and R. pyridinivorans JCM 10940 T [8,44,45], and vanAB genes comparable to R. jostii RHA1 as R. aetherivorans BCP1 [46,47] were able to grow on VA and its O-demethylation product, PCA (Figure 4), suggesting that all these strains contain proficient VanOD-encoded genes and a functional PCA pathway ( Figure 1D). We also included in our growth profile assays marine-isolated Rhodococcus strains MS13 and H-CA8f as control [48,49], since they apparently do not harbor VanOD-encoded genes, even though they carry the classical PCA pathway [50], as revealed by BLAST searches and confirmed by growth on PCA as a sole carbon and energy source (Figure 4). Accordingly, both Rhodococcus strains of marine origin were unable to grow on VA (Figure 4), confirming the previous bioinformatic survey that revealed the absence of VanOD encoded genes, and suggesting that VA degradation activity could be linked to Rhodococcus species found mainly in soil or freshwater environments, probably correlated to lignin depolymerization [51][52][53]. Strains belonging to Rhodococcus genus were grown in mineral salt medium with 5 mM of several meta-methoxylated aromatic acids (methoxylated substrates; left side) and its O-demethylated products (right side) as sole carbon and energy sources. Shading indicates optical density (OD) at 600 nm after 40 h (average of three biological replicates). It is worth mentioning that O-demethylated products related to syringate (3-O-methylgallate and gallate) and homovanillate (homoprotocatechuate) were rapidly oxidized in solution yielding an intense dark brown color on the medium, precluding determination of optical density, and consequently were excluded of the study.
Remarkably, R. ruber strains DSM 43338 T and Chol-4 that harbor vanAB genes close related to R1 homologues were also able to use 3-MB and 5-MS, and its putative demethylation products, 3-HB and gentisate, respectively, as sole carbon and energy sources (Figure 4), suggesting that their VanOD enzymes are able to act on both meta-methoxylated substrates, and that they harbor the corresponding putative downstream pathways ( Figure 1D). No-tably, R. pyridinivorans JCM 10940 T containing R1-like vanAB genes was unable to grow on 3-MB and 5-MS, and was also unable to grow on 3-HB and gentisate (Figure 4), which is in accordance with the absence of genes encoding 3-HB 6-hydroxylase and gentisate 1,2-dioxygenase enzymes. This suggests that lack of a functional 3-HB/gentisate pathway might impair its growth on such meta-methoxylated substrates, regardless of the presence of a proficient VanOD. Conversely, despite strains RHA1 and BCP1 harbor RHA1-like vanAB genes and contain a functional 3-HB/gentisate pathway, both were unable to grow on 3-MB and 5-MS (Figure 4), suggesting that the VanOD harbored by these Rhodococcus strains does not support the O-demethylation activities toward these meta-methoxylated substrates. These results could indicate that the in vivo range of substrate acceptance for R1-like and RHA1-like VanOD enzymes is not the same. Finally, all Rhodococcus strains tested were unable to use the remaining VA analogs assayed as sole carbon and energy sources, including those where -COOH group is replaced by -CH2COOH as 3-methoxyphenylacetate or homovanillate (Figure 4).

Two-Component Rieske-Type VanOD of Rhodococcus Species Are Allocated in Two Divergent Phylogenetic Clades
The existence of at least two distinct VanOD types in Rhodococcus species prompted us to evaluate the distribution of each kind in this genus and other actinobacterial and proteobacterial genomes. For that purpose, we chose as gene marker the VanA product, coding the oxygenase component of the enzyme, from R. jostii RHA1 [11], R. ruber R1, and also Pseudomonas sp. HR199 as bona fide representative of proteobacterial VanOD [14]. Then, we conducted a search in the non-redundant protein sequences database of GenBank as of September 2022, selecting VanA from bacterial species displaying at least 60% amino acid identity in order to establish phylogenetic relationships. As a result of a high number of redundant VanA sequences, we selected one representative VanA homologue belonging to each genus identified. The resulting VanA phylogenetic tree showed three clearly divergent groups, in which a precise partition was perceived between a proteobacterial clade (called type I), including the well-known VanA homologues from Acinetobacter baylyi ADP1, Pseudomonas sp. HR199, and Comamonas testosteroni BR6020 [14,16,18]; and an actinobacterial clade (type II), including the aforementioned VanA homologues from R. jostii RHA1 and Streptomyces sp. NL15-2K [11,19] (Figure 5). Interestingly, a distinct third clade was detected (type III), internally partitioned in two subclades including homologues from Proteobacteria (type IIIA) and Actinobacteria (type IIIB) ( Figure 5). The last one included VanA from R. ruber R1, revealing that the VanOD reported in this work is the first member of this clade whose functionality and substrate range is analyzed in detail. It should be noted that a closer inspection of each clade reveals a predominance of βand γ-proteobacterial VanA homologues among members of the type I clade, meanwhile only homologues from Actinobacteria representatives were found in type II (data not shown). Conversely, a prevalence of VanA homologues from α-proteobacteria subclass representatives (IIIA) in conjunction with Actinobacteria (IIIB) were observed in type III (data not shown).
In order to gain a deeper understanding of the phylogenetic relationships between VanOD enzymes from Rhodococcus species, an additional phylogenetic tree was constructed including VanA homologues from a broader range of Rhodococcus species representatives ( Figure 6). Similar to what was previously observed, it was shown that VanA from Rhodococcus species are grouped either in conjunction with VanA from RHA1 strain (type II clade) or VanA from R1 strain (type III clade). The number of VanA homologues from Rhodococcus species grouped in each clade was roughly similar, suggesting that both types of VanOD are numerically relevant in this actinobacterial genus. No VanA homologue of the Rhodococcus species considered in this study was located out of these clades.  [33] based on sequence alignments calculated using MAFFT software [36] is shown with SH-like approximate likelihood ratio support values (n = 1000) given at each node (values >70% are shown). Light orange, Actinobacteria phyla; green, Proteobacteria phyla.
In order to gain a deeper understanding of the phylogenetic relationships between VanOD enzymes from Rhodococcus species, an additional phylogenetic tree was constructed including VanA homologues from a broader range of Rhodococcus species representatives ( Figure 6). Similar to what was previously observed, it was shown that VanA from Rhodococcus species are grouped either in conjunction with VanA from RHA1 strain (type II clade) or VanA from R1 strain (type III clade). The number of VanA homologues from Rhodococcus species grouped in each clade was roughly similar, suggesting that both types of VanOD are numerically relevant in this actinobacterial genus. No VanA homologue of the Rhodococcus species considered in this study was located out of these clades. Figure 5. Evolutionary relationships among VanA homologues from bacteria. Maximum likelihood topology provided by IQ-TREE software [33] based on sequence alignments calculated using MAFFT software [36] is shown with SH-like approximate likelihood ratio support values (n = 1000) given at each node (values >70% are shown). Light orange, Actinobacteria phyla; green, Proteobacteria phyla.

Conclusions
Given the current interest in O-demethylation reactions for lignin conversion into renewable chemicals [54], this study aimed to shed light on Rhodococcus enzymes acting on meta-methoxylated aromatic compounds such as VA, one of the most prominent lignin-derived phenolics. This work revealed that Rhodococcus genus harbors at least two divergent types of VanOD-encoding genes represented by vanAB from Rhodococcus jostii RHA1 (Type II) and vanAB from Rhodococcus ruber R1 (Type III). Most interestingly, the VanOD from R1 strain is responsible for catabolism of additional meta-methoxylated phenolics such as 3-MB and 5-MS, as inferred from growth tests and resting cell assays of a heterologous strain expressing R1 vanAB genes, and from the substrate utilization pattern of Rhodococcus strains harboring close homologues of this enzyme. This expanded substrate specificity would be advantageous for metabolic engineering endeavors focused on bioconversion process toward renewable chemicals based on microbial demethylation of lignin monomers.

Data Availability Statement:
The data supporting the conclusions of this work are included within the manuscript and there were no large datasets generated or analyzed during the current study.