An E. coli codon-optimized, N-terminal hexahistidine-tagged SSEG_08523 gene encoding SsDesB was synthesized and subcloned into pET28a (+) by GenScript Corp. (Piscataway, NJ). N-hydroxycadaverine was purchased from Enamine Ltd. (Kyiv, Ukraine). All other materials were purchased from Sigma-Aldrich (St. Louis, MO) or Thermofisher Scientific (Waltham, MA) and used without further purification.

Microplate kinetic assays were monitored on an Epoch BioTek microplate spectrophotometer (Winooski, VT). The SsDesB extinction coefficient was determined using an Agilent 8453 diode array UV-vis spectrophotometer (Santa Clara, CA). High resolution electrospray ionization mass spectrometry (ESI-MS) was performed on an Acquity Ultra Performance Liquid Chromatography (UPLC) I-Class System in tandem with a Waters Xevo quadrupole-time of flight (q-TOF) mass spectrometer. Oxygen consumption assays were monitored on an Oxygraph Plus system (Hansatech; Norfolk, UK). DNA sequencing was performed by Eurofins Genomics (Louisville, KY).

The SsDesB gene cloned into pET28a (+) was transformed into Escherichia coli BL21 Star (DE3) One ShotⓇ cells (Invitrogen; Carlsbad, CA) according to the manufacturer’s instructions. Cells were plated on Luria-Bertani (LB)-agar containing 50 μg/mL kanamycin and grown overnight at 37°C. A single colony was used to inoculate LB broth (100 mL) containing 50 μg/mL kanamycin, which was grown overnight at 37°C at 200 rpm. The overnight culture was used to inoculate 1-L of LB broth containing 50 μg/mL kanamycin, which was grown at 30°C at 200 rpm. Cells were induced with 1 mM isopropyl β-D-1 thiogalactopyranoside (IPTG) once an optical density at 600 nm (OD₆₀₀) of 0.6 was reached and grown for four additional hours before being harvested at 4,500 rpm and stored at -80°C.

Cells were lysed twice via the French Press at 20,000 psi after incubation with lysozyme (1 mg/ml) and protease inhibitor cocktail (P8849; Sigma-Aldrich, St. Louis, MO) 50 mM phosphate buffer (pH 8) containing 300 mM sodium chloride and 10 mM imidazole. After centrifugation at 9,100 rpm for 1 hr at 4°C, the supernatant was filtered with a 0.22 μM Millex-GS syringe filter (EMD Millipore; Burlington, MA) and incubated with HisPur^TM Ni-NTA resin (Thermofisher; Waltham, MA). The resin was equilibrated with 50 mM phosphate buffer (pH 8) containing 300 mM sodium chloride and 20 mM imidazole. Protein was eluted in 50 mM phosphate buffer (pH 8) containing 300 mM sodium chloride, 125 mM imidazole, and 2.5% glycerol. Using a 10 K Amicon Ultra centrifugal filter unit (Thermofisher Scientific; Waltham, MA), protein was buffer-exchanged into 100 mM phosphate buffer containing 1 mM dithiothreitol and 10% glycerol and stored at -80°C.

SsDesB (156 μM) was heated in triplicate at 95°C for 15 min and clarified by centrifugation at 13,000 rpm for 10 min at room temperature. After the supernatant was removed and reheated again, its absorbance was measured over a range of 200–800 nm. Using Beer’s Law (A = εcl) and the extinction coefficient of FAD (11,300 M^-1cm^-1), the extinction coefficient of DesB was determined, as the concentration of the enzyme should be equivalent to the concentration of free FAD after denaturation. Furthermore, the percent of FAD bound was also determined using Eq (1).

Steady-state kinetic parameters for NAD(P)H in oxidation assays were determined by incubating SsDesB (0.7 μM) with 50 μM FAD, and 10 mM cadaverine at 25°C. Assays were initiated after 5 min with varying concentrations of NAD(P)H (5.5 μM–1 mM), resulting in a total volume of 250 μL. The decrease in absorbance at 340 nm was monitored using an Epoch BioTek microplate spectrophotometer. The initial rate at which the enzyme oxidized NAD(P)H over time was determined using Beer’s Law (ε_NAD(P)H = 6,220 M^-1cm^-1 at 340 nm) and an NAD(P)H standard curve (to determine the quality of cofactor and pathlength of the microplate).

Molecular oxygen consumption consumed by SsDesB was monitored using a Hansatech Oxygraph (Norfolk, England, UK). Steady-state kinetic parameters for substrates were obtained by incubating substrate of various concentrations, 100 mM phosphate buffer, pH 8, SsDesB (7.4 μM), and 50 μM FAD for 5 min at 25°C with stirring. Reactions were then initiated by the addition of 0.7 mM NADPH (700 μL total volume). Steady-state kinetic parameters for NADPH were obtained by incubating 100 mM phosphate buffer, pH 8, SsDesB (7.4 μM), 50 μM FAD, and 0.7 mM cadaverine for 5 min incubation with stirring at 25°C. Assays were initiated with various NADPH concentrations (9.1 μM–9.3 mM).

Standard assays (400 μL total volume) consisting of 100 mM phosphate buffer pH 8, 50 μM FAD, 2.12 μM SsDesB, 10 mM cadaverine, and 0.7 mM NADPH were incubated at 25°C. Assays with boiled enzyme as well as no enzyme were also prepared. After 30 s and 72 s, aliquots were removed from assays and the amount of hydrogen peroxide produced was quantified via the QuantiChrom^TM peroxide assay kit (BioAssay Systems; Hayward, CA) according to the manufacturer’s instructions. A hydrogen peroxide standard curve was used to quantify the amount of hydrogen peroxide produced by SsDesB.

The SsDesB gene was amplified from E. coli codon-optimized DNA obtained from GenScript Corp. (Piscataway, NJ) (SSEG_08523, pET28a (+)) by polymerase chain reaction (PCR) using the following oligonucleotide primers: 5’-GGC TCG GAG AAC CTG TAC TTC CAG ACC GCG CGT CCG GAG -3’(PE-3112) and 5’-GGG GAC CAC TTT GTA CAA GAA AGC TGG GTT ATT ACA CGC TAA ACT CCT GGA ACG C-3’ (PE-3113). The PCR amplicon was then used as the template for a second PCR amplification with primer PE-277 (5’-GGG GAC AAG TTT GTA CAA AAA AGC AGG CTC GGA GAA CCT GTA CTT CCA G -3’) and PE-3113. The amplicon from the second PCR, coding for SsDesB (T2-V425) with an N-terminal tobacco etch virus (TEV) recognition site (ENLYFQ/T), was recombined into the entry vector pDONR 221 (Life Technologies; Grand Island, NY) using Gateway cloning technology to generate the entry clone pGL3067, and the nucleotide sequence was confirmed by Sanger sequencing [27]. The modified SsDesB gene was then recombined from pGL3067 into pDEST527 to generate the expression vector pGL3070. This plasmid produces an N-terminal hexahistidine-tagged SsDesB that can be cleaved by TEV protease to yield the SsDesB enzyme [25]. The protein was expressed in the E. coli strain Rosetta2(DE3). Cultures were grown to mid-log phase (OD₆₀₀ ~ 0.5) at 37°C in Luria Bertani broth containing 0.2% glucose, 100 μg ml^-1 ampicillin, and 30 μg ml^-1 chloramphenicol. Overexpression of the SsDesB fusion protein was induced with 1 mM isopropyl β-D-1-thiogalactopyranoside for 18–20 h at 18°C. Cells were pelleted by centrifugation and stored at -80°C.

All purification steps were performed at 4–8°C. E. coli cell paste (10 g) was resuspended in 200 mL of ice-cold lysis buffer containing 50 mM Tris-HCl pH 7.4, 200 mM NaCl, 25 mM imidazole, and cOmplete, EDTA-free protease-inhibitor cocktail tablets (Roche Diagnostics; Mannheim, Germany). Cells were lysed by passing through an APV-1000 homogenizer (Invensys APV Products, Albertslund, Denmark) at 69 MPa three times, and the lysate was centrifuged for 30 min at 15,500 rpm. The supernatant was filtered by vacuum through a 0.2 μm polyethersulfone membrane and applied to a 5-mL HisTrap FF column (GE Healthcare, Piscataway, NJ) pre-equilibrated with lysis buffer. The column was washed to baseline with lysis buffer and protein was eluted using a linear gradient from 25 mM to 500 mM imidazole. Fractions containing the His₆-SsDesB fusion protein were combined and concentrated using an Ultracel 30 kDa ultrafiltration disc (EMD Millipore, Billerica, MA). The concentrated protein was diluted with 50 mM Tris pH 7.4, 200 mM NaCl buffer to achieve an imidazole concentration of 25 mM. The His₆-SsDesB fusion protein was then digested with 5 mg polyhistidine-tagged TEV protease overnight [27]. The digest was applied to a second HisTrap column (2 × 5 mL) pre-equilibrated with lysis buffer. The deep-yellow colored fractions containing SsDesB were combined and half was concentrated to 12–15 mg/mL in the lysis buffer mentioned above. Aliquots were flash frozen in liquid nitrogen and stored at -80°C for use in structural studies. The protein was judged to be >90% pure by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and the molecular weight was confirmed by electrospray ionization mass spectrometry. The remaining portion of protein was incubated overnight at 4°C with 10 mM DTT. The sample was concentrated, as above, and fractionated on a HiPrep 26/60 Sephacryl S-300 HR column (GE Healthcare; Piscataway, NJ) equilibrated with a buffer containing 25 mM Tris-HCl pH 7.4, 150 mM NaCl, and 2 mM tris(2-carboxyethyl)phosphine (TCEP). The peak fractions of recombinant SsDesB were combined and concentrated to 12–15 mg/mL. Aliquots were flash frozen in liquid nitrogen and stored at -80°C for use in biochemical studies. This protein was judged to be >95% pure by SDS-PAGE using a Coomassie stain and its molecular weight was confirmed by electrospray ionization mass spectrometry.

Using SsDesB without an N-terminal hexahistidine tag, the amount of N-hydroxylated product formed in SsDesB assays at 25°C was determined using a modified Csaky iodine oxidation assay [4, 28]. In a 96-well plate, SsDesB (1 μM) was incubated in 100 mM phosphate buffer, pH 8.0 with 50 μM FAD and varied amounts of substrate (cadaverine or putrescine) for 5 min before reactions were initiated with 0.7 mM NAD(P)H (400 μL total volume). For assays in which NAD(P)H concentrations were varied (0.031–4 mM), the substrate concentration was 10 mM. Hydroxylamine standards (9.4–300 μM) in 100 mM phosphate buffer, pH 8.0, were also prepared. Aliquots of enzyme assays (61.5 μL) as well as standards were removed and quenched with perchloric acid (0.03 N). Samples were then centrifuged at 7,000 rpm for 3 min at room temperature before the supernatant (47 μL) was removed and added to equal volumes of 10% w/v sodium acetate (47 μL) and 0.6% w/v sulfanilic acid in 25% acetic acid (47 μL). Samples were then incubated with 0.07% w/v iodine in glacial acetic acid for 15 min with shaking before the addition of 19 μL of sodium thiosulfate (0.02 N) and 19 μL of 1-napthylamine (0.6% w/v). After 45 min of shaking at 25°C, 96-well plates were read at 562 nm using an Epoch Biotek microplate spectrophotometer.

Standard assays (400 μL total volume) consisting of 100 mM phosphate buffer pH 8, 50 μM FAD, 1 μM SsDesB (without an N-terminal hexahistidine tag), 10 mM substrate, and 0.7 mM NADPH were incubated at 25°C. The following substrates were assayed: cadaverine, putrescine, L-lysine, 1,3-diaminopropane, spermidine, L-ornithine, n-butylamine, and 3-dimethylaminopropylamine.

Standard assays (400 μL) consisting of 100 mM phosphate buffer pH 8, 10 mM substrate, SsDesB (1.0 μM, without an N-terminal hexahistidine tag) and 0.7 mM NADPH were incubated at 25°C. Assays without substrate were also prepared as a negative control. Aliquots (50 μL) were removed after 10 min, quenched with twice the volume of HPLC-grade acetonitrile (Thermofisher; Waltham, MA), and chilled at -20°C for 10 min. After centrifugation at 12,500 rpm for 3 min at room temperature, the supernatant was incubated with 50 μL of 100 mM borate buffer pH 8 followed by the addition of 20 μL of 10 mM Fmoc-Cl dissolved in LC/MS grade methanol. After 5 min, samples were incubated with 20 μL of 0.1 M 1-adamantylamine in 1:1 acetonitrile: water for an additional 10 min to remove excess Fmoc-Cl. Samples were then analyzed by LC/MS. Using a flow rate of 0.6 mL/min, samples (1 μL) were injected onto an Acquity UPLC BEH C18 column (2.1 × 50 mm × 1.7 μm) (Waters; Milford, MA) with an Acquity precolumn filter (2.1 mm × 0.2 μm) attached. The column was warmed to 40°C and UPLC separation was achieved using a 5–100% gradient of acetonitrile: water with 0.1% formic acid over 9 min. ESI-MS in sensitivity mode was used to obtain exact mass measurements of Fmoc-derivatized substrates and products. The following ESI settings were used: 3 kV capillary voltage, 45 V sample cone voltage, 120°C source temperature, 550°C desolvation temperature, 50 L/hr cone gas flow, and 800 L/hr desolvation gas flow. A leucine enkephalin reference (Waters; Milford, MA) was used for the lock spray with a 30 s reference scan frequency, 45 V reference cone voltage, 6 eV collision energy, and a 99.9 dynamic range enhancement setting.

Kinetic data were analyzed in Kaleidagraph (Synergy; Reading, PA) and Prism 8 (GraphPad; San Diego, CA). The k_cat and K_M of SsDesB with specific substrates were determined by fitting initial rate data to the Michaelis-Menten Eq (2). The substrate inhibition constant (K_i) was determined for cadaverine and NAD(P)H (with the exception of oxygen consumption initial rate data with varied NADPH) by fitting the initial rate data to the Haldane Eq (3). All reaction rates were measured such that no more than 10% product formed in enzyme assays. See the S1 File for fitted data.

Purified SsDesB (12 mg/mL in 50 mM Tris-HCl, pH 7.4, 200 mM NaCl, and 25 mM imidazole) was screened for crystals using several sparse-matrix crystallization screens from Hampton Research, Microlytic, Qiagen, and Molecular Dimensions using a Gryphon crystallization robot (Art Robbins Instruments; Sunnyvale, CA). Optimization screens of the initial crystallization screening hits were carried out using the hanging-drop vapor diffusion method in Easy-Xtal 15-well plates (Qiagen; Germantown, MD). Additional optimization screens were completed using the Hampton Research Additive screen. Crystals of SsDesB for data collection were obtained by mixing 2.5 μL of protein solution (12.1 mg/mL) with 2 μl well solution (0.1 M Hepes pH 7.5, 0.2 M sodium chloride, 25% w/v polyethylene glycol 3350), and 0.5 μL of 0.1 M taurine and sealing over 500 μL of well solution. The trays were incubated at 4°C. Yellow, plate-like crystals appeared within 1 week. A single crystal was retrieved from a drop using a Litholoop (Molecular Dimensions; Maumee, OH) and transferred to a 1 μl drop of Paratone-N where excess mother liquor was whisked away from the crystal. The crystal was removed with a Litholoop and flash-cooled by plunging it into liquid N₂.

Crystals of SsDesB in complex with NADP⁺ were obtained by incubating 12.1 mg/mL of protein with 1 mM NADP⁺ (Sigma-Aldrich, St. Louis, MO) for 2 hrs at 4°C. Precipitated material was removed by centrifugation for 10 min and 2.5 μL of the SsDesB:NADP⁺ mixture was mixed with 2.0 μL of well solution (0.1 M Hepes pH 7.5, 0.2 M sodium chloride, 25% w/v polyethylene glycol 3350) and 0.5 μL (0.1 M sarcosine) and sealed over 500 μL of well solution. After 1 hr, the drops were streak-seeded with a whisker using previous crystals of SsDesB as a seed source. Yellow, plate-like crystals appeared within 2 days. A single crystal for data collection was retrieved using a Litholoop, transferred to a 1 μL drop of Paratone-N, where excess mother-liquor was whisked away from the crystal, and the crystal was flash-cooled by plunging into liquid N₂.

X-ray diffraction data for crystals of SsDesB and SsDesB complexed with NADP⁺ were collected on the 22-ID and 22-BM beamlines, respectively, of the SER-CAT facilities at the Advanced Photon Source, Argonne National Laboratory, Argonne, Illinois. For the SsDesB crystal, 360 images were collected using a wavelength of 1.0000 Å, an oscillation angle of 1.0°, a crystal to detector distance of 300 mm, and an exposure time of 0.25 s. Data were collected from a single crystal of SsDesB complexed with NADP⁺ using a wavelength of 1.0000 Å, a crystal-to-detector distance of 300 mm, an oscillation angle of 1.0°, and an exposure time of 20 s. A total of 180 frames of data were collected. All X-ray diffraction images were processed with HKL-3000 [29].

The 2.37 Å structure of SsDesB in complex with NADP⁺ was solved by molecular replacement with the program PHASER [30] in the PHENIX [31] suite and using the structure of E. amylovora DfoA [2] (PDB entry 5o8r, chain A, 53% sequence identity) as a search model after removing all solvent and ligand molecules. Based on the Matthews coefficient of 2.36 ų Da^-1 and solvent content of 47.9%, a search for 8 molecules in the asymmetric unit was performed [32–34]. Iterative rounds of model-rebuilding were performed manually using Coot [35] followed by additional automated model adjustment using the PDBredo server [36]. The PDB coordinate files for the FAD and NADP⁺ ligands were prepared using Molinspiration (https://molinspiration.com), and the ligand restraint files used during refinements were prepared with the program eLBOW [37] in PHENIX. Refinements were performed with phenix.refine [38]. Water molecules were automatically located with Coot, manually inspected, and refined with phenix.refine.

The structure of SsDesB at 2.86 Å resolution was solved by molecular replacement with Phaser using chain A of the SsDesB-NADP⁺ complex structure after removing all solvent and ligand molecules and searching for 8 molecules in the asymmetric unit (2.30 ų Da^-1 and solvent content of 46.5%). Iterative rounds of manual model rebuilding were performed using Coot followed by refinement with phenix.refine. All structures were validated using the validation tools available in Coot and also with the Molprobity server [39]. Refinement statistics and model validation are outlined in Table 4. The coordinates and structure factor files were deposited in the Protein Data Bank under accession codes 6XBB and 6XBC.

The purification yielded 30–40 mg of a 51.7 kDa N-terminal hexahistidine tagged protein from 1 L of culture (S4A Fig in S1 File). The UV-visible absorbance spectrum of SsDesB indicated the presence of the FAD cofactor bound to the purified enzyme with characteristic absorbance maxima at 370 nm and 457 nm (S4B Fig in S1 File). The extinction coefficient at 457 nm was 14,478 M^-1cm^-1 and ~15% of flavin was estimated to copurify with SsDesB.

Kinetic parameters of SsDesB NAD(P)H oxidation activity was determined by monitoring the decrease in the absorbance of NAD(P)H at 340 nm (Table 1). At a constant concentration of cadaverine (10 mM), SsDesB oxidized NADPH with an apparent k_cat of 0.28 ± 0.01 s^-1 and k_cat/K_M of 43 ± 11 mM^-1s^-1. NADH was oxidized with an apparent k_cat of 0.38 ± 0.09 s^-1 and k_cat/K_M of 3.3 ± 1.3 mM^-1s^-1, indicating a preference for oxidizing NADPH. There was also evidence of substrate inhibition when assayed with either cadaverine or NAD(P)H at concentrations above 0.7 mM.

The activity of SsDesB was assessed by monitoring its rate of oxygen consumption with NAD(P)H, FAD, and various substrates. Substrate-independent oxidation was observed as SsDesB consumed slightly more oxygen (1.5-fold more) with cadaverine (Fig 2). Similar kinetic parameters were determined using the oxygen consumption assay compared to those obtained from NAD(P)H oxidation assays (Table 2). At a constant cadaverine concentration of 10 mM and varying NADPH concentrations, SsDesB consumed molecular oxygen with an apparent k_cat of 0.24 ± 0.01 s^-1 and k_cat/K_m of 2.9 ± 0.4 mM^-1s^-1. These values are somewhat consistent with those determined in NADPH oxidation assays. When the concentration of NADPH was held constant and cadaverine concentrations were varied, SsDesB consumed molecular oxygen with an apparent k_cat of 0.20 ± 0.01 s^-1 and k_cat/K_m of 4.3 ± 0.1 mM^-1s^-1. When putrescine concentrations were varied at constant NADPH concentrations, SsDesB consumed molecular oxygen with a similar apparent k_cat of 0.29 ± 0.01 s^-1 but lower catalytic efficiency of 0.62 ± 0.15 mM^-1s^-1. QuantiChrome assays detected peroxide formation in only assays with enzyme (11 μM in 30 s and 20 μM in 72 s) and none in control assays without enzyme or with boiled enzyme (S5 Fig in S1 File). The presence of peroxide indicated that SsDesB does not use all of its reduced molecular oxygen to form an N-hydroxylated product.

Using SsDesB without a hexahistidine tag at a constant concentration of NADPH, the enzyme N-hydroxylated cadaverine with an apparent k_cat of 0.048 ± 0.004 s^-1 and catalytic efficiency of 2.5 ± 1.2 mM^-1 s^-1 (Table 3). However, the apparent k_cat value was approximately 6-fold lower than those determined in oxygen consumption and NADPH assays (Fig 3). N-hydroxylated product was confirmed by LC/MS in SsDesB assays derivatized with Fmoc-Cl and not in assays without enzyme (Fig 4). At a constant NADPH concentration, SsDesB also N-hydroxylated putrescine with a higher apparent k_cat of 0.034 ± 0.006 s^-1 but lower catalytic efficiency of 0.31 ± 0.23 mM^-1s^-1, further indicating a preference for N-hydroxylating cadaverine over putrescine. The apparent K_M of cadaverine (31 μM) with SsDesB was 35-fold lower than that of putrescine (1100 μM). When the NADH concentration was varied, the apparent k_cat value was ~2-fold higher than that when NADPH was varied, however the apparent K_M was ~4-fold larger, indicating a preference for binding NADPH over NADH. At high cadaverine and NAD(P)H concentrations, substrate inhibition was observed. To rule out product inhibition, we assayed SsDesB in the presence of increasing concentrations of N-hydroxycadaverine and there was no decrease in enzyme activity (S10 Fig in S1 File). Thus, the initial rate data were fit to Eq 3 when inhibition was observed. Products were confirmed via LC/MS analysis of SsDesB assays with cadaverine detected the generation of new compounds, such as N-hydroxycadaverine, after derivatization with Fmoc-Cl that were not present in control assays without enzyme (Fig 4).

SsDesB N-hydroxylated cadaverine, putrescine, spermidine, and L-lysine substrates (Fig 5). L-ornithine, 1,3-diaminopropane, and 3-dimethylaminopropylamine were not N-hydroxylated. The m/z values of these products are consistent with N-hydroxycadaverine ([M+H]⁺ m/z calcd. for C₃₅H₃₄N₂O₅H 563.25; found 562.2563), N-hydroxyputrescine ([M+H]⁺ m/z calcd. for C₃₄H₃₂N₂O₅H: 549.23; found 549.2389), N-hydroxylysine ([M+Na]⁺ calcd. for C₃₆H₃₄N₂O₇Na: 629.23; found 629.2264), and N-hydroxyspermidine ([M+H]⁺ m/z calcd. for C₃₇H₃₉N₃O₅H: 606.29; found 606.2968). See SI for extracted ion chromatograms and high resolution mass spectral data (S11–S15 Figs in S1 File). With the exception of assays with spermidine that yielded two stereoisomers of N-hydroxylated products, four stereoisomers of products were detected in all assays (S10–S13 and S15 Figs in S1 File). Assays were also performed with SsDesB containing an N-terminal hexahistidine tag was assayed, and no N-hydroxylated products were detected in product formation assays containing SsDesB with L-lysine. A slight difference in activity was only observed with L-lysine when it was assayed with the hexahistidine tagged SsDesB, suggesting that the affinity tag interferes with the enzyme N-hydroxylating L-lysine.

Crystal structures of SsDesB were obtained for the FAD-bound holoenzyme and the ternary complex with bound FAD and NADP⁺ at 2.86 Å and 2.37 Å, respectively. Data collection and processing statistics are presented in Table 4. SsDesB crystallized as a homotetramer with each protomer consisting of two dinucleotide-binding Rossman-fold domains [40]: an FAD-binding domain (red) and an NADPH-binding domain (blue) (Fig 6). In both structures, each molecule in the asymmetric unit contained bound FAD and in the SsDesB-FAD-NADP⁺ complex, a bound NADP⁺ molecule was clearly resolved in each protomer. Based on a query of the Protein Data Bank (PDB) using PDBefold [41, 42], the overall structure of the SsDesB resembles those of previously determined crystal structures of the flavin-dependent monooxygenases E. amylovora DfoA (PDB entry 5O8P, r.m.s.d. 0.85 Å over 407 aligned residues) [43], the ornithine hydroxylase KtzI from Kutzneria sp. 744 (PDB entry 4TM1, r.m.s.d. 1.74 Å over 364 aligned residues) [5], the P. aeruginosa ornithine hydroxylase PvdA (PDB entry 3S5W, r.m.s.d 1.96 over 361 aligned residues) [44], and the A. fumigatus ornithine hydroxylase, SidA, (PDB entry 5CKU, r.m.s.d 2.08 Å over 362 aligned residues) (Fig 7 and S1 Table in S1 File).

In the SsDesB-FAD complex structure, the FAD molecule is bound in an elongated conformation with the fully planar isoalloxazine ring positioned at the interface of the two dinucleotide-binding domains and a large area of the molecule, particularly the adenine dinucleotide portion, is exposed to solvent. During purification, after passage through the gel filtration column, fractions containing SsDesB lost the deep yellow color, suggesting that FAD was lost during the process. For our crystallization efforts, we used concentrated protein eluted from an IMAC column in which the aliquots retained a deep yellow color. However, similar crystallization hits could also be obtained after adding exogenous FAD to the protein aliquots after passage through a size exclusion column and then concentrating the protein. Loss of FAD during purification was also reported with other homologs [3]. In the holoenzyme, FAD is held in the active site pocket primarily by hydrogen bonding interactions involving S43, E42, K44, H50, P390, and W49 along the flavin adenine nucleotide backbone while residues Q61 and L392 contribute direct hydrogen bonds to the planar isoalloxazine ring primarily via the backbone amide nitrogen atoms (Fig 8).

There are no significant global structural changes observed upon binding of the NADP⁺ molecule (r.m.s.d. = 0.26 over 2741 aligned atoms). The higher resolution 2.37 Å structure of the ternary complex shows that the NADP⁺ cofactor is held in the active site pocket in an elongated manner via several direct hydrogen bonds to SsDesB and a network of water-mediated hydrogen bonding bridges between SsDesB residues and NADP⁺ (Fig 9). The electron density is well-defined for both the FAD and NADP⁺ cofactors, including the nicotinamide ring, in all 8 protomers in the asymmetric unit. The nicotinamide ring is held in position by hydrogen bonding interactions between the nicotinamide N7N atom and main chain carbonyl oxygen atom of H59 and the N5 atom of the FAD isoalloxazine ring. The nicotinamide O7N atom participates in water-mediated hydrogen bonding interactions via water 679 to the E202 side chain and H50 main chain carbonyl oxygen atom. The side chain of Q61 also contributes to stabilizing interactions via a hydrogen bond between the side chain OE1 oxygen atom and the O2D ribose oxygen atom. The phosphate moiety of NADP⁺ is held in place by a salt bridge between the O2X oxygen and the side chain of K264, and the R232 side chain to the O2X and O1X oxygen atoms. In addition, there is a hydrogen bond between the O3X oxygen and the side chain of S224.

In comparison to the holoenzyme structure, upon NADP⁺ binding, several structural adjustments are observed among the residues surrounding the NADP⁺ binding pocket. At the interface of the FAD isoalloxazine ring and NADP⁺ nicotinamide ring, the side chain of Q61 shifts away from the FAD isoalloxazine to accommodate the binding of the nicotinamide ring of NADP⁺, which in turn, induces a shift in the rotamer positions of the nearby L230, E231, and Y232 side chains (Fig 10). Additionally, the D391 side chain is now engaged in a water-mediated hydrogen bond bridge between the side chain OD2 oxygen atom and the isoalloxazine ring O2 atom via water 691 (Fig 9). The side chain of T62 is also hydrogen bonded to the isoalloxazine ring N3 nitrogen atom (Fig 9). A shift in the conformation of the loop consisting of residues 261–269 is observed which avails space for NADP⁺ binding and also properly positions the side chain of K264 to engage in an ionic interaction with the phosphate moiety of NADP⁺. The side orientation of R223 also shifts and is now properly positioned for an ionic interaction with the NADP⁺ phosphate and is also stacked against the terminal adenine moiety.