Metabolic pathway of 6-aminohexanoate in the nylon oligomer-degrading bacterium Arthrobacter sp. KI72: identification of the enzymes responsible for the conversion of 6-aminohexanoate to adipate

Abstract
Arthrobacter sp. strain KI72 grows on a 6-aminohexanoate oligomer, which is a by-product of nylon-6 manufacturing, as a sole source of carbon and nitrogen. We cloned the two genes, nylD1 and nylE1, responsible for 6-aminohexanoate metabolism on the basis of the draft genomic DNA sequence of strain KI72. We amplified the DNA fragments that encode these genes by polymerase chain reaction using a synthetic primer DNA homologous to the 4-aminobutyrate metabolic enzymes. We inserted the amplified DNA fragments into the expression vector pColdI in Escherichia coli, purified the His-tagged enzymes to homogeneity, and performed biochemical studies. We confirmed that 6-aminohexanoate aminotransferase (NylD1) catalyzes the reaction of 6-aminohexanoate to adipate semialdehyde using α-ketoglutarate, pyruvate, and glyoxylate as amino acceptors, generating glutamate, alanine, and glycine, respectively. The reaction requires pyridoxal phosphate (PLP) as a cofactor. For further metabolism, adipate semialdehyde dehydrogenase (NylE1) catalyzes the oxidative reaction of adipate semialdehyde to adipate using NADP+ as a cofactor. Phylogenic analysis revealed that NylD1 should be placed in a branch of the PLP-dependent aminotransferase sub III, while NylE1 should be in a branch of the aldehyde dehydrogenase superfamily. In addition, we established a NylD1/NylE1 coupled system to quantify the aminotransferase activity and to enable the conversion of 6- aminohexaoate to adipate via adipate semialdehyde with a yield of > 90%. In the present study, we demonstrate that 6- aminohexanoate produced from polymeric nylon-6 and nylon oligomers (i.e., a mixture of 6-aminohexaoate oligomers) by nylon hydrolase (NylC) and 6-aminohexanoate dimer hydrolase (NylB) reactions are sequentially converted to adipate by metabolic engineering technology.

Introduction
Nylons are synthetic polymers that contain recurring amide groups in their main polymer chains. Due to their high strength, elasticity, and chemical and thermal resistance, ny- lons have been used for the production of various fibers and plastics. Particularly, nylon-6 (PA6) and nylon-66 (PA66) ac- count for approximately 90% of the total production of syn-unit of 6-aminohexaoate (Ahx), whereas two monomeric units, hexamethylenediamine and adipate, are alternatively combined in PA66 (Travis 1998; McIntyre 2005). PA6 is in- dustrially obtained by the ring cleavage polymerization of ε- caprolactam. A major route of ε-caprolactam production in- volves oxime formation with NH2OH (prepared from NH3 and H2O2) and a Beckmann rearrangement using cyclohexa- none as the starting material in the presence of a strong acid, which generates large amounts of salts as by-products (Travis 1998). The chemical reaction responsible for adipate production also requires high temperature and pressure and produces hazardous substances such as nitrous oxide as a by-product (McIntyre 2005). Attempts to develop the bio- based production of adipate (Chen and Nielsen 2013; Draths and Frost 1994; Polen et al. 2013) and Ahx (Sattler et al. 2014; Turk et al. 2016) using synthetic biology and metabolic engi- neering have been made as an approach for the construction of a sustainable production system. Moreover, environmental considerations suggest that nylon recycling should be manda- tory. However, most nylon wastes are currently disposed of by burning or dumping, although the chemical conversion of PA6 to ε-caprolactam has been reported (Chen et al. 2010; Iwaya et al. 2006). Biochemical studies on the relevant enzymes are important to achieve the biotechnological production and recycling of nylons.

Arthrobacter sp. KI72 is a bacterium that can grow on an Ahx oligomer (designated as “nylon oligomer”) as the sole carbon and nitrogen source (Okada et al. 1983). The strain harbors three different plasmids, pOAD1, pOAD2, and pOAD3 (Kato et al. 1995). Previous biochemical studies have revealed that three enzymes NylA, NylB, and NylC encoded on pOAD2 are responsible for the degradation of the Ahx oligomer to monomers (Negoro 2000) (Fig. 1a). The Ahx- cyclic dimer hydrolase (NylA; EC3.5.2.12), a member of the amidase signature hydrolase family, specifically hydrolyzes one of the two amide bonds in the Ahx-cyclic dimer to gen- erate an Ahx linear dimer (Yasuhira et al. 2010). Ahx dimer hydrolase (NylB; EC3.5.1.46), a member of the penicillin- recognizing family of serine-reactive hydrolases, hydrolyzes the Ahx oligomers by an exo-type mechanism (Negoro et al. 2005, 2007). The Ahx oligomer hydrolase (NylC; EC3.5.-.-) degrades the Ahx-cyclic and Ahx-linear oligomers with a de- gree of polymerization greater than three by an endo-type mechanism (Negoro 2000; Yasuhira et al. 2007a; Negoro et al. 2012). The thermostabilized NylC protein variant de- grades thin-layered PA6 (thickness 0.26 μm) almost completely at a constant reaction rate (Nagai et al. 2014).The compound 4-aminobutyrate (γ-aminobutyrate: GABA), which contains shorter methylene chains than Ahx, is a non-protein amino acid produced by the de- carboxylation of L-glutamate by glutamate decarboxyl- ase in most eukaryotic and prokaryotic organisms. GABA plays an important role as a neurotransmitter in mammalian cells (Macdonald and Olsen 1994). Stress in plants initiates a calmodulin-dependent signal transduc- tion pathway, in which increased cytosolic Ca2+ acti- vates glutamate decarboxylase (Shelp et al. 1999). The degradation of GABA begins with its transamination to succinate semialdehyde ( SSA) catalyzed by 4- aminobutyrate aminotransferase (EC 2.6. 1.19) (Schneider et al. 2002) (Fig. 1b). The SSA is subse- quently oxidized to succinate by NAD(P)+-dependent succinate semialdehyde dehydrogenase (EC 1.2.1.16)(Fig. 1e) or reduced to γ-hydroxybutyrate by succinate semialdehyde reductase (EC 1.1.1.79) (Kockelkorn and Fuchs 2009). However, the enzymes responsible for the metabolisms of Ahx (containing longer methylene chains than GABA) have not been identified so far.

We have recently reported the draft genomic sequence of Arthrobacter strain KI72 (Takehara et al. 2017). In the present study, we screened two putative genes (nylD1 and nylD2) for the production of 6-aminohexanoate amino- transferase and 20 putative genes (nylE1–nylE20) for the production of adipate semialdehyde dehydrogenase from a homology search of the genomic sequence of strain KI72. We selected two genes (nylD1 and nylE1) from the phylogenic analysis, purified and characterized their gene products, and established a coupled reaction system com- posed of NylD1 and NylE1that enables the continuous monitoring of the amino transferase activity and complete conversion of 6-aminohexanoate to adipate.6-Aminohexanoate (Ahx), pyridoxal phosphate (PLP), and L-glutamate were purchased from Nacalai Tesque, Inc. (Kyoto, Japan). Adipate, succinate, pyruvate, glyoxylate, L-alanine, and glycine were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). NAD+, and NADPH were purchased from Oriental Yeast Co., Ltd. (Tokyo, Japan). 4-Aminobutyrate (GABA) was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). α-Ketoglutarate (α-KG) was purchased from Sigma-Aldrich Co. LLC (St. Louis, USA). Adipate semialdehyde (ASA) and succinate semi- aldehyde (SSA) were chemically synthesized in our lab- oratory. The details are described in the supplementary text. The structure of the synthesized compounds was confirmed by an NMR analysis. We stored the ASA and SSA samples by freezing them at − 20 °C and pre- pared the substrate solution just before use.The bacterial strains, plasmids, and enzymes used in this study are described in Table S1. Arthrobacter sp. KI72 strains (NBRC 14590, National Institute of Technology and Evaluation, Japan) were grown at 30 °C in 100 ml of LB medium. Whole genome DNA was prepared from strain KI72 by the conventional phenol/chloroform extraction meth- od (Sambrook and Russell 2001). NylD1 and NylE1 were designed to be expressed as the His-Tag fused proteins.

The NylD1 and NylE1 genes were amplified from Arthrobacter sp. KI72 genomic DNA by PCR using the primers described in Table S2. The amplified fragments were inserted into a pColdI expression vector (Takara Bio Inc. Shiga, Japan), and the li- gated DNA was used to transform Escherichia coli BL21 (DE3) cells. The resulting plasmids, pCold-NylD1 and pCold-NylE1, were constructed for the expression of the re- combinant proteins. E. coli strains harboring the hybrid plas- mid were grown at 37 °C in LB medium containing ampicillin (100 μg/ml). Plasmid DNA was prepared by the conventional alkaline extraction method (Sambrook and Russell 2001)monomeric units of 6-aminohexanoate, adipate semialdehyde, and adipate are illustrated as light blue, purple, and red closed circles, respectively. Terminal amino N, carboxyl C, and aldehydic C are shown as light blue, red, and black letters, respectively. b, c Reactions with GABA (b) or Ahx (c) catalyzed by NylD1 are shown. d, e Reactions from ASA (d) or SSA (e) catalyzed by NylE1 are shownThe phylogenetic tree was constructed using the Clustal W pro- grams (http://www.ebi.ac.uk/tools/clustalw2, Larkin et al. 2007).To purify the His-tagged NylD1 and NylE1 enzymes, E. coli cells obtained by centrifugation (14,500×g for 10 min, 4 °C) were lysed by sonication (20 kHz, 30 s × 12 times) in 10 mL of 100 mM potassium phosphate buffer (KPB, pH 7.0). Cell extracts obtained by centrifugation were used as crude enzyme solutions. The enzymes were purified from the supernatant using TALON metal affinity resin (Takara Bio USA, Inc. California, USA) following the manufacturer’s instructions. Following purification, the enzymes were identified by SDS- PAGE, and the protein bands were visualized by staining with Coomassie Blue. The fraction containing NylD1 and NylE1 was dialyzed overnight against 100 mM KPB (pH 7.0). The protein concentration was determined with the Bradford assay using bovine serum albumin as the standard.For the quantitative assay of the semialdehyde dehydrogenase (NylE) activity, the enzyme was incubated at 30 °C with0.1 mM synthetic semialdehyde (ASA or SSA) and 0.1 mM NADP+ in 100 mM KPB (pH 7.8).

The reaction was initiated by the addition of NylE1, and the rate of the oxidative reaction was monitored as the change in the NADPH concentration, as indicated by the absorbance at 340 nm (A340) (ε = 6200 M−1 cm−1) using a spectrometer (V-730BIO, Jasco, Japan). For the qualitative detection of the aminotransferase (NylD) activity by thin layer chromatography (TLC), the en- zyme was incubated at 30 °C with 5 mM of the amino donor (Ahx or GABA), 5 mM of the amino acceptor (α-KG), and0.1 mM PLP in 100 mM KPB (pH 7.8). One microliter of the reaction mixture was spotted on a thin-layer silica gel plate (TLC silica gel 60, Merck Millipore) and developed with sol- vent I (1-propanol:ethyl acetate:water:25% ammonia solu- tion = 4.49:0.75:4.49:0.28). The reaction product on the TLC plate was then detected by spraying it with a 0.2% ninhydrin solution in butanol saturated with water as described previous- ly (Negoro et al. 2005).For quantitation of the Ahx aminotransferase activity, a NylD1 reaction was performed in the presence of a great ex- cess of NylE1 enzyme activity (i.e., the NylD1/NylE1 coupled system). NylD1 and NylE1 were incubated at 30 °C with0.2 mM of the amino donor (Ahx), 0.25 mM of the amino acceptor (α-KG, pyruvate, or glyoxylate), 0.1 mM PLP, and0.25 mM NADP+ in 100 mM KPB (pH 7.8) (i.e., the standard assay conditions for the coupled system). The rate of the re- action was monitored at A340. Aminotransferase activity was calculated by subtracting the background level of the activity (without NylD1) from the activity (in the coupled system) (see Fig. 7). For a reaction using a higher concentration of Ahx in the TLC analysis, the enzyme reactions were performed in1.0 mL of a reaction mixture containing 2 mM Ahx,2.5 mM of the amino acceptor, 0.1 mM PLP, 2.5 mM NADP+, and 100 mM KPB buffer (pH 7.8). Aliquots of 30 μl were sequentially removed, and the reaction products were visualized by TLC. The NADPH concentration was de- termined from the A340.

Results
We planned to clone the genes responsible for Ahx metabo- lism in Arthrobacter sp. strain KI72, assuming that the Ahx is converted to adipate by enzymes analogous to those involved in GABA metabolism (Fig. 1). On the basis of the protein assignment performed for the draft genomic sequence com- posed of 105 contig sequences [4,568,574 base pairs (bp) in total length] (Takehara et al. 2017), we screened two putative enzymes (NylD1 and NylD2) that were homologous to the GABA aminotransferases on the genome of Arthrobacter sp. KI72 (Fig. 2). We also identified 20 putative semialdehyde dehydrogenases (NylE1–NylE20), which should be classified in the aldehyde dehydrogenase (ALDH) superfamily (Fig. 3). In the following sections, we describe the phylogenic relation- ship of these enzymes.PLP-dependent aminotransferases have been phylogenetically classified in four classes (Alexander et al. 1994, Mehta et al. 1993) but have recently been further divided into five sub- groups (Hwang et al. 2005). Subgroups I and II include aro- matic and aspartic acid transaminases, subgroup IV includes the branched-chain transaminases, and subgroup V includes the serine and histidinol phosphate transaminases. Phylogenic analyses revealed that NylD1 and NylD2 should be placed in subgroup III, which includes GABA aminotransferase (EC 2.6.1.19), ornithine aminotransferase (OAT; EC2.6.1.13), ω- amino acid-pyruvate aminotransferase (OAPT) (EC2.6.1.18), acetylornithine aminotransferase (ACOAT) (EC2.6.1.11), 7,8- diaminopelargonic acid aminotransferase (DAPAT) (EC2.6.1. 62), 2,2-dialkylglycine decarboxylase (DGD) (EC4.1.1.64), and glutamate-1-semialdehyde aminomutase (GSAT; EC5.4. 3.8) (Table S3).

Both NylD1 and NylD2 are phylogenetically related to GABA aminotransferase. NylD1 had 95.0% overall homology to GABA aminotransferase (GabT) from Arthrobacter aurescens TC1 (Ara-GABT), which degrades s-triazine-like pesticide compounds (Strong et al. 2002; Fig. 2). In addition, GABA aminotransferase from Rhodococcus (Rho-GABT), Mycobacterium (Myc-GABT), Streptomyces griseoruber (Str-GABT), and E. coli (Eco-GABT) had 73.0, 62.5, 54.6, and 42.8% homology to NylD1, respectively. In contrast, NylD2 had a relatively large sequence disparity from NylD1 (49.8% homology), although NylD1 and NylD2 are isozymes from the same bacterial strain. Ornithine amino- transferase from E. coli (Eco-OAT) had a different substrate specificity and a much greater sequence disparity in the phylogenic relationship (34.9% homology to NylD1). DAPAT, 7,8-diaminopelargonic acid aminotransferase from E. coli; Bur- DGD, 2,2-dialkylglycine decarboxylase from Burkholderia cepacia; Pig- GABT, GABA aminotransferase from pig; Mes-GSAT, glutamate-1- semialdehyde aminomutase from Mesorhizobium sp. NylD1 and NylD2 are identified on different loci: NylD1 on contig 23 and NylD2 on contig22. A detailed description of enzymes, protein ID, and the references used for the phylogenic analysis are shown in Table S3. b The amino acid sequence of NylD1 was aligned with the sequences of Ara-GABT and Eco-GABT. The Lys residue responsible for PLP binding (marked by a blue arrow) was conserved among the three enzymes. Amino acid residues are shown as one-letter code The aldehyde dehydrogenase (ALDH) superfamily comprises diverse protein families distributed among various eukaryotic and prokaryotic organisms (Perozich et al. 1999).

Approximately 20,000 genes for the predicted enzymes have been phylogenetically classified into at least 13 families in the ALDH superfamily (Sophos and Vasiliou 2003). The phylogenic analyses suggest that NylE2, NylE4, NylE5, and NylE8 should be classified in the succinate semialdehyde dehydrogenase (SSALDH) family, which has been identified in various organisms (Fig. 3, Table S4). In contrast, NylE1 had the highest homology (63.5%) with adipate semialdehyde (6- oxohexanoate) dehydrogenase (the chnE gene from Rhodococcus, Rho-ChnE) (Fig. S2; Iwaki et al. 1999). We suggest that NylE1 should be placed with NylE3, NylE6, and ChnE at a new branch designated the “adipate semialdehyde dehydrogenase (ASALDH) subfamily” distinguished from the other subfamilies (Fig. 3). As described below, functional analysis of the purified enzyme demonstrates that NylE1 is specific for Ahx metabolism rather than GABA metabolism. The remaining aldehyde dehydrogenases, NylE12 and NylE13, should be placed in the γ-aminobutyraldehyde de- hydrogenase (ABALDH) subfamily, NylE17 and NylE18 in the α-ketoglutarate semialdehyde dehydrogenase (AKGSALDH) subfamily, NylE7 and NylE9 in the 5- carboxymethyl-2-hydroxymuconate semialdehyde dehydro- genase (CHMSALDH) subfamily, NylE11 in the betaine al- dehyde dehydrogenase (BALDH) subfamily, NylE14 and NylE15 in the methylmalonyl semialdehyde dehydrogenase (MMSALDH) subfamily, NylE20 in the Tugor ALDH sub- family, and NylE19 in the γ-glutamyl semialdehyde (GGSALDH) subfamily. NylE10 should be classified in a “Group X” subfamily as suggested by Taniyama et al. (2012). Regarding NylE1–NylE20, no enzymes classified in the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) subfamily were found.To quantitatively compare the catalytic function of the various NylD and NylE enzymes, efficient gene expression followed by the purification of the gene product and construction of an accurate enzyme assay system to analyze the catalytic activi- ties are required. Further studies were focused on the enzymes NylD1 and NylE1, and we evaluated their catalytic functions to determine the physiological roles of these enzymes.We amplified the DNA fragments containing the nylD1 gene by PCR using strain KI72 genomic DNA as a tem- plate and synthetic oligonucleotide primers (Table S2). We expressed NylD1 as His-tagged proteins using the pColdI expression vector (Fig. 4a). Similarly, we cloned and expressed NylE1 as the His-tagged enzyme (Fig. 4b). Nucleotide sequencing of the DNA fragments cloned in the pCold-NylD1 or pCold-NylE1 vectors revealed that no mutations were integrated during the PCR and cloning processes. We purified the His-tagged enzymes from the cell extracts of the E. coli clones.

The purified enzymes produced a single protein band on SDS-polyacrylamide gel electrophoresis corresponding to molecular weights of approximately 47 kDa (NylD1) and 49 kDa (NylE1) that corresponded well to the theoretical values expected from the nucleotide sequences (Fig. 4c).NylD1 We performed the enzyme reaction using an amino donor (Ahx or GABA) and amino acceptor (α-KG) to con- firm the catalytic function of the enzyme. We identified the spot corresponding to L-glutamate in the reaction mixture by TLC (Fig. 5a). Although the intensity of the spot was weak, the presence of the expected reaction product demon- strates that NylD1 has aminotransferase activity either for Ahx or GABA. In a reaction using glyoxylate as the amino acceptor, the conversion of Ahx to glycine was approxi- mately 50% after a 3-h reaction, as indicated by the intensity of the spots identified by ninhydrin, which were found to be nearly at a similar level (Fig. 5e). Thus, the increase in the reaction apparently stopped due to the reverse reaction of NylD1. However, as stated below, we found that the NylD1 reaction coupled with NylE1 significantly enhanced the con- version (Fig. 5b–d).plasmids expressing NylD1 andNylE1. a, b DNA fragments coding for nylD1 (a) or nylE1 (b) genes were amplified by PCR using the primers described in Table S2. The amplified mixtures were sampled at 30-min intervals (up to 180 min), and the reaction products were analyzed by TLC. As a control experiment, transamination from Ahx to glyoxylate was investigated in the absence of NylE1 and NADP+ (e). Authentic Ahx, GABA, L-Glu, Gly, and L-Ala were spotted on TLC plates. f To quantify the rate of NADPH formation, the aliquots were suitably diluted, and the A340 was measured. α-KG (red circles), glyoxylate (blue triangles), pyruvate (green squares). Time course of the reaction using glyoxylate in the absence of NylE1 and NADP+ was shown as orange circles. The initial reaction rate was estimated from the dashed line for each substrate (Table 2). NylE1 We assayed the dehydrogenase activity of NylE1 using chemically synthesized ASA or SSA as the substrate and NADP+ as a coenzyme (Fig. 6). We estimated the initial reac- tion rate by monitoring the absorbance of NADPH at 340 nm (A340). After the reaction was started, NADPH was formed from either ASA or SSA.

However, the specific activity of ASA (18 μmol/min (U)/mg-protein) is approximately 500- fold of the activity of SSA (0.034 U/mg-protein). We con- firmed that the initial reaction rate is proportional to the en- zyme concentration up to 0.0016 mg/ml (for ASA) and 0.3 mg/ml (for SSA) (Fig. 6c, d). These results demonstrate that NylE1 has a much higher specific activity with the sub- strate responsible for Ahx/adipate metabolism than for GABA/succinate metabolism. However, it should be noted that the reaction apparently stops when approximately 20% of the NADP+ is converted to NADPH (Fig. 6a). We estimate that the net concentrations of ASA or SSA involved in the reaction seemed to be decreased by spontaneous oxidation and/or polymerization reactions in aqueous solution due to the instability of the semialdehyde. However, as stated below, the coupled reaction using NylD1/NylE1 almost completely converts the Ahx to adipate. Therefore, we have concluded that ASA formed by the reaction with NylD1 is successively converted to adipate by the reaction with NylE1.To test whether NylE1 catalyzes the reductive reaction from adipate to ASA, we performed the reactions using purified NylE1 (0.05 mg/ml: 31-fold concentrated enzyme used for the oxidative reaction), 0.2 mM adipate, and 0.25 mM NADPH in KPB buffer (pH 7.8). However, we found no detectable con- version of NADPH to NADP+ by A340 even after the reaction ran for 30 min (Fig. 6e). On the basis of these findings, we have concluded that NylE1 is highly specific for the oxidative activ- ity of ASA and uses NADP+ as a cofactor.Construction of NylD1/NylE1 coupled reaction systemTo establish a quantitative assay system for Ahx aminotrans- ferase activity, we investigated the applicability of a NylD1/ NylE1 coupled system using the purified enzymes (NylD1, NylE1), substrates (amino-donor and amino-acceptor), and coenzymes (NADP+ and PLP) (Fig. 7a). As stated below, we demonstrate that the reaction system has great advantages for monitoring the aminotransferase activity.We confirmed that NylD1 has aminotransferase activity either for Ahx or GABA using α-KG as an amino acceptor (Fig. 5a).

Especially in the coupled reaction where the reaction components in b) was continued up to 150 min, and the time course of the formation of NADPH is shown. d–f Time course of NylD1/NylE1 coupled reaction system at different enzyme concentrations. The coupled reaction was performed by changing the ratio of NylD1 and NylE1 activity in the standard assay conditions. Enzyme activity was assayed using α-KG (d), glyoxylate (e), or pyruvate (f) as amino acceptor. NylD1/NylE1 = 0.05/0.1 mg/ml (1); 0.05/0.05 mg/ml (2); 0.025/0.05 mg/ml (3); 0.0125/0.05 mg/ml (4); 0/0.05 mg/ml (5). g Aminotransferase activity was calculated by subtracting the background level ([NylD1] = 0, line 5) from the reaction rate in the complete system. The activity was plotted for the protein concentrations of NylD1. α-KG (red circles); glyoxylate (blue triangles); pyruvate (green squares) rate of NylE1 is much greater than that of NylD1, Ahx should be immediately converted to adipate without the accumulation of ASA. Therefore, the system makes it possible to monitor the activity continuously via the A340. Since the high specific activity of NylE1 for ASA is suitable for quantification of the Ahx aminotransferase activity, we analyzed the activity of the coupled system. The reaction proceeds at constant rate up to 55% conversion (approximately 0.11 mM NADPH at 60 min) in a reaction mixture containing all of the essential compo- nents (Fig. 7c). Moreover, even if the reaction is started by using a two-fold greater amount of NylE1 enzyme, the reac- tion rate was barely affected (Fig. 7d, curves 1 and 2). For the reaction using NylD1/NylE1 = 0.05/0.05 mg/ml (Fig. 7d curve 2), the NylE1 activity for the synthetic ASA is calculated to be0.9 U/ml from a specific activity of 18 U/mg (Fig. 6). However, the observed activity in NylD1/NylE1 coupled reac- tion is approximately 0.0033 U/ml (specific activity =0.066 U/mg, Tables 1 and 2), indicating that the rate-limiting Table 2 6-Aminohexanoate aminotransferase activity by NylD1/NylE1 coupled system. Specificity of amino acceptor Amino acceptor Specific activity (μmol/min/mg) step in the sequential reaction is the NylD1 reaction.

Therefore, we have concluded that aminotransferase reaction occurs in the presence of a great excess of the semialdehyde dehydrogenase (NylE) activity.In a reaction system where either NylD1, NylE1, Ahx, α- KG, or NADP+ was absent, the activity drastically decreased to a background level (approximately 1.2–3.8% of the activity of the complete system; Table 1, Fig. 7b). However, even in a reaction mixture lacking PLP, approximately 26% of the ac- tivity is retained. The implication of this result is discussed below. Additionally, we found that DTT or EDTA has no detectable effect on the reaction rate (Table 1), although some bacterial semialdehyde dehydrogenases are stabilized and/or activated by DTT (Sanchez et al. 1989) but inhibited by the metal chelating agent EDTA (de Carvalho et al. 2011).To confirm the acceptor specificity in the aminotransferase reaction, we tested the activity of NylD1 using pyruvate or glyoxylate as the amino acceptor instead of α-KG (Fig. 7e, f). We found that NylD1 utilizes αKG and glyoxylate as the amino acceptor at almost similar levels, whereas the activity of pyruvate was found to be 30% (assay 1) to 70% (assay 2) of the level of αKG (Table 2). A TLC analysis of the reaction mixtures revealed that the amino acids glutamate, alanine, andEnzyme reactions were performed under the standard assay conditions for a NylD/NylE coupled reaction system (0.2 mM Ahx, 0.25 mM α-KG,0.1 mM PLP, and 0.25 mM NADP+ in 100 mM KPB, pH 7.8) containing 1 mM DTT and 1 mM EDTA as a common component. Details are de- scribed in the “Materials and methods” section. In Tables 1 and 2, enzyme samples obtained by independent purification experiments were used Enzyme activities were quantified by two assays, which differed in the concentrations of amino donor (Ahx), amino acceptor (α-KG, pyruvate, glyoxylate), and NADP+ . Assay 1, 0.2 mM Ahx; 0.25 mM of the amino acceptor; 0.25 mM NADP+ . Assay 2, 2 mM Ahx; 2.5 mM of the amino acceptor; 2.5 mM NADP+ . In Tables 1 and 2, enzyme samples obtained by independent purification experiments were usedglycine are produced from Ahx and oxo-acids during the course of the reaction (Fig. 5). With αKG or glyoxylate as the amino acceptor, approximately 50% of the conversion was achieved after the reaction runs for 60 min, and Ahx is almost completely converted to glutamate and glycine after the reaction runs for 180 min (Fig. 5b, c). In a control exper- iment using glyoxylate and NylD1 without NylE1, approxi- mately 50% of the conversion was achieved after 180 min (Fig. 5e). Thus, NylD1 utilizes the three oxo-acids as amino acceptors, although the activity was found to be in the order of glyoxylate, α-KG, and pyruvate (Table 2). We have conclud- ed from these findings that the lower level of conversion caused by the reversible reaction of NylD1 is enhanced by coupling it to the NylE1 reaction.

Discussion
We have confirmed that 6-nylon monomer, 6-aminohexanoate (Ahx), is converted to adipate by sequential reactions of Ahx aminotransferase (NylD1) and ASA dehydrogenase (NylE1) in Arthrobacter sp. strain KI72. An enzyme assay using a NylD1/NylE1 coupled system revealed that NylD1 requires PLP as a cofactor for the aminotransferase activity. Generally, PLP-dependent enzymes are involved in the bio- synthesis of amino acids and amino acid-derived metabolites, in which PLP acts as a coenzyme for the transamination, de- amination, racemization, and decarboxylation reactions (Eliot and Kirsch 2004; Percudani and Peracchi 2003; Hwang et al. 2005; Phillips 2015; Steffen-Munsberg et al. 2015; Schiroli and Peracchi 2015).GABA transferases from pig (Pig-GABT) (Storici et al. 2004), E. coli (Eco-GABT) (Liu et al. 2004), and Arthrobacter aurescens (Ara-GABT) (Bruce et al. 2012) have been structurally and biochemically characterized. The ε- amino group of the active site Lys generally forms a Schiff base with the aldehydic carbon of PLP (Eliot and Kirsch 2004) at the catalytic center of the PLP-dependent enzymes. Since approximately 26% of the activity is retained if the PLP is left out of the reaction mix (Table 1), we estimate that a portion of PLP is bound to the enzyme even after purification has been carried out. An X-ray crystallographic analysis of Ara-GABT aminotransferase complexed with PLP has revealed that the ε- amino group of Lys295 formed a Schiff base (Bruce et al. 2012). Lys268 in the E. coli enzyme (Eco-GABT) is identified to be connected to PLP (Liu et al. 2004). From the alignment of the amino acid sequences, the active site Lys295 in Ara- GABT and Lys268 in Eco-GABT is conserved as Lys295 in NylD1 (Fig. 2b). Thus, the conserved PLP-binding site and presence of Ahx- and GABA-aminotransferase activity of NylD1 demonstrate that the unnatural amino acid Ahx is me- tabolized as an analog of physiological substrate GABA via a similar catalytic mechanism as reported for most PLP- dependent aminotransferases.Of the 20 putative aldehyde dehydrogenases (NylE1– NylE20) identified in Arthrobacter strain KI72, NylE1 had the highest homology (63.5%) with adipate semialdehyde (6-oxohexanoate) dehydrogenase (i.e., the chnE gene from Rhodococcus) (Fig. 3; Iwaki et al. 1999). Actually, since the specific activity of NylE1 for ASA (18 U/mg) is approximate- ly 500-fold of the activity of SSA (0.034 U/mg) (Fig. 6), we think that NylE1 is specific for Ahx/adipate metabolism rather than GABA/succinate metabolism, as stated above.

The gabT (GABA amino transferase) and gabD (succinate semialdehyde dehydrogenase) genes responsible for GABA metabolism constitutes an operon in the E. coli chromosome (Bartsch et al. 1990). In contrast, the nylD1 and nylE1 genes are located on the contig sequences no. 23 and 1, respectively. This result indicates that nylD1 and nylE1 genes are not linked in the genomic DNA of strain KI72 (Figs. 2 and 3). It should be noted that nylD2, and most of the nylE2–nylE20 genes are distributed at various loci on the chromosomal DNA of strain KI72. In contrast, the nylBC genes responsible for the hydro- lysis of Ahx oligomers and nylon-6 are located on a plasmid in Arthrobacter, but these genes are chromosomal in Agromyces and Kocuria (Yasuhira et al. 2007a). Moreover, we have sug- gested that at least two genetic recombinations, i.e., the IS6100-mediated transposition of nylB/nylC gene region and a recombination that generates hybrid gene structures for the analogous nylB’/nylC’ regions, are involved in the genetic organization in the alkalophilic nylon-oligomer degradative bacterium Agromyces (Yasuhira et al. 2007b). In addition, a plasmid-encoded NylA from Arthrobacter and Pseudomonas and ω-laurolactam hydrolase from Cupriavidus , Rhodococcus, and Sphingomonas are found to have 98–99% overall homology, although these strains have been classified into different genera (Yasuhira et al. 2010). These results may imply that bacterial strains have evolved a novel metabolic pathway for nylon-6-related compounds by assembling the responsible genes into the parental genomic DNA.We have found that 6-aminohexanoate oligomers are al- most completely converted to the monomers (Ahx) by the three nylon oligomer-degrading enzymes, NylA, NylB, and NylC (Negoro 2000, Negoro et al. 2005, 2007, 2012; Nagai et al. 2014). In addition, the present study has revealed that the yield for the conversion of Ahx to adipate is estimated to be > 90% on the basis of the stoichiometry of the overall reaction in the coupled system (Figs. 5 and 7). On the basis of these findings, we demonstrate that bio-based conversion from wastes/byproducts of nylon-6 to adipate (nylon-66 monomer) is possible, when the NADPH generated by the NylE reaction is re-oxidized to NADP+ by coupling a suitable oxidoreductase reaction and the enzyme activity is high enough at a Aminocaproic practical level.