Identification of the Gene Responsible for Lignin-Derived Low-Molecular-Weight Compound Catabolism in Pseudomonas sp. Strain LLC-1

Pseudomonas sp. strain LLC-1 (NBRC 111237) is capable of degrading lignin-derived low-molecular-weight compounds (LLCs). The genes responsible for the catabolism of LLCs were characterized in this study using whole-genome sequencing. Despite the close phylogenetic relationship with Pseudomonas putida, strain LLC-1 lacked the genes usually found in the P. putida genome, which included fer, encoding an enzyme for ferulic acid catabolism, and vdh encoding an NAD+-dependent aldehyde dehydrogenase specific for its catabolic intermediate, vanillin. Cloning and expression of the 8.5 kb locus adjacent to the van operon involved in vanillic acid catabolism revealed the bzf gene cluster, which is involved in benzoylformic acid catabolism. One of the structural genes identified, bzfC, expresses the enzyme (BzfC) having the ability to transform vanillin and syringaldehyde to corresponding acids, indicating that BzfC is a multifunctional enzyme that initiates oxidization of LLCs in strain LLC-1. Benzoylformic acid is a catabolic intermediate of (R,S)-mandelic acid in P. putida. Strain LLC-1 did not possess the genes for mandelic acid racemization and oxidation, suggesting that the function of benzoylformic acid catabolic enzymes is different from that in P. putida. Genome-wide characterization identified the bzf gene responsible for benzoylformate and vanillin catabolism in strain LLC-1, exhibiting a unique mode of dissimilation for biomass-derived aromatic compounds by this strain.


Introduction
Microorganisms capable of catabolizing aromatic compounds play an important role in degrading aromatic biomass compounds in the environment, major components of which are lignin-derived low-molecular-weight compounds (LLCs) produced by lignin depolymerization. They are composed of aldehydes, ketones, and organic acids. Aromatic LLCs are typically catabolized through the upper and lower ring fission pathways, where the aromatic rings are cleaved and subsequent metabolites enter the central carbon metabolism [1].
Many genes coding for the LLC metabolic pathways have been isolated and characterized. The upper and lower ring fission pathways are encoded by different loci and are controlled by independent transcriptional regulators [2,3]. The genes coding for lower pathways, such as potocatechate, benzoate, and gentisate, are ubiquitous and well conserved in many aerobic bacteria [2]. Recent advances in DNA sequencing technology have enabled the identification of additional genes involved in LLC catabolism; however, most of these remain to be studied in terms of their function and regulation.

Bacterial Strains, Plasmid, and Cultivation
The bacterial strains used in this study are presented in Table 1. Pseudomonas sp. strain LLC-1 (NBRC 111237) was isolated from soil in Miyazaki, Japan [15]. P. putida NBRC 14164 was obtained from the Biological Resource Center of the National Institute of Technology and Evaluation (NBRC), Tokyo, Japan. For the growth of Pseudomonas strains, a basal salts medium (BSM) containing (in grams per liter) K 2 7.0) was used. BSM was supplemented with 0.1% (w/v) of various aromatic compounds. The bacterial strains were grown with shaking (120 rpm) at 30 • C and the growth was monitored by optical density measurements at 600 nm. For DNA isolation or freezing stock preparation, Luria-Bertani (LB) medium (Bacto Trypton, 10 g; yeast extract, 5 g; and NaCl, 10 g per liter, pH 7.0) was used. The Escherichia coli cells were grown with shaking (160 rpm) at 37 • C in Terrific broth (TB) medium (1.2% Bacto Tryptone, 2.4% yeast extract, 0.5% glycerol, 0.17 M KH 2 PO 4 -0.72 M K 2 HPO 4 10%(v/v)) supplemented with ampicillin sodium (100 µg/mL). Isopropyl-β-D-thiogalactpyranoside (IPTG) was added to a final concentration of 0.2 mM to induce gene expression under the control of the lac promoter in the plasmid vector.

Genome Sequencing and Computational Analysis
The whole genome sequence of strain LLC-1 was determined as previously reported [16]. The reads obtained by the two systems were assembled using Newbler v.2.8 (Roche, Basel, Switzerland). The genome sequences were annotated using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) [17] and Rapid Annotations using Subsystems Technology (RAST) server v.2.0 [18]. Identification of coding genes was performed using BLAST and BLASTX searches [19]. Sequence comparisons were performed using EasyFig v.2.1 [20], and a map was generated using drawGeneArrows3. The 16S rRNA gene sequences were aligned computationally using the ClustalW algorithm, as described. Phylogenetic trees were generated using the neighbor-joining method with the MEGA X program [21]. The trees were evaluated through bootstrap resampling (500 replicates).

Cloning and Expression of bzf Gene Cluster
DNA isolation, Southern blot analysis, PCR, DNA sequencing, and other DNA manipulations were performed according to standard procedures as described [23]. A sequence encoding the bzfC gene (1.58 kb) was amplified by PCR using primers VSDH-6 (5 -GGCCATTGCCCTTATCCAGGTG-3 ) and VSDH-7 (5 -GTGGTCGAAAGGCTGCATGAATG-3 ), and inserted into pMD20, using the Mighty TA-cloning Kit (Takara Bio Inc., Kusatsu, Japan), yielding the plasmid pMDBzfC. A 1.58 kb XbaI-BamHI fragment was excised from pMDBzfC and inserted into pBluescriptKS+ using the Ligation-Convenience Kit (Nippon Gene, Tokyo, Japan), yielding the plasmid pKSBzfC. An 8.5 kb fragment was amplified from LLC-1 genomic DNA using the primers VSDH-13 (5 -GGTATCGATAAGCTTGATCCAGTCATGAGTGGTTTTCCAGG-3 ) and VSDH-14 (5 -GGGCTGCAGGAATTCGATCTGTTGGCTTGTTCGGGCAATAC-3 ) and PrimeSTAR GXL DNA Polymerase (Takara Bio Inc.). The PCR fragment was treated with Cloning Enhancer and inserted into pBluescriptKS+, digested with EcoRV, using an In-Fusion ® HD Cloning Kit (Clontech Laboratories, Mountain View, CA, USA) to yield the plasmid pKSBzf1. The consistency of the sequence of each plasmid with the genomic sequence was checked by DNA sequencing. All plasmids were transformed into E. coli JM109 cells.

Metabolism of LLCs by Bacterial Strains
E. coli cells grown in TB were harvested by centrifugation and suspended to a turbidity of 2.0 at 600 nm in 50 mM phosphate buffer (pH 7.5) containing 20 mM glucose. Cell suspensions (20 mL) in 200 mL flasks were incubated at 30 • C on a rotary shaker (120 rpm) in the presence of substrates (1 mM). After 12 h incubation, the pH of the sample was adjusted to 2.0-3.0 with 1 M HCl and then extracted with 1.5-fold volumes of ethyl acetate. The organic extracts were dried over anhydrous sodium sulfate and concentrated under vacuum at 37 • C. The residual extracts were dried with anhydrous sodium sulfate and concentrated using a rotary evaporator. The samples were dissolved in methanol and subsequently subjected to gas chromatography-mass spectrometry (GC-MS) analysis.

Analytical Procedures
GC-MS analysis was performed on a Shimadzu model QP-5050A (Shimadzu Co., Kyoto, Japan) platform equipped with a Phenomenex ZB-5 capillary column (inside diameter: 0.25 mm; film thickness: 0.25 mm; length: 30 m) (Phenomenex, Torrance, CA, USA). The column temperature was programmed at 80 • C for 2 min, from 80 to 200 • C at 15.0 • C/min, from 200 to 240 • C at 10.0 • C/min. For trimethylsilyl (TMS) derivatization, samples were dried under vacuum and treated with 30 µL of N-Methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) and 7.5 µL of pyridine at room temperature for 30 min. Oxygen consumption during the metabolism of aromatic compounds by bacteria was measured as described previously [15], and dissolved dioxygen was monitored using a Clark-type polarographic O 2 electrode (Yellow Springs Instruments Model 5300A, Yellow Springs, OH, USA). To achieve this, 3 mL of cell suspension in a chamber was incubated at 30 • C on a magnetic stirrer in the presence of the substrate.

Phylogenetic Analysis of Strain LLC-1
A phylogenetic tree of strain LLC-1 and various Pseudomonas species is shown in Figure 1. The 16S rRNA of strain LLC-1 was almost identical to that of P. putida GB-1 and P. putida F1 (99.9%), with one or two base substitutions, compared with that of P. putida GB-1 or P. putida F1. From the topology of the phylogenetic tree, strain LLC-1 clustered clearly with the P. putida species (Figure 1), with the exception of P. putida NBRC14164, which is distant from other P. putida strains [22]. Comparison of the genome sequences available in the RAST data sets revealed that P. putida GB-1 is the closest neighbor of strain LLC-1, followed by P. putida F1 [16]. These results are consistent with the 16S rRNA phylogenetic analysis. The 16S rRNA of other species exhibited < 99.4% similarity to that of strain LLC-1, showing a distant relationship with strain LLC-1. Oxygen consumption during the metabolism of aromatic compounds by bacteria was measured as described previously [15], and dissolved dioxygen was monitored using a Clark-type polarographic O2 electrode (Yellow Springs Instruments Model 5300A, Yellow Springs, OH, USA). To achieve this, 3 mL of cell suspension in a chamber was incubated at 30 °C on a magnetic stirrer in the presence of the substrate.

Phylogenetic Analysis of Strain LLC-1
A phylogenetic tree of strain LLC-1 and various Pseudomonas species is shown in Figure 1. The 16S rRNA of strain LLC-1 was almost identical to that of P. putida GB-1 and P. putida F1 (99.9%), with one or two base substitutions, compared with that of P. putida GB-1 or P. putida F1. From the topology of the phylogenetic tree, strain LLC-1 clustered clearly with the P. putida species (Figure 1), with the exception of P. putida NBRC14164, which is distant from other P. putida strains [22]. Comparison of the genome sequences available in the RAST data sets revealed that P. putida GB-1 is the closest neighbor of strain LLC-1, followed by P. putida F1 [16]. These results are consistent with the 16S rRNA phylogenetic analysis. The 16S rRNA of other species exhibited < 99.4% similarity to that of strain LLC-1, showing a distant relationship with strain LLC-1.  Pseudomonas sp. strain LLC-1 (NZ_LUVY00000000), P. putida KT2440 (AE015451), P. putida F1 (CP000712), P. putida GB-1 (CP000926), P. putida NBRC14164 (AP013070), P. aeruginosa PAO1(NC_002516), P. mendocina ymp (CP000680), P. fluorescens Pf-5 (CP000076), P. stutzeri A1501 (NC_009434), P. syringae pv tomato str DC3000 (AE016853) and P. emtomophila L48 (CT573326). The numbers in parentheses indicate variants found in one strain. Multiple sequence alignment outputs of the 16S rRNA sequence (1528 bp) were used to generate neighbor-joining phylogenetic trees using MEGA X [21]. The bar indicates expected nucleotide substitutions per site. Numbers indicate the percentage occurrence of the branch in the bootstrapped trees of 500 replicates.

Detection of Catabolic Genes for Aromatic Compounds in Strain LLC-1
Catabolic genes for various aromatic acids were identified from the draft genome sequence of strain LLC-1: anthranilate catabolic atn gene, benzoate catabolic ben gene, catechol catabolic cat gene, 4-hydroxybenzoic acid catabolic pob gene, 4-hydroxyphenylacetic acid catabolic, hpa gene, homogentisate catabolic hmg gene, phenylacetate catabolic paa gene, protocatechuate catabolic pca gene, and vanillate catabolic van genes were found. Aromatic compound catabolic genes found in the genome sequence of strain LLC-1 are shown in Table S1. Protocatechuate, homogentisate, and benzoate are known as key intermediates of funneling pathways for aromatic ring degradation. The van gene encodes a two-component vanillate-O-demethylase (VanAB), and a regulatory component involved in oxidative demethylation of vanillic acid. The pca gene encodes β-ketoadipate pathway enzymes following ring-cleavage reactions, which subsequently enter the tricarboxylic acid cycle.
Despite the close phylogenetic relationship to P. putida, strain LLC-1 lacked certain genes usually found in the P. putida genome; strain LLC-1 does not possess the fer gene, which encodes the ferulic acid metabolic pathway widely distributed in the genome of P. putida [14], and the vdh gene encoding an NAD + -dependent aldehyde dehydrogenase that specifically oxidizes its catabolic intermediate, vanillin. In addition, strain LLC-1 was deficient in genes encoding the isoeugenol (iem) catabolic pathways detected in the genome of a minor group of P. putida [12]. We attempted to search the draft genome sequence of the LLC-1 strain for a gene encoding vanillin dehydrogenase, which converts vanillin to vanillic acid. Annotation by RAST detected 47 open reading frames (ORFs) identified as aldehyde dehydrogenase. Two of these were benzaldehyde dehydrogenases. A BLASTX search was performed on one of the open reading frames present in the vicinity of the van operon encoding vanillate demethylase. The ORF showed 65% identity with the vdh gene encoding vanillin dehydrogenase of P. putida strain IE27 [12], which is a component of the isoeugenol catabolic gene cluster. The ORF identified as benzaldehyde dehydrogenase (ORF4) was adjacent to the LysR family of transcriptional regulators (ORF1), benzoylformate decarboxylase (ORF2), putative major facilitator superfamily (MFS) transporters that are related to benzoate transporter BenK of Acinetobacter sp. [24] (ORF3), and outer membrane OprD family porin protein (ORF5), which is related to the putative phenylacetic acid uptake porin PhaK of P. putida [25] (Figure 2). A BLASTX search showed that ORF2 exhibited 84% and 63% identity at the amino acid level to thiamine diphosphate-dependent benzoylformate decarboxylase from P. fluorescens (PfBFDC) [26] and P. putida (MdlC) [27], respectively. MdlC is known to be involved in mandelic acid metabolism in mandelic acid-degrading P. putida. From these results, it was presumed that this gene cluster is involved in the transport and catabolism of benzoylformic acid. The gene clusters encoded by these five ORFs were named bzfR, bzfA, bzfB, bzfC, and bzfD, as shown in Figure 2.
Genes closely related to the bzf gene cluster in P. fluorescens Pf-5 [26] and P. aeruginosa [28] have been reported, including benzoylformate decarboxylase, MFS transporter, aldehyde dehydrogenase, and porin protein. The gene organization of the regions corresponding to bzfRABCD of strain LLC-1 is shown in Figure 2. The five components of the bzf gene cluster showed 78-81% identity with those of P. fluorescens Pf-5, and 73-80% identity with those of P. aeruginosa PAO1, respectively, with the exception of bzfB whose homolog was not found for Pf-5. From the genome sequences of P. putida F1, P. putida GB-1, and P. putida KT2440, a gene cluster showing identity with the bzf gene cluster was not detected ( Figure 2). Genes 2020, 11, x FOR PEER REVIEW 6 of 13

Biotransformation of LLCs by E.coli Cells Expressing bzf Gene Cluster
We cloned the bzfABCD gene cluster (8.5 kb locus adjacent to the van operon) and expressed it in E. coli to investigate the function of the bzf gene products. The results of GC-MS analysis of the reaction products after the degradation of various aromatic carboxylic acids by resting cells of recombinant E. coli JM109 (pKSBzf1) expressing the bzfABCD gene are shown in Figure 3. When benzoylformic acid was used as a substrate, in the vector control obtained by incubating E. coli carrying pBluescript KS+, the degradation of benzoylformic acid did not occur. In contrast, E. coli cells carrying pKSBzf1 completely converted benzoylformic to benzoic acid. The bzf gene cluster was suggested to be involved in benzoylformic acid catabolism. Although BzfA is presumed to decarboxylate vanillic acid to guaiacol, the reaction products vanillic acid and the by-product vanillyl alcohol were detected, whereas guaiacol was not detected when E. coli cells carrying pKSbzf1 were incubated with vanillin. Therefore, decarboxylation by BzfA is likely to be benzoylformatic acidspecific.  Supplemental Table S1.
BzfC, a structural gene of the bzf gene cluster, has been identified as an NAD + -dependent dehydrogenase. It is proposed to be involved in the oxidation of benzaldehyde produced by the decarboxylation of benzoylformic acid. Strain LLC-1 lacked the ferulic acid catabolic fer gene, and the

Biotransformation of LLCs by E.coli Cells Expressing bzf Gene Cluster
We cloned the bzfABCD gene cluster (8.5 kb locus adjacent to the van operon) and expressed it in E. coli to investigate the function of the bzf gene products. The results of GC-MS analysis of the reaction products after the degradation of various aromatic carboxylic acids by resting cells of recombinant E. coli JM109 (pKSBzf1) expressing the bzfABCD gene are shown in Figure 3. When benzoylformic acid was used as a substrate, in the vector control obtained by incubating E. coli carrying pBluescript KS+, the degradation of benzoylformic acid did not occur. In contrast, E. coli cells carrying pKSBzf1 completely converted benzoylformic to benzoic acid. The bzf gene cluster was suggested to be involved in benzoylformic acid catabolism. Although BzfA is presumed to decarboxylate vanillic acid to guaiacol, the reaction products vanillic acid and the by-product vanillyl alcohol were detected, whereas guaiacol was not detected when E. coli cells carrying pKSbzf1 were incubated with vanillin. Therefore, decarboxylation by BzfA is likely to be benzoylformatic acid-specific.

Biotransformation of LLCs by E.coli Cells Expressing bzf Gene Cluster
We cloned the bzfABCD gene cluster (8.5 kb locus adjacent to the van operon) and expressed it in E. coli to investigate the function of the bzf gene products. The results of GC-MS analysis of the reaction products after the degradation of various aromatic carboxylic acids by resting cells of recombinant E. coli JM109 (pKSBzf1) expressing the bzfABCD gene are shown in Figure 3. When benzoylformic acid was used as a substrate, in the vector control obtained by incubating E. coli carrying pBluescript KS+, the degradation of benzoylformic acid did not occur. In contrast, E. coli cells carrying pKSBzf1 completely converted benzoylformic to benzoic acid. The bzf gene cluster was suggested to be involved in benzoylformic acid catabolism. Although BzfA is presumed to decarboxylate vanillic acid to guaiacol, the reaction products vanillic acid and the by-product vanillyl alcohol were detected, whereas guaiacol was not detected when E. coli cells carrying pKSbzf1 were incubated with vanillin. Therefore, decarboxylation by BzfA is likely to be benzoylformatic acidspecific.  Supplemental Table S1.
BzfC, a structural gene of the bzf gene cluster, has been identified as an NAD + -dependent dehydrogenase. It is proposed to be involved in the oxidation of benzaldehyde produced by the decarboxylation of benzoylformic acid. Strain LLC-1 lacked the ferulic acid catabolic fer gene, and the  Supplemental Table S2.
BzfC, a structural gene of the bzf gene cluster, has been identified as an NAD + -dependent dehydrogenase. It is proposed to be involved in the oxidation of benzaldehyde produced by the decarboxylation of benzoylformic acid. Strain LLC-1 lacked the ferulic acid catabolic fer gene, and the vdh gene encoding an NAD + -dependent aldehyde dehydrogenase specific for vanillin. It was confirmed Genes 2020, 11, 1416 7 of 13 that bzfC shows 65% identity with the vdh gene, encoding the vanillin dehydrogenase component of the isoeugenol catabolic gene cluster in P. putida IE27, as described above. Therefore, we assumed that bzfC is involved in the oxidization of both benzaldehyde and vanillin. We investigated the transformation of benzaldehyde, vanillin, isovanillin, and syringaldehyde in E. coli cells expressing the bzfC gene. In the vector control, E. coli carrying pBluescript KS+ reduced all aldehydes investigated to the corresponding alcohol completely (Figure 4), a reaction assumed to be catalyzed by benzaldehyde reductase from E. coli host cells [29]. After the conversion of benzylaldehyde and vanillin for 12 h, in reaction mixtures containing recombinant E. coli carrying plasmid pKSBzfC, benzoic acid, and vanillic acid instead of benzyl alcohol and vanillyl alcohol were detected ( Figure 4A,B). The conversion of isovanillin by bzfC-expressing E. coli resulted in the formation of the corresponding carboxylic acids, whereas part of syringaldehyde was transformed into syringic acid ( Figure 4C,D). From these results, it was confirmed that oxidization by aldehyde dehydrogenase encoded by bzfC (BzfC) is not limited to benzaldehyde produced by benzoylformic acid decarboxylization, and is expanded to vanillin, isovanillin, and syringaldehyde.
Genes 2020, 11, x FOR PEER REVIEW 7 of 13 vdh gene encoding an NAD + -dependent aldehyde dehydrogenase specific for vanillin. It was confirmed that bzfC shows 65% identity with the vdh gene, encoding the vanillin dehydrogenase component of the isoeugenol catabolic gene cluster in P. putida IE27, as described above. Therefore, we assumed that bzfC is involved in the oxidization of both benzaldehyde and vanillin. We investigated the transformation of benzaldehyde, vanillin, isovanillin, and syringaldehyde in E. coli cells expressing the bzfC gene. In the vector control, E. coli carrying pBluescript KS+ reduced all aldehydes investigated to the corresponding alcohol completely (Figure 4), a reaction assumed to be catalyzed by benzaldehyde reductase from E. coli host cells [29]. After the conversion of benzylaldehyde and vanillin for 12 h, in reaction mixtures containing recombinant E. coli carrying plasmid pKSBzfC, benzoic acid, and vanillic acid instead of benzyl alcohol and vanillyl alcohol were detected ( Figure 4A, B). The conversion of isovanillin by bzfC-expressing E. coli resulted in the formation of the corresponding carboxylic acids, whereas part of syringaldehyde was transformed into syringic acid ( Figure 4C, D). From these results, it was confirmed that oxidization by aldehyde dehydrogenase encoded by bzfC (BzfC) is not limited to benzaldehyde produced by benzoylformic acid decarboxylization, and is expanded to vanillin, isovanillin, and syringaldehyde.   Supplemental Table S2.

Induction of Benzoylformic Acid Metabolism
The metabolic ability of resting cells of strain LLC-1 grown in a medium containing benzoylformic acid or pyruvic acid as the sole carbon source was compared by measuring oxygen consumption rates ( Figure 5). Strain LLC-1 grown on benzoylformic acid as the sole carbon source showed a higher oxygen uptake rate in samples incubated with benzoylformate, benzaldehyde, and benzoic acid, compared to cells grown on pyruvate as a carbon source. This result indicated that strain LLC-1 can metabolize benzoylformic acid and its catabolic intermediate benzaldehyde and benzoic acid more efficiently under induction by benzoylformic acid. coli cells carrying pKSBzfC (black line) or E. coli cells carrying pBluescriptKS+ (yellow line). Peak 3, benzyl alcohol; 4, benzoic acid; 5, vanillyl alcohol; 6, vanillic acid; 7, isovanillyl alcohol, 8; isovanillic acid; 9, syringic alcohol; 10, syringic acid. The products were derivatized with MSTFA and identified by comparing the retention time and mass spectra of authentic standards. m/z values of each peak are listed in Supplemental Table S1.

Induction of Benzoylformic Acid Metabolism
The metabolic ability of resting cells of strain LLC-1 grown in a medium containing benzoylformic acid or pyruvic acid as the sole carbon source was compared by measuring oxygen consumption rates ( Figure 5). Strain LLC-1 grown on benzoylformic acid as the sole carbon source showed a higher oxygen uptake rate in samples incubated with benzoylformate, benzaldehyde, and benzoic acid, compared to cells grown on pyruvate as a carbon source. This result indicated that strain LLC-1 can metabolize benzoylformic acid and its catabolic intermediate benzaldehyde and benzoic acid more efficiently under induction by benzoylformic acid.

Growth on Mandelic Acid
Benzoylformic acid, a substrate for benzoylformate decarboxylase encoded by bzfA, is known to be a catabolic intermediate in the mandelate pathway encoded by the mdl gene of P. putida [27]. The bzfA genes showed 63% identity at the amino acid level with the mdlC genes, but the homologous gene for mdlA encoding mandelate racemase and the mdlB gene encoding (S)-mandelate dehydrogenase were not present in the genome of strain LLC-1. The growth of strain LLC-1 and the mandelic acid-degrading bacterial strain, P. putida NBRC 14164, in a medium containing (R,S)mandelic acid or benzoylformic acid as the sole carbon source were examined ( Figure 6). Strain LLC-1 and P. putida NBRC 14164 were able to grow on benzoylformic acid as a sole carbon source, but strain LLC-1 was not able to grow on mandelic acid. From these results, it was concluded that strain LLC-1 does not produce mandelate racemase and mandelate dehydrogenase, which transforms mandelate to benzoylformic acid.

Growth on Mandelic Acid
Benzoylformic acid, a substrate for benzoylformate decarboxylase encoded by bzfA, is known to be a catabolic intermediate in the mandelate pathway encoded by the mdl gene of P. putida [27]. The bzfA genes showed 63% identity at the amino acid level with the mdlC genes, but the homologous gene for mdlA encoding mandelate racemase and the mdlB gene encoding (S)-mandelate dehydrogenase were not present in the genome of strain LLC-1. The growth of strain LLC-1 and the mandelic acid-degrading bacterial strain, P. putida NBRC 14164, in a medium containing (R,S)-mandelic acid or benzoylformic acid as the sole carbon source were examined ( Figure 6). Strain LLC-1 and P. putida NBRC 14164 were able to grow on benzoylformic acid as a sole carbon source, but strain LLC-1 was not able to grow on mandelic acid. From these results, it was concluded that strain LLC-1 does not produce mandelate racemase and mandelate dehydrogenase, which transforms mandelate to benzoylformic acid. Genes 2020, 11, x FOR PEER REVIEW 9 of 13

Discussion
In this study, the draft genome sequence of strain LLC-1 was used to characterize the bzf gene cluster involved in the catabolism of LLCs. The sequence of the 16S rRNA gene indicated that strain LLC-1 is phylogenetically close to P. putida (Figure 1), but the overall genomic structure exhibited variation between strain LLC-1 and P. putida. LLC-1 does not possess the fer gene, involved in the ferulic acid metabolic pathway widely distributed in the genome of P. putida [11], and the vdh genes encoding a dehydrogenase that specifically oxidizes its catabolic intermediate, vanillin. In addition, strain LLC was deficient in genes encoding the isoeugenol (iem) catabolic pathways detected in the genome of a sub-group of P. putida [12], as well as the mdl gene encoding the mandelic acid degradation pathway [22,27]. On the other hand, strain LLC-1 carried the bzf gene cluster encoding benzoylformic acid catabolic enzymes and transporters that do not exist in the P. putida genome.
In this study, the bzf gene cluster of strain LLC-1 was shown to be involved in the conversion of benzoylformic acid to benzoic acid (Figure 7). Benzoic acid can be degraded to tricarboxylic acid cycle intermediates via the β-ketoadipate pathway encoded by the pca gene in the genome of strain LLC-1 (Table S1) [30]. The first ORF at the end of the bzf gene cluster was identified as a transcriptional regulator (bzfR). Benzoylformic acid-metabolizing enzymes were induced by benzoylformic acid in strain LLC-1 ( Figure 5), and bzfR is a putative transcriptional regulator of benzoylformic acid catabolic enzymes. We previously reported that vanillin-catabolizing enzymes are not induced by vanillin in strain LLC-1 [15]. This may be due to poor interaction between the putative transcriptional regulator BzfR and vanillin or its metabolite, vanillic acid. This transcriptional regulator is likely to be benzoylformic acid-specific. The second ORF (bzfA) in the bzf gene cluster showed 84% or 63% identity at the amino acid level with thiamine diphosphate-dependent benzoyldormate decarboxylase from P. putida (MdlC) or P. fluorescence (PfBFDC), respectively [26,27]. Threedimensional structures of both MdlC (PDB entry code: 1BFD) [31] and PfBFDC (6QSI) [26] are available. They are enzymes whose reaction mechanisms have been precisely investigated based on their structure [32,33]. The results of the biotransformation study ( Figure 3) and comparison of sequences with these known benzoylformate decarboxylases (Figure 2) supported the hypothesis that BzfA is an enzyme that decarboxylates benzoylformic acid. The fourth ORF (bzfC) located between bzfB and bzfD showed 65% identity at the amino acid level with the mdlD gene encoding benzaldehyde dehydrogenase of P. putida [27]. It is obvious that bzfA and bzfC are homologous to

Discussion
In this study, the draft genome sequence of strain LLC-1 was used to characterize the bzf gene cluster involved in the catabolism of LLCs. The sequence of the 16S rRNA gene indicated that strain LLC-1 is phylogenetically close to P. putida (Figure 1), but the overall genomic structure exhibited variation between strain LLC-1 and P. putida. LLC-1 does not possess the fer gene, involved in the ferulic acid metabolic pathway widely distributed in the genome of P. putida [11], and the vdh genes encoding a dehydrogenase that specifically oxidizes its catabolic intermediate, vanillin. In addition, strain LLC was deficient in genes encoding the isoeugenol (iem) catabolic pathways detected in the genome of a sub-group of P. putida [12], as well as the mdl gene encoding the mandelic acid degradation pathway [22,27]. On the other hand, strain LLC-1 carried the bzf gene cluster encoding benzoylformic acid catabolic enzymes and transporters that do not exist in the P. putida genome.
In this study, the bzf gene cluster of strain LLC-1 was shown to be involved in the conversion of benzoylformic acid to benzoic acid ( Figure 7). Benzoic acid can be degraded to tricarboxylic acid cycle intermediates via the β-ketoadipate pathway encoded by the pca gene in the genome of strain LLC-1 (Table S1) [30]. The first ORF at the end of the bzf gene cluster was identified as a transcriptional regulator (bzfR). Benzoylformic acid-metabolizing enzymes were induced by benzoylformic acid in strain LLC-1 ( Figure 5), and bzfR is a putative transcriptional regulator of benzoylformic acid catabolic enzymes. We previously reported that vanillin-catabolizing enzymes are not induced by vanillin in strain LLC-1 [15]. This may be due to poor interaction between the putative transcriptional regulator BzfR and vanillin or its metabolite, vanillic acid. This transcriptional regulator is likely to be benzoylformic acid-specific. The second ORF (bzfA) in the bzf gene cluster showed 84% or 63% identity at the amino acid level with thiamine diphosphate-dependent benzoyldormate decarboxylase from P. putida (MdlC) or P. fluorescence (PfBFDC), respectively [26,27]. Three-dimensional structures of both MdlC (PDB entry code: 1BFD) [31] and PfBFDC (6QSI) [26] are available. They are enzymes whose reaction mechanisms have been precisely investigated based on their structure [32,33]. The results of the biotransformation study ( Figure 3) and comparison of sequences with these known benzoylformate decarboxylases (Figure 2) supported the hypothesis that BzfA is an enzyme that decarboxylates benzoylformic acid. The fourth ORF (bzfC) located between bzfB and bzfD showed 65% identity at the amino acid level with the mdlD gene encoding benzaldehyde dehydrogenase of P. putida [27]. It is obvious that bzfA and bzfC are homologous to mdlA and mdlC, respectively. However, the overall structures of the bzf and mdl gene clusters are completely different. MdlA and mdlC are encoded in the opposite direction to each other and are located on a clearly different transcription unit. It is presumed that the mdl gene cluster underwent reorganization during evolution. Transporter membrane proteins are usually required for the uptake of aromatic carboxylic acids into cells through the cell membrane [34]. The BenK-like MFS transporter encoded by the third ORF, bzfB, and the PhaK-like outer membrane OprD family porin protein encoded by the fifth ORF, bzfD, are putative transporters specific for benzoylformic acid (Figure 7). They may transport benzoylformic acid in strain LLC-1, as well as in recombinant E. coli expressing the bzf gene cluster. On the other hand, vanillin is proposed to be incorporated via simple diffusion in strain LLC-1 and recombinant E. coli cells due to its hydrophobic nature [35].
Genes 2020, 11, x FOR PEER REVIEW 10 of 13 mdlA and mdlC, respectively. However, the overall structures of the bzf and mdl gene clusters are completely different. MdlA and mdlC are encoded in the opposite direction to each other and are located on a clearly different transcription unit. It is presumed that the mdl gene cluster underwent reorganization during evolution. Transporter membrane proteins are usually required for the uptake of aromatic carboxylic acids into cells through the cell membrane [34]. The BenK-like MFS transporter encoded by the third ORF, bzfB, and the PhaK-like outer membrane OprD family porin protein encoded by the fifth ORF, bzfD, are putative transporters specific for benzoylformic acid (Figure 7). They may transport benzoylformic acid in strain LLC-1, as well as in recombinant E. coli expressing the bzf gene cluster. On the other hand, vanillin is proposed to be incorporated via simple diffusion in strain LLC-1 and recombinant E. coli cells due to its hydrophobic nature [35].  Benzoylformic acid is known to be an intermediate of the mandelic acid catabolic pathway of P. putuda. Benzoylformic acid is produced through the isomerization and oxidation of mandelic acid and then undergoes decarboxylation and oxidization to form benzoic acid. Strain LLC-1 did not utilize (R,S)-mandelates as a carbon source ( Figure 6). In addition, ORFs coding for enzymes involved in the metabolism of mandelic acid was not found. Therefore, the role of benzoylformic acid catabolic enzymes of strain LLC-1 is different from that of P. putida, where strain LLC-1 assimilates benzoylformic acid, and not as a metabolic intermediate of other carbon sources. Wei  Benzoylformic acid is known to be an intermediate of the mandelic acid catabolic pathway of P. putuda. Benzoylformic acid is produced through the isomerization and oxidation of mandelic acid and then undergoes decarboxylation and oxidization to form benzoic acid. Strain LLC-1 did not utilize (R,S)-mandelates as a carbon source ( Figure 6). In addition, ORFs coding for enzymes involved in the metabolism of mandelic acid was not found. Therefore, the role of benzoylformic acid catabolic enzymes of strain LLC-1 is different from that of P. putida, where strain LLC-1 assimilates benzoylformic acid, and not as a metabolic intermediate of other carbon sources. Wei et al. recently studied the role of the benzoylformate decarboxylase of P. fluorescens Pf-5 in the degradation of lignin depolymerization fragments. It was confirmed that benzoylformate decarboxylase does not act on guaiacil-glycerol derivatives, an allyl C3 lignin-degrading fragment, and specifically decarboxylates benzoylformic acid or 4-hydroxybenzoylformic acid, which represent aryl C2 substrates [26]. It was proposed that 4-hydroxybenzoylformic acid is formed during the degradation of lignin via oxidation of a keto-aldehyde precursor. Our results provided insight into the role of the benzoylformate pathway encoded by the bzf gene cluster of strain LLC-1 in lignin degradation via the formation of 4-hydroxybenzoylformic acid.
The result that E. coli expressing the bzfC gene oxidized vanillin and syringaldehyde to the corresponding acids indicated that BzfC is a multifunctional enzyme that initiates oxidization of LLCs in strain LLC-1. The substrates of BzfC were not limited to benzaldehyde, which is a catabolic intermediate of benzoylformic acid. This result was consistent with our previous study in which the same substrates were transformed into their corresponding aromatic acids as dead-end products by strain LLC-1 [15]. Alternative use of BfzC in benzoylformate and vanillin catabolism may be due to adaptation to environmental niches. There are two other reasons to explain the role of BzfC in vanillin catabolism. First, the amino acid sequence encoded by the bzfC gene showed relatively high identity (65%) with that of vanillin dehydrogenase (Vdh) of P. putida IE27 [12]. Second, the bzf gene cluster is flanked by the van operon, which encodes a downstream vanillate-metabolizing enzyme (vanAB) and its transcriptional regulator (vanR). Cross-regulation between bzf and van genes may be feasible when the two genes are located at short distances from each other [36,37].
There are examples of studies that attempted to produce vanillin from ferulic acid by constructing a vanillin dehydrogenase-deficient microbial strain [13,14]. Construction of mutant strains deficient in bzfB, bzfC, or bzfD will be useful to determine the role of bzf in vanillin catabolism and benzoylformic acid transportation.

Conclusions
In this study, we revealed that BzfC has a wide substrate range and is an important enzyme that initiates the oxidation of aromatic biomass (LLC) in strain LLC-1. Genome-wide characterization demonstrated that the strain LLC-1 catabolized benzoylformic acid and vanillin via a metabolic pathway that is independent of ferulic acid, eugenol, isoeugenol, or mandelic acid catabolic pathways, exhibiting a unique mode of dissimilation for the metabolic pathways related to biomass-derived aromatic compounds.