Identification of a Mouse Short-chain Dehydrogenase/Reductase Gene, Retinol Dehydrogenase-similar: FUNCTION OF NON-CATALYTIC AMINO ACID RESIDUES IN ENZYME ACTIVITY

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2003 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 278, No. 41, Issue of October 10, pp. 40079 –40087, 2003 Printed in U.S.A.

Identification of a Mouse Short-chain Dehydrogenase/Reductase Gene, Retinol Dehydrogenase-similar FUNCTION OF NON-CATALYTIC AMINO ACID RESIDUES IN ENZYME ACTIVITY* Received for publication, May 9, 2003, and in revised form, July 10, 2003 Published, JBC Papers in Press, July 10, 2003, DOI 10.1074/jbc.M304910200

Min-Sun Song, Weiguo Chen‡, Min Zhang, and Joseph L. Napoli§ From the Department of Nutritional Sciences and Toxicology, University of California, Berkeley, California 94720

* This work was supported by National Institutes of Health Grant DK36870. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EBI Data Bank with accession number(s) AY046408. ‡ Present address: Div. of Allergy and Immunology, Dept. of Pediatrics, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave., Cincinnati, OH 45229. § To whom correspondence should be addressed: Dept. of Nutritional Sciences and Toxicology, University of California, 119 Morgan Hall, MC3104, Berkeley, CA 94720. Tel.: 510-642-0809; Fax: 510-642-0535; E-mail: [email protected]. This paper is available on line at http://www.jbc.org

The short-chain dehydrogenase/reductase (SDR)1 gene superfamily encodes ⬃100 bacterial, plant, and animal members related through a limited number of conserved residues that determine structure, provide for cofactor binding, and catalyze dehydrogenation/reduction (1–3). SDR family members do not always share substantial amino acid identities and have relatively few strictly conserved residues. Animal SDR catalyze intermediary metabolism and activation/inactivation of nuclear receptor ligands such as prostaglandins, retinoids, and steroid hormones. SDR tend to have multifunctional catalytic abilities: they can catalyze reactions with dissimilar substrates and/or can recognize functional groups in different loci of the same substrates. An apparent subgroup of the SDR superfamily consisting of phylogenetically related enzymes catalyzes dehydrogenation of all-trans-retinol, cis-retinols, and androgens or reduction of retinals. Functions of this subgroup could include serving in the visual cycle, generating the endocrine factors all-trans-retinoic acid and 9-cis-retinoic acid, reducing retinal produced by carotenoid metabolism, and/or reactivating 5␣-androstane-3␣,17␣-diol (3␣-adiol) into dihydrotestosterone (4, 5). Of the mouse SDR in the retinoid subfamily, RDH1 has widespread expression and seems to have the highest catalytic efficiency for all-trans-retinol dehydrogenation; mouse 17␤HSD9 catalyzes all-trans-retinol dehydrogenation about an order of magnitude less efficiently than RDH1, and RNase protection assays reveal 17␤-HSD9 mRNA expression only in liver (6, 7). Other mouse SDR such as CRAD1 and CRAD3 catalyze 9-cis-retinol dehydrogenation much more efficiently than alltrans-retinol dehydrogenation and have weak, if any, activity with all-trans-retinol (8, 9). CRAD2 has very low efficiency for all-trans-retinol and even lower efficiency for 9-cis-retinol (10). RDH4 catalyzes all-trans-retinol dehydrogenation at least 2 orders of magnitude less efficiently than RDH1 (11, 12). The mouse SDR retSDR1, RRD, and PSDR1 function as reductases that convert all-trans-retinal into all-trans-retinol, but do not catalyze dehydrogenation of all-trans-retinol (13–16). Therefore, although an array of SDR catalyze retinoid metabolism, RDH1 represents the only SDR thus far in the mouse with high catalytic efficiency for all-trans-retinol dehydrogenation and widespread tissue expression, with expression initiating early during embryogenesis. Here we report cDNA cloning of a novel mouse SDR gene and determination of its mRNA expression pattern and its chromo1 The abbreviations used are: SDR, short-chain dehydrogenase(s)/ reductase(s); 3␣-adiol, 5␣-androstane-3␣,17␣-diol; HSD, hydroxysteroid dehydrogenase; CRAD, cis-retinol/androgen dehydrogenase; RDH, retinol dehydrogenase(s); RDH-S, retinol dehydrogenase-similar; mRDH, mouse retinol dehydrogenase; rRDH, rat retinol dehydrogenase; RDH-E, retinol dehydrogenase-epidermal; RACE, rapid amplification of cDNA ends; E17, embryonic day 17.

40079

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

We report a mouse short-chain dehydrogenase/reductase (SDR), retinol dehydrogenase-similar (RDH-S), with intense mRNA expression in liver and kidney. The RDH-S gene localizes to chromosome 10D3 with the SDR subfamily that catalyzes metabolism of retinoids and 3␣-hydroxysteroids. RDH-S has no activity with prototypical retinoid/steroid substrates, despite 92% amino acid similarity to mouse RDH1. This afforded the opportunity to analyze for functions of non-catalytic SDR residues. We produced RDH-S⌬3 by mutating RDH-S to remove an “additional” Asn residue relative to RDH1 in its center, to convert three residues into RDH1 residues (L121P, S122N, and Q123E), and to substitute RDH1 sequence G208FKTCVTSSD for RDH-S sequence F208FLTGMASSA. RDH-S⌬3 catalyzed all-trans-retinol and 5␣-androstane-3␣,17␣-diol (3␣-adiol) metabolism 60 –70% as efficiently (Vm/Km) as RDH1. Conversely, substituting RDH-S sequence F208FLTGMASSA into RDH1 produced a chimera (viz. C3) that was inactive with all-trans-retinol, but was 4-fold more efficient with 3␣-adiol. A single RDH1 mutation in the C3 region (K210L) reduced efficiency for all-trans-retinol by >1250-fold. In contrast, the C3 area mutation C212G enhanced efficiency with all-trans-retinol by ⬃2.4-fold. This represents a >6000fold difference in catalytic efficiency for two enzymes that differ by a single non-catalytic amino acid residue. Another chimera (viz. C5) retained efficiency with alltrans-retinol, but was not saturated and was weakly active with 3␣-adiol, stemming from three residue differences (K224Q, K229Q, and A230T). The residues studied contribute to the substrate-binding pocket: molecular modeling indicated that they would affect orientation of substrates with the catalytic residues. These data report a new member of the SDR gene family, provide insight into the function of non-catalytic SDR residues, and illustrate that limited changes in the multifunctional SDR yield major alterations in substrate specificity and/or catalytic efficiency.

40080

mRDH1 and mRDH-S

FIG. 1. Strategy for cloning a cDNA encoding RDH-S. A 1890-bp segment of RDH-S was cloned and sequenced. Based on these data, an ⬃1100-bp fragment representing the coding region was amplified from mouse e17 cDNA and cloned to create pcDNA3/RDH-S.

EXPERIMENTAL PROCEDURES

cDNA Cloning of Mouse RDH-S (mRDH-S)—Mouse E17 poly(A⫹) RNA (Sigma) was subjected to reverse transcription at 42 °C for 60 min using random hexamers. PCR was done with primers 5⬘-TTYATHACNGGNTGYGAYTCNGGNTTYGGN (forward) and 5⬘-RTCNGARAANGCYTCNANNCCRWAYTTNGA (reverse) for 35 cycles at 94 °C for 45 s, 52 °C for 45 s, and 72 °C for 1 min. The 468-bp products were cloned into pGEM-T (Promega), sequenced, and analyzed using NTI Vector Suite 5.5 (InforMax). Mouse E17 cDNA (Marathon-ReadyTM cDNA, Clontech) was used as template for rapid amplification of cDNA ends (RACE) to extend the 5⬘- and 3⬘-ends of one product. 5⬘-RACE was done with antisense primer R1 (5⬘-GACGTTGACTACACGGCCCCTT) with PCR conditions of one cycle at 94 °C for 1.5 min, five cycles at 94 °C for 40 s and 72 °C for 3 min, five cycles at 94 °C for 40 s and 70 °C for 3 min, and 25 cycles at 94 °C for 40 s and 68 °C for 3 min, followed by 68 °C for 5 min. Antisense primer R2 (5⬘-AGTTCTGTTTGTTCATCCACTGAC) was used for nested PCR with the first RACE products as templates. PCR conditions were as follows: one cycle at 94 °C for 1.5 min and 35 cycles at 94 °C for 40 s and 61 °C for 1 min, followed by 72 °C for 5 min. Antisense primers R3 (5⬘-TGAAGACATACTTGTCTTGGAGG) and R4 (5⬘-TGTCCACAGTCCCACCAGTGGTA; nested for primer R3) were used to extend the 5⬘-region further. 3⬘-RACE was done with sense primers F1 (5⬘-TCTCCGTCCCCTTGGGTCTCAGTC) and F2 (5⬘-CATCATGGGTCGAGTGTCACTTCA; nested for primer F1) using the PCR conditions described above for primers R1 and R2, respectively. Further extension was achieved with sense primers F3 (AGTGTCACTTCATGGTAA) and F4 (GGGGTTCTTCATTCACACAT) with PCR conditions of 35 cycles at 94 °C for 1 min, 52 °C for 1 min, and 72 °C for 1 min. To extend the 3⬘-untranslated region even further, sense primers F4 and F5 (GTGTCCCTTCTCTGAAACCAGA; nested with primer F4) were used. Adapter-specific primer AP1 (CCATCCTAATACGACTCACTATAGGGC) was used with primers R1, R3, F1, and F3; adapter-specific primer AP2 (ACTCACTATAGGGCTCGAGCGGC) was used with primers R2, R4, F2, F4, and F5. RACE products were gel-purified, cloned into pGEM-T, and sequenced. We amplified a 1.1-kb

cDNA containing the coding region with sense primer F6 (5⬘-AACCATGTGGCTCTACCTAGTAC) and antisense primer R5 (5⬘-AATGAGATGAAATGTGTGAATGA) using mouse e17 cDNA. This PCR product was used as template with sense primer F7 (5⬘-CCCGGAATTCGCCGCCACCATGTGGCTCTACC; underlining indicates an EcoRI site, and double underlining indicates a Kozak sequence) and antisense primer R6 (5⬘-CCGCTCGAGTCAGAGGGCTTTCTCA; underlining indicates an XhoI site). The PCR product was gel-purified, digested with EcoRI/ XhoI, and cloned into pcDNA3 to construct pcDNA3/RDH-S. Radiation Hybrid Mapping—A mouse radiation hybrid panel was purchased from Research Genetics. A forward primer (CCCTGGGTAGGAGGTTCAGTCCCT) and a reverse primer (ACAGGTAAATTCTGTTCATGGGCTT) were selected from the 3⬘-untranslated sequence of the mRDH-S cDNA for amplification. PCR was done for 35 cycles at 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s. The radiation hybrid PCR experiment was repeated, and the sequence of the amplified 350-bp fragment was confirmed by sequencing. The PCR results were sent for analysis to the Whitehead Institute/Massachusetts Institute of Technology Center for Genome Research.2 Northern Blotting—A probe from the 3⬘-untranslated region of mRDH-S was amplified from the cDNA with the primers used for radiation hybrid mapping. The PCR product was cloned into vector pGEM-T, sequenced, and labeled with [32P]dCTP using the RadPrime DNA labeling system (Invitrogen). The probe was hybridized overnight at 68 °C with a mouse multiple-tissue blot (Clontech) according to the manufacturer’s protocol. Mouse ␤-actin cDNA was used as a control. Blots were exposed to x-ray film with an intensifying screen at ⫺70 °C for 1 day. Mouse RDH1 (mRDH1) Mutants—mRDH-S and mRDH1 cDNAs were amplified using forward primer 5⬘-CGGGATCCACCATGTGGCTCTACCT-3⬘, with an engineered BamHI restriction site and a Kozak sequence, and reverse primer 5⬘-CGGAATTCTCAGAGGGCTTTCTCAGGCT-3⬘, containing an EcoRI site. PCR products were digested with BamHI and EcoRI and cloned into pcDNA3. We used restriction sites to create chimeras/mutants A7P, C1, C2, C2⫹N, C4, and C6. Chimeras C3 and C5 were made by PCR mutagenesis. Sequences were confirmed by DNA sequencing. Site-directed Mutagenesis—Mutagenesis was done using the BamHI/ XbaI sites of pcDNA3/mRDH1. Splicing by overlap extension was used in PCR amplification with Pfu polymerase of an mRDH1 region with primers containing the desired mutation (17). The primers were 18 or 21 nucleotides long with the mutation located centrally: K210L, 5⬘CACACAAGTTAGGAAGCCGCC-3⬘ (forward) and 5-GGCGGCTTCCTAACTTGTGTG-3⬘ (reverse); C212G, 5⬘-ACTTGTCACGCCAGTCTTGAA-3⬘ (forward) and 5⬘-TTCAAGACTGGCGTGACAAGT-3⬘ (reverse); V213M, 5⬘-ACTACTTGTCATACAAGT-3⬘ (forward) and 5⬘-ACTTGTATGACAAGTAGT-3⬘ (reverse); T214A, 5⬘-ACTACTAGCCACACAAGT-3⬘ (forward) and 5⬘-ACTTGTGTGACAAGTAGT-3⬘ (reverse); and D217A, 5⬘-TGATAGCCTAGCACTACTTGT-3⬘ (forward) and 5⬘ACAAGTAGTGCTAGGCTATCA-3⬘ (reverse). After the first round of PCR, amplified fragments were gel-purified, and each was used as 2 Available at www.genome.wi.mit.edu/cgi-bin/mouse_rh/rhmapauto/rhmapper.cgi.

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

somal localization. This SDR, termed retinol dehydrogenasesimilar (RDH-S), shares 92–95% amino acid similarity with mouse RDH1, CRAD1, and CRAD3, but does not have enzyme activity with their prototypical substrates. This afforded the opportunity to investigate the residues that influence substrate specificity and catalytic efficiency for retinoids and androgens. We prepared RDH1/RDH-S chimeras and mutants of RDH-S and RDH1 and determined their catalytic activities with retinoids and steroids. We show that minor alterations activate RDH-S, that one-residue changes in RDH1 have a large impact on efficiency for retinol, and that relatively limited changes alter the specificity and/or catalytic efficiency for either retinol or 3␣-adiol. These data provide new insight into the amino acid residues that contribute to the substrate specificity and catalytic efficiency of SDR and present the possibility that RDH-S may have a regulatory rather than an enzymatic function.

mRDH1 and mRDH-S

40081

TABLE I Nuclcotide and amino acid sequences of RDH-S and related SDR HSE, hydroxysteroid enzyme; SDR-O, SDR-orphan; RDH-TBE, RDH-tracheobronchial epithelial; retSDR1, retinol SDR1; pr-RDH, photoreceptor RDH; RRD, retinol reductase. Amino acid homology Species

SDR (gene)

Identity %

RDH-S CRAD1 (Rdh6) CRAD3 (Rdh9) RDH1 (Rdh1) RoDH2 CRAD2 (Rdh7) RoDH3 RoDH1 RDH-E/RDH4 HSE/RDH 17␤-HSD9 (Rdh8) 17␤-HSD6 SDR-O 9,11-cis-RDH (Rdh4) 9,11-cis-RDH (Rdh5) 3␣-HSD/RDH-TBE retSDR1 retSDR1 pr-RDH RRD

GenBank™/EBI accession no.

100 94 93 92 90 86 85 85 78 71 70 70 63 60 60 55 45 44 44 42

100 93 95 92 90 88 86 86 81 93 75 73 65 65 63 64 32 32 33 28

AY046408 AF030513 AF372838 AY028928 U33500 AF056194 U33501 U18762 NM003708/AF086735 U89281/AF223225 AF103797 U89280 AY04434 AF013288 U43559 AF343729/AY017349 AF061743 AF061741 AF229845 AB045132

template for overlapping PCR extension using forward primer 5⬘GCTCTAGATCAGAGGGCTTTCTC-3⬘ and reverse primer 5⬘-CGGGATCCCCACCATGTGGCTC-3⬘. BamHI/XbaI digestions were done on the PCR fragments from the second round. The fragments were gelpurified and ligated into pcDNA3. PCR was done with 30 cycles of 92 °C for 45 s, 45 °C for 1 min, and 72 °C for 1 min. Mutations were verified by DNA sequencing. Expression of RDH—CHO-K1 cells (American Type Culture Collection, Manassas, VA) were cultured at 37 °C in Ham’s F-12 medium supplemented with 10% fetal calf serum and 1% penicillin/streptomycin. Cells were transfected with pcDNA3 constructs (8 ␮g/100-mm plate) with LipofectAMINE, harvested 24 h later, and lysed as described (6). The supernatant obtained from centrifuging the lysates at 800 ⫻ g for 10 min was used for enzyme and Western blot analyses. Supernatants from CHO-K1 cells transfected with pcDNA3 or pFLAGCMV5a (Sigma) served as controls. Protein concentrations were determined by the dye-binding method (18). Enzyme Assays—Protein amounts and reaction times were used in the linear range to obtain initial velocity values. Kinetic analyses were done with two to three replicates per point at 37 °C in 0.25 ml of 50 mM Hepes, 150 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, and 2 mM NAD⫹ (pH 8.0) with 2– 4 ␮g of protein for 8 –15 min for all-trans- and 9-cisretinol. Analyses of 3H-labeled steroids (20,000 dpm/reaction) were done with 1–3 ␮g of protein for 1.5–2 min. Each curve was obtained at least twice. Retinoids were quantified by high-performance liquid chromatography. 3H-Labeled steroids (40 –101 Ci/mmol) were separated by thin-layer chromatography, detected by autoradiography, and quantified using liquid scintillation counting as described (6 –9). Kinetic data were fit by nonlinear regression analysis with Prism Version 3.0 (GraphPAD Software, Inc). FLAG-tagged Fusion Proteins—To produce FLAG epitope (DYKDDDDK)-tagged RDH, expression vectors were used as templates with forward primer 5⬘-CGGAATTCCACCATGTGGCTCTACCTGGTT (containing a Kozak sequence) and reverse primer 5⬘-GGGGTACCGAGGGCTTTCTCAGG, which moved the stop codon to place the FLAG sequence in-frame at the C terminus. PCR products were digested with EcoRI and KpnI and cloned into pFLAGCMV5a. Western Blotting—Ten ␮g of cell lysate protein were separated by 12% SDS-PAGE and transferred onto nitrocellulose membranes. Membranes were incubated with rabbit anti-mRDH1 antibody (1:5000 (v/v); raised against RDH1 peptide D217RLSSNTKMIWDKASSEVK) in phosphate-buffered saline/Tween. For FLAG-tagged proteins, mouse antiFLAG tag monoclonal antibody M2 (Sigma) was used at a dilution of 1:5000 (v/v) in phosphate-buffered saline/Tween. Blots were incubated with either anti-rabbit or anti-mouse secondary antibodies conjugated with alkaline phosphatase (Promega). Bands were visualized using nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate. Molecular Modeling—mRDH1 was modeled using Swiss-Model Version 36.0002. Details of the modeling procedure have been described,

%

100 91 90 89 85 80 80 79 70 65 66 63 51 50 48 47 22 21 19 16

including global alignment, model relaxation for possible steric overlaps of the side chains, and energy minimization (19 –21). The BLASTP2 program found all similarities using pairwise alignment of mRDH1 with sequences of three-dimensional structures from the ExNRL-3D Database. The SIM program selected all templates with amino acid sequence identities above 25% and projected model size larger than 20 residues. This step also detected domains that could be modeled based on unrelated templates. This generated ProModII (Version 3.5) input files. Global alignment was done using the loaded templates. ProModII then generated all models using the ExPDB Database. Energy minimization of all models was done with the Gromos96 program. RESULTS AND DISCUSSION

cDNA Cloning of RDH-S—We sequenced 14 clones encoding SDR after reverse transcription-PCR of mouse E17 embryo mRNA using primers from conserved regions (F33ITGCDSGFG and S179K(Y/F)G(I/V/L/F/M)EAFSD) of retinoid/steroid-metabolizing SDR. Eight encoded CRAD2 (10); two encoded 17␤HSD9 (7); three encoded an orphan SDR (22); and one encoded a novel SDR, similar to RDH1, which we named RDH-S. 5⬘- and 3⬘-RACE generated the full-length coding sequence of mRDH-S. The 3⬘-untranslated region was extended further with two rounds of 3⬘-RACE to generate a cDNA of 1890 bp (Fig. 1). The open reading frame encodes a deduced protein of 318 amino acid residues with high homology to other SDR that catalyze retinoid/steroid metabolism (Table I). mRDH-S differs in only 33 residues from RDH1, many of them conservative substitutions (Fig. 2). Notably, RDH-S has the six peptide motifs characteristic of retinoid/steroid-metabolizing SDR (1). Nineteen of the 23 amino acid residues conserved in ⬃70% of SDR occur within these six motifs. Conserved residues include the cofactor-binding sequences T35GX3GXG and N111NAG in the first and second motifs, the catalytic sequence S164X11YX3K in the fourth and fifth motifs, part of the substrate-binding domain in the sixth motif, and the oligomerization domain in the third motif. An additional amino acid, Asn, occurs in RDH-S between residues 173 and 174 (RDH1 numbering), which potentially could change the secondary/tertiary structure relationship between catalytic residues Ser164, Tyr176, and Lys180. RDH-S mRNA Expression—Northern blot analysis was done with a 3⬘-untranslated region probe. mRDH-S mRNA was intensely expressed in liver and kidney, but expression was not observed in the six other tissues analyzed (Fig. 3). Four sizes of

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

Mouse Mouse Mouse Mouse Rat Mouse Rat Rat Human Human Mouse Rat Mouse Mouse Human Human Mouse Human Human Mouse

Similarity

Nucleotide identity

40082

mRDH1 and mRDH-S

FIG. 3. Expression of mRDH-S mRNA. The upper panel shows a Northern blot of RDH-S mRNA expression. The lower panel shows ␤-actin mRNA expression, with ␣- or ␥-actin in heart, skeletal muscle, and testis.

mRNA occurred in liver, with the major ones at 2.4, 3.4, and 4.4 kb. The 3.4-kb mRNA was not expressed in kidney. RDH-S Chromosomal Localization—Radiation hybrid mapping indicated that mRDH-S locates to mouse chromosome 10D3 between markers D10Mit269 and D10Mit271 and near CRAD1 (Rdh6) and CRAD2 (Rdh7). The RDH-S gene maps

close to seven other members of the subfamily (Fig. 4). Lack of mRDH-S Enzyme Activity—mRDH-S, which shares 92% amino acid similarity and 89% identity with mRDH1 (Fig. 2 and Table I), did not catalyze metabolism of the major substrates recognized by mRDH1. These include all-trans-retinol, 9-cis-retinol, 3␣-adiol, and androsterone (data not shown). The lack of RDH-S activity was surprising because none of the residue differences occurred in obviously crucial sections (see above), and many were conservative substitutions. These limited differences between RDH1 and RDH-S afforded the opportunity to distinguish the residues in SDR that contribute to activity with retinoids and steroids. Distinct Requirements for Steroid Versus Retinoid Activity— The non-conservative differences provide obvious starting points for evaluating the effects of substituting RDH-S residues into RDH1. The RDH1 mutant A7P behaved enzymatically similar to RDH1; therefore, no further analysis was done with this mutant. Substituting the section of RDH-S from residues 117 to 147 into RDH1 produced chimera C1, with nine total and eight non-conservative residue differences from RDH1 (Figs. 2 and 5). C1 had no detectable activity with all-trans-retinol or 3␣-adiol (Fig. 6, A and B, bars 4). C1 mutant L121P/S122N/ Q123E, i.e. C1/PNE, was made because sequence L121SQ in C1 differs substantially from P121NE conserved in mRDH1, rat RoDH1–3, and the human ortholog RDH-E. C1/PNE had partial activity with all-trans-retinol and 3␣-adiol (Fig. 6, A and B, bars 5). Mutating any two of the three C1 residues produced chimeras/mutants without activity (C1/PN, C1/PE, and C1/NE) (Fig. 6, A and B, bars 6 – 8). Chimera C2 differs from RDH1 in four residues, with one non-conservative change. C2 had par-

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

FIG. 2. mRDH-S and mRDH1 amino acid sequences. White letters on a black background denote substitutions in RDH-S relative to RDH1: non-conservative differences are shown in boldface italics. The boxes labeled with Roman numerals denote the six peptide motifs conserved in retinoid/steroid-metabolizing SDR. Lines above sequences indicate regions of RDH-S substituted into RDH1 to create chimeras, with the specific chimera indicated by the CX notation. The arrowhead shows insertion of an Asn residue in RDH-S between RDH-1 residues 173 and 174.

mRDH1 and mRDH-S

40083

FIG. 4. mRDH-S chromosomal locus. mRDH-S localizes to mouse chromosome 10D3 near the genes that encode the other members of the SDR subfamily that recognize retinoids and steroids as substrates. CRAD-L refers to a CRAD-like protein without retinoid/steroid activity. CRAD-t refers to a truncated mRNA that does not have a complete fourth exon. The chevrons indicate the directions of the genes. SDR-O, SDR-orphan.

tial activity with both substrates (Fig. 6, A and B, bars 9). Chimera C2 with Asn inserted between residues 173 and 174 of RDH1, i.e. C2⫹N, was inactive with both substrates (Fig. 6, A and B, bars 10). Chimera C3 differs in six residues from mRDH1, with three non-conservative changes. C3 was not active with all-trans-retinol, but catalyzed 3␣-adiol metabolism at a higher rate than RDH1 (Fig. 6, A and B, bars 11). Chimera

C4, which includes the RDH-S residues of C3 as well as six additional differences, four of which are non-conservative, was inactive with both all-trans-retinol and 3␣-adiol (Fig. 6, A and B, bars 12). Consequently, only all-trans-retinol activity requires the specific C3 area residues of RDH1, but 3␣-adiol activity requires the residues in the C-terminal part of C4, i.e. Lys224, Lys229, and Ala230. Chimera C5, which partially over-

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

FIG. 5. Schematic of RDH1 and RDH-S chimeras. The open bars indicate amino acid residues from RDH1; the filled bars indicate residues from RDH-S. The numbers above the bars (other than 1 and 317) indicate the beginning of the inserted sequence followed by the return to the RDH1 or RDH-S template sequence, e.g. in C1, 117 indicates the first residue of the RDH-S section inserted into RDH1, and 148 indicates the return to RDH1 sequence. Therefore, the insertion in this case represents RDH-S residues 117–147. In C1/PNE, 121 indicates the beginning of RDH1 residues 121–123 inserted into the section of RDH-S represented in C1, and 124 indicates the return to RDH-S residues. In C1/PE, 122 indicates Ser122 of RDH-S inserted between Pro121 and Glu123 of RDH1.

40084

mRDH1 and mRDH-S

laps with C4, had activity with both all-trans-retinol and 3␣adiol, but not as high as that of RDH1 (Fig. 6, A and B, bars 13), indicating compensation for the deleterious effects of the residues in the C terminus of C4. Chimera C6, which has the C-terminal 76 residues of RDH-S substituted into RDH1, catalyzed a higher rate of metabolism with all-trans-retinol and a lower rate with 3␣-adiol (Fig. 6, A and B, bars 14). This is intriguing because, with one exception (N249S), the nine residue differences are conservative in the C6 sections of RDH1 and RDH-S. Activity Requires Three Regions—These results obtained above indicated that only three areas in RDH-S abolished enzyme activity: sequence L121SQ, the Asn insertion, and the segment from Phe208 through Ala217. Therefore, RDH-S was mutated to remove the Asn residue; to make the conversions L121P, S122N, and Q123E; and to replace F208FLTGMASSA with G208FKTCVTSSD of RDH1. This construct, RDH-S⌬3, catalyzed the metabolism of both all-trans-retinol and 3␣-adiol at 60 –70% of the rate of RDH1 (Fig. 6, A and B, bars 15). Each of these three changes was made independently in RDH-S, i.e. chimera C7, deletion of only the “additional” Asn residue; chimera C8, substitution of RDH1 P121NE for RDH-S L121SQ; and chimera C9, substitution of RDH-S F208FLTGMASSA with RDH1 G208FKTCVTSSD. In addition, the two remaining combinations of any two changes were made in RDH-S, i.e. chimera C10, RDH-S L121SQ/P121NE and deletion of the Asn residue; and chimera C11, RDH-S F208FLTGMASSA/G208FKTCVTSSD

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

FIG. 6. Activities and expression of RDH and mutants. The activities of RDH1/RDH-S chimeras were compared with the activity of RDH1. Assayed were mRDH1 (bars 1), mock (bars 2), mRDH-S (bars 3), C1 (bars 4), C1/PNE (bars 5), C1/PN (bars 6), C1/PE (bars 7), C1/NE (bars 8), C2 (bars 9), C2⫹N (bars 10), C3 (bars 11), C4 (bars 12), C5 (bars 13), C6 (bars 14), and RDH-S⌬3 (bars 15). Protein amounts were varied (0, 2, 5, 10, 20, and/or 30 ␮g). Differences at each protein level were averaged. Data are presented as means ⫾ S.D. (n ⬎ 3). A, alltrans-retinol; B, 3␣-adiol. Western blotting was done with the 800 ⫻ g supernatants of transfected CHO-K1 cells. C, anti-RDH1 peptide antiserum was used on protein from cells expressing mRDH1 (lane 1), mock (lane 2), mRDH-S (lane 3), C1 (lane 4), C1/PNE (lane 5), C1/PN (lane 6), C1/NE (lane 7), C2 (lane 8), C2⫹N (lane 9), C3 (lane 10), and C6 (lane 11). D, anti-FLAG antiserum was used on protein from cells expressing mRDH1-FLAG (lane 1), mock (lane 2), mRDH-S-FLAG (lane 3), and RDH-S⌬3-FLAG (lane 4).

and deletion of the Asn residue. All of these chimera were inactive. Changes Occur in Both kcat and Km—Values of kinetic constants were determined for the most active chimeras and mutants (Table II). Converting RDH-S into RDH-S⌬3 created a catalytically active protein with lower Km values compared with those of RDH1 with both all-trans-retinol and 3␣-adiol, albeit one less efficient than RDH1 because of lower Vm values. With the exception of RDH-S⌬3, the mutations affected activity with all-trans-retinol and 3␣-adiol much differently. C1/PNE showed major (⬃10-fold or greater) increases in Km values for both all-trans-retinol and 3␣-adiol, but maintained an efficient kcat only with all-trans-retinol. C3 had no detectable activity with all-trans-retinol, but had ⬃4-fold higher efficiency with 3␣-adiol compared with RDH1, predominantly because of a 5-fold lower Km value. C5 showed substantial efficiency with all-trans-retinol, but was unsaturated kinetically with 3␣adiol. The activity of C6 with all-trans-retinol was only 20% as efficient as that of RDH1, despite a Vm value ⬃4-fold higher, but was not saturated kinetically with 3␣-adiol. These results seem remarkable because most of the residue differences in each construct are conservative relative to RDH1, consistent with a large impact of a very few non-conservative changes. RDH1 Residues 208 –217 Contribute to Retinol Specificity— C3 was assayed with 9-cis-retinol and androsterone because mRDH1 catalyzes metabolism of both (6). C3 had weak activity with 9-cis-retinol, just as it did with all-trans-retinol, but substantial activity with androsterone, just like with 3␣-adiol, which reinforces the conclusion that retinoid recognition requires the specific C3 residues of RDH1, whereas steroid activity can better tolerate substitution (Fig. 7). The Km values of RDH1 and C3 for NAD⫹ were 30 ⫾ 0.9 and 40 ⫾ 9 ␮M, respectively, and those for NADP⫹ were 3.7 ⫾ 0.5 and 18 ⫾ 4.7 mM, respectively (data not shown). Thus, residues 208 –217 do not seem to contribute to NAD⫹ binding or function, and protein folding seems normal in C3. RDH1 Single-residue Mutants—We performed site-directed mutagenesis on each non-conservative difference among RDH1 residues 208 –217. Two single mutations had opposite effects on all-trans-retinol dehydrogenase activity: K210L and C212G. Mutant K210L was at least 2500-fold less efficient than RDH1 (Fig. 8). This coincides with strict conservation of an Arg/Lys210 residue in the SDR with measurable retinol dehydrogenase activity, viz. mRDH1, mouse 17␤-HSD9, human RDH-E, and rRDH1–3. CRAD2, which has weak activity with all-transretinol, with a Vm/Km ⬃2–3 orders of magnitude less than that of RDH1, also has a conserved Arg residue at position 210. In contrast, CRAD1 and CRAD3 have barely detectable or no activity with all-trans-retinol and have a Leu residue at position 210, like RDH-S. These data suggest that Arg/Lys210 is necessary but not sufficient for all-trans-retinol dehydrogenase activity. The RDH1 mutation C212G had a 2.4-fold increase in efficiency produced by an increase in the Vm value, without marked impact on the Km value, and presented the only mutation with increased efficiency for all-trans-retinol (Table III). This difference does not appear to reflect a general principle because CRAD1 and CRAD3 have a Gly residue at position 212, and each has negligible or undetectable activity with all-transretinol. In addition, SDR active with all-trans-retinol, such as rRDH1–3 and mouse 17␤-HSD9, have Asp, Glu, and Asn residues at position 212. The D217A mutant had increased Km and Vm values, but maintained the same overall efficiency as RDH1. rRoDH1, rRoDH3, mouse 17␤-HSD9, and human RDH-E have a Glu residue at position 217, but rRoDH2 has a Val residue, consist-

mRDH1 and mRDH-S

40085

TABLE II Kinetic values of RDH1 and RDH1/RDH-S chimeras with all-trans-retinol and 3␣-adiol Substrate RDH

mRDH1 RDH-S⌬3 C1/PNE C3 C5 C6 a b

3␣-Adiol

All-trans-retinol Km

Vm

␮M

nmol/min/mg

4.7 ⫾ 2.2 1.8 ⫾ 0.6 40 ⫾ 9 NDa 3.4 ⫾ 2.8 64 ⫾ 18

2.4 ⫾ 0.3 0.5 ⫾ 0.3 2.2 ⫾ 0.02 NDa 1.5 ⫾ 0.1 9.1 ⫾ 1.7

Vm/Km

0.5 0.3 0.05 0.4 0.1

Km

Vm

␮M

nmol/min/mg

9 ⫾ 0.9 6.1 ⫾ 0.1 ⬎20b 1.8 ⫾ 0.3 ⬎20b ⬎20b

50 ⫾ 9 26 ⫾ 1.5 43 ⫾ 9

Vm/Km

6 4 23

ND, no activity detected (⬍0.02 nmol/min/mg of protein). Reactions were not saturated at 20 ␮M substrate.

ent with the tolerance observed here in the D217A mutant. Mutants V213M and T214A had reduced efficiencies stemming from increased Km values, despite a 3-fold increase in the Vm value in the case of V213M. Most SDR with measurable retinol dehydrogenase activity have Val213 (RDH1, rRoDH1–3, mouse/ human RDH4/5, and human RDH-E); 17␤-HSD9 provided the exception, with a Met residue like that of RDH-S, but 17␤HSD9 showed much lower efficiency than RDH1. Each SDR with measurable retinol dehydrogenase activity has a Thr residue at position 214, including 17␤-HSD9, consistent with the noted negative effect of the T214A mutation. These data and activity comparisons among the retinoid dehydrogenase SDR indicate that loss of all-trans-retinol-metabolizing activity with chimera C3 could have been caused by K210L alone, but probably was exacerbated by V213M and T214A. Differences in Expression Are Not Responsible for Differences in Activities—Western blotting confirmed the expression of RDH1 and the lack of cross-reactivity of the anti-RDH1 antibody with RDH-S (Fig. 6C, lanes 1 and 3). Because the antiRDH1 antibody does not recognize RDH-S, C-terminally FLAG-tagged chimeras were made of RDH-S and RDH1. Neither the anti-FLAG nor anti-RDH1 antibody reacted with endogenously expressed CHO-K1 cell proteins (Fig. 6, C and D, lanes 2). The chimeras/mutants tested, including the FLAGtagged proteins, were expressed to the same extent as wildtype RDH1. The enzyme activity of RDH1-FLAG was similar to

FIG. 8. All-trans-retinol dehydrogenase activities of RDH1 single-residue mutants. Upper panel, mRDH1 (●), C212G (E), and K210L (䡺) were assayed with all-trans-retinol. Data are from representative experiments. Lower panel, Western blot analyses were carried out with anti-mRDH1 antibody on CHO-K1 cell supernatants expressing mRDH1 (lane 1), mock (lane 2), C3 (lane 3), prestained molecular mass marker (lane 4), K210L (lane 5), C212G (lane 6), V213M (lane 7), T214A (lane 8), and D217A (lane 9). TABLE III Kinetic values of mRDH1 point mutants with all-trans-retinol Mutation

K210L C212G V213M T214A D217A

Km

Vm

␮M

nmol/min/mg

⬎400 5.4 ⫾ 0.4 27 ⫾ 0.6 23 ⫾ 0.4 10 ⫾ 4.2

⬍0.07 6.4 ⫾ 0.1 6.2 ⫾ 2.1 2.3 ⫾ 0.02 4.8 ⫾ 0.2

Vm/Km

Vm/Km relative to RDH1

⬍0.0002 1.2 0.2 0.1 0.5

⬍0.0004 2.4 0.4 0.2 1

that of RDH1, demonstrating that the C-terminal FLAG addition did not affect catalytic properties (data not shown). The single-residue mutants also had expression levels similar to those of RDH1 (Fig. 8, lower panel). These data exclude protein expression differences as a major contributor to changes in Vm values. Molecular Modeling—We generated a molecular model based on three-dimensional structures of soluble SDR using an optimal amino acid sequence alignment. The model relied on x-ray structures of human 17␤-HSD type I complexed with cofactor and/or substrate (39.8% identity; Protein Data Bank codes 1FDS, 1A27, and 1EQUB) (23–25), 20␤-HSD (51.5%; Protein Data Bank code 1HU4) (26), and ␤-ketoacyl-(acyl-car-

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

FIG. 7. Activity of chimera C3. Assays were done with 10 ␮M 9-cis-retinol (A) or androsterone (B) for 10 min with lysates of cells transfected with mRDH1 (●), C3 (E), or RDH-S (䡺) or mock-transfected (f).

40086

mRDH1 and mRDH-S

rier-protein) reductase (41% identity; Protein Data Bank codes 1I01E and 1I01B) (27). Alignment of RDH1 and other membrane-associated SDR (11␤-HSD types I and II) with several soluble SDR revealed that the membrane-associated SDR have N-terminal extensions beyond the core SDR structure to provide for membrane insertion and orientation (28). N-terminal extensions would not affect the core three-dimensional SDR structure. The six motifs characteristic of retinoid/steroid-metabolizing SDR start at RDH1 residue 35 and occupy the same relative positions in soluble and membrane-bound SDR (28). Because soluble SDR do not have the N-terminal extension, the extension was not included in the model: modeling began with RDH1 residue 28 (Fig. 9). The model suggests that elimination of activity by the Asn insertion in RDH-S between Gly173 and Gly174 results from a change in the orientation of the catalytic residue Tyr176 and perhaps Lys180 and the relationship between the two. It also suggests that the ␣-helix and ␤-sheet in the area represented in chimera C1 secure the substrate near cofactor-binding residues (Gly36, Gly40, and Gly42) and catalytic residues (Ser164, Tyr176, and Lys180). The P121NE sequence in this area appears to be essential for alignment. The LSQ substitutions in chimera C1 and in RDH-S likely eliminate activity by changing the orientation of the internal ␤-sheet (P121L) and by altering substrate binding (NE to SQ). The model reveals the juxtaposition of RDH1 residues 208 –217 with the catalytic residues and suggests that the differences in this area in RDH-S, i.e. the residues of chimera C3, alter substrate alignment with the catalytic residues. Thus, creating RDH-S⌬3 from RDH-S likely produces enzyme activity by affecting substrate alignment with catalytic residues.

REFERENCES 1. Bailey, T. L., Baker, M. E., and Elkan, C. P. (1997) J. Steroid Biochem. Mol. Biol. 62, 29 – 44 2. Baker, M. E. (2001) Mol. Cell. Endocrinol. 171, 211–215 3. Duax, W. L., Ghosh, D., and Pletnev, V. (2000) Vitam. Horm. 58, 121–148 4. Napoli, J. L. (2000) Prog. Nucleic Acid Res. 63, 139 –188 5. Napoli, J. L. (2001) Mol. Cell. Endocrinol. 171, 103–109 6. Zhang, M., Chen, W., Smith, M. S., and Napoli, J. L. (2001) J. Biol. Chem. 276, 44083– 44090 7. Su, J., Lin, M., and Napoli, J. L. (1999) Endocrinology 140, 5275–5284 8. Chai, X., Zhai, Y., and Napoli, J. L. (1997) J. Biol. Chem. 272, 33125–33131 9. Zhuang, R., Lin, M., and Napoli, J. L. (2002) Biochemistry 41, 3477–3483 10. Su, J., Chai, X., Kahn, B., and Napoli, J. L. (1998) J. Biol. Chem. 273, 17910 –17916 11. Gamble, M. V., Shang, E., Zott, R. P., Mertz, J. R., Wolgemuth, D. J., and Blaner, W. S. (1999) J. Lipid Res. 40, 2279 –2292 12. Gamble, M. V., Mata, N. L., Tsin, A. T., Mertz, J. R., and Blaner, W. S. (2000) Biochim. Biophys. Acta 1476, 3– 8 13. Haeseleer, F., Huang, J., Lebioda, L., Saari, J. C., and Palczewski, K. (1998) J. Biol. Chem. 273, 21790 –21799

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

FIG. 9. Molecular model of mRDH1. Nt indicates residue 28 at the N terminus, and Ct shows the C-terminal residue. Green indicates backbone residues that were not the focus of this work. Black depicts the cofactor-binding residues Gly36, Gly40, and Gly42. Orange depicts the catalytic residues Ser164, Tyr176, and Lys180. Dark blue depicts chimera C1 with the P121NE sequence depicted in yellow. Yellow shows Gly173 and Gly174, which are interrupted by an Asn residue in RDH-S, indicated by asn. Magenta depicts chimera C3, with Lys210 and Cys212 shown in yellow. Red depicts residues 224 –242 of chimera C5. Cyan depicts residues 243–316, which compose chimera C6 and the latter residues of C5.

Lack of C4 activity reinforced the observations with C3 and C5 (C3 had extraordinary specificity and 4-fold enhanced efficiency for 3␣-adiol, whereas C5 showed specificity for retinol) because C4 contains the mutations of C3 and the section of C5 that does not overlap with C6 (RDH-1 residues Lys224, Lys229, Ala230, Val234, and Lys235). A “stick” depiction of the two single mutations in the C3 region with the largest effects (K210L and C212G) indicated that their side chains do not protrude into the active-site cavity (data not shown). Direct steric interactions in the active-site pocket therefore seem unlikely. Notably, SDR with substantial all-trans-retinol dehydrogenation activity (e.g. rRDH1–3, human RDH-E, and 17␤-HSD9) have a basic residue (Lys or Arg) at position 210. Those with stronger cisretinol activity (e.g. CRAD1 and CRAD3) have a Leu residue at position 210. These data show that the Lys residue may be necessary but not sufficient for all-trans-retinol dehydrogenase activity. The results with C5 seem more remarkable because the impact of the five residue differences in the section of C5 that does not overlap with C6 had to overcome the deleterious influence of the residue differences that C5 and C6 have in common. The two conservative differences in the C4 end of C5 (V234I and K235R) draw attention to the probable disproportionate influence of the other three (K224Q, K229Q, and A230T) and indicate that these three residues preserve the all-trans-retinol activity of chimera C5. Examination of other retinoid/3␣-adiol-metabolizing SDR confirms that retinol (but not 3␣-adiol) activity tolerates these variations well. For example, CRAD3 with Gln224, Gln229, and Thr230 has a 3-fold decrease in efficiency with 3␣-adiol relative to RDH1 (9). Overall, the model suggests that residues in the C4 end of C5 influence the orientation of the C3 residues and that the C3 residues affect catalytic residue/substrate orientation and perhaps substrate binding. Concluding Summary—We have reported the occurrence of a new member of the mammalian SDR superfamily, mRDH-S. Characterization of RDH-S was instrumental in developing data to help define the function of SDR residues that contribute to the substrate-binding pocket. The data from contrasting RDH-S and RDH1 demonstrate that minor modifications to the multifunctional SDR can cause major changes in catalytic efficiency and/or substrate specificity. The function of RDH-S remains unknown. Possibly, RDH-S catalyzes metabolism of ligands for “orphan” nuclear receptors. The Asn insertion in RDH-S argues against this, however, because it would perturb the orientation of catalytic residues. RDH-S may have a regulatory function, i.e. its expression may regulate metabolism by binding potential substrates and/or products and/or by serving as a regulatory factor (dominant-negative). In addition, expression of the mRNA does not guarantee translation: the mRNA may have a regulatory function. We are investigating the latter possibility.

mRDH1 and mRDH-S 14. Moore, S., Pritchard, C., Lin, B., Ferguson, C., and Nelson, P. S. (2002) Gene (Amst.) 292, 149 –160 15. Kedishvili, N. Y., Chumakova, O. V., Chetyrkin, S. V., Belyaeva, O. V., Lapshina, E. A., Lin, D. W., Matsumura, M., and Nelson, P. S. (2002) J. Biol. Chem. 277, 28909 –28915 16. Lei, Z., Chen, W., Zhang, M., and Napoli, J. L. (2003) Biochemistry 42, 4190 – 4196 17. Horton, R. M., Ho, S. N., Pullen, J. K., Hunt, H. D., Cai, Z., and Pease, L. R. (1993) Methods Enzymol. 217, 270 –279 18. Bradford, M. M. (1976) Anal. Biochem. 72, 248 –254 19. Guex, N., and Peitsch, M. C. (1997) Electrophoresis 18, 2714 –2723 20. Peitsch, M. C. (1996) Biochem. Soc. Trans. 24, 274 –279 21. Peitsch, M. C. (1995) Bio/Technology 13, 658 – 660

40087

22. Chen, W., Song, M.-S., and Napoli, J. L. (2002) Gene (Amst.) 294, 141–146 23. Ghosh, D., Pletnev, V. Z., Zhu, D.-W., Wawrzak, Z., Duax, W. L., Pangborn, W., Labrie, F., and Lin, S.-X. (1995) Structure 3, 503–513 24. Azzi, A., Rehse, P. H., Zhu, D.-W., Campbell, R. I., Labrie, F., and Lin, S.-X. (1996) Nat. Stuct. Biol. 3, 665– 668 25. Breton, R., Housset, D., Mazza, C., and Fontecilla-Camps, J. C. (1996) Structure 4, 905–915 26. Ghosh, D., Weeks, C. M., Grochulski, P., Duax, W. L., Erman, M., Rimsay, R. L., and Orr, J. C. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 10064 –10068 27. Fisher, M., Kroon, J. T., Martindale, W., Stuitje, A. R., Slabas, A. R., and Rafferty, J. B. (2000) Structure Fold Des. 15, 339 –347 28. Wang, J., Bongianni, J. K., and Napoli, J. L. (2001) Biochemistry 40, 12533–12540

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

Identification of a Mouse Short-chain Dehydrogenase/Reductase Gene, Retinol Dehydrogenase-similar: FUNCTION OF NON-CATALYTIC AMINO ACID RESIDUES IN ENZYME ACTIVITY Min-Sun Song, Weiguo Chen, Min Zhang and Joseph L. Napoli J. Biol. Chem. 2003, 278:40079-40087. doi: 10.1074/jbc.M304910200 originally published online July 10, 2003

Access the most updated version of this article at doi: 10.1074/jbc.M304910200 Alerts: • When this article is cited • When a correction for this article is posted

This article cites 28 references, 8 of which can be accessed free at http://www.jbc.org/content/278/41/40079.full.html#ref-list-1

Downloaded from http://www.jbc.org/ by guest on February 6, 2016

Click here to choose from all of JBC's e-mail alerts