Cardiolipin drives the catalytic activity of GPX4 on membranes: Insights from the R152H mutant

Abstract

The aim of this study was to examine, in biochemical detail, the functional role of the Arg152 residue in the selenoprotein Glutathione Peroxidase 4 (GPX4), whose mutation to His is involved in Sedaghatian-type Spondylometaphyseal Dysplasia (SSMD). Wild-type and mutated recombinant enzymes with selenopcysteine (Sec) at the active site, were purified and structurally characterized to investigate the impact of the R152H mutation on enzymatic function. The mutation did not affect the peroxidase reaction's catalytic mechanism, and the kinetic parameters were qualitatively similar between the wild-type enzyme and the mutant when mixed micelles and monolamellar liposomes containing phosphatidylcholine and its hydroperoxide derivatives were used as substrate. However, in monolamellar liposomes also containing cardiolipin, which binds to a cationic area near the active site of GPX4, including residue R152, the wild-type enzyme showed a non-canonical dependency of the reaction rate on the concentration of both enzyme and membrane cardiolipin. To explain this oddity, a minimal model was developed encompassing the kinetics of both the enzyme interaction with the membrane and the catalytic peroxidase reaction. Computational fitting of experimental activity recordings showed that the wild-type enzyme was surface-sensing and prone to "positive feedback" in the presence of cardiolipin, indicating a positive cooperativity. This feature was minimal, if any, in the mutant. These findings suggest that GPX4 physiology in cardiolipin containing mitochondria is unique, and emerges as a likely target of the pathological dysfunction in SSMD.

Keywords: Cardiolipins; GPX4; Sedaghatian-type spondylometaphyseal dysplasia; Selenium.

Graphical abstract

Fig. 1

GPX4^R152H purification and characterization. A: final purification on BSP affinity chromatography of recombinant Sec GPX4^R152H produced in E. coli. Absorbance profile at 280 nm shows three main peaks eluting from the affinity column, while GPX4^R152H activity is present only in the fractions corresponding to the first peak. B: mass spectrometry analysis of purified GPX4^R152H. MS spectra and deconvolution obtained from fractions with the highest activity show that GPX4^R152H is almost 100% homogeneous; Z +29 and Z +22 indicate the mass with 29 or 22 charges. C: far-UV CD spectra of wt GPX4 and GPX4^R152H. Recorded spectra reveal a difference in the secondary structure content.

Fig. 2

Computational analysis of BSP and cardiolipin polar head interaction with wild-type and GPX4^R152H. A: molecular docking of BSP (green) against wt GPX4 (yellow)/GPx4^R152H (orange). B: molecular docking of cardiolipin (cyan) against wt GPX4 (yellow)/GPx4^R152H (orange). The interaction of a hydroperoxide group of cardiolipin floating to the surface and interacting with Sec46 is also reported (see also ref [18]). The most important interacting residues are highlighted. Hyd Loop indicates the hydrophobic loop containing Ile129 and Leu130. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

SPR analysis of wild-type or GPX4^R152H binding over TOCL and DOPC bilayers. Equilibrium bindings (R_eq) are measured for various concentrations (5 × 10⁻⁸ to 2.5 × 10⁻⁶ M) of wt GPX4 and GPX4^R152H on a DOPC and on a DOPC:TOCL (8:2) bilayer in 25 mM phosphate buffer, pH 7.8. R_eq values reported in figure are calculated subtracting (R_eq)_DOPC from (R_eq)_DOPC:TOCL [18].

Fig. 4

Initial rate of wild-type or GPX4^R152H activity on liposomes. Total liposome phospholipid concentration is 0.4 mM and SLPCOOH concentration in the test is 0.05 mM. A and B: increasing concentrations of wild-type GPX4 (3.2 × 10⁻⁹ M − 6.4 × 10⁻⁸ M) or GPX4^R152H (2.8 × 10⁻⁹ M − 5.5 × 10⁻⁸ M). A: SLPC-SLPCOOH liposomes; simple linear regression (continuous line) and 95% confidence intervals (dotted lines) for slopes are reported (R² = 0.9397 for wt GPX4 and R² = 0.9322 for GPX4^R152H); B: SLPC-SLPCOOH liposomes containing 20% molar fraction of TOCL; simple linear regression (continuous line) and 95% confidence intervals (dotted lines) are reported (R² = 0.9744) for GPX4^R152H, a second order polynomial fit is reported (R² = 0.9949) for wild-type GPX4. C: SLPC-SLPCOOH liposomes containing Increasing molar fraction of TOCL (0%–20%); the enzyme concentrations for wild-type GPX4 and GPX4^R152H are 4.6 × 10⁻⁹ M and 1.6 × 10⁻⁸ M, respectively.

Fig. 5

Examples of the progression curves of reaction rates at different concentration of the wild-type GPX4 or the R152H mutant. A: 9.5 × 10⁻⁹ M and 6.4 × 10⁻⁸ M wild-type GPX4; B: 8.3 × 10⁻⁹ M and 5.5 × 10⁻⁹ M GPX4^R152H. SLPC-SLPCOOH liposomes contained 20% molar fraction of TOCL.

Fig. 6

Computational fitting of the progression curves of the reaction. The absorbance (continuous lines) of the peroxidatic reactions on liposomes with the corresponding fit (dotted lines) are reported. Wild-type GPX4 (A) or GPX4^R152H (B) were added in a range of concentration (5.1 × 10⁻⁸ M-6.4 × 10⁻⁹ M for wild-type GPX4 and 5.5 × 10⁻⁸ M-2.9 × 10⁻⁹ M for GPX4^R152H) to SLPC-SLPCOOH liposomes. In panels C and D the wild-type and the mutant enzymes respectively were used at the same range of concentration on SLPC-SLPCOOH liposomes also containing 20% molar fraction of TOCL. Lower panels (E-F-H-H-I) show the kinetic parameters obtained by fitting of the model with the traces reported in panels A-B-C-D (wild-type GPX4 green triangles, GPX4^R152H red dots). Colour gradient indicates increasing enzyme concentration. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 7

Effect of TOCL on the progression curves of wild-type or GPX4 ^R152H activity. Continuous traces indicate the relative absorbance [a.u.], dotted lines the corresponding fit. A: wt GPX4; B: GPX4^R152H used at a concentration 4.6 × 10⁻⁹ M and 1.59 × 10⁻⁸, respectively. TOCL fractional ratio was - 0, 3, 11 and 20%, and darker color indicates higher fractional ratio. Panels C to F report the kinetic parameters at different TOCL fraction, as obtained from the fit of the traces in the first two panels (wt GPX4 green triangles, GPX4^R152Hred dots). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)