| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Nutritional Science and Toxicology, University of California, Berkeley, California, USA
3Correspondence: Nutritional Science and Toxicology, 119 Morgan Hall, MC#3104, University of California, Berkeley, CA 94720-3104, USA. Email: jna{at}berkeley.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: retinoic acid • retinol • short-chain dehydrogenase/reductase
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4
5)
. Biosynthesis of atRA from retinol proceeds through two steps: reversible retinol conversion into retinal, followed by irreversible retinal oxidation into atRA (6)
. Interest in SDRs (short-chain dehydrogenase/reductase) as retinol dehydrogenases involved in atRA biosynthesis was prompted by the demonstration that microsomes produce retinal when presented with either holo-CRBP (retinol bound to cellular retinol binding protein, type I), the physiological form of intracellular retinol, or "free" retinol (7)
, and the complementary demonstrations that the rates of retinal generation from holo-CRBP by microsomes supercede those in cytosol by 10- to 50-fold and that microsomes harbor 80–90% of retinal generating capacity (enzyme units) from holo-CRBP (8)
. Holo-CRBP, but not apo-CRBP, crosslinks to microsomal SDR only in the presence of pyridine nucleotide cofactor, revealing an ordered bisubstrate relationship between the holo-binding protein and SDR (9)
. This enabled cloning of rat Rodh2 and its mouse ortholog Rdh1 (10
, 11)
. Continuing research into SDRs has identified numerous family members with retinoid metabolizing activity in vitro, several of which are potential all-trans-retinol dehydrogenases in human, rat, and mouse (12
13
14
15
16
17
18
19)
. Mouse Rdh1/rat Rodh2 present promising candidates for first-step enzymes, because they have the highest efficiency (Vm/Km) of any SDR with all-trans-retinol, contribute to a reconstituted path of atRA biosynthesis in intact cells, and exhibit widespread expression starting as early as embryo day (e) 7.5 (11)
.
Identification, cloning, and characterization of four retinal dehydrogenase (Raldh) isozymes have provided crucial insight into the enzymology of the second step of atRA biogenesis and enabled physiological studies of Raldh functions (20
21
22
23
24)
. Gene knockout studies have been reported for three of the four Raldhs (25
26
27)
. Raldh1-null mice show no severe phenotype but have greater endogenous retinol and retinyl esters (RE) levels than WT (wild type), suggesting retinoid sparing. Raldh3-null mice are born in Mendelian frequency but die by postnatal day 8 from choanal atresia caused by persistence of nasal fins, which leads to respiratory distress. Raldh2-null mice have the most severe phenotype, dying in midgestation without forming limb buds, fully elongating the anterioposterior axis, or fully developing their hearts and frontonasal regions. These combined biochemical, physiological, and genetic data have provided broad insight into atRA biosynthesis during embryogenesis.
Although vertebrate development requires atRA, excess causes teratogenic effects, i.e., RA embryopathy (28
29
30)
. Cyp26A1, 26B1, 26C1, and 2C39 impart additional control over atRA effects by catalyzing its degradation (31
32
33
34
35
36)
. For example, mice deficient in cyp26a1 die in mid to late gestation with numerous defects, including spina bifida, and abnormalities in the kidneys, cervical vertebrae, and rostral hindbrain, which resemble the teratogenic effects of atRA (37
38
39)
.
Here we show that rdh1 functions in retinoid homeostasis in vivo under physiological conditions. Rdh1-null mice spare retinol and RE, have decreased liver cyp26A1 mRNA, and grow larger with greater adiposity than WT mice, when restricted in dietary vitamin A. Rdh1 and cyp26a1 mRNA have inversely related expression in the mouse embryo from e7.5 to e18.5 and in liver from e12.5 to P2M, suggesting paired modulation of atRA homeostasis. These are the first genetic data consistent with a candidate retinol dehydrogenase affecting vitamin A-related function or gene expression and reveal an unexpected contribution of Rdh1 to size and adiposity.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 kb upstream of exon 1. A neomycin cassette (pGK-Neo) flanked by LoxP sites was inserted into an AclI site 800 base pairs downstream of exon 1. The thymidine kinase cassette (pKO-TK, Stratagene) was cloned into the 3'-end of the targeting construct as a negative selection marker. The targeting construct was linearized with NotI and electroporated into E14 embryonic stem cells (129SVJ background). Cells were selected with G418 and ganciclovir. Homologous recombination was verified by Southern blotting with flanking probes P1 and P2. A pCMV-Cre plasmid was electroporated into ES cell clones to generate ES cells with either complete or conditional knockouts. Southern blotting with probe P3 identified desired clones. ES cell clones containing complete knockout alleles were injected into C57BL/6J blastocysts.
Genotyping
Genotyping was done with tail tip DNA by PCR and/or Southern blot. Samples were lysed overnight at 50°C with proteinase K (5 µl of 10 mg/ml) in 150 µl of 0.45% Nonidet P-40 and 0.45% Tween-20 added to an autoclaved buffer of 50 mM KCl, 10 mM Tris, 2 mM MgCl2, and 0.1 mg/ml gelatin, pH 8.0, and then were denatured (95°C for 10 min) and centrifuged (12,000 g for 10 min). Supernatants (1 µl) were used for PCR. The WT allele was amplified with forward and reverse primers, 5'-CCCTCTTGAAGTAAGAAGGTA-3' and 5'-ACTCCCAGAGTAGATATGAG-3', respectively, to produce a 1.3 kb product. The null allele was amplified with forward and reverse primers 5'-CCCTCTTGAAGTAAGAAGGTA-3' and 5'-CATTGTTGATGGGATTGCAAGC-3', respectively, to produce a 0.5 kb product. For Southern blot, genomic DNA was prepared with DNeasy Tissue Kit (Qiagen, Valencia, CA, USA), digested with EcoRI, and probed as described above.
Reverse transcription PCR and real-time quantitative PCR
Total RNA was isolated with Trizol reagent (Invitrogen, Carlsbas, CA, USA) and treated with DNase treatment and removal reagent (DNA-free; Ambion, Austin, TX, USA). RNA (1.5 µg) was reverse transcribed with SuperScript III kit (Invitrogen), using oligo(dT)20 for initiating cDNA synthesis. TaqMan quantitative real-time PCR assays were done with a Perkin-Elmer ABI PRISM 7900 sequence detection system (SDS), using SDS version 2.0 software (Applied Biosystems, Foster City, CA, USA). Primers and fluorogenic TaqMan probes were designed with Primer Express version 2.0 software and synthesized by custom services (Applied Biosystems). Sequences of primers, probes, and PCR conditions are available on request.
Animals and diets
Male chimeras were bred with C57BL/6 females to establish strains with a mixed genetic background (129/C57). These heterozygotes were either bred to produce homozygous mice with hybrid backgrounds or backcrossed with the C57BL/6 strain for five generations to obtain a C57BL/6 background. Experiments were done with backcrossed mice, unless noted otherwise. Dams were placed on one of five diets at mating, and pups were weaned onto the same diets as their dams: stock mouse chow (
30 IU vitamin A/g) or purified American Institute of Nutrition (AIN) 93G diets with 4, 0.6, 0.1 IU vitamin A/g or a vitamin A-deficient diet (VAD) with no added vitamin A.
Retinoid analyses
Retinol, RE, and atRA were quantified as described, with modifications (40)
. Briefly, tissues were harvested under yellow light, placed on ice immediately, minced, and homogenized in cold phosphate-buffered saline. One milliliter of 0.025 N KOH/ethanol was added to 0.5 ml of 25% homogenate, followed by extraction with 10 ml hexane. The top (hexane) layer containing retinol and RE was removed, and the solvent was evaporated with nitrogen. Residues from kidney, brain, testis, and serum were dissolved in 150 µl acetonitrile. Liver residues were dissolved in 250–500 µl methanol or 500-1000 µl 2-propanol, because of high retinol and RE concentrations. Retinol and RE were resolved by reverse-phase HPLC and quantified by their UV absorbance at 325 nm. The HPLC column was eluted at 1 ml/min with 11% water/methanol for 4.5 min and then by a linear gradient for 3.5 min to 100% methanol. The 100% methanol wash was maintained for 0.5 min, and then a linear gradient to 5% methanol/95% dichloroethane was generated over 2 min. These conditions were held for an additional 7 min. Retinol eluted at 4.1 min, and RE eluted at 11.4 min. The injection volume was 100 µl for all samples, with the exception of liver samples for RE analysis. To quantify liver RE accurately, a separate 10 µl injection was done. To recover RA, 4 N HCl was added to the aqueous phase (bottom layer) and RA was extracted with another 10 ml of hexane. The solvent was evaporated under a stream of nitrogen, and the residues were dissolved in 60 µl of acetonitrile. RA was quantified by a liquid chromatography/tandem mass spectrometry assay (LC/MS/MS; ref 40
).
Histology
Tissues were placed in cold PBS (pH 7.4), fixed in 4% paraformaldehyde overnight at 4°C, embedded in paraffin, and sliced into 6 µm sections. Slides were stained with hematoxylin and eosin (H&E) and mounted in Permount (Fisher Scientific, Pittsburg, PA, USA).
Carcass analyses
Carcasses were processed for water, total fat, and ash in the laboratories of Craig Warden and Judy Stern, UC-Davis, as described previously (41)
.
Food intake
Food intake per cage was recorded every 2 days for 4 wk for three cages of 6 month-old WT (4–5 per cage) and rdh1-null (2–4 per cage) male mice fed a VAD diet.
Statistics
Data are presented as means ± SE and were analyzed using two-tailed, unpaired Student's t tests in GraphPad Prism 4 or Microsoft Excel.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
12 kb on mouse chromosome 10D3 (42)
. Exon 1 encodes the N-terminal membrane-targeting domain, essential for anchoring Rdh1 in the endoplasmic reticulum facing the cytoplasm. Exon 1 also houses the ATG start site and the cofactor-binding site. All three are crucial for producing an active enzyme (43
, 44)
. We designed the targeting vector to delete the proximal promoter region and exon 1 (Fig. 1
A) and used PCR (data not shown) and Southern blotting to confirm the outcome (Fig. 1B
). Positive ES cell lines (129SVJ) were transfected with pCMV-Cre to delete targeted regions. We selected two types of Cre-treated ES cells: ones with complete null alleles (Fig. 1C
, arrowhead) and others with conditional null alleles (Fig. 1C
, arrow). ES cell clones with complete null alleles were injected into C57BL/6 blastocysts to generate chimeras. Male chimeras with germ line expression were bred with C57BL/6 females to establish strains with a mixed genetic background. Heterozygous mice were either intercrossed to produce homozygous mice with a hybrid background or backcrossed with C57BL/6 strain mice for five generations. Heterozygous backcrossed mice were intercrossed to produce homozygous null offspring. Loss of both copies of the WT rdh1 allele was confirmed by PCR (data not shown) and Southern blot hybridization (Fig. 1D
). We confirmed lack of rdh1 mRNA in liver, kidney, and testis by reverse transcriptase (RT)-PCR (Fig. 1E
).
|
Rdh1 inactivation does not impair embryonic development
Rdh1-null mice fed diets containing vitamin A in marginal (0.6 IU vitamin A/g diet), recommended (4 IU vitamin A/g diet), or copious (30 IU vitamin A/g diet) amounts were born in Mendelian frequency (Table 1
; ref 45
). Overall, WT and null littermates seemed indistinguishable in initial appearance. Virgin rdh1-null mice produced the same number and sex distribution of offspring as WT mice (data not shown). Breeding male and female rdh1-mull mice produced apparently healthy offspring. H&E staining revealed no obvious differences in the morphologies of multiple tissues in the rdh1-null mice compared to WT (Fig. 2
).
|
|
Rdh1 inactivation spares endogenous retinol and RE
We quantified steady-state concentrations of RE, retinol, and atRA in tissues of male and female rdh1-null and WT mice fed diets with 0.6, 4, and 30 IU/g of vitamin A at 6–7 wk of age. No differences occurred in endogenous atRA between male rdh1-null and WT littermates in serum and the tissues assayed, regardless of vitamin A content of the diet, measured with an LC/MS/MS assay specific for atRA and sensitive to 5 fmol. Nor were differences noted between the two genotypes in tissue retinol or RE concentrations in liver (Fig. 3
), kidney, brain, or testis (not shown) of male mice fed a diet with 30 IU vitamin A/g.
|
Rdh1-null male mice fed a diet with 4 IU vitamin A/g had liver retinol levels 2-fold greater than WT littermates but showed no differences in kidney, brain, or testis levels (not shown) and displayed no differences in RE of liver (Fig. 3)
or other tissues (not shown) relative to WT. Rdh1-null mice fed marginal dietary vitamin A, i.e., 0.6 IU/g, had 40% higher RE levels in liver, 2-fold greater retinol in liver, and 40% greater retinol in kidney relative to WT but had no differences in brain or testis. Results similar to these were obtained with female mice fed diets with all three levels of vitamin A (not shown). Therefore, inactivation of rdh1 spares liver retinol, when diet contains the recommended amount of vitamin A, and spares both retinol and RE, with diets restricted in vitamin A.
Effect of rdh1 inactivation on liver cyp26a1 mRNA and protein
CYP26A1, B1, C1, and 2C39 contribute to vitamin A homeostasis by catabolizing atRA (31
32
33
34
35
36)
. Therefore, we determined whether decreases in their expression helped compensate for loss of rdh1. Cyp26a1 mRNA expression decreased 2.5-fold in livers of rdh1-null mice fed the 4 IU vitamin A/g diet, relative to WT littermates (Fig. 4
A). Cyp26b1 and cyp2c39 expression did not change. The mRNA of a fourth RA inactivation enzyme, cyp26c1, was not detected in mouse liver. Western blot showed a corresponding 2.5-fold decrease in Cyp26A1 in rdh1-null mice (Fig. 4B
).
|
Quantitative real-time PCR confirmed lower expression of cyp26a1 in livers of rdh1-null mice fed the 4 IU vitamin A/g diet, extended the observation to mice fed the 0.6 IU/g diet, and revealed an effect of dietary vitamin A on cyp26a1 expression (Fig. 4C
). In both rdh1-null and WT mice, increasing dietary vitamin A from 0.6 to 4 IU/g caused 8-fold increases in liver cyp26a1 mRNA, i.e., maintained expression differences between null vs. WT mice. Cyp26a1 mRNA expression in kidney, brain, and testis did not differ significantly between rdh1-null and WT mice fed the 4 IU/g diet (not shown), consistent with the retinoid concentration data, which revealed liver as the primary site affected by rdh1 inactivation in mice fed 4 IU vitamin A/g diet. Feeding the copious vitamin A diet (30 IU/g) induced a 160-fold increase in cyp26a1 mRNA in the rdh1-null mice, and a 60-fold increase in the WT mice—resulting in indistinguishable expression between both genotypes. Quantitative real-time PCR also showed that the amount of vitamin A in the diet affects expression of cyp26b1 and confirmed that rdh1 inactivation does not.
Relative expression of rdh1, cyp26a1, and cyp26b1
We assessed rdh1 expression relative to cyp26a1 and cyp26b1 expression in mice fed the 4 IU vitamin A/g diet by RT-PCR and quantitative real-time PCR. The mouse embryo expressed rdh1 mRNA weakly on e7.5 but increased expression continuously until e18.5. In contrast, the embryo expressed cyp26a1 intensely on e7.5 but decreased expression through e18.5. Cyp26b1 mRNA expression did not correlate with either rdh1 or cyp2al expression (Fig. 5
A). Real-time quantitative PCR confirmed an inverse relationship between rdh1 and cyp26a1 mRNA in the embryo (Fig. 5B
).
|
Mouse liver expressed rdh1 mRNA most intensely on e12.5, the first day that a well-formed liver can be dissected, but expressed cyp26a1 weakly (Fig. 5C
). Liver cyp26a1 expression increased markedly after birth, with the largest change noted between 1 and 2 months. Once again, cyp26b1 mRNA expression in the liver did not correlate with either rdh1 or cyp26a1 expression. Real-time quantitative PCR confirmed an inverse relationship between rdh1 and cyp26a1 mRNA signals in liver (Fig. 5D
).
Effects of rdh1 inactivation on growth
We next determined whether rdh1 deficiency influenced growth, a prototypical function of vitamin A. Male and female rdh1-null mice raised on a diet with copious vitamin A (30 IU/g) did not differ statistically in weight gain with age relative to WT (Fig. 6
A). Male and female rdh1-null mice produced by dams fed either a diet low in vitamin A (0.1 IU/g) or a VAD diet since mating, and then weaned onto the same diet, gained more weight than WT (Fig. 6B, C
). For example, by 13 wk the null males fed either the low vitamin A diet or the VAD weighed 5 g more than the WT males fed the same diets, and 5 g more than the rdh1-null mice fed a diet with 30 IU vitamin A/g, i.e., null males fed diets restricted in vitamin A weighed 18% more than WT. At 33 wk of age, the male rdh1-null mice fed the VAD weighed 10.6 g (37%) more than WT (39.4±8.5, n=10, vs. 28.8±4.6, n=13; P<0.001) and the female rdh1-null mice weighed 5.8 g (23%) more than WT (30.7±3.8, n=19 vs. 24.9±6.8, n=14; P<0.004). In addition, the male null mice were 11% longer than WT from the end of the nose to the base of the tail: 9.4 ± 0.3 (n=6) vs. 8.5 ± 0.1 cm (n=5), P<0.0001).
|
Multiple tissues in male rdh1-null mice weighed significantly more than in WT (Fig. 7
, top). All fat pads measured were larger in null mice. Epididymal, retroperitoneal, and brown fat pads weighed 1.8-fold more in null than in WT mice. Larger increases were noted in mesentary, femoral, and inguinal fat pads (3.7, 2.7, and 2.8-fold increases, respectively). Kidney and liver increased 25 and 58%, respectively, whereas testis showed no gain in weight. Even though larger in null mice, kidney contributed 22% less to total body weight. Liver weight was proportional to the increase in total weight and testis contributed 37% less to total body weight (Fig. 7
, bottom). Epididymal, retroperitoneal, and brown fat pads, while larger in null mice, contributed the same percent to total body weight as in WT. In contrast, mesentery, femoral, and inguinal fat pads were not only larger in null relative to WT mice but were disproportionately so, normalized to total body mass, with increases of 2.3, 1.7, and 1.4-fold, respectively, relative to WT.
|
Carcasses were measured for water, total fat, and ash content after dehydration and extraction, and an adiposity index (AI) was calculated. Rdh1-null male mice carcasses had 15% less water than WT carcasses and 28% less ash, but 59% more fat, and a 39% greater AI (Fig. 8
). Increased size and adiposity were evident visually (Fig. 9
). Each genotype had the same food intake: rdh1-null male mice fed the VAD consumed 2.9 ± 0.5 g/day/mouse, whereas WT mice consumed an average 2.6 ± 0.2 g/day/mouse, between 24 and 28 wk of age.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Although multiple mouse orthologs/homologs of these two rat genes cluster on chromosome 10D3, including rdh1, rdh6, rdh9, rdhS, and 17ß-HSD9, only rdh1 and 17ß-HSD9 have appreciable activity with all-trans-retinol, and rdh1 functions 20-fold more efficiently (Vm/Km) than 17ß-HSD9. We therefore elected to inactivate rdh1 to evaluate its contribution to retinoid metabolism in vivo.
Lack of a phenotype in the rdh1-null mice fed a diet with 30 IU vitamin A/g might reflect bountiful vitamin A status. Stock diets contain 8-fold more than the 4 IU vitamin A (retinol plus RE)/g recommended by the AIN for rodent chow, which itself is more than the 2.4 IU/g diet considered adequate for laboratory mice by the National Research Council (45
, 46)
. Enzymes and binding proteins that evolved specifically to maximize retinoid use may become unnecessary in animals fed diets with copious amounts of vitamin A, especially the laboratory mouse, which lives in an artificial environment that lacks much of the stress of living in the wild.
Decreasing dietary vitamin A revealed effects of inactivating rdh1. The 2-fold increase in liver retinol concentrations of null vs. WT mice fed a diet with 4 IU vitamin A/g demonstrated that eliminating Rdh1 activity spares retinol. Reducing dietary vitamin A content further to 0.6 IU/g revealed an additional retinol sparing effect of Rdh1 inactivation in kidney and a RE sparing effect in liver and confirmed that the diet with 4 IU/g contained ample vitamin A. Liver serves as the quantitatively major site of retinoid storage, a site of atRA biosynthesis, and a major regulator of retinoid homeostasis: liver supports the retinoid needs of extrahepatic tissues by secreting retinol into blood. Therefore, liver would reflect retinol sparing throughout an organism. The decrease in liver RE stores of the rdh1-null mice fed the 0.6 IU/g diet (7 mM) would have been difficult to quantify in mice fed 4 and 30 IU vitamin A/g diets, because of their very large liver RE stores. RE concentrations in the mM range, even with 0.6 IU vitamin A/g diet, illustrate the capacity of mice to store retinol, and highlight the copious vitamin A content of stock diets, which increase liver RE to >0.5 M.
The 30% increase in liver RE of rdh1-null mice vs. WT fed 0.6 IU vitamin A/g diet provides insight into the degree of retinoid metabolism by Rdh1. Overall, the diet-dependent retinoid sparing indicate that the lower the dietary vitamin A content, the more significant the Rdh1 contribution. This insight complements insights generated by mutating retinoid binding-proteins. For example, CRBP type II-null mouse pups succumb within 24 h of birth, when borne by dams fed a diet marginal in vitamin A (47)
; the CRBP-null mouse has a 6-fold faster rate of vitamin A elimination than WT, showing optimization of retinol storage and use by CRBP (48)
, reflecting the scarcity of vitamin A in nature (49
, 50)
.
Inverse expression of rdh1 and cyp26A1 mRNA (Fig. 5)
suggests a relationship to maintain retinoid homeostasis. Multiple reports have focused on Cyp26, especially A1, as a modulator of atRA concentrations that generate patterns or gradients of atRA to effect precise developmental outcomes (52
53
54)
. The present data suggest that Cyp26A1 and Rdh1 function in opposition to exacerbate differences in atRA concentrations. Selective down-regulation of cyp26A1 in the rdh1-null mouse during vitamin A restriction reveals another dimension of a relationship and seems consistent with autocompensation to spare atRA (Fig. 10
). This latter phenomenon appears unique in retinoid homeostasis. For example, neither the raldh2-null mouse nor the cyp26A1-null mouse autocompensated (sufficiently?) for abnormal atRA concentrations: each mutation was embryonic lethal (25
, 37)
. Notably, however, some mice with double mutations of cyp26A1 and raldh2 survived to adulthood, confirming reduced Cyp26A1 activity can offset decreases in atRA biosynthesis (51)
.
|
Increases in length, weight, and adiposity were unexpected for a mutation that affects retinoid homeostasis. Notably, these effects became significant early in the lives of the animals and continued during periods when the mice would not have been atRA-deficient—as indicated by continued growth and then by maintenance of body weight (Fig. 6)
. The increases occurred relative to WT mice fed the same vitamin A-restricted diets and relative to both WT and null mice fed a diet with copious vitamin A. Thus, these surprising effects disclose a unique function for rdh1 during limited vitamin A intake—the usual dietary state of most vertebrates.
Other retinol dehydrogenases must provide retinal and compensate for rdh1 inactivation. The nature of the enzyme(s) that compensate are also under study. The xenobiotic metabolizing medium-chain alcohol dehydrogenases (ADHs) have been proposed as retinol dehydrogenases, because they recognize "free" retinol in vitro (55
56
57)
. Gene knockouts studies, however, do not provide support for ADH metabolizing retinoids under physiological conditions, because Adh1, 3, or 4-null mice, or dual Adh1 and 4-null mice do not display phenotypes related to lack of retinoid function or show compensatory gene expression changes related to retinoid metabolism or function (58
, 59)
. To date, the only data consistent with Adh metabolizing retinol in vivo have been generated with a single extraordinarily high dose (50 mg retinol/kg) in mice, which would overcome controls imposed by retinoid binding-proteins (58)
. Because retinoids are scarce in nature, and CRBP prevents Adh from metabolizing retinol (8
, 60)
, such an overwhelming dose does not model a physiological process. Continuing research into SDRs has identified several candidates as potential all-trans-retinol dehydrogenases in the path of atRA biosynthesis in addition to Rdh1, such as Rdh10, retSDR8, and 17ß-HSD9 (12
13
14
15
16
17
18
19)
. We are attempting to determine which of these, if any, compensate for loss of Rdh1.
Unexpected insights that emerged from altering dietary vitamin A were the relationships among liver retinoids and between liver retinoids and cyp26 mRNA expression (Figs. 3
4
5)
. Relative to WT mice fed a diet with 30 IU vitamin A/g, livers of WT mice fed a diet with 4 IU vitamin A/g showed a 6-fold decrease in RE, a 4-fold decrease in retinol, a 63-fold decrease in cyp26a1 mRNA, a 10-fold decrease in cyp26b1 mRNA but a 2-fold increase in atRA (not shown). Decreasing the amount of dietary vitamin A from 4 to 0.6 IU/g in WT mice resulted in further decreases in liver RE (6-fold), retinol (5-fold), and cyp26A1 mRNA (8-fold), no further decrease in cyp26b1 mRNA, but a 2-fold increase in atRA. Clearly, atRA does not necessarily reflect retinol concentrations, and cyp26 mRNA expression correlates directly with RE and retinol concentrations but inversely with atRA. These results reveal that complex mechanisms regulate cyp26 expression in the intact animal, likely reflecting total retinoid status and perhaps dependent on the mRNA and/or proteins involved in maintaining retinoid homeostasis. These data also revealed differences in the sensitivities of cyp26a1 and cyp26b1 to vitamin A status.
In summary, this study establishes a physiological contribution of rdh1 to retinoid metabolism and indicates autocompensation through down-regulation of cyp26A1 as a mechanism of survival of the rdh1-null mouse. This study also denotes rdh1 as a heretofore-unrecognized modulator of animal length, weight and adiposity. The mechanisms of these effects are under study.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Current address: Cell Sciences and Development, SAFC Biosciences, Sigma-Aldrich, St. Louis, MO 63103, USA.
Received for publication January 17, 2007. Accepted for publication March 8, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-hydroxysteroid dehydrogenases from rat and human prostate. J. Biol. Chem. 272,15959-15966-hydroxysteroids. J. Biol. Chem. 273,19778-19785-hydroxysteroid/retinoid short-chain dehydrogenase/reductase, 17ß-HSD9. Endocrinology 140,5275-5284-hydroxysteroid dehydrogenase: unique tissue distribution and catalytic properties. J. Biol. Chem. 276,22278-22286This article has been cited by other articles:
|
|
 |
|
  D. C. Berry and N. Noy All-trans-Retinoic Acid Represses Obesity and Insulin Resistance by Activating both Peroxisome Proliferation-Activated Receptor {beta}/{delta} and Retinoic Acid Receptor Mol. Cell. Biol., June 15, 2009; 29(12): 3286 - 3296. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Y. S. Shih, M. A. Kane, P. Zhou, C. L. E. Yen, R. S. Streeper, J. L. Napoli, and R. V. Farese Jr. Retinol Esterification by DGAT1 Is Essential for Retinoid Homeostasis in Murine Skin J. Biol. Chem., February 13, 2009; 284(7): 4292 - 4299. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. V. Belyaeva, M. P. Johnson, and N. Y. Kedishvili Kinetic Analysis of Human Enzyme RDH10 Defines the Characteristics of a Physiologically Relevant Retinol Dehydrogenase J. Biol. Chem., July 18, 2008; 283(29): 20299 - 20308. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
