Short-Chain Dehydrogenase NcmD Is Responsible for the C-10 Oxidation of Nocamycin F in Nocamycin Biosynthesis

Nocamycins I and II, featured with a tetramic acid scaffold, were isolated from the broth of Saccharothrix syringae NRRL B-16468. The biosynthesis of nocamycin I require an intermediate bearing a hydroxyl group at the C-10 position. A short chain dehydrogenase/reductase NcmD was proposed to catalyze the conversion of the hydroxyl group to ketone at the C-10 position. By using the λ-RED recombination technology, we generated the NcmD deletion mutant strain S. syringae MoS-1005, which produced a new intermediate nocamycin F with a hydroxyl group at C-10 position. We then overexpressed NcmD in Escherichia coli BL21 (DE3), purified the His6-tagged protein NcmD to homogeneity and conducted in vitro enzymatic assays. NcmD showed preference to the cofactor NAD+, and it effectively catalyzed the conversion from nocamyin F to nocamycin G, harboring a ketone group at C-10 position. However, NcmD showed no catalytic activity toward nocamyin II. NcmD achieved maximum catalytic activity at 45°C and pH 8.5. The kinetics of NcmD toward nocamycin F was investigated at 45°C, pH 8.5 in the presence of 2 mM NAD+. The Km and kcat values were 131 ± 13 μM and 65 ± 5 min−1, respectively. In this study, we have characterized NcmD as a dehydrogenase, which is involved in forming the ketone group at the C-10 position of nocamycin F. The results provide new insights to the nocamycin biosynthetic pathway.

The gene cluster responsible for nocamycins biosynthesis has been identified from S. syringae NRRL B-16468 in 2017 (Mo et al., 2017b). The skeleton of nocamycins is assembled by a hybrid type I polyketide synthases (PKS) and non-ribosomal peptide synthetase (NRPS) system (Mo et al., 2017b). The gene cluster for nocamycins consists of 21 open reading frames (ORFs), which includes five genes encoding for type I PKS, one gene encoding for NRPS, one gene encoding for a Dieckmann cyclase NcmC, one gene encoding for a SDR NcmD, two genes encoding for cytochrome P450 oxidases NcmO and NcmG, one gene encoding for a glycoside dehydratase NcmE, one gene encoding for a SAM dependent methyltransferase NcmP, and five genes encoding for regulatory proteins (Mo et al., 2017b). Among the 21 ORFs, the functions of several genes have been investigated. The results of gene disruption experiment indicated that the cytochrome P450 oxidase NcmG was involved in formation of furan ring (Mo et al., 2017b). Furthermore, gene disruption and biochemical assays clearly demonstrated that the SAM-dependent methyltransferase NcmP was responsible for formation of the carboxylic methyl ester (Mo et al., 2017a).
The disparity between nocamycins I and II is the ketone or hydroxyl group at C-10 position, and the similar structure has been observed in tirandamycin B and its biosynthetic intermediate tirandamycin E (Figure 1). For tirandamycins, a FAD dependent dehydrogenase TrdL/TamL has been shown to catalyze the formation of ketone group at C-10 position (Carlson et al., 2011;Mo et al., 2011). However, for nocamycins, the enzyme for formation of ketone group at C-10 position remains unclear. We initially inactivated the gene ncmL encoding for a FAD-dependent protein, and resultant research demonstrated that it exerted no impact on production of nocamycins (Mo et al., 2017b). By carefully examining the biosynthetic gene cluster, the SDR NcmD is proposed to be the candidate to catalyze the conversion from hydroxyl to ketone at C-10 position. In the current study, we have established that NcmD acts as a dehydrogenase and catalyzes the formation of C-10 ketone group in nocamycin biosynthetic pathway by using in vivo gene disruption and in vitro biochemical assays.

Bacteria, Medium, and Culture Conditions
Saccharothrix syringae NRRL B-16468 was used as producer of nocamycins (Mo et al., 2017b). Escherichia coli DH5α was used as host for general clone. Escherichia coli BL21(DE3) was used as host for protein expression. Saccharothrix syringae NRRL B-16468 and its derivative strain were maintained on ISP4 agar medium. The medium contained 1% soybean flour, 3% glycerol, 0.2% NaCl, and 0.2% CaCO 3 , pH 7.0 was used for fermentation of S. syringae and its derivative strain. All cultures for S. syringae were incubated at 28°C. The E. coli strains were cultured by using Luria-Bertani (LB) agar or liquid medium with appropriate antibiotics.

Fermentation and Analysis of Mutant Strain Saccharothrix syringae MoS-1005
S. syringae wild type and mutant strain S. syringae MoS-1005 were cultured by using the method described previously (Mo et al., 2017b). After 7 days culture, the broth was extracted by ethyl acetate, subsequently, the organic phase was collected and evaporated into dryness, re-dissolved in methanol, and subject to high performance liquid chromatography (HPLC) analysis. Analytical HPLC was performed on Waters 2699 HPLC system (Waters Technologies Inc., United States) equipped with a PDA detector and a Welch Ultimate AQ-C18 ODS column (250 × 4.60 mm, 5 μm). The mobile phase contained solvent A and B. Solvent A consisted of 15% CH 3 CN in water supplemented with 0.1% formic acid. Solvent B consisted of 85% CH 3 CN in water supplemented with 0.1% formic acid. Samples were eluted with a linear gradient from 5 to 90% solvent B in 20 min, followed by 90 to 100% solvent B for 5 min, then 100% solvent B for 3 min, at a flow rate of 1 ml/ min under UV detection at 355 nm.

MoS-1005
For fermentation of S. syringae MoS-1005 in a large scale, 8 L liquid media were used by using a two-step fermentation procedure as described previously (Mo et al., 2017b). After incubation, the culture broths were collected and centrifuged. The supernatant broth was extracted by ethyl acetate for three times and the mycelia were extracted by methanol for three times. Then, the entire organic solvents were evaporated into dryness to yield crude extracts. The crude extracts were dissolved in a mixture of CH 2 Cl 2 :CH 3 OH (1:1), and then mixed with appropriate amount of silica gel (100-200 mesh, Qingdao Marine Chemical Corporation, China). The samples were applied on normal phase silica gel chromatography column and eluted with CH 2 Cl 2 :CH 3 OH (100:0-50:50) to give 10 fractions, and all of them were analyzed by HPLC. Fractions 4 and 5 containing the target compound were used for further purification on reverse phase C-18 silica gel (YMC, Japan) by using medium-pressure liquid chromatography (MPLC, Agela corporation, China). The fractions containing the target compound were combined and further purified by Sephadex LH-20 (GE healthcare, Sweden) gel filtration chromatography to afford the purified nocamycin F.

Heterologous Production and Purification of Recombinant Protein NcmD
The NcmD gene was amplified from cosmid p5-C-9 by using PCR with primers NcmD-expF (5'-TAATAATCATATGCGCGAG TTGACCGACCG-3', underline is NdeI site) and NcmD-expR (5'-ATATGGATCCCTAGTCGGTCGGCGCGGCG-3', underline is BamHI site). The resultant PCR products were digested by NdeI and BamHI, and then inserted into pET-28a (+) vector digested with the same restriction enzymes to yield plasmid pMosD. After verification of the inserted gene fragment by sequencing, the plasmid pMosD was introduced into E. coli BL21(DE3) for protein expression.
Escherichia coli BL21 (DE3) strain carrying plasmid pMosD was grown in LB medium with 50 μg/ml kanamycin at 37°C to an OD 600 = 0.6. Then, isopropyl-β-D-thiogalactopyranoside (IPTG) was added into the culture at a final concentration of 0.1 mM to induce the expression of NcmD for 18 h at 20°C. Subsequently, the cells were collected, centrifuged, and re-suspended in binding buffer (50 mM Tris-Cl buffer, 500 mM NaCl, and 10 mM imidazole, pH 7.9), and sonicated on ice. Cellular debris was removed by centrifugation (12,000 rpm, 30 min, 4°C). The supernatant was further purified by nickel-nitrilotriacetic acid (Ni-NTA) affinity chromatography according to the manufacturer's protocol (Novagen, CA, United States). The purified protein was desalted by PD-10 column (GE Healthcare, United States) according to the manufacturer's instructions. The purified protein NcmD was finally stored in 50 mM Tris-Cl buffer (pH 8.0) with 10% glycerol at −80°C for further enzymatic assays.
by adding substrate nocamycin F. At 2, 4, and 6-min, 30 μl reaction mixtures were taken, and then 60 μl ice-cold methanol was added and rigorously mixed by vortex. After centrifuge, 30 μl liquids were used for further HPLC analysis and quantified by a standard curve. The reaction rates were calculated and confirmed to be linear. The kinetics data were fitted to the Michaelis-Menten equation using Origin8.0 software. For each concentration of substrate, three replicates were conducted. For elucidating the structure of the products obtained from the reactions with nocamycin F, 5 ml reaction mixtures containing 2 μM NcmD, 2 mM NAD + and 6 mg nocamycin F were performed at pH 8.5, 40°C for 8 h. Then, the reaction mixtures were extracted by 10 ml ethyl acetate for three times. The target product was purified by preparative HPLC for NMR and high-resolution mass spectral (HR-MS) analysis.
Spectroscopy Analysis of New Produced Nocamycin Derivatives 1 H and 13 C NMR spectra were recorded at 25°C on Bruker AV 500 instruments. LC-HR-MS data were acquired on a Waters micro MS Q-Tof spectrometer or a Thermo MAT95XP high resolution mass spectrometer.

Bioinformatics Analyses of NcmD
Within nocamycin biosynthetic gene cluster, the gene NcmD encoding for the SDR located next to the NRPS gene NcmB. NcmD shows identity to a series of SDRs, including 40% identity to BatM originated from kalimantacin/batumin-related polyketide antibiotic biosynthetic pathway (Mattheus et al., 2010), and 31% identity to clavulanic acid dehydrogenase (CAD) involved in clavulanic acid biosynthetic pathway (MacKenzie et al., 2007). Bioinformatics analyses revealed that NcmD shared the conserved motifs of classical SDR, namely, Rossmann fold NAD(P)-binding motif T 12 G 13 (X) 3 G 17 XG 19 , conserved catalytic triad S 142 (X) 12 Y 155 ( X) 3 K 159 , N 89 NAG 92 motif and P 187 G (X) 3 T 192 motif (Figure 2; Filling et al., 2002;Persson and Kallberg, 2013;Gräff et al., 2019). In addition, an Asn 115 residue (N) frequently served as an additional active site residue to make a catalytic triad is also conserved in NcmD (Figure 2). These bioinformatics analyses have demonstrated that NcmD is a classical SDR. As shown in Figure 2, an acidic residue aspartate (Asp, D) other than a basic residue occupies the position 37, indicating that the cofactor for NcmD is likely to be NAD + , which has been confirmed by other NAD + -preference SDRs such as FgaDH, PDH, 1YDE, and TQS88917 (Lukacik et al., 2007;Wallwey et al., 2010;Hofmann et al., 2016;Shanati and Ansorge-Schumacher, 2020).

Construction, Analysis of NcmD Mutant Strain Saccharothrix syringae MoS-1005 and Isolation of the New Nocamycin Analog
To investigate the exact role played by NcmD in nocamycin biosynthetic pathway, we firstly inactivated NcmD by replacing partial internal NcmD with aac(3)IV gene cassette through λ-RED recombination technology and generate ΔNcmD mutant strain ( Figure 3A). After verifying the genotype of S. syringae MoS-1005 (Figure 3B), the mutant strain was cultured, and the broth was extracted for HPLC analysis. The results of HPLC revealed that three peaks with retention time of 15.1, 16.9, and 19.8 min showed UV absorption characteristics of nocamycin were detected in S. syringae MoS-1005 ( Figure 3C). The peaks at 15.1 and 19.8 min showed the same retention time to nocamycin II and nocamycin I, respectively. Subsequently, the molecular mass of these three peaks were detected by LC-HR-MS.  To elucidate the structure of the new nocamycin analog produced by S. syringae MoS-1005, we purified nocamycin F from 8 L culture broth by using chemical isolation methods. The structure of nocamycin F was determined by multiple spectroscopic data analyses (Figure 4). As revealed from HR-MS (Supplementary Figure S1-C), the molecular formula of nocamycin F was determined as C 25 H 35 NO 7 , with an oxygen atom more than that of nocamycin III (Mo et al., 2017b), suggesting that nocamycin F should be a hydroxylation congener of nocamycin III. The one-dimensional NMR data exhibited marked similarities to those of nocamycin III, except that the signal of a methylene group (CH 2 -10, δ C 23.9) in nocamycin III was absent from the NMR spectra of nocamycin F (Table 1). Instead, another oxygenated CH group (CHOH-10, δ C 68.8/δ H 4.59) could be observed in the spectra of nocamycin F, also as established by the COSY and HMBC data (Supplementary Figure S2).

In vitro Characterization of NcmD in Nocamycin Biosynthetic Pathway
For in vitro characterization of NcmD activity, the protein NcmD was produced as a N-terminal His 6 -tagged protein in heterologous host E. coli BL21(DE3) carrying the plasmid pMoS-1005. The His 6 -tagged protein NcmD was purified to homogeneity by using Ni-NTA affinity chromatography, and SDS-PAGE analysis displayed that it had an expected molecular weight (the calculated molecular mass for NcmD is 30.078 kDa; Figure 5A). Subsequently, the catalytic properties of NcmD were investigated. At first, we detected the preference of NcmD toward cofactors NAD + and NADP + . In the presence of NAD + , NcmD can efficiently catalyze substrate nocamycin F to a new compound with less polarity (Figure 5B). Whereas, in the reaction mixture with NADP + instead of NAD + , only about 25% nocamycin F was transformed to a new compound. These results demonstrated that NcmD preferred NAD + as the cofactor, which was consistent with our initial bioinformatics analyses.
To determine the structure of the new product converted from nocamycin F, we conducted a large-scale reaction of NcmD toward nocamycin F, leading to purification of the new  Figure S3], two hydrogen less than that of nocamycin F. The amount of 6 obtained here was not enough for 2D NMR analysis; however, the 13 C NMR spectra still give us sufficient structural information of 6. By comparison of 13 C NMR data of 6 and nocamycin F, 6 showed one more carbonyl carbon (C=O, δ C 208.8) and three oxygenated carbons (δ C 77.8, 79.1, and 81.4), indicating that one hydrogen group in nocamycin F should be oxidized to the carbonyl group (Supplementary Figure S4). From the biosynthesis view, the difference between nocamycin F and compound 6 should be at C-10 position. For these reasons, we proposed the structure of 6 as shown in Figure 4, and it was termed as nocamycin G. Thus, NcmD was proposed to act as a dehydrogenase to catalyze the formation of ketone moiety at C-10 position.
Considering the same hydroxyl moiety harbored by nocamycin II at C-10 position, we then investigated whether NcmD can also accept nocamycin II as substrate and catalyze the conversion from nocamycin II to nocamycin I. Unfortunately, no conversion from nocamycin II to nocamycin I was observed ( Figure 5B). This result demonstrated that nocamycin II was not the substrate of NcmD.
The effect of pH and temperature on NcmD properties was also investigated. For pH in the range from 6.0 to 11.0, NcmD was found to achieve maximum catalytic activity at pH 8.5, and in the range of pH 7.5-10.0, NcmD can retain activity in a high level, indicating NcmD was tolerant to pH values ( Figure 5C). As for temperature, NcmD showed robust activity in the range of 37-50°C, and the optimum temperature for NcmD was 45°C ( Figure 5C). Finally, we measured the steadystate kinetic parameters of NcmD toward nocamcyin F under 45°C, pH 8.5. In the presence of 2 mM NAD + , the K m and k cat values of NcmD were 131 ± 13 μM and 65 ± 5 min −1 , respectively ( Figure 5D).

DISCUSSION
Nocamycins belong to a small family of tetramic acid compounds bearing bicyclic ketone structure. Nocamycins demonstrate excellent antibacterial activity, especially against some anaerobic bacteria Tsunakawa et al., 1980;Bansal et al., 1982). Recently, we have identified the gene cluster responsible for nocamycin biosynthesis from a rare actinomycete S. syringae (Mo et al., 2017b). Through manipulating the gene cluster, we have generated several nocamycin analogs and characterized several gene functions involved in nocamycin biosynthetic pathway (Mo et al., 2017a,b). In this study, through in vivo gene inactivation and in vitro enzymatic assays, the SDR NcmD has been assigned to be involved in formation of ketone group at C-10 position, leading to generate nocamycin G from nocamycin F. Short-chain dehydrogenases/reductases have been classified into seven families and the classical type is the most prominent. NcmD shows the conserved motifs belonging to the classical SDR subfamily. The majority (about 60%) of classical SDRs are expected to prefer NADP(H) (Gräff et al., 2019). For example, the SDR CAD has been shown to use NADPH as cofactor to catalyze the conversion from clavaldehyde to clavulanic acid (MacKenzie et al., 2007). The SDR Cro013448 (KP411011.1) from Catharanthus roseus recruits NADP(H) as cofactor in the biosynthetic pathway of plant monoterpene indole alkaloid vitrosamine (Stavrinides et al., 2018). However, the NAD + dependent SDRs have also been found. FgaDH involved in ergot alkaloid fumigaclavine C biosynthetic pathway originating from Aspergillus fumigatus employs NAD + as cofactor to catalyze the conversion from chanoclavine-I to chanoclavine-I aldehyde (Wallwey et al., 2010). Pseudoephedrine dehydrogenase (TQS88917) is a NAD + dependent SDR and it catalyzes the oxidation of converted (S, S)-(+)-pseudoephedrine and (S, R)-(+)-ephedrine to (S)-and (R)-methcathinone (Shanati and Ansorge-Schumacher, 2020). For classical SDRs, the aspartic acid residue at standard position 37 has been described as a determinant of NAD(H) specificity (Belyaeva et al., 2015;Gräff et al., 2019). The current results confirm that Asp37-containing NcmD prefers NAD + (Kallberg et al., 2002), which has been confirmed by resultant in vitro enzymatic assays. Generally, the catalytic activities decrease significantly when the SDRs are not compatible with cofactors, which have been observed from NcmD enzymatic assays, and similar results are also revealed from other SDRs (Moon et al., 2012;Takase et al., 2014;Cao et al., 2019;Gmelch et al., 2020).
Structurally, nocamycins show high similarity to tirandamycins. Both tirandamycin B and nocamycin I harbor a ketone moiety at C-10 position. However, the proteins involved in synthesizing this moiety are significantly different. For tirandamycin B, a FAD dependent dehydrogenase TrdL catalyzes the conversion from hydroxyl moiety to ketone group at C-10 position, meanwhile, TrdL displays a flexible substrate spectrum (Mo et al., 2011). Whereas for nocamycin, the SDR NcmD shows substrate selectivity. Though NcmD was inactivated in the strain S. syringae MoS-1005, we still detected the accumulation of nocamycin I and nocamycin II, two major metabolites produced by the wild type S. syringae. Nocamyin II is a major metabolite in S. syringae wild type, which indicates the other tailoring enzymes such as NcmG, NcmO, and NcmP show flexible substrate selectivity, thus, it is rational that nocamycin II can be detected in S. syringae MoS-1005. For production of nocamycin I in S. syringae MoS1005, two possible pathways are proposed (Figure 6). Firstly, an unknown enzyme located in the genome elsewhere can compensate for the function of NcmD and catalyze the conversion from nocamycin F to nocamycin G, which then undergoes several tailoring steps to generate nocamycin I. Secondly, an unknown enzyme can catalyze the transformation from nocamycin II to nocamycin I. No matter which strategy has been employed, nocamycin I can be produced in S. syringae MoS1005. Taking these results together, we envision a plausible tailoring process from the formation of the intermediate nocamycin F, which is shown as Figure 6.
In summary, we generated NcmD deletion mutant strain S. syringae MoS-1005 and identified an important intermediate nocamycin F from this mutant strain. In vitro enzymatic assays have demonstrated that the NAD + dependent SDR NcmD acts as a dehydrogenase and it is involved in formation of ketone moiety at C-10 position. However, NcmD shows substrate preference and it only displays catalytic activity toward nocamycin F. The results presented in this study provide new insights into nocamycin biosynthetic pathway.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
XM and SY designed the experiment. XM, SY, and FD analysis the data and wrote the paper. XM and HZ performed the experiments.