Regulates the Expression of Retinoic Acid-responsive Genes in H

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 286, NO. 15, pp. 13550 –13560, April 15, 2011 © 2011 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

Retinol Dehydrogenase 10 but Not Retinol/Sterol Dehydrogenase(s) Regulates the Expression of Retinoic Acid-responsive Genes in Human Transgenic Skin Raft Culture*□ S

Received for publication, September 1, 2010, and in revised form, February 18, 2011 Published, JBC Papers in Press, February 23, 2011, DOI 10.1074/jbc.M110.181065

Seung-Ah Lee1, Olga V. Belyaeva1, Lizhi Wu, and Natalia Y. Kedishvili2 From the Department of Biochemistry and Molecular Genetics, Schools of Medicine and Dentistry, University of Alabama at Birmingham, Birmingham, Alabama 35294 Retinoic acid is essential for skin growth and differentiation, and its concentration in skin is controlled tightly. In humans, four different members of the short-chain dehydrogenase/reductase (SDR) superfamily of proteins were proposed to catalyze the rate-limiting step in the biosynthesis of retinoic acid (the oxidation of retinol to retinaldehyde). Epidermis contains at least three of these enzymes, but their relative importance for retinoic acid biosynthesis and regulation of gene expression during growth and differentiation of epidermis is not known. Here, we investigated the effect of the four human SDRs on retinoic acid biosynthesis, and their impact on growth and differentiation of keratinocytes using organotypic skin raft culture model of human epidermis. The results of this study demonstrate that ectopic expression of retinol dehydrogenase 10 (RDH10, SDR16C4) in skin rafts dramatically increases proliferation and inhibits differentiation of keratinocytes, consistent with the increased steady-state levels of retinoic acid and activation of retinoic acid-inducible genes in RDH10 rafts. In contrast, SDRs with dual retinol/sterol substrate specificity, namely retinol dehydrogenase 4 (RoDH4, SDR9C8), RoDH-like 3␣-hydroxysteroid dehydrogenase (RL-HSD, SDR9C6), and RDH-like SDR (RDHL, SDR9C4) do not affect the expression of retinoic acid-inducible genes but alter the expression levels of several components of extracellular matrix. These results reveal essential differences in the metabolic contribution of RDH10 versus retinol/sterol dehydrogenases to retinoic acid biosynthesis and provide the first evidence that non-retinoid metabolic products of retinol/sterol dehydrogenases affect gene expression in human epidermis.

All-trans-retinoic acid is a small lipophilic molecule derived from vitamin A that regulates the expression of ⬎530 different genes in various types of cells and tissues through binding to

* This work was supported by National Institute on Alcohol Abuse and Alcoholism Grant AA12153 (to N. Y. K.). The University of Alabama-Birmingham Targeted Metabolomics and Proteomics Laboratory is supported by the National Institutes of Health Grants RR19231, P30 CA13148, P50 AT00477, U54 CA100949, P30 AR50948, and P30 DK079337. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Fig. 1. 1 Both authors contributed equally to this work. 2 To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Genetics, Schools of Medicine and Dentistry, University of Alabama-Birmingham, 720 20th St. South, 440B Kaul Genetics Bldg., Birmingham, AL 35294. Tel.: 205-996-4023; Fax: 205-934-0758; E-mail: nkedishvili@ uab.edu.

13550 JOURNAL OF BIOLOGICAL CHEMISTRY

nuclear transcription factors (retinoic acid receptors (RARs))3 (1). Skin is one of the best characterized targets of retinoid action (2, 3) and contains all of the major components of retinoid signaling and metabolic machinery such as retinoid receptors RAR␥ and RXR␣ (4); cellular retinoic acid binding protein (5) and cellular retinol binding protein (6, 7); retinaldehyde dehydrogenase (8); lecithin retinol acyltransferase (9); and cytochrome P450 (10). Keratinocytes can synthesize retinoic acid in situ from plasma-derived retinol (11–13); the reported concentration of retinoic acid in the epidermal cells is very low (ⱕ20 nM), and it is controlled strictly (14). Concentrations that exceed the optimal range suppress differentiation and promote hyperproliferation, whereas concentrations below this range lead to formation of orthokeratotic epithelium (15). Retinoic acid is synthesized from retinol in two steps; first, retinol is reversibly oxidized to retinaldehyde, and then retinaldehyde is oxidized irreversibly to retinoic acid. The oxidation of retinol to retinaldehyde is the rate-limiting step in retinoic acid biosynthesis (16). Recent studies suggested that this step is catalyzed by the members of the short-chain dehydrogenase/ reductase (SDR) superfamily of proteins (17) (for SDR nomenclature, see Ref. 18). In humans, four different SDRs were implicated in the biosynthesis of retinoic acid. Three of these enzymes, namely retinol dehydrogenase 4 (RoDH4, SDR9C8), RoDH-like 3␣-hydroxysteroid dehydrogenase (RL-HSD, SDR9C6), and RDH-like SDR (RDHL, also known as DHRS9, SDR9C4) share significant sequence similarity with one another and belong to the same branch of the SDR phylogenetic tree (18, 19). Besides the retinol dehydrogenase activity, all three of these human enzymes exhibit high activity toward 3␣-hydroxysteroids and were proposed to catalyze the back conversion of inactive 5␣-androstane-3␣,17␤-diol to the potent androgen dihydrotestosterone (20) and to oxidize and inactivate the bioactive neurosteroid allopregnanolone (21). In addition, RL-HSD was shown to exhibit a 3(␣3␤)-hydroxysteroid epimerase activity, converting 3␣-hydroxysteroids into 3␤-hydroxysteroids (22). The fourth SDR enzyme that was shown to catalyze the oxidation of retinol for retinoic acid biosynthesis is retinol dehydrogenase 10 (RDH10, SDR16C4) (23, 3

The abbreviations used are: RAR, retinoic acid receptor; RDH or RoDH, retinol dehydrogenase; HSD, hydroxysteroid dehydrogenase; RL-HSD, RoDH-like 3␣-hydroxysteroid dehydrogenase; RDHL, retinol dehydrogenase-like); RSDH, retinol/sterol dehydrogenase; SDR, short-chain dehydrogenase/reductase; PHK, primary human keratinocyte; ADT, androsterone; ALLO, allopregnanolone.

VOLUME 286 • NUMBER 15 • APRIL 15, 2011

Functional Analysis of Retinoid-active SDRs in Human Skin 24). RDH10 shares little similarity with the retinol/sterol dehydrogenases (abbreviated here as RSDHs) described above and belongs to a different branch of the SDR phylogenetic tree. It is not yet known whether RDH10 is active toward hydroxysteroids or any other substrates besides retinoids. Data from this and other laboratories (25–27) indicate that the human epidermis contains at least three of the retinoidactive SDRs. However, their relative roles in the biosynthesis of retinoic acid and regulation of gene expression are not known. In part, this is due to technical difficulties associated with analyzing functions of enzymes that are expressed at very low levels in human cell lines and tissues and their generally low enzymatic activity. Furthermore, the use of mouse models has been limited by the redundancy of RoDH-like SDR homologs in mice and the lack of orthologs for some of the human genes (19). In this study, we took advantage of the ex vivo human organotypic skin culture to examine the contribution of human SDRs to retinoic acid biosynthesis and their impact on gene expression. Human organotypic skin culture is very similar to human skin in its morphology and metabolism because human foreskin keratinocytes are grown at the liquid-air interface, a system that recreates fully differentiated squamous epithelium (28). Keratinocytes placed on top of the collagen bed receive moisture and nutrients through the support matrix, growing upwards and forming a “raft” culture. The keratinocytes in this raft culture proliferate, stratify, differentiate, and form layers just like normal skin. Importantly, this model recreates the complex process of epidermal differentiation that involves the temporal and spatial regulation of a large number of key molecules (29); gene expression pattern in raft cultures is very similar to that seen in whole foreskin (30, 31). The genetic make-up of skin raft culture can be manipulated using retrovirus-mediated gene expression. This aspect of skin raft culture model was utilized in the present study to investigate the individual contribution of the four human SDR enzymes to the biosynthesis of retinoic acid and regulation of gene expression during the growth and differentiation of human epidermis. The results of this study reveal important differences in the physiological roles of the four human SDRs.

EXPERIMENTAL PROCEDURES Retroviral Vectors—The cDNAs encoding human RDH10, RoDH4, RL-HSD, and RDHL were PCR-amplified from clones described previously using the primers containing BamHI and EcoRI restriction sites. Sequences of primers are as follows: RDH10, 5⬘-GTG GAT CCA TGA ACA TCG TGG TGG AGT TCT TCG-3⬘ (forward) and 5⬘-GAG AAT TCA GAT TCT TAG ATT CCA TTT TTT GCT TCA T-3⬘ (reverse); RL-HSD, 5⬘-ATC GGA TCC ATG TGG CTC TAC CTG GCA GCC T-3⬘ (forward) and 5⬘-ATC GAA TTC TTA GAC TGC CTG GGC TGG TTT GG-3⬘ (reverse); RDHL, 5⬘-ATC GGA TCC ATG CTC TTT TGG GTG CTA GGC CT-3⬘ (forward) and 5⬘-ATC GAA TTC TCA CAC TGC CTT GGG ATT AGC C-3⬘ (reverse); and RoDH4, 5⬘-ATC GGA TCC ATG TGG CTC TAC CTG GCG GTT TTC-3⬘ (forward) and 5⬘-ATC GAA TTC TCA TAG AGC CTT GGC CGG GCT TG-3⬘ (reverse). The PCR products were digested with BamHI and EcoRI endonucleases and cloned into the corresponding sites of Moloney APRIL 15, 2011 • VOLUME 286 • NUMBER 15

murine leukemia retroviral vector pBabe puro under the control of the retroviral long terminal repeat promoter. All of the constructs were verified by sequencing. Production of Retroviruses—Retroviral vectors were introduced into an ecotropic cell line Bosc23 by electroporation. Culture media from electroporated cells containing ecotropic retroviruses were used to infect GP⫹envAM12 cells (ATCC, Manassas, VA), an NIH3T3-derived amphotropic packaging cell line. Twenty-four hours after infection, the cells were placed into selection medium containing 1 ␮g/ml puromycin for 6 days. The resistant cells were allowed to reach confluence to obtain high titer retroviruses. The retrovirus-containing media from the producer cells was used to infect primary human keratinocytes (PHKs) (32). Preparation of Transgenic Organotypic Skin Rafts—Neonatal foreskins were obtained from the Newborn Nursery of the University of Alabama at Birmingham Hospital in compliance with University of Alabama at Birmingham Institutional Review Board regulations. Epidermal raft cultures were prepared as described previously (32). Briefly, PHKs were isolated from freshly collected neonatal foreskins and cultured in DermaLife calcium-free medium (Lifeline Cell Technology, Walkersville, MD). Freshly collected retrovirus-containing media from producer cells were used to infect PHKs. Infected PHKs were selected with 1.5 ␮g/ml puromycin for 2 days and then cultured in DermaLife calcium-free medium until confluency. Retrovirus-transduced PHKs (4 ⫻ 105 cells/ml) were seeded onto a dermal equivalent consisting of collagen with embedded Swiss 3T3 J2 fibroblasts (kindly provided by Dr. Louise T. Chow, Department of Biochemistry and Molecular Genetics, University of Alabama, Birmingham). After 3 days, skin equivalents were lifted onto stainless steel grids and cultured at the medium-air interface using raft culture medium prepared from Dulbecco’s modified Eagle’s medium, Ham’s F12 medium, and bovine fetal serum, which was supplemented with cholera toxin, insulin, apo-transferrin, hydrocortisone-21, and human epidermal growth factor as described previously (32). The raft cultures were allowed to stratify and differentiate for 10 –11 days, whereupon they were harvested for analysis. Twelve hours before harvest, the medium was supplemented with BrdU (50 ␮g/ml) to mark cells in S phase. The rafts were harvested, fixed in 10% buffered formalin, and embedded in paraffin. Alternatively, epithelial tissues were separated manually from the collagen bed and used for protein extraction. Immunohistochemistry and H&E Staining—Paraffin-embedded skin rafts were cut into 5-␮m sections, mounted on Superfrost/Plus slides (Fisher Scientific, Pittsburgh, PA), and then deparaffinized and rehydrated by a series of washes in containers with decreasing concentrations of ethanol (95, 85, 70, 50, and 30%). For H&E staining, the sections were incubated with Weigert’s hematoxylin (Poly Scientific, Bay Shore, NY), dehydrated through graded ethanol, restained with eosin (Fisher Scientific), and mounted with Permount (Fisher Scientific). For immunohistochemical staining, the sections were pretreated with sodium citrate buffer (pH 6.0) at 95 °C for 10 min to unmask antigens, followed by incubation in 3% hydrogen peroxide for 20 min to block endogenous peroxidase activity. The slides were then washed in PBS (pH 7.2) and incubated JOURNAL OF BIOLOGICAL CHEMISTRY

13551

Functional Analysis of Retinoid-active SDRs in Human Skin with primary antibodies as follows: a 1:100 dilution of BrdU antibodies (Invitrogen); and a 1:100 dilution of filaggrin antibodies (Leica Microsytems, Inc. Bannockburn, IL). After incubation with primary antibodies at 4 °C overnight, samples were rinsed with PBS and reincubated with a 1:50 dilution of biotinylated secondary antibodies (anti-rabbit, anti-mouse immunoglobulins) from the SuperSensitive Link-Label-IHC detection kit manufactured by Biogenex (San Ramon, CA) for 30 min at room temperature. After washing with PBS several times, the sections were incubated for 30 min with a 1:50 dilution of streptavidin-conjugated horseradish peroxidase followed by incubation with 3,3⬘-diaminobenzidine tetrahydrochloride from Turbbo DAB kit (Innovex Biosciences, Richmond, CA) as the chromogenic substrate. Sections were counterstained with Weigert’s hematoxylin (Poly Scientific), dehydrated through graded ethanol, cleared in xylene, and mounted with Permount (Fisher Scientific). All sections were analyzed at a 20⫻ magnification using AxioImager A2 microscope equipped with an AxioCam camera and AxioVision image capture software (Carl Zeiss MicroImaging, Inc., Thornwood, NY). Western Blotting—The epidermis from raft cultures was peeled off collagen beds and homogenized on ice in PBS with 20% glycerol containing a mixture of protease inhibitors: aprotinin (1 ␮g/ml), leupeptin (1 ␮g/ml), pepstatin A (1 ␮g/ml), and phenylmethylsulfonyl fluoride (50 ␮g/ml). Lysates were cleared by centrifugation at 16,100 ⫻ g for 15 min at 4 °C. Protein concentrations were quantified using Bio-Rad DC protein assay (Bio-Rad). Samples were resolved by electrophoresis in 15% SDS-PAGE and transferred to a nitrocellulose membrane for subsequent probing with polyclonal antibodies against RoDH4 (1:5000 dilution) (33), RL-HSD (1:5000 dilution) (34), RDHL (1:5000 dilution) (21), or with affinity-purified polyclonal antibody against RDH10 (1:500 dilution) (24) in conjunction with the appropriate HRP-conjugated secondary antibody and ECLenhanced chemiluminescence detection system (GE Healthcare). Gel loading was determined by reprobing with HRP-conjugated monoclonal ␤-actin antibody (1:5000; Sigma-Aldrich). Activity Assays—Organotypic cultures of PHKs expressing RDH10, RoDH4, RL-HSD, and RDHL were prepared essentially as described above with minor modifications. Retrovirustransduced PHKs (4 ⫻ 105 cells/ml) were seeded onto Costar Transwells (Corning Life Sciences, Lowell, MA), which were placed into 24-well plates. At confluency, transwells containing PHKs were transferred onto dermal equivalents, and skin rafts were cultured as described above. For analysis of retinoid metabolism, culture medium was supplemented with 2 ␮M alltrans-retinol under reduced light 24 h before harvest. Rafts were peeled off transwells and homogenized in PBS on ice. Culture media were collected separately. All manipulations were done under reduced light. Retinoids were extracted into hexane and separated by reversed phase HPLC using a Waters Alliance separation module and 2996 Photodiode array detector (35). Chromatographic peaks were identified by comparing retention times and spectra against retinoid standards and quantified as described previously (36). Total retinoids were normalized by tissue weight and protein amount. For steroid activity assays, commercially available radiolabeled steroids (PerkinElmer Life Sciences; ⬃40 – 60 Ci/mmol


each) were diluted with cold steroids (Steraloids, Inc. Newport, RI) dissolved in Me2SO. For analysis of activities in raft cultures, steroid stock solutions were added to skin raft culture medium containing fetal bovine serum. Steroid reference standards were generated by incubating Sf9 microsomes expressing RLHSD with androsterone (ADT) or allopregnanolone (ALLO) in phosphate buffer (90 mM KH2PO4, 40 mM KCl, pH 7.4) containing fatty acid-free BSA equimolar to steroids as described previously (34). After a 24-h incubation, media and cells were collected separately as described above and extracted with 7 volumes of dichloromethane. The organic layer was evaporated under a stream of nitrogen and dissolved in 50 ␮l of fresh dichloromethane. Steroids were separated by development in toluene: acetone (4:1) on silica gel TLC plates (Sigma-Aldrich, St. Louis, MO). TLC plates containing 3H-labeled steroids were exposed to a PhosphorImager tritium screen (GE Healthcare) overnight, and the intensity of the bands was calculated using ImageQuant program (version 5.0). Products of each reaction were identified by comparison to reference steroids (34). LC/MS-MRM Analysis—Samples were separated by gradient reversed phase high performance liquid chromatography. The gradient was generated with a Shimadzu Prominence Ultra Fast Liquid Chromatography system consisting of a vacuum degasser, binary pump, and a temperature-controlled autosampler. The injection volume was 80 ␮l, and the samples were maintained in the autosampler at 4 °C. The separation was performed on a Develosil RP Aqueous 2.00 mm ⫻ 150 mm (5 ␮) column with a 2.00 mm ⫻ 30 mm guard column. Mobile phase A was 0.1% formic acid in water; mobile phase B was acetonitrile ⫹ 0.1% formic acid. The gradient was as follows: 0 –3 min, 70% B; 4 – 6 min, 95% B; and 6.5–10 min, 70% B. The flow rate was 0.75 ml/min. Data were acquired with an Applied Biosystems API-4000 triple-quadrupole mass spectrometer equipped with atmospheric pressure chemical ionization operating in positive ion mode. The mass spectrometer and HPLC were controlled with Analyst (version 1.4.2) in multiple reaction monitoring (MRM) mode. The mass transitions used to monitor all-trans-retinoic acid were m/z 301 (M ⫹ H)⫹ to the m/z 123, 105, and 81 fragments. The dwell time was 40 ms for all mass transitions. The atmospheric pressure chemical ionization-optimized conditions include the following: curtain gas, 10; collision gas, 12; nebulizer gas, 70; nebulizer current, 4; and source temperature, 350. All gases were nitrogen. q-PCR—RNA from rafts was isolated using TRIzol reagent (Invitrogen). First-strand cDNA was synthesized using SuperScript III kit (Invitrogen). cDNA was purified using a MiniPrep DNA purification kit (Qiagen). Real-time PCR was performed in a LightCycler威 480 instrument (Roche Applied Science) using LightCycler威 480 SYBR Green I Master Mix (Roche Applied Science) with 0.5 ␮M primers and 2.5–25 ng of purified cDNA per reaction. Sequences of the primers are available by request. Levels of transcripts were determined using relative quantification method (37). Three empty vector-transduced and three SDR-expressing rafts were included in each q-PCR experiment. To evaluate the significance of differences in expression levels of each transcript between control and SDRVOLUME 286 • NUMBER 15 • APRIL 15, 2011

Functional Analysis of Retinoid-active SDRs in Human Skin

FIGURE 1. Western blot analysis of SDR expression in skin rafts. A, three skin rafts were peeled off collagen beds, combined, and homogenized using a Dounce homogenizer. The microsomal fraction was isolated by differential centrifugation and used for Western blotting (100 ␮g). The loading control for the microsomal fraction was P450 reductase. Total homogenate of HEK293 cells expressing recombinant RDH10 (RDH10 st.) was used as a molecular weight marker for RDH10 (30 ␮g). B–D, total homogenates prepared from single rafts were used for Western blotting (50 ␮g); the loading control was ␤-actin. Microsomes from Sf9 cells (0.3– 0.5 ␮g) expressing corresponding SDRs were used as markers of the molecular masses.

expressing rafts, an unpaired t test was performed, and the twotailed p value was determined using GraphPad InStat (version 3.00; GraphPad Software, San Diego California).

RESULTS Expression of Human SDRs in Transgenic Skin Rafts—Previous studies have indicated that the retinoid-active SDRs are expressed naturally in human skin (25–27), suggesting that skin organotypic culture is a physiologically relevant model for analyzing functions of these enzymes. We have confirmed the presence of RoDH4, RDHL, and RDH10 transcripts in human organotypic raft cultures by RT-PCR (supplemental Fig. 1). However, the protein levels of most of these enzymes were below detection limit by Western blotting (Fig. 1). Therefore, we examined their functions by increasing the levels of individual proteins through retrovirusmediated ectopic expression of the corresponding cDNAs. PHKs from neonatal foreskin infected with retroviruses were plated on a collagen-fibroblast matrix and then lifted to the air-liquid interface to generate stratified and differentiated epithelium. To confirm the ectopic expression of the SDRs, epidermis was separated manually from collagen beds and processed for immunoblotting. As shown in Fig. 1, protein levels of RDH10, RoDH4, and RL-HSD, which were below detection limit by available antisera in skin rafts infected with empty virus (Fig. 1, A–C), increased significantly following transduction with the corresponding recombinant retroviruses. The levels of RDHL protein, which was visible in mock-transduced rafts, increased further in transgenic skin infected with RDHL-expressing retrovirus (Fig. 1D). This analysis confirmed successful expression of the SDR proteins in transgenic skin. Histological Analysis of Skin Rafts—As PHKs proliferated and formed stratified epithelium, significant differences were APRIL 15, 2011 • VOLUME 286 • NUMBER 15

noted in the appearance of rafts expressing RDH10 compared with other rafts. RDH10 rafts appeared to be softer and more transparent than mock-transduced rafts or rafts expressing RoDH4, RL-HSD, or RDHL. To evaluate the morphology of transgenic epidermis, rafts were sectioned and stained with H&E. This analysis revealed striking differences in the histology of RDH10 rafts compared with other rafts (Fig. 2). RoDH4, RLHSD, RDHL, and mock-transduced rafts had a well developed histological appearance with clearly visible spinous, granular, and cornified layers. In contrast, the granular layer in RDH10 rafts was developed poorly, and the cornified layer was much thinner than in control cultures or completely absent. Extending the time of RDH10 rafts in culture from 11 to 16 days did not produce a more differentiated epidermis (data not shown), indicating that the differentiation was not simply delayed but was profoundly impaired. Immunohistochemical staining for BrdU incorporation revealed increased proliferation of basal cells in RDH10 rafts (Fig. 2). Furthermore, the expression of keratinocyte differentiation marker filaggrin, which was detected in a continuous band in the granular layer of RSDH rafts and mock-transduced rafts, was severely reduced or completely absent in RDH10 rafts (Fig. 2), indicating an overall inhibition of keratinocyte differentiation and stratification/cornification of the epidermis. Analysis of Retinoic Acid-responsive Genes—Retinoic acid is known to induce proliferation of basal keratinocytes and to inhibit their differentiation (2, 3). To determine whether the abnormal differentiation pattern in RDH10 rafts was associated with up-regulation of retinoic acid-responsive genes, we carried out quantitative RT-PCR analysis of several well known targets of retinoic acid such as STRA6, RAR␤, lecithin retinol acyltransferase, CRBP1, CRABP2, DHRS3, and CYP26A1. This analysis revealed significant up-regulation of these genes in RDH10 rafts (Fig. 3A). A corresponding increase in protein levels was confirmed by Western blot analysis for CRBP1 and DHRS3 (Fig. 3B). In contrast, the retinoic acid-responsive genes were not up-regulated in skin rafts expressing RoDH4, RL-HSD, or RDHL (Fig. 3A). This result suggested that an increase in the steady-state levels of retinoic acid occurred in RDH10 rafts, but not in the rafts expressing RoDH4, RL-HSD, or RDHL enzymes. Analysis of Retinoic Acid Production—To examine retinoic acid biosynthesis in RDH10 rafts versus mock-infected rafts, epidermis was grown in transwell inserts to exclude the contribution of collagen-embedded fibroblasts to the metabolism of retinol. Prior to the addition of retinol, transwell inserts with stratified epidermis were lifted off collagen beds and placed into multiwell culture plates containing medium supplemented with retinol or vehicle. After an overnight incubation, the rafts were peeled off transwells, homogenized, and extracted for HPLC analysis following the protocol designed to maximize the yield of retinoic acid (38). This analysis revealed that RDH10 rafts that were grown in regular culture medium and did not receive additional retinol contained a peak that was not detectable in control rafts (Fig. 4A) and had a retention time and spectral properties identical to those of all-trans-retinoic acid (Fig. 4, B, peak 2 and inset 2, and C). Another peak detected in RDH10 rafts but not in control rafts had spectral properties of JOURNAL OF BIOLOGICAL CHEMISTRY

13553


FIGURE 2. Analysis of skin rafts morphology. Skin rafts embedded in paraffin were sectioned and stained with H&E for analysis of skin morphology, with BrdU antibodies for analysis of cell proliferation, and with filaggrin antibodies for analysis of cell differentiation as described under “Experimental Procedures.”

all-trans-3,4-didehydro-retinoic acid (Fig. 4B, peak 1 and inset 1) (39). The presence of all-trans-3,4-didehydro-retinoic acid in skin epidermis is consistent with the previous reports (40 – 42); however, a conclusive identification of this peak was not possible in the absence of commercially available all-trans-3.4-didehydro-retinoic acid standard.


The identity of the peak that had the retention time and spectral properties of all-trans-retinoic acid was examined further by LC/MS-MRM analysis. Three fragment ions that had the highest signal to noise ratio in MRM analysis were selected to monitor all-trans-retinoic acid. As shown in Fig. 4D, the three mass transitions eluted at the same retention time and in VOLUME 286 • NUMBER 15 • APRIL 15, 2011


FIGURE 3. q-PCR analysis of retinoic acid-inducible gene expression. A, q-PCR analysis was performed on RNA isolated from three individual rafts. Control groups are represented by white bars and experimental groups are represented by gray bars. Error bars represent standard error of the mean (S.E.). Levels of transcripts were determined using the relative quantification method (24). The two-tailed p value was determined using GraphPad InStat (version 3.00; GraphPad Software, San Diego, CA). Statistically significant changes are indicated by *, p ⬍ 0.05); **, p ⬍ 0.01; or ***, p value ⬍ 0.0001. STRA6, stimulated by retinoic acid gene 6; RAR␤, retinoic acid receptor ␤; LRAT, lecithin retinol acyltransferase; CRBP1, cellular retinol binding protein type 1; CRABP2, cellular retinoic acid binding protein type 2; DHRS3, dehydrogenase/reductase member 3 (also known as retSDR1); CYP26B1, cytochrome P450 26B1; CYP26A1, cytochrome P450 26A1; RAR␥, retinoic acid receptor gamma; PDK1, 3-phosphoinositide-dependent kinase 1. B, Western blot analysis of CRBP1 and DHRS3 expression in skin rafts transduced with empty virus (mock) and RDH10 retrovirus (RDH10). Cytosolic fractions were used for the detection of CRBP1, and microsomal fractions were used for detection of DHRS3 with ␤-actin as the loading control for cytosol, and cytochrome P450 reductase was used for microsomes essentially as described in the legend to Fig. 1. Rabbit polyclonal antiserum custom made against the CRBP1-intein fusion protein by Cocalico Biologicals, Inc. (Reamstown, PA) was used at a 1:1000 dilution. Rabbit polyclonal antibodies against human DHRS3 from Proteintech Group, Inc. (catalog no. 15393-1-AP; Chicago, IL) were used at a 1:2000 dilution.

approximately the same ratio for both the retinoic acid standard and the RDH10 skin raft sample, confirming the presence of all-trans-retinoic acid in RDH10 rafts. The amount of endogenous all-trans-retinoic acid in RDH10 rafts grown in regular medium was estimated to be ⬃0.15 pmol/mg cellular protein. An overnight incubation in medium supplemented with 2 ␮M retinol resulted in further increase in the amount of retinoic acid in RDH10 rafts (up to 0.52 pmol/mg cellular protein) but not in control rafts. This end point activity assay demonstrated that RDH10 was catalytically active toward retinol and that overexpression of APRIL 15, 2011 • VOLUME 286 • NUMBER 15

RDH10 resulted in the overall increase in the steady-state levels of retinoic acid, consistent with the up-regulation of retinoic acid-responsive genes. Activity Assays with 3␣-Hydroxysteroids—To exclude the possibility that the lack of up-regulation of retinoic acid target genes in RSDH rafts was due to the lack of catalytic activity of recombinant RoDH4, RL-HSD, and RDHL proteins, we incubated the respective skin rafts with 3␣-hydroxysteroids, the alternative substrates recognized by these enzymes. ADT was added to the media of RoDH4 and RL-HSD rafts, as this is the preferred 3␣-hydroxysteroid substrate for RoDH4 and RLHSD enzymes (33, 34), whereas RDHL rafts were incubated with ALLO, which is a better substrate for RDHL (21). RDH10 rafts were included in this experiment to determine whether RDH10 exhibits a 3␣-hydroxysteroid dehydrogenase activity similarly to the three other SDRs. TLC analysis of the reaction products demonstrated that RDH10 was inactive toward androsterone, but skin rafts expressing RoDH4 and RL-HSD efficiently converted ADT to 5␣-androstanedione (5␣-dione) and epiandrosterone (Fig. 5). The formation of epiandrosterone in RL-HSD skin rafts was in agreement with the previously described 3(␣3␤)-hydroxysteroid epimerase activity of RL-HSD (22). As expected, RoDH4 converted ADT to 5␣-dione, which was further reduced to epiandrosterone by the endogenous 3␤-ketosteroid reductase activity described previously (22). RDHL rafts showed increased conversion of ALLO to 5␣-dihydroprogesterone (12% versus 9% in mock-transduced rafts) and epiallopregnanolone (18% versus 15% conversion), also consistent with the previous report (21). These experiments demonstrated that all three retinol/sterol dehydrogenases expressed in skin rafts were catalytically active. Gene Expression and Morphology of Rafts Treated with Retinol—RoDH4 and RL-HSD both exhibit higher Km values for all-trans-retinol than RDH10 (24, 34, 43). To test a possibility that RoDH4, RL-HSD, and RDHL could contribute to retinoic acid production at a higher concentration of retinol than that provided by the standard medium, we grew skin rafts transduced with each of the four SDRs as well as control rafts in medium supplemented with 2 ␮M retinol. Interestingly, supplementation with retinol resulted in epithelial morphology very similar to that observed for RDH10 rafts that were grown in regular medium (Fig. 6), suggesting that the additional retinol was utilized efficiently by the endogenous retinol dehydrogenase(s) present in the original PHKs, resulting in increased levels of retinoic acid. This was further confirmed by q-PCR analysis, which showed significant induction of retinoic acidresponsive genes in all rafts grown in the presence of additional retinol including rafts infected with empty virus (Fig. 7, upper panel). However, importantly, the rafts expressing RoDH4, RLHSD, or RDHL did not show any greater up-regulation of retinoic acid-responsive genes than mock-transduced rafts (data not shown), as would be expected if these enzymes could function as retinol dehydrogenases at higher concentrations of retinol. On the other hand, RDH10 rafts that were grown in the presence of additional retinol showed further increase in RAR␤ expression compared with mock-transduced rafts (Fig. 7, lower panel). This suggested that, although most of the retinoic acidJOURNAL OF BIOLOGICAL CHEMISTRY

13555

Functional Analysis of Retinoid-active SDRs in Human Skin responsive genes reached their maximum expression levels in rafts treated with retinol, RAR␤ was not yet fully activated, and overexpression of RDH10 resulted in further increase in the level of retinoic acid and, consequently, in the expression level of RAR␤. This result demonstrated that while RDH10 rafts responded to retinol supplementation by fully activating RAR␤, neither RoDH4, nor RL-HSD, nor RDHL rafts displayed activation of retinoic acid-responsive genes over that seen in mocktransduced rafts even when provided with a higher concentration of retinol. Identification of Novel Gene Targets for Retinol/Sterol Dehydrogenase and RDH10—Having established that the retinol/ sterol dehydrogenases are catalytically active in skin rafts but are not involved in the biosynthesis of retinoic acid, we examined whether their activities have an impact on the expression of genes other than those regulated by retinoic acid. Microarray data analysis revealed that the expression of a number of genes was altered in skin rafts expressing RoDH4, RL-HSD, or RDHL that were grown in regular skin raft culture medium. To confirm these findings, we carried out q-PCR analysis of several identified genes. As shown in Fig. 8, protease inhibitor Kazal type 6 (SPINK6), fibroblast activation protein (FAP), and uroplakin 1B (UPK1B) were up-regulated, whereas corneodesmosin (CDSN) was slightly down-regulated in RoDH4 rafts. RDHL rafts displayed statistically significant down-regulation of fibronectin 1 (FN1), CDSN, and anterior gradient homolog 2 (AGR2) genes. The AGR2 gene was also down-regulated in RLHSD rafts. This analysis revealed some overlap in the genes targeted by the RSDHs, consistent with their overlapping substrate specificity. Interestingly, several novel genes were identified as retinoic acid targets in RDH10 rafts. Mucin 21 (MUC21) and ␥-aminobutyric acid (GABA) receptor subunit ␲ (GABRP) were very strong responders with 40-fold and 28-fold increases in the levels of mRNA, respectively. FN1, AGR2, and periostin (POSTN) were up-regulated 2– 4-fold, whereas SPINK6 and solute carrier family 7, (cationic amino acid transporter, y⫹ system) member 11 (SLC7A11) were down-regulated ⬃2-fold. The opposite changes in the expression of SPINK6 and AGR2 observed for RDH10 rafts compared with rafts expressing RoDH4 or RDHL/RL-HSD further emphasized the differences in the physiological impact of RDH10 versus RSDHs. Together, these data suggested that RDH10 and RSDHs have different metabolic substrates and different sets of target genes in human skin rafts.

FIGURE 4. Analysis of retinoic acid production in RDH10 skin rafts. Rafts infected with empty virus (A) or with RDH10 expression construct (B) were incubated with 2 ␮M retinol for 24 h. Three rafts of each kind were combined, and retinoids were extracted for reversed phase HPLC analysis. Peak 2 in B was identified as all-trans-retinoic acid (atRA) based on the elution time of all-transretinoic acid standard (C) and the absorption spectrum of the peak (B, inset 2) characteristic of all-trans-retinoic acid. Spectral analysis of peak 1 (B, inset 1) suggests that this peak represents all-trans-3,4-didehydro-retinoic acid.


DISCUSSION Human epidermis contains several retinoid-active SDRs (25–27); however, their relative roles in the regulation of retinoic acid-responsive genes in skin remain unknown. Here, for the first time, we compared the contribution of individual human SDRs to retinoic acid biosynthesis and regulation of gene expression during proliferation, stratification, and differentiation of organotypic skin raft cultures. D, MRM chromatograms of standard all-trans-retinoic acid solution (upper panel) and of RDH10 skin raft extract (lower panel). The arrows indicate the peaks corresponding to specific fragment ions from mass transitions m/z 301/ 123, m/z 301/105, and m/z 301/81. Cps, counts per second; AU, absorbance units.



FIGURE 5. Analysis of steroid metabolism by SDR skin rafts. Rafts expressing RDH10, RL-HSD, RDHL, or mock-transfected with pBabe vector (Mock) were incubated with tritiated ADT or ALLO for 3–24 h. Steroids were extracted and separated by TLC. The plates were exposed to PhosphorImager tritium screen overnight. Numbers indicate individual skin raft samples. For a positive control, 1 ␮g of RDHL-expressing Sf9 microsomes (ms) was incubated with 1 ␮M ALLO for 30 min at 37 °C. ADT, androsterone (3␣-hydroxy-5␣-androstan-17-one); 5␣-Dione, 5␣-androstanedione (5␣-androstan-3,17-dione); Epi-ADT, epiandrosterone (3␤-hydroxy-5␣-androstan-17-one; 5␣-DHP, 5␣-dihydroprogesterone (5␣-pregnane-3,20-one; ALLO, allopregnanolone (3␣-hydroxy-5␣-pregnan-20one); Epi-ALLO, epiallopregnanolone (3␤-hydroxy-5␣-pregnan-20-one).

FIGURE 6. Effect of retinol on skin raft morphology. Skin rafts were grown for 11 days in the absence (control) or presence of additional 2 ␮M all-trans-retinol (retinol). Sections of paraffin-embedded rafts were analyzed by H&E staining and immunostaining for BrdU and filaggrin as described in the legend to Fig. 2.

The results of this study show that all four SDRs ectopically expressed in skin rafts are catalytically active, but RDH10 has a very different effect on skin morphology than RoDH4, RL-HSD, or RDHL. Skin rafts ectopically expressing RDH10 exhibit both greater proliferation of basal cells, as indicated by the increased incorporation of BrdU and severe inhibition of terminal differentiation, evidenced by the reduced expression of filaggrin (the marker of terminally differentiated keratinocytes). Increased APRIL 15, 2011 • VOLUME 286 • NUMBER 15

proliferation of basal cells and inhibition of terminal differentiation are the hallmark features of the effects of retinoids on epithelial morphology (2, 3), and they are consistent with the higher steady-state levels of retinoic acid detected in RDH10 skin rafts. In normal human skin, the level of retinoic acid is undetectable or detectable only at trace amounts although topical application of retinol produces such biological effects as epidermal JOURNAL OF BIOLOGICAL CHEMISTRY

13557


FIGURE 7. Effect of retinol on RA-responsive gene expression in mocktransduced versus RDH10 rafts. Mock-transduced or RDH10-expressing rafts were grown in the presence of additional 2 ␮M retinol. Retinoic acidinducible gene expression was analyzed by q-PCR as described in the legend to Fig. 3. Gene expression levels were compared between mock-transduced rafts grown in the absence (Mock, white bars) or presence of retinol (ROL) (Mock ⫹ ROL, gray bars) (upper panel); and between mock-transduced rafts grown in the presence of retinol (Mock ⫹ ROL, white bars) and RDH10 rafts grown in the presence of retinol (RDH10 ⫹ ROL, gray bars). Statistically significant changes are indicated by *, p ⬍ 0.05. Error bars represent standard error of the mean (S.E.). RALDH2, retinaldehyde dehydrogenase 2; STRA6, stimulated by retinoic acid gene 6; RAR␤, retinoic acid receptor ␤; LRAT, lecithin retinol acyltransferase; CRBP1, cellular retinol binding protein type 1; CRABP2, cellular retinoic acid binding protein type 2; DHRS3, dehydrogenase/reductase member 3 (also known as retSDR1); CYP26B1, cytochrome P450 26B1; CYP26A1, cytochrome P450 26A1; RAR␥, retinoic acid receptor gamma; PDK1, 3-phosphoinositide-dependent kinase 1. Note that there is a further increase in the expression of RAR␤ in RDH10 rafts after treatment with retinol. In contrast, skin rafts expressing RoDH4, RL-HSD, or RDHL did not exhibited increase in retinoic acid-responsive genes greater than that observed in mock-transduced rafts (not shown).

thickening and an increase in expression of cellular retinoic acid binding protein type II (14). Thus, the conversion of retinol to retinoic acid in human skin is regulated tightly (44). Our data suggest that RDH10 is expressed in human skin at a very low level, and the results of this study highlight the importance of the low level of RDH10 by demonstrating that overexpression of RDH10 results in significant up-regulation of retinoic acid target genes and disruption of normal epithelial growth and differentiation. It is important to note that this phenotype is observed in regular raft culture medium, without supplementation with retinol, and it is consistent with the high affinity of RDH10 for retinol as a substrate (Km ⫽ 0.035 ␮M) (24). Thus, it is clear that the amount of RDH10 in skin rafts controls the amount of retinoic acid produced from the physiologically relevant levels of retinol. RoDH4 and RL-HSD have lower affinities for retinol (Km values of 1.1 ␮M and 3.2 ␮M, respectively) than RDH10, which might indicate that these enzymes function as retinol dehydrogenases at higher levels of retinol. However, supplementation of the growing RoDH4, RL-HSD, or RDHL skin rafts with 2 ␮M retinol does not lead to a greater up-regulation of retinoic acidinducible genes in these rafts compared with that observed in mock-infected rafts, indicating that even at a higher retinol


FIGURE 8. Quantitative RT-PCR analysis of gene expression in SDR skin rafts. q-PCR analysis was performed as described in the legend to Fig. 3. Control groups are represented by white bars, and experimental groups are represented by gray bars. Statistically significant changes are indicated by *, p value ⬍ 0.05 or **, p value ⬍ 0.01. Error bars represent standard error of the mean (S.E.). IGFL1, IGF-like family member 1; SPINK6, serine peptidase inhibitor, Kazal type 6; THBS1, thrombospondin 1; FAP, fibroblast activation protein ␣; POSTN, periostin, osteoblast-specific factor; MUC21, mucin 21; UPK1B, uroplakin 1B; TGFBI, transforming growth factor, ␤-induced, 68 kDa; GABRP, ␥-aminobutyric acid (GABA) A receptor ␲; FN1, fibronectin 1; TGFB1, transforming growth factor, ␤1; AGR2, anterior gradient homolog 2 (Xenopus laevis); TOB2, transducer of ERBB2, 2; CDSN, corneodesmosin; SLC7A11, solute carrier family 7, (cationic amino acid transporter, y⫹ system) member 11.

concentration, these enzymes have no effect on retinoic acid biosynthesis in the reconstructed epidermis. Together, these observations suggest that out of four candidate SDRs only RDH10 can function as a retinol dehydrogenase in human skin in vivo. Although the RSDHs do not increase the expression of retinoic acid target genes, elevated levels of these proteins in skin rafts grown in regular medium do affect the expression levels of other genes. Here, we report for the first time that the activities of RoDH4, RL-HSD, and RDHL induce changes in the mRNA levels of SPINK6 (serine protease inhibitor Kazal type 6), FAP (fibroblast activation protein), UPK1B (uroplakin 1B), FN1 (fibronectin 1), AGR2 (anterior gradient homolog 2) and CDSN (corneodesmosin). The relatively small changes in the expression levels of these genes could be due to limited availability of SDR substrates in skin raft culture medium. Changes in some of these transcripts are specific for certain SDRs, but other genes are affected similarly in at least two of the SDR-expressing rafts, suggesting an overlap in their gene targets. For example, the 3-fold up-regulation of FAP (also called FAP␣ or seprase) appears to be specific for RoDH4 rafts. FAP is a cell surface serine protease that has emerged as a marker of reactive fibroblasts in tumors but may also play a role as a tumor suppressor (reviewed in Ref. 45). Another gene that is up-regulated specifically in RoDH4 rafts is UPK1B (uroplakin 1B). UPK1B codes for a structural protein VOLUME 286 • NUMBER 15 • APRIL 15, 2011

Functional Analysis of Retinoid-active SDRs in Human Skin that is a terminal differentiation component of the mammalian bladder membrane and is the only uroplakin that is expressed in tissues other than the urothelium (reviewed in Ref. 46). SPINK6 is significantly up-regulated in RODH4 rafts but exhibits similar upward tendency in RL-HSD and RDHL rafts. SPINK6 is induced during keratinocyte differentiation and is thought to function as the selective inhibitor of kallikrein-related peptidases, which play a central role in skin desquamation by cleaving corneodesmosomes (47). Examples of gene expression changes that overlap between at least two of RSDH rafts are the down-regulation of AGR2 in rafts expressing RDHL and RL-HSD and down-regulation of CDSN in RoDH4 and RDHL rafts. AGR2 (also known as AG2, HAG-2, or GOB-4) is a protein disulfide isomerase that was shown to be essential for production of mucus (48). Interestingly, in contrast to RSDH rafts, this gene was up-regulated in RDH10 rafts. CDSN is an extracellular component of corneodesmosomes that is specific to desmosomes of epithelia undergoing cornification. CDSN displays adhesive properties and is proteolyzed sequentially as corneocytes migrate toward the skin surface. Deletion of Cdsn gene in mice results in desmosomal breaks at the interface between the living and cornified layers (49, 50). Thus, non-retinoid products of RoDH4, RDHL, and RL-HSD activities appear to regulate directly or indirectly several components of the extracellular matrix important for structural integrity, adhesion, and movement of cells. In general, rafts expressing RSDHs seem to exhibit up-regulation of genes associated with terminal differentiation of keratinocytes (e.g. SPINK6 and UPK1B) as opposed to RDH10 rafts which exhibit up-regulation of genes associated with increased proliferation. Interestingly, microarray analysis of gene expression in RDH10 skin rafts revealed a number of previously unrecognized genes that respond to RDH10 activity and, presumably, to retinoic acid. Mucin 21 (MUC21, also known as epiglycanin) was identified as one of the strongest responders with a 40-fold increase in expression. Mucin 21 was the first mucin associated with the malignant behavior of carcinoma cells and was later shown to prevent cell adhesion (51, 52). The up-regulation of MUC21 is consistent with the fragile nature and abnormal morphological appearance of RDH10 rafts. The second gene highly up-regulated in RDH10 rafts is GABRP (peripheral type GABA receptor subunit ␲). Although GABA is known to function primarily as an inhibitory neurotransmitter in the mature central nervous system, GABA receptors are also present in nonneural tissues, including cancer. GABRP is abundant in the uterus and was detected in breast (53, 54) and lung (55) but rarely in the brain. In agreement with our detection of GABRP expression in skin raft culture, keratinocyte growth factor and an air-liquid interface culture system were shown to significantly up-regulate the expression of GABRP mRNA and protein (55). GABRP appears to mediate the growth-promoting effect of GABA (56). This is consistent with the retinoic acid-induced increase in GABRP expression and proliferation of basal keratinocytes. Among the genes negatively regulated by RDH10 activity are SPINK6, which is up-regulated in RSDH rafts, and solute carrier family 7 (cationic amino acid transporter, y⫹ system), APRIL 15, 2011 • VOLUME 286 • NUMBER 15

member 11 (SLC7A11 or xCT). The SLC7A11 gene encodes the xCT subunit of the heterodimeric x(c)(⫺) cystine/glutamate antiporter that was implicated in GSH-based chemoresistance because it mediates cellular uptake of cystine/cysteine for sustenance of intracellular GSH levels (57, 58). In skin, SLC7A11 is required for melanogenesis (reviewed in Ref. 59). In contrast to RDH10 rafts, expression of SLC7A11 does not change significantly in RSDH rafts. Thus, analysis of gene expression in RDH10 versus RSDH rafts revealed a number of important differences. First, the retinoic acid-inducible genes are up-regulated in RDH10 rafts but not in RSDH rafts, indicating that RDH10 but not RSDH is involved in retinoic acid biosynthesis in human epidermis. Second, a number of genes exhibit opposite changes in RDH10 versus RSDH rafts, suggesting that they are regulated by different biologically active metabolites. Finally, some genes exhibit similar changes in RSDH rafts, suggesting a potential overlap in the metabolic products of different RSDHs, consistent with their overlapping substrate specificity. Together, these results demonstrate that RDH10, but not RSDHs, functions as a bona fide retinol dehydrogenase in a growing and differentiating human organ and provide the first evidence that the activities of RSDHs regulate the expression of non-retinoid target genes in human epidermis. Identification of the extracellular matrix components as the major targets of RSDHs in skin should facilitate the search for the true physiological substrates of these enzymes. Acknowledgments—We thank Dr. Louise Chow (University of Alabama School of Medicine, Birmingham, AL) and members of the Chow laboratory, especially Nick Genovese, Hsu-Kun (Wayne) Wang, and Eun-Young Kho for generously sharing expertise in preparation of human skin rafts. We also thank Doyle Ray Moore, 2nd, at University of Alabama at Birmingham Targeted Metabolomics & Proteomics Laboratory for help with LC/MS-MRM analysis of retinoic acid.

REFERENCES 1. Mangelsdorf, D., Umesono, K., and Evans, R. M. (1994) in The Retinoids: Biology, Chemistry, and Medicine (Sporn, M. B., Roberts, A. B., and Goodman, D. S., eds), pp. 319 –350, Raven Press, New York 2. Gudas, L. J., Sporn, M. B., and Roberts, A. (1994) in The Retinoids: Biology, Chemistry, and Medicine (Sporn, M. B., Roberts, A. B., and Goodman, D. S., eds), pp. 443–520, Raven Press, New York 3. Fisher, G. J., and Voorhees, J. J. (1996) FASEB J. 10, 1002–1013 4. Kang, S. (2005) Cutis 75, 10 –13 5. Chatellard-Gruaz, D., Randolph, R. K., Hagens, G., Saurat, J. H., and Siegenthaler, G. (1998) J. Lipid Res. 39, 1421–1429 6. Busch, C., Siegenthaler, G., Vahlquist, A., Nordlinder, H., Sundelin, J., Saksena, P., and Eriksson, U. (1992) J. Invest. Dermatol. 99, 795– 802 7. Karlsson, T., Virtanen, M., Sirsjo¨, A., Rollman, O., Vahlquist, A., and To¨rma¨, H. (2002) Exp. Dermatol. 11, 143–152 8. Cheung, C., Smith, C. K., Hoog, J. O., and Hotchkiss, S. A. (1999) Biochem. Biophys. Res. Commun. 261, 100 –107 9. Baron, J. M., Heise, R., Blaner, W. S., Neis, M., Joussen, S., Dreuw, A., Marquardt, Y., Saurat, J. H., Merk, H. F., Bickers, D. R., and Jugert, F. K. (2005) J. Invest. Dermatol. 125, 143–153 10. Pavez Lorie`, E., Chamcheu, J. C., Vahlquist, A., and To¨rma¨, H. (2009) Arch. Dermatol. Res. 301, 475– 485 11. To¨rma¨, H., Rollman, O., and Vahlquist, A. (1999) Arch. Dermatol. Res. 291, 339 –345 12. Randolph, R. K., and Simon, M. (1993) J. Biol. Chem. 268, 9198 –9205

JOURNAL OF BIOLOGICAL CHEMISTRY

13559

Functional Analysis of Retinoid-active SDRs in Human Skin 13. Siegenthaler, G., Saurat, J. H., and Ponec, M. (1990) Biochem. J. 268, 371–378 14. Kang, S., Duell, E. A., Fisher, G. J., Datta, S. C., Wang, Z. Q., Reddy, A. P., Tavakkol, A., Yi, J. Y., Griffiths, C. E., Elder, J. T., et al. (1995) J. Invest. Dermatol. 105, 549 –556 15. Asselineau, D., Bernard, B. A., Bailly, C., and Darmon, M. (1989) Dev. Biol. 133, 322–335 16. Napoli, J. L., and Race, K. R. (1987) Arch. Biochem. Biophys. 255, 95–101 17. Pare´s, X., Farre´s, J., Kedishvili, N., and Duester, G. (2008) Cell. Mol. Life Sci. 65, 3936 –3949 18. Kallberg, Y., Oppermann, U., and Persson, B. (2010) FEBS J. 277, 2375–2386 19. Belyaeva, O. V., and Kedishvili, N. Y. (2006) Genomics 88, 820 – 830 20. Biswas, M. G., and Russell, D. W. (1997) J. Biol. Chem. 272, 15959 –15966 21. Chetyrkin, S. V., Belyaeva, O. V., Gough, W. H., and Kedishvili, N. Y. (2001) J. Biol. Chem. 276, 22278 –22286 22. Belyaeva, O. V., Chetyrkin, S. V., Clark, A. L., Kostereva, N. V., SantaCruz, K. S., Chronwall, B. M., and Kedishvili, N. Y. (2007) Endocrinology 148, 2148 –2156 23. Wu, B. X., Chen, Y., Chen, Y., Fan, J., Rohrer, B., Crouch, R. K., and Ma, J. X. (2002) Invest. Ophthalmol. Vis. Sci. 43, 3365–3372 24. Belyaeva, O. V., Johnson, M. P., and Kedishvili, N. Y. (2008) J. Biol. Chem. 283, 20299 –20308 25. Markova, N. G., Pinkas-Sarafova, A., Karaman-Jurukovska, N., Jurukovski, V., and Simon, M. (2003) Mol. Genet. Metab. 78, 119 –135 26. Jurukovski, V., Markova, N. G., Karaman-Jurukovska, N., Randolph, R. K., Su, J., Napoli, J. L., and Simon, M. (1999) Mol. Genet. Metab. 67, 62–73 27. Karlsson, T., Vahlquist, A., Kedishvili, N., and To¨rma¨, H. (2003) Biochem. Biophys. Res. Commun. 303, 273–278 28. Banerjee, N. S., Chow, L. T., and Broker, T. R. (2005) in Human Papillomaviruses: Methods and Protocols, Methods in Molecular Medicine (Davy, C., and Doorbar, J., eds), pp. 187–202, Humana Press, Totowa, NJ 29. Fuchs, E., and Raghavan, S. (2002) Nat. Rev. Genet. 3, 199 –209 30. Wilson, J. L., Dollard, S. C., Chow, L. T., and Broker, T. R. (1992) Cell Growth Differ. 3, 471– 483 31. Bernard, F. X., Pedretti, N., Rosdy, M., and Deguercy, A. (2002) Exp. Dermatol. 11, 59 –74 32. Cheng, S., Schmidt-Grimminger, D. C., Murant, T., Broker, T. R., and Chow, L. T. (1995) Genes Dev. 9, 2335–2349 33. Gough, W. H., VanOoteghem, S., Sint, T., and Kedishvili, N. Y. (1998) J. Biol. Chem. 273, 19778 –19785 34. Chetyrkin, S. V., Hu, J., Gough, W. H., Dumaual, N., and Kedishvili, N. Y. (2001) Arch. Biochem. Biophys. 386, 1–10 35. Molotkov, A., Ghyselinck, N. B., Chambon, P., and Duester, G. (2004) Biochem. J. 383, 295–302 36. Belyaeva, O. V., Korkina, O. V., Stetsenko, A. V., Kim, T., Nelson, P. S., and Kedishvili, N. Y. (2005) Biochemistry 44, 7035–7047 37. Pfaffl, M. W. (2001) Nucleic Acids Res. 29, e45 38. Napoli, J. L., and Horst, R. L. (1998) Methods Mol. Biol. 89, 29 – 40


39. Hoover, F., Gundersen, T. E., Ulven, S. M., Michaille, J. J., Blanchet, S., Blomhoff, R., and Glover, J. C. (2001) J. Comp. Neurol. 436, 324 –335 40. Karlsson, T., Vahlquist, A., and To¨rma¨, H. (2010) J. Dermatol. Sci. 57, 207–213 41. Andersson, E., Bjo¨rklind, C., To¨rma¨, H., and Vahlquist, A. (1994) Biochim. Biophys. Acta 1224, 349 –354 42. Sani, B. P., Venepally, P. R., and Levin, A. A. (1997) Biochem. Pharmacol. 53, 1049 –1053 43. Lapshina, E. A., Belyaeva, O. V., Chumakova, O. V., and Kedishvili, N. Y. (2003) Biochemistry 42, 776 –784 44. Tang, X. H., Vivero, M., and Gudas, L. J. (2008) Exp. Cell Res. 314, 38 –51 45. Pure´, E. (2009) Expert Opin. Ther. Targets. 13, 967–973 46. Wu, X. R., Kong, X. P., Pellicer, A., Kreibich, G., and Sun, T. T. (2009) Kidney Int. 75, 1153–1165 47. Meyer-Hoffert, U., Wu, Z., Kantyka, T., Fischer, J., Latendorf, T., Hansmann, B., Bartels, J., He, Y., Gla¨ser, R., and Schro¨der, J. M. (2010) J. Biol. Chem. 285, 32174 –32181 48. Park, S. W., Zhen, G., Verhaeghe, C., Nakagami, Y., Nguyenvu, L. T., Barczak, A. J., Killeen, N., and Erle, D. J. (2009) Proc. Natl. Acad. Sci. U.S.A. 106, 6950 – 6955 49. Leclerc, E. A., Huchenq, A., Mattiuzzo, N. R., Metzger, D., Chambon, P., Ghyselinck, N. B., Serre, G., Jonca, N., and Guerrin, M. (2009) J. Cell Sci. 122, 2699 –2709 50. Matsumoto, M., Zhou, Y., Matsuo, S., Nakanishi, H., Hirose, K., Oura, H., Arase, S., Ishida-Yamamoto, A., Bando, Y., Izumi, K., Kiyonari, H., Oshima, N., Nakayama, R., Matsushima, A., Hirota, F., Mouri, Y., Kuroda, N., Sano, S., and Chaplin, D. D. (2008) Proc. Natl. Acad. Sci. U.S.A. 105, 6720 – 6724 51. Yi, Y., Kamata-Sakurai, M., Denda-Nagai, K., Itoh, T., Okada, K., IshiiSchrade, K., Iguchi, A., Sugiura, D., and Irimura, T. (2010) J. Biol. Chem. 285, 21233–21240 52. Itoh, Y., Kamata-Sakurai, M., Denda-Nagai, K., Nagai, S., Tsuiji, M., IshiiSchrade, K., Okada, K., Goto, A., Fukayama, M., and Irimura, T. (2008) Glycobiology 18, 74 – 83 53. Hedblom, E., and Kirkness, E. F. (1997) J. Biol. Chem. 272, 15346 –15350 54. Zafrakas, M., Chorovicer, M., Klaman, I., Kristiansen, G., Wild, P. J., Heindrichs, U., Knu¨chel, R., and Dahl, E. (2006) Int. J. Cancer 118, 1453–1459 55. Jin, N., Narasaraju, T., Kolliputi, N., Chen, J., and Liu, L. (2005) Cell Tissue Res. 321, 173–183 56. Takehara, A., Hosokawa, M., Eguchi, H., Ohigashi, H., Ishikawa, O., Nakamura, Y., and Nakagawa, H. (2007) Cancer Res. 67, 9704 –9712 57. Chintala, S., Li, W., Lamoreux, M. L., Ito, S., Wakamatsu, K., Sviderskaya, E. V., Bennett, D. C., Park, Y. M., Gahl, W. A., Huizing, M., Spritz, R. A., Ben, S., Novak, E. K., Tan, J., and Swank, R. T. (2005) Proc. Natl. Acad. Sci. U.S.A. 102, 10964 –10969 58. Sato, H., Tamba, M., Kuriyama-Matsumura, K., Okuno, S., and Bannai, S. (2000) Antioxid. Redox. Signal. 2, 665– 671 59. Lo, M., Wang, Y. Z., and Gout, P. W. (2008) J. Cell. Physiol. 215, 593– 602
