Efficacy of aldose reductase inhibitors is affected by oxidative stress induced under X-ray irradiation

Abstract

Human aldose reductase (hAR, AKR1B1) has been explored as drug target since the 1980s for its implication in diabetic complications. An activated form of hAR was found in cells from diabetic patients, showing a reduced sensitivity to inhibitors in clinical trials, which may prevent its pharmacological use. Here we report the conversion of native hAR to its activated form by X-ray irradiation simulating oxidative stress conditions. Upon irradiation, the enzyme activity increases moderately and the potency of several hAR inhibitors decay before global protein radiation damage appears. The catalytic behavior of activated hAR is also reproduced as the K_M increases dramatically while the k_cat is not much affected. Consistently, the catalytic tetrad is not showing any modification. The only catalytically-relevant structural difference observed is the conversion of residue Cys298 to serine and alanine. A mechanism involving electron capture is suggested for the hAR activation. We propose that hAR inhibitors should not be designed against the native protein but against the activated form as obtained from X-ray irradiation. Furthermore, since the reactive species produced under irradiation conditions are the same as those produced under oxidative stress, the described irradiation method can be applied to other relevant proteins under oxidative stress environments.

Figure 1

Activation of hAR with irradiation. (a) Evolution of hAR normalized mean activity (red circles) and radius of gyration, Rg (gray triangles), as a function of dose. At 6 mM D,L-glyceraldehyde substrate concentration, enzymatic activity increases with irradiation dose until hAR undergoes major structural changes due to radiation damage. (b) Inhibitory potency of Zenarestat (black bars), JF0048 (gray bars), Epalrestat (light gray bars) and Tolrestat (white bars) at 0 kGy, 1 kGy and 8 kGy doses. Tested concentrations close to the IC50 value were used for each compound. The non-normalized specific activities of the inhibitors are found in Supp. (c) Effect of irradiation on hAR kinetics. Native enzyme without irradiation loses 50% of its activity due to excess-substrate inhibition at 6 mM d, l-glyceraldehyde (red circles). Irradiation with a dose of 2 kGy increases the value of the Michaelis constant K_M towards substrate by near one order of magnitude (7 times) while the inhibition disappears (blue circles).

Figure 2

Absence of specific radiation damage in the binding pocket and in the catalytic tetrad. (a) Difference Fourier map (F_obs,20 - F_obs,1, α_calc,1) of the most irradiated dataset, D20. The NADPH (pink) and a citrate molecule (yellow) are bound to the binding pocket. The residues of the catalytic tetrad are shown in cyan and Cys298 in navy blue. No signs of specific radiation damage are observed in the vicinity of the binding-pocket surface, except for the Cys298 residue. The map was contoured at 0.56 e/ų (σ = ±5). (b) Detail of the difference Fourier maps (F_obs,10 - F_obs,1, α_calc,1) (top row) and (F_obs,20 - F_obs,1, α_calc,1) (bottom row) on the catalytic tetrad (Asp43, Lys77, His110 and Tyr48) at 1.2 Å resolution, calculated using the amplitudes of datasets D10 or D20 and the initial data set D1 and the phases α_calc,1 of D1. The maps do not reveal signs of specific radiation damage in the residues of the catalytic tetrad. The D10-D1 and D20-D1 difference Fourier maps of the catalytic tetrad were contoured at 0.33 e/ų (σ = ±4.97 and σ = ±2.97, respectively).

Figure 3

Specific changes in Cys298 as a function of dose. (a) Difference Fourier maps (F_obs,n - F_obs,1, α_calc,1) on Cys298 residue at 1.2 Å resolution, calculated between each dataset n and the initial dataset using the phases of the fresh, not previously irradiated dataset, α_calc,1. The difference Fourier maps are contoured at two different electron density sigma levels (0.145 e/ų for the first row, 0.267 e/ų for the second row) for clarity. The maps have sigma levels ranging from σ = ±9.00 (D2-D1 map) to σ = ±2.53 (D8-D1 map) for the first row, and from σ = ±3.99 (D10-D1 map) to σ = ±2.41 (D20-D1 map) for the second row. Electron densities colored in green and red represent positive and negative sigma levels, respectively. Red blobs are indicative of the loss of electron density due to movements and desulfuration of Cys298; green blobs pointed by blue arrows show the increase of electron density due to displacements of Cys298 and green blobs pointed by green arrows reveal the appearance of a new atom in the vicinity of Cys298. (b) Electron density maps at 0.92–1.17 Å resolution on the Cys298 residue as a function of dose. In blue, 3Fo-2Fc electron density maps at σ = 1.00; in green, Fo-Fc electron density maps at σ = 2.50. (c) The Cys298/Ser298 amino acid is well fit in the 2Fo-Fc electron density map at σ = 1 for the D20 dataset. The Fo-Fc map at σ = ±2.50 does not show any electronic density. (d) Occupancy-related electron density maxima versus dataset. Red circles: occupancy-related electron density of Sγ atom decreases as Cys298 is desulfurated and the Sγ atom moves; green squares: occupancy-related electron density of Oγ atom increases as the Ser298 is formed; pink circles: loss of the electron density of the Sγ atom due to the desulfuration, calculated as the difference between the two previous electron densities; gray triangles: difference between the electron densities related to the desulfuration of Cys298 (pink circles) and the formation of the Ser298 (green squares). (e) Distances measured in the difference Fourier maps from the Cys298-Cβ atom to the center of the green blobs indicated by green arrows in (a) versus dataset. The top and bottom dashed lines represent the tabulated distances Cβ-Sγ and Cβ-Oγ in cysteine and serine residues, respectively.

Figure 4

LC-MS/MS analysis. LC-MS/MS spectra for non-irradiated protein (b) and for the irradiated protein at 1 kGy subjected to modifications at position 298 from cysteine to serine (b) and alanine (c). Characteristic peaks from the peptide fragmentation are colored in red (bn), blue (yn) and green (precursor). Non-characteristic peaks from the sequence are marked in gray. For clarity, only peaks that contain information about modifications are labeled. Only peaks above 1% of the base peak and with match tolerance of 0.6 Da are represented.

Figure 5

Changes on the topology of the hAR pocket entry channel and steric collisions of activated hAR with inhibitors. The conversion of Cys298 into Ser298 modifies the entrance of the substrate-binding pocket. (a) Entrance of the binding pocket of the native form of hAR. The sulfur atom of the Cys298 residue is marked in yellow. (b) Entrance of the binding pocket of the irradiated structure, which shows the Oγ-Ser298 atom, marked in red. In both cases, the nicotinamide ring of the cofactor is shown in pink color. (c–f) Interaction of the studied inhibitors with Cys298 and Ser298. The structure of the D20 dataset (in cyan) is superimposed with the crystallographic structures of hAR including the inhibitors Zenarestat (c, purple, PDB 1IEI), JF0048 (d, magenta, PDB 4XZH), Epalrestat (e, orange, PDB 4JIR) and Tolrestat (f, gray, PDB 1AH3). In all cases the inhibitors, designed against the native form, show severe steric collisions with Oγ-Ser298 but not with Sγ-Cys298. Sulfur atoms are displayed in yellow whereas oxygen atoms are displayed in red.

Figure 6

Indication that electrons are the main agent for the activation of hAR and proposed mechanism. (a) hAR activity in the presence of 10 mM uridine (red circles) and in the presence of 10 mM uridine plus 40 mM sodium nitrate (gray circles) as a function of irradiation dose. (b) Number of non-thiolated cysteines per protein molecule as a function of dose, as measured by the Ellman’s assay. Accessible cysteines (in the absence of urea) are shown in gray, and buried cysteines (revealed in the presence of 8 M urea) are shown in yellow. (c) Scheme of the proposed mechanism for the C298S and C298A conversion induced by X-ray irradiation or oxidative stress. Carbon atoms are colored in cyan, carbon radicals in purple, oxygen atoms in red, and sulfur atoms in yellow. Hydrogen atoms are not shown for clarity.