Molecular characterization of a mouse short chain dehydrogenase/reductase active with all-trans-retinol in intact cells, mRDH1.

Metabolic activation of retinol (vitamin A) via sequential actions of retinol and retinal dehydrogenases produces the active metabolite all-trans-retinoic acid. This work reports cDNA cloning, enzymatic characterization, function in a reconstituted path of all-trans-retinoic acid biosynthesis in cell culture, and mRNA expression patterns in adult tissues and embryos of a mouse retinol dehydrogenase, RDH1. RDH1 represents a new member of the short chain dehydrogenase/reductase superfamily that differs from other mouse RDH in relative activity with all-trans and cis-retinols. RDH1 has a multifunctional catalytic nature, as do other short chain dehydrogenase/reductases. In addition to retinol dehydrogenase activity, RDH1 has strong 3alpha-hydroxy and weak 17beta-hydroxy steroid dehydrogenase activities. RDH1 has widespread and intense mRNA expression in tissues of embryonic and adult mice. The mouse embryo expresses RDH1 as early as 7.0 days post-coitus, and expression is especially intense within the neural tube, gut, and neural crest at embryo day 10.5. Cells cotransfected with RDH1 and any one of three retinal dehydrogenase isozymes synthesize all-trans-retinoic acid from retinol, demonstrating that RDH1contributes to a path of all-trans-retinoic acid biosynthesis in intact cells. These characteristics are consistent with RDH1 functioning in a path of all-trans-retinoic acid biosynthesis starting early during embryogenesis.

Naturally occurring retinoids (vitamin A and its metabolites) function beyond their well known contributions to vision, conception, growth, and epithelial differentiation. Vertebrates require retinoids for the development of numerous embryonic structures (e.g. limbs, nervous system, heart, kidney), the immune response, and control of intermediary metabolism (1)(2)(3)(4). Retinol metabolites serve as endocrine factors that bind to and activate the RAR␣, -␤, and -␥, and the RXR␣, -␤, and -␥ members of the nuclear hormone receptor superfamily (5,6). Mechanisms for the pleiotropic actions of retinoids are provided by the number of retinoid receptors, their many cell-specific isoforms (generated by differential promoter use and alternative splicing), and heterodimerization of RXR with several other nuclear receptors (7). Receptor expression patterns alone, however, do not explain fully the complex temporal and spatial effects of retinoids during embryogenesis and during postnatal development and growth (8).
Metabolism activates vitamin A (retinol) by producing atRA, 1 an endogenous RAR ligand generated both centrally and locally (9,10). Two sequential reactions produce atRA from all-trans-retinol; they are reversible and rate-limiting dehydrogenation into all-trans-retinal catalyzed by RDH and irreversible and perhaps rate-determining dehydrogenation of alltrans-retinal catalyzed by RALDH. Members of two classes of alcohol dehydrogenases have been proposed to serve as RDH (11,12). The medium chain dehydrogenases, ADH classes I and IV, convert retinol into retinal in vitro. These enzymes belong to a family that largely detoxifies xenobiotics and have kinetic characteristics more consistent with xenobiotic clearance than for producing endocrine factors (13). ADHI expression first appears at e10.5 and does not correlate well with sites of atRA synthesis in the mouse embryo, prompting the conclusion that its involvement appears unlikely in embryonic atRA biosynthesis (14). ADHIV shows more widespread expression than ADHI, but it is not expressed in diverse areas of atRA biosynthesis and use in the adult (15). ADHIV is expressed episodically during mouse embryogenesis; it is notably absent at various loci and times of atRA need. These considerations indicate that ADH1 and ADHIV do not function universally to generate endocrine levels of atRA.
Pursuit of universal RDH candidates has identified several previously unknown members of the SDR superfamily (11). This phylogenetically diverse enzyme family consists of numerous mammalian members that regulate the concentrations of estrogens, androgens, glucocorticoids, and prostaglandins (16). Three SDR isozymes, RoDH1, -2, and -3, have been identified as RDH candidates in the rat. Each shares similar substrate specificities and has greater activity for all-trans-retinol than for cis-retinols (17). Each has a distinct pattern of mRNA expression. The precise functions of each have not been established in part because mouse orthologs have not been identified. As for retinoid-specific receptors and binding proteins, an RDH physiologically significant in atRA biosynthesis should be expressed across species with high amino acid sequence homology and ligand/substrate specificity. Attempts to identify RDH in mice have resulted in the cloning and characterization of either cis-retinoid preferring SDR or retinal reductases rather than mouse RoDH orthologs (18 -22). This suggests that mouse enzymes significant to the atRA endocrine system have not been identified or that the mouse generates atRA fundamentally differently from the rat.
This work reports the cDNA cloning, enzymatic characterization, and mRNA expression patterns in tissues and cells of a mouse RDH, RDH1. RDH1 functions as a dehydrogenase for atRA biosynthesis in intact cells and has greater activity with all-trans-retinol than with cis-retinols, in contrast to the previously described mouse CRAD isozymes. RDH1 represents the all-trans-retinol dehydrogenase of most widespread mRNA expression in mouse starting as early as e7 and continuing expression in multiple tissues of the adult. RDH1 functions in intact cells in conjunction with any one of three RALDH isozymes to generate atRA from all-trans-retinol. No cDNA sequence for RDH1 has been found in any public data base, and reverse transcription-PCR with mouse embryo and adult mRNA using several sets of degenerate primers failed to identify related SDR with activity for all-trans-retinol. These data suggest that RDH1 contributes to atRA biosynthesis in the mouse starting early during gestation and that the rat and mouse may differ in SDR use for RA biogeneration.

MATERIALS AND METHODS
cDNA Encoding Mouse RDH1-Degenerate primers for reverse transcription-PCR were designed from conserved amino acid sequences of rat RoDH and mouse CRAD members of the SDR superfamily. Reverse transcription was done with the Moloney murine leukemia virus reverse transcriptase system (Life Technologies, Inc.) with random hexamers primer (Promega) and mouse e17 mRNA. PCR used 1 cycle at 94°C for 2 min, 35 cycles at 94°C for 30 s, 52°C for 30 s, 72°C for 1 min, followed by 72°C for 7 min. The ϳ550-bp PCR product was gel-purified and cloned into pGEM-T (Promega).
A cDNA of the coding region was amplified with sense primer F8 (5Ј-GCAGCTGTGTAAACCATGTGGC) and antisense primer R4 (5Ј-CTCTCTCACACTCTACCTACATC) using mouse e17 mRNA (Sigma) reverse transcription product. Reverse transcription was done with the SUPERSCRIPT TM II RNase H Ϫ reverse transcriptase system (Promega) according to manufacturer's protocol. The PCR program had 1 cycle at 94°C for 1.5 min; 35 cycles at 94°C for 40 s, 61°C for 1 min, and 72°C for 1.5 min; followed by 72°C for 5 min. This PCR product was used as template with the sense primer F9 (5Ј-CCGGAATTCGCCGC-CACCATGTGGCTCTACC) (underlining indicates the EcoRI site; double underlining indicates the Kozak sequence) and the antisense primer R5 (5Ј-CCGCTCGAGTCAGAGGGCTTTCTCA) (the XhoI site is underlined). The PCR product was gel-purified, digested with EcoRI and XhoI, and cloned into the EcoRI/XhoI sites of pcDNA3 (Invitrogen) to construct pcDNA3/RDH1. Expression of RDH1-CHO-K1 cells (ATCC) were cultured at 37°C in Ham's F-12 medium supplemented with 10% fetal calf serum. Cells were transfected with 8 g/100-mm plate of pcDNA3/RDH1 using Li-pofectAMINE (Life Technologies, Inc.) and harvested 24 h later. Cells were lysed by sonication in 20 mM Hepes, 150 mM KCl, 1 mM EDTA, 10% sucrose, and 2 mM dithiothreitol, pH 7.5. The supernatant obtained from centrifuging the lysate at 800 ϫ g for 10 min was used for enzyme assays. Similar supernatants from CHO-K1 cells transfected with pcDNA3 served as controls. Protein concentrations were determined by the method of Bradford (23). Alternatively, CHO-K1 cells were transfected with combinations of pcDNA3, pcDNA3/RDH1, and pcDNA3/ AHD2 or pcDNA3/RALDH2 or pcDNA3/ALDH6. Twenty-four h after transfection the medium was replaced with 5 ml of fresh medium. All-trans-retinol was added in 5 l of Me 2 SO, and incubation was continued for 30 min. Cells and medium were analyzed by HPLC to quantify atRA (24).
Enzyme Assays-Kinetic data were generated from retinoids with 5 g of CHO cell supernatant protein for 5 min and with steroids using 1 g of protein and 3 min. These values were in the linear ranges of protein and time (initial velocity conditions) and produced easily quantifiable amounts of products. Kinetic data were fit with the nonlinear regression program GraphFit 4. RDH assays were done in triplicate at 37°C in 0. 25  The aqueous phase was extracted twice with 2.5 ml of hexane. The organic phase was evaporated with nitrogen, and residues were dissolved in 0.1 ml of hexane. Retinal oximes were quantified by HPLC as described (18). Steroid dehydrogenase assays were done with [ 3 H] steroids (40 -101 Ci/mmol, 20,000 dpm/reaction) under the same conditions. Reactions were quenched with dichloromethane (3 ml), and steroids were extracted as described (18). Organic phases were evaporated with nitrogen, and the residues were dissolved in 30 l of ethanol and applied to 1) aluminum oxide thin-layer chromatography plates developed with chloroform/ethyl acetate (3/1, v/v) to analyze conversion of androsterone into androstanedione, 3␣-adiol into dihydrotestosterone, dihydrotestosterone into androstanedione, and testosterone into androstanedione; 2) silica plates developed with the same mobile phase to analyze conversion of ␤-estradiol into estrone; 3) silica plates developed with chloroform/ethanol (96/4; v/v) to analyze conversion of corticosterone into 11-dehydrocorticosterone. [ 3 H]Steroids were detected by autoradiography. Radioactive zones were excised and counted with a liquid scintillation counter.
Northern Blots-Mouse liver and kidney total RNA were extracted with Trizol reagent (Life Technologies, Inc.), and ϳ25 g from each were run on a formaldehyde gel, transferred to a nylon membrane, and probed with the 3Ј-UTR probe described below. Prehybridization was done in 10 ml of prehybridization solution (50% formamide, 5ϫ saline/ sodium phosphate/EDTA, 10ϫ Denhardt's solution, 2% SDS, and 100 g/ml denatured salmon sperm DNA) at 42°C overnight. Hybridization was done with labeled cDNA probe at 42°C for 24 h. The blot was washed three times in 1ϫ SSC (1ϫ SSC ϭ 0.15 M NaCl and 0.015 M sodium citrate) with 0.1% SDS for 30 min at 42°C followed by 2 washes in 0.1ϫ SSC with 0.1% SDS for 30 min at 65°C and was exposed to x-ray film. After stripping the mRDH1 probe, the blot was reprobed with a ␤-actin probe (CLONTECH).
Northern blots also were done with the mouse Multiple Tissue Northern blot, which contains 2 g of poly(A ϩ ) RNA per lane on a Nylon membrane (CLONTECH). Two cDNA probes were hybridized to the blot separately according to the manufacturer's protocol. One probe was the same 3Ј-UTR probe described below. The second was a 5Ј-end probe corresponding to nucleotides 6 -298. Both hybridizations were done at 68°C for 1 h. Each time the blot was washed three times in 2ϫ SSC with 0.05% SDS at room temperature for 30 min followed by two washes in 0.1ϫ SSC with 0.1% SDS for 40 min at 50°C and exposed to x-ray film overnight. Blots were reprobed with ␤-actin.
RNA Dot Blots-A 3Ј-UTR probe corresponding to nucleotides 2392-2679 was labeled with 32 P using the RadPrime DNA-labeling system (Promega). The probe was purified by a Micro Bio-Spin 30 column (Bio-Rad) and hybridized with a mouse RNA Master Blot TM according Boldface also denotes 21 amino acids conserved in at least 70% of SDR. Boxed areas indicate the conserved motifs characteristic of SDR. The underlined TAA in the 5Ј-UTR shows the in-frame stop codon before the first ATG.
to the manufacturer's protocol (CLONTECH). Briefly, the membrane was prehybridized for 30 min at 65°C in 10 ml of Expresshyb solution. Hybridization was done at 65°C overnight with a cDNA probe mix that included 30 g of Cot-1 DNA and 150 g of sheared salmon testes DNA. The blot was washed 5 times in 2ϫ SSC (0.3 M NaCl, 0.03 M sodium citrate) with 1% SDS at 65°C for 20 min followed by two 20-min washes in 0.1ϫ SSC with 0.5% SDS at 55°C for 20 min. The dot blot was exposed to x-ray film at Ϫ70°C for 48 h with an intensifying screen.
In Situ Hybridization-Mouse embryos (NIH Swiss, Harlan Sprague-Dawley, Indianapolis, IN) were used. The morning of plug detection was designated e0.5. In situ hybridization was done on paraformaldyhyde-fixed, paraffin-embedded tissue sections using 35 Slabeled sense and antisense riboprobes as described (25). The RDH1 probe spanned nucleotides 1-1199. After their development, tissue sections were counterstained with propidium iodide, and images were documented under dark-field and epifluorescent illumination.

RESULTS AND DISCUSSION
cDNA Cloning-To obtain mouse RDH1, degenerate reverse transcription-PCR was done with mRNA from e17 mice. Prim-ers were designed from conserved regions of RoDH/CRAD members of the SDR superfamily. A 550-bp cDNA (fragment A) amplified by primers F1 and R1 had 89 and 85% nucleotide identity, respectively, with mouse CRAD1 and rat RoDH2 (Fig.  1). The sequence of this partial cDNA was used to design primers for 5Ј-and 3Ј-RACE. The two 5Ј-RACE primers, R2 and R3, produced a ϳ700-bp product (fragment B). The two 3Ј-RACE primers, F2 and F3, produced a 900-bp product (fragment C). Additional primers, F4 and F5, extended the 3Ј-UTR ϳ1.5-kb (fragment D), whereas primers F6 and F7 extended the cDNA out to poly(A) mRNA (fragment E). Primers F8 and R4 then were used to amplify a 1020-bp cDNA with a 951-base pair open reading frame that encoded 317 amino acid residues (fragment F). The insert to construct pcDNA3/RDH1 was amplified from fragment F with primers F9 and R5. None of these FIG. 3. RDH1 substrate recognition. Substrate screening was done with 800 ϫ g supernatants of transfected CHO-K1 cells. A, activity with all-trans-retinol (striped bars) or 9-cis-retinol (solid bars). Conditions were as follows. 1, pcDNA3 (mock transfection); 2, pcDNA3/ RoDH1 (rat RoDH1); 3-5, pcDNA3/mRDH1. Cofactors used were as follows. 1-3, 2 mM NAD ϩ ; 4, 2 mM NADP ϩ ; 5, none. Assays were done for 15 min with 50 g of protein and 5 M substrate. B, steroid dehydrogenase characteristics of RDH1 are as follows. 1, 3-adiol (two products were observed, the more abundant was DHT; the less abundant was androstanedione); 2, androsterone; 3, DHT; 4, testosterone; 5, estradiol; 6, corticosterone. Assays were done for 5 min with 2 g of protein, 2 M substrate, and 2 mM NAD ϩ .  constructs nor any of the fragments was found in any public data base.
To determine whether additional all-trans-retinol recognizing SDR occur, five primer pairs (F1/R1; F1/R7; F10/R1; F11/ R1; F11/R7) were used independently for reverse transcription-PCR with e17 mouse embryo mRNA. Note that although the PCR conditions were not varied, two different degenerate forward primers, F1 and F11, were designed from the "signature" SDR motif (GXWGXVNNAG). In addition, two different degenerate reverse primers, R1 and R7, were designed from the same RoDH/CRAD motif (PRTXYSXGWD). A third forward primer, F10, was designed from the SDR cofactor binding motif (FIT-GCDSGFG). Seventy-three of the clones generated were sequenced, 13-16 produced by each primer pair. Thirty-seven of the 73 clones sequenced encoded SDR. Only three encoded RDH1. Twenty-one encoded CRAD2; two encoded 17␤-HSD9; two encoded RDH4; five encoded two previously unknown SDR with no activity for retinol 2 ; two encoded a previously unknown enzyme, CRAD3, with activity for 9-and 11-cis-retinol but little or no activity for all-trans-retinol. 3 To probe further for additional RDH candidates, nested PCR was done with mRNA from e7 and e11 embryos, embryonic heart, adult testis, and adult small intestine. The first round was done with the primer mixture F10 (designed from the SDR cofactor binding motif), R1 and R9. The two degenerate reverse primers, R1 and R9, were designed from the CXEHALTX sequence conserved in RoDH/CRAD. The second round was done with the mixture of degenerate primers: F10, F11, R8, and R9. Sixty-seven of the clones produced were sequenced. RDH1 was amplified from e7 embryo (5 of 30 clones sequenced), e11 embryo (4 of 10 clones sequenced), and testis (2 of 10 clones sequenced). RDH1 was not amplified from embryonic heart or from adult small intestine. RDH4, a 9-and 11-cis-retinol-metabolizing SDR, accounted for 22 of the 30 clones sequenced from e7 mice, 5 of the 10 sequenced from e11 mice, and all 10 sequenced from embryonic heart. The remaining clones sequenced encoded CRAD2, CRAD3, and 17␤-HSD9. No other SDR were found that were active with all-trans-retinol.
Deduced Amino Acid Sequence-The protein deduced from pcDNA3/RDH1 has a calculated molecular mass of ϳ36 kDa and includes 6 peptides with high sequence similarity (73-100%) to the 6 motifs characteristic of SDR (26). Nineteen of the 23 amino acids conserved in ϳ70% of SDR occur within these 6 motifs. These motifs include the conserved cofactor binding region, G 36 X 3 GXG, and catalytic residues S 164 X 11 YX 3 K (Fig. 2). The deduced amino acid sequence also shares similarity with other members of the RoDH/CRAD subfamily of SDR but less similarity with PR-RDH, retSDR, 3␤-HSD, and 11␤-HSD (Table I). RDH1 has an 18-amino acid hydrophobic sequence at its N terminus bounded by four hydrophilic amino acids, R 19 ERQ, in common with other SDR that metabolize retinoids/steroids (17). This sequence targets the enzymes to the smooth endoplasmic reticulum and is essential for enzymatic activity. 4 Enzymatic Activity-The supernatant from RDH1-transfected CHO cells was screened for enzyme activity with arbitrarily chosen amounts of protein and time. Under these conditions, the supernatant generated 745 Ϯ 90 pmol of all-transretinal, 292 Ϯ 6 pmol of 9-cis-retinal (mean Ϯ S.D., n ϭ 3) (Fig.  3), and no detectable 13-cis-retinal (not shown) from 5 M all-trans-, 9-cis-, or 13-cis-retinol and 2 mM NAD ϩ , respectively. RDH1 had 8 -11-fold higher activity with NAD ϩ versus NADP ϩ and had greater activity with all-trans-retinol than rat RoDH1, assuming equivalent transfection efficiencies. Both mouse RDH1 and rat RoDH1 had higher activity with all-trans-retinol versus 9-cis-retinol, but RoDH1 showed greater discrimination. RDH1 reached maximum activity at pH 8 and retained the same activity at pH values of 8.5 and 9. Activity at pH 7.5 was 54% that at pH 8 (data not shown). The most active steroid substrate tested was 3-adiol. RDH1 first converted 3-adiol via 3␣-HSD activity into DHT (290 Ϯ 73 pmol) and then converted the DHT via 17␤-HSD activity into androstanedione (35 Ϯ 4 pmol). The second most active steroid substrate tested was androsterone, which RDH1 converted via 3␣-HSD activity into androstanedione (175 Ϯ 15 pmol). 17␤-HSD activity of RDH1 was much less than 3␣-HSD activity, as shown by much lower catalysis of DHT, testosterone, and estradiol into their 17-oxo products androstanedione (44 Ϯ 14 pmol), androstanedione (33 Ϯ 10 pmol), and estrone (26 Ϯ 6 pmol), respectively. 11␤-HSD activity with corticosterone was detectable, but low, producing dehydrocorticosterone (19 Ϯ 2 pmol).
RA Synthesis in Intact Cells-To determine whether RDH1 contributes to RA biosynthesis in intact cells, cotransfections were done with RDH1 and three mouse RALDH isozymes, AHD2, RALDH2, and ALDH6 (27)(28)(29)(30)(31)(32). Mock transfections with pcDNA3 or transfection with RDH1 alone or with any one of the RALDH cDNAs alone produced no detectable RA from all-trans-retinol (data not shown). Combination of RDH1 with any one of the RALDH isozymes generated atRA from all-transretinol (Fig. 5). RA biosynthesis increased with increasing amounts of transfected DNA; small increases in vector amounts resulted in substantial increases in RA biosynthesis. Mouse RALDH2 seemed somewhat more efficient than the mouse ortholog of human ALDH6, whereas AHD2 was less efficient. Consistent with its activity in whole cells, purified recombinant RALDH1/AHD2 functions about 20% as efficiently (V/K 0.5 ) as RALDH2 in vitro (11). Thus, these results reflect the V m /K 0.5 values of the three RALDHs (relative values of 100, 70, and 20 for RALDH2, ALDH6, and AHD2, respectively), demonstrating the predictive potential of RALDH enzymology in vitro (11,32) and suggesting that the activity differences shown here did not stem from transfection efficiency differences.
mRNA Expression-Northern blotting revealed a major ϳ4.2-kb transcript in mouse liver and kidney 24 h after expo- sure to a 3Ј-UTR cDNA probe (Fig. 6). Longer exposure (48 h) revealed 1.35-and 0.7-kb transcripts. RDH/CRAD tend to have high nucleotide sequence homology, sometimes even in the 3Ј-UTR, including CRAD1 and RDH1. Indeed, we have reported that CRAD1 has transcripts of 4.4, 3.5, 3.0, and 2.7 kb in mouse liver (18). When we noticed the homology between CRAD1 and RDH1 in nucleotides 2900 -3459 of the 3Ј-UTR (RDH1 numbering), mouse liver was reanalyzed with a CRAD1 probe that does not share homology with RDH1. Use of the CRAD1 unique probe revealed only transcripts of 4.4, 3.5, and 2.7 kb. To minimize the possibility that the RDH1 3Ј-UTR probe produced data through cross-reaction with a heterologous SDR although it does not share homology with CRAD1, we probed a mouse Multiple Tissue Northern blot (CLONTECH) independently with the 3Ј-UTR probe and with a 5Ј-end RDH1 cDNA probe. Both probes produced identical results, showing intense 4.2-kb transcripts and less intense transcripts of 1.35 and 0.7 kb in liver, testis, kidney, heart, brain, and in e7, e11, e15, and e17 day mice (data not shown).
Dot blot analysis with an RNA master blot confirmed and extended the Northern blot data (Fig. 7). RDH1 mRNA expression was detected in all tissues examined. Expression was intense in pancreas, testis, skeletal muscle, brain, heart, submaxillary gland, liver, and kidney. Expression was detected in eye, lung, smooth muscle, thyroid, thymus, ovary, prostate, epididymis, and uterus. E7, e11, e15, and e17 mice expressed RDH1, in agreement with the Northern blot data.
Cellular mRNA Expression Patterns-In situ hybridization results were consistent with the Northern and dot blot data (Fig. 8). RDH1 transcripts were most abundant during stages of early organogenesis in the mouse, from e7.5 to e10.5. RDH1 mRNA was expressed throughout the embryo and extraembryonic regions, with particular enrichment within the neural plate at e7.5 and within the neural tube, gut, neural crest, and Rathke's pouch at e10.5. By e14.5, RDH1 expression reflected its adult pattern. Transcripts were abundant in the developing eye, ventral neural regions, cartilage, liver, and lung but appeared relatively less intense elsewhere. Sense controls were done at all stages, and all gave background signals.
Concluding Summary-This report identifies the first alltrans-retinol RDH/SDR candidate in the mouse and suggests that mouse and rat may differ in the initial step of RA biosynthesis from retinol. An enzyme physiologically significant in atRA biosynthesis should be expressed by all vertebrates and should be well conserved among orthologs. Attempts to identify mouse SDR candidates for atRA biosynthesis, however, have not identified all-trans-retinol-favoring enzymes and instead have revealed SDR more efficient with cis-retinols than all-trans-retinol, including CRAD1, -2, and -3 and RDH4 (18,19,21). Failure to identify mouse RDH with preferential activity for all-trans-retinol cast doubt on the universal importance of SDR to RA biosynthesis. Identifying an enzyme in the mouse that catalyzes retinal production from retinol represents a crucial step to understanding atRA biosynthesis, provides opportunity to take advantage of the studied embryonic development of the mouse, and provides a convenient knock-out opportunity.
No other efficient all-trans-RDH have been identified in mouse either here in our analysis of multiple clones produced from diverse templates with different sets of redundant primers or in previous work, and no public data base includes mouse RDH1 or portions thereof. Yet multiple cis-retinoid RDH and retinal reductases have been identified (Table II). Mouse 17␤-HSD9 has weak all-trans-retinol RDH activity, ϳ5% of RDH1, but as its name implies, is closer phylogenetically to 17␤-HSD than to RDH. In contrast to 17␤-HSD9, RDH have little or no detectable 17␤-HSD activity. Three rat RoDH SDR isozymes have been identified that manifest their most efficient RDH activity with all-trans-retinol, and their cDNAs have been cloned (34 -36). Two, RoDH1 and RoDH2, show widespread mRNA, but only liver expresses the third, RoDH3. These were the first SDR identified as candidates for contributing to the physiological path of RA biogenesis. Mouse RDH1 differs from the rat enzymes in several respects. Three closely related rat enzymes occur with amino acid similarities between 88 and 99%. The rat enzymes were detected by Northern blot only in liver and were detected in extra-hepatic tissues with RNase protection assays or in situ hybridization (34 -36, 37), in contrast to the widespread mRNA expression of RDH1 detectable by Northern blot. Potential human homologues also have been cloned using probes from RoDH1 (38,39). A human SDR, hRDH-E, shares ϳ80% amino acid similarity with the three rat isozymes but has expression limited to liver and epidermis. A second putative human RoDH shares less amino acid similarity with the rat enzymes (ϳ72%) and may function primarily in steroid metabolism (33,40). So far only RDH-E and the RoDHlike HSD have been identified in human. Future work should clarify RDH that are othologous in rat, mouse, and human and should confirm whether additional RDHs occur in mouse.
Factors consistent with RDH1 functioning to biosynthesize atRA include its expression during early embryogenesis and in all known vitamin A target tissues probed and its ability to provide retinal to each of the RALDH isozymes for the biosynthesis of atRA in intact cells.