Biochemical data from the characterization of a new pathogenic mutation of human pyridoxine-5'-phosphate oxidase (PNPO)

PNPO deficiency is responsible of severe neonatal encephalopathy, responsive to pyridoxal-5′-phosphate (PLP) or pyridoxine. Recent studies widened the phenotype of this condition and detected new genetic variants on PNPO gene, whose pathogenetic role and clinical expression remain to be established. One of these mutations, Arg116Gln, is of particular interest because of its later onset of symptoms (beyond the first months of life) and its peculiar epileptic manifestations in patients. This protein variant was expressed as recombinant protein in E coli, purified to homogeneity, and characterized with respect to structural and kinetic properties, stability, binding constants of cofactor flavin mononucleotide (FMN) and product (PLP) in order to define the molecular and structural bases of its pathogenicity. For interpretation and discussion of reported data, together with the description of clinical studies, refer to the article [1] (doi: 10.1016/j.ymgme.2017.08.003).

kinetic properties, stability, binding constants of cofactor flavin mononucleotide (FMN) and product (PLP) in order to define the molecular and structural bases of its pathogenicity.
For interpretation and discussion of reported data, together with the description of clinical studies, refer to the article [

Data
The data presented in this paper refer to structural and functional characterization of a new pathogenic PNPO mutant form, related to the onset of a peculiar epileptic status in patients. Fig. 1 shows insights into the active site structure of modeled human PNPO (Arg116Gln mutant form). Figs. 2-5 show the characterization of Arg116Gln PNPO mutant.

Molecular modeling
The three-dimensional structure of human PNPO (PDB Code: 1NRG) [2] was used as a starting point to model the Arg116Gln mutant (Fig. 1). The mutation on protein structure was carried out using the "Mutate model" script [3]. Procheck [4] was used to monitor the stereochemical quality of the  [2]. The two protein subunits are shown in cyan end blue. Secondary structures are shown as cartoon, while Arg116 and Glu143 residues are depicted as sticks. FMN and PLP in the active site (A) are shown as sticks, in orange and yellow color, respectively. Location of PLP in the secondary tight binding site (B)(shown as two different conformers) results from the superimposition with E. coli PNPO three-dimensional structure in the complex with PLP (PDB 1G79; [12]). model, ProsaII [5] to measure the overall protein quality. Prediction of protein stability was carried out using DUET server [6].

Expression and purification
PNPO forms were purified following a simplified procedure derived form a previous protocol [2]. Briefly, supernatant from cell lysis was directly loaded onto a HisTrap FPLC column and eluted with imidazole (0-250 mM) in 50 mM Tris-HCl pH 7.6, 300 mM NaCl. Spectrum of Arg116Gln mutant is shown in Fig. 2. Apo-enzyme was prepared as previously described [2].

Kinetic studies
Enzymatic assays were performed as previously described [7]. Experimental determinations are depicted in Fig. 3. K m and k cat values are shown in Table 1

Circular dichroism measurements
Far-UV CD measurements were carried out at 20 and 40°C in 50 mM sodium Hepes buffer pH 7.6, 150 mM NaCl with a protein concentration of 4 μM. The relative spectra are shown in Fig. 4B.

PLP binding equilibrium
Analyses took advantage of protein intrinsic fluorescence quenching observed upon PLP binding as previously described [8]. PLP binding curves are shown in [PNPO PLP ] represents the concentration of PNPO PLP complex.
The fraction in Eq. (1) corresponds to the fraction of enzyme-bound PLP at equilibrium ½PNPO PLP eq ½PNPO .
PNPO PLP ½ eq was derived from the equation for the dissociation constant of the binding equilibrium, as one of the two solutions of the quadratic equation

FMN binding
Dissociation constant for FMN binding to PNPO forms were analyzed by FMN fluorescence quenching observed upon binding of the cofactor to apo-PNPO [9]. The saturation curves are shown in Figure 3A of Ref [1]. The relative K d values are reported in Table 1 of Ref [1]. Data were analyzed according to Eq. (2), in which F rel is the measured relative fluorescence at 525 nm, F 0 is fluorescence in the absence of apo-PNPO, F inf is fluorescence at infinite apo-PNPO concentration, [APO] is the total apo-enzyme concentration, [FMN] stands for the total cofactor concentration and K d is the dissociation constant of the equilibrium APO þ FMN⇌HOLO The fraction in Eq. (2) corresponds to the fraction of enzyme-bound FMN at equilibrium ½HOLO eq ½FMN .
HOLO eq Â Ã was derived from the equation for the dissociation constant of the binding equilibrium, as one of the two solutions of the quadratic equation

Thermal denaturation
Protein samples (4 μM), were heated from 20 to 90°C monitoring the dichroic activity at 220 nm [10]. Denaturation curves were fitted to Eq. (3), in which θ 220 is the measured ellipticity at 220 nm, Δθ is the maximum ellipticity change, T is the temperature expressed in Celsius, T m is the apparent melting temperature, and n is the steepness of the sigmoid curve [10]: Denaturation curves are shown in Figure 3B of Ref [1].

Stoichiometry of PNPO PLP complexes
Wild type or Arg116Gln PNPO (150 μM) were mixed with 450 μM PLP and incubated at 30°C for one hour. Samples were run on P-6DG Bio-Gel. Fractions were analyzed by absorption spectra to detect the presence of protein and PLP. Protein concentration was measured by Bradford assay. PLP concentration was measured as in [11].

Transfer of tightly bound PLP
Protein samples saturated with PLP (PNPO PLP complexes) were mixed at a final concentration of 2 µM with equimolar amount of apo-human cytosolic serine hydroxymethyltransferase (hcSHMT). At various time intervals, transfer of PLP was determined by measuring hcSHMT catalytic activity [11]. PLP transfer kinetics are shown in Figure 5 of Ref [1].