Retinoid processing in vivo : characterization and structure-function analysis of retinol dehydrogenases

Abstract

Retinoids (vitamin A derivatives) are vital nutrients that regulate a variety of physiological processes, including embryonic development, postnatal growth, reproduction, cell differentiation as well as maintenance of the immune system and vision. Most of the physiological functions of retinoids are believed to be controlled by retinoic acid (RA) altered gene expression via two classes of nuclear hormone receptors, the RARs and RXRs. However, in the eye, 11-cis retinal acts as the visual chromophore that can be activated by light leading to a visual response.

The active retinoids, RA and 11-cis retinal, are generated from a common precursor, retinol. RA is synthesized from retinol in a two-step process, first by a retinol dehydrogenase (RDH) to retinal, and further by a retinal dehydrogenase to RA. To generate 11-cis retinal, retinol is first isomerized to 11-cis retinol and then oxidized by RDHs to 11-cis retinal. To date 20 different RDHs have been identified. In this study we focused on the characterization and structure-function analysis of two RDHs to expand the knowledge of the cellular pathways generating active retinoids.

The murine RDH4 oxidizes cis-retinols into corresponding aldehydes. We revealed that RDH4 was in fact the murine homologue of the human RDH5, based on the amino acid similarity, substrate specificity and protein expression patterns. These enzymes are thought to have a role in chromophore generation in the visual cycle. Genetical evidence for the role of RDH5 in 11-cis retinal synthesis was provided from mutations found in the RDH5 gene that caused fundus albipunctatus, a stationary form of night blindness. Intriguing was that both RDH4 and RDH5 enzyme domains faced the endoplasmic reticulum (ER) lumen. Later we showed that CRAD1, an enzyme closely related to RDH5, had a similar membrane topology. Considering the similarity of the microsomal RDHs, this suggests a lumenal orientation for most if not all of the RDHs.

The analysis of the two RDHs, CRAD1 and RDH5, revealed that they had short cytosolic tails of 6-7 amino acids. To examine the significance of these tails, we generated mutants with deleted tails. A reporter assay was developed to measure the enzyme activities of RDHs in transfected cells. Strikingly the tail deletion mutants lacked enzymatic activity in the cellular reporter assay, while they retained their activity in vitro. This led us to examine the significance of the cytosolic tails for subcellular localization. Indeed, the deletion mutants had abnormal cellular localizations, indicating that the reporter assay detects intracellular alterations.

To map the essential amino acid residues, further mutations were generated in the cytosolic tail. Results generated with the reporter assay from a deletion series, alanine scan and chimeric proteins suggested that a conserved proline residue is important for enzymatic activity in vivo. The abolished activity of CRAD1 tail deletion mutant could not be rescued by relocating the enzyme to the ER with a well-characterized double lysine ER retention signal. Therefore, the CRAD1 tail might associated with other proteins necessary for enzyme activity in vivo.