Biochemical properties of purified human retinol dehydrogenase 12 (RDH12): catalytic efficiency toward retinoids and C9 aldehydes and effects of cellular retinol-binding protein type I (CRBPI) and cellular retinaldehyde-binding protein (CRALBP) on the oxidation and reduction of retinoids

Abstract

Retinol dehydrogenase 12 (RDH12) is a novel member of the short-chain dehydrogenase/reductase superfamily of proteins that was recently linked to Leber's congenital amaurosis 3 (LCA). We report the first biochemical characterization of purified human RDH12 and analysis of its expression in human tissues. RDH12 exhibits approximately 2000-fold lower K(m) values for NADP(+) and NADPH than for NAD(+) and NADH and recognizes both retinoids and lipid peroxidation products (C(9) aldehydes) as substrates. The k(cat) values of RDH12 for retinaldehydes and C(9) aldehydes are similar, but the K(m) values are, in general, lower for retinoids. The enzyme exhibits the highest catalytic efficiency for all-trans-retinal (k(cat)/K(m) approximately 900 min(-)(1) microM(-)(1)), followed by 11-cis-retinal (450 min(-)(1) mM(-)(1)) and 9-cis-retinal (100 min(-)(1) mM(-)(1)). Analysis of RDH12 activity toward retinoids in the presence of cellular retinol-binding protein (CRBP) type I or cellular retinaldehyde-binding protein (CRALBP) suggests that RDH12 utilizes the unbound forms of all-trans- and 11-cis-retinoids. As a result, the widely expressed CRBPI, which binds all-trans-retinol with much higher affinity than all-trans-retinaldehyde, restricts the oxidation of all-trans-retinol by RDH12, but has little effect on the reduction of all-trans-retinaldehyde, and CRALBP inhibits the reduction of 11-cis-retinal stronger than the oxidation of 11-cis-retinol, in accord with its higher affinity for 11-cis-retinal. Together, the tissue distribution of RDH12 and its catalytic properties suggest that, in most tissues, RDH12 primarily contributes to the reduction of all-trans-retinaldehyde; however, at saturating concentrations of peroxidic aldehydes in the cells undergoing oxidative stress, for example, photoreceptors, RDH12 might also play a role in detoxification of lipid peroxidation products.

Figure 1

Purification of RDH12–His₆. Samples of RDH12-containing fractions were analyzed by SDS–PAGE and stained by Coomassie R-250. Lane 1, Sf9 microsomes containing RDH12–His₆ (16 μg); lane 2, DHPC extract of Sf9 microsomes (16 μg); lane 3, flow through the Ni²⁺–NTA resin (~16 μg); and lane 4, purified RDH12–His₆ (0.42 μg). Positions of molecular weight markers are indicated on the left.

Figure 2

Oxidation of holo-CRBPI by RDH12. Purified RDH12–His₆ (0.7 μg) was incubated with the indicated concentrations of holo-CRBPI, and the products were extracted and analyzed by normal-phase HPLC as described under Experimental Procedures. HPLC traces were measured at 350 nm. Peaks were identified as follows: 1, 13-cis-retinal; 2, 9-cis-retinal; 3, all-trans-retinal; 4, 13-cis-retinol; 5, 9-cis-retinol; 6, all-trans-retinol. The result shown is representative of at least three independent experiments.

Figure 3

CRBPI inhibition of 9-cis-retinol oxidation by RDH12. Purified RDH12–His₆ (0.02 μg) was incubated with 1.6 μM 9-cis-retinol and 1 mM NADP⁺ in the presence of 2 μM or 10 μM apo-CRBPI. Control reactions contained the same concentrations of BSA. Products were analyzed by normal-phase HPLC. Peaks were identified as follows: 1, 9-cis-retinal; 2, 9-cis-retinol; 3, all-trans-retinol.

Figure 4

Effect of DTT on the isomerization of free retinol. An aqueous 10 μM stock of all-trans-retinol was prepared from the ethanol stock by solubilization with equimolar BSA in the reaction buffer and subsequent sonication for 10 min. Equal amounts (500 μL) of solubilized all-trans-retinol were aliquoted into eight tubes. DTT was added to four out of eight tubes to the final concentration of 50 mM. Pairs of tubes (with and without DTT) were incubated for 24 h under different conditions as follows: (A) at 4 °C in the dark (covered with aluminum foil), (B) at 4 °C in the light (not covered), (C) at room temperature (RT) in the dark, and (D) at room temperature in the light. HPLC chromatograms were extracted at 325 nm. At this wavelength, 13-cis-retinol and 9-cis-retinol have an identical absorbance; therefore, the corresponding peaks can be compared directly. Peaks are numbered as follows: 1, 13-cis-retinol; 2, all-trans-retinol; 3, 9-cis-retinol.

Figure 5

Isomerization of retinol in the presence of BSA or apoCRBPI. (A) HPLC spectrum of the ethanol stock of all-trans-retinol used in experiments B–D. (B) Twenty micromolar all-trans-retinol was solubilized using CRBPI and incubated at 4 °C in the light. After 2 days, retinoids were extracted with hexane and analyzed by HPLC. (C) Same as B, except CRBPI was substituted for BSA. (D) Same as B after 4 weeks of incubation. Retinoids in the control mixture were largely degraded after 4 weeks (data not shown). HPLC chromatograms were extracted at 325 nm. Peaks are numbered as follows: 1, 13-cis-retinol; 2, all-trans-retinol; 3, 9-cis-retinol. Inset: absorbance spectrum of peak 3 with a λ_max of 322 nm, which matches the λ_max of standard 9-cis-retinol in the mobile phase of HPLC.

Figure 6

RDH12 activity toward all-trans-retinal in the presence of CRBPI. The formation of all-trans-retinol from 2 μM all-trans-retinal (atRal) was measured in the presence of increasing concentrations of CRBPI (0–40 μM). The 0.5-mL reactions were carried out with 0.3 μg of RDH12-containing Sf9 microsomes for 10 min at 37 °C. The result shown is representative of at least three independent experiments.

Figure 7

RDH12–His₆ activity toward 11-cis-retinoids in the presence of CRALBP. Effect of CRALBP on the conversion of 11-cis-retinol (black bars) or 11-cis-retinal (white bars) by RDH12 was determined at 1 μM substrate and 0.5–4 μM CRALBP. The amount of enzyme used in the oxidative reactions was 0.04 μg. In the reductive direction, 0.01 μg of the purified RDH12–His₆ was used. The reactions were performed in 1 mL volume with 1 mM NADP⁺ or NADPH as cofactors; samples were incubated at 37 °C for 10 min.

Figure 8

Human tissue expression profile of RDH12 and CRBPI. SYBR green real-time quantitative RT-PCR was carried out using 200 pg of cDNA from the indicated tissues or 12.5 pg of retina cDNA and primers specific for human RDH12, CRBPI, and GAPDH. RDH12 and CRBPI expression was normalized to GAPDH and presented as a relative expression or fold difference (Δ) to the expression level in the liver for each gene, which was set to one. Data represent the average ± standard deviation of the triplicate real-time PCR within one experiment and are representative of three other experiments.