Human Brain Short Chain l-3-Hydroxyacyl Coenzyme A Dehydrogenase Is a Single-domain Multifunctional Enzyme

Human brain short chainl-3-hydroxyacyl-CoA dehydrogenase (SCHAD) was found to catalyze the oxidation of 17β-estradiol and dihydroandrosterone as well as alcohols. Mitochondria have been demonstrated to be the proper location of this NAD+-dependent dehydrogenase in cells, although its primary structure is identical to an amyloid β-peptide binding protein reportedly associated with the endoplasmic reticulum (ERAB). This fatty acid β-oxidation enzyme was identified as a novel 17β-hydroxysteroid dehydrogenase responsible for the inactivation of sex steroid hormones. The catalytic rate constant of the purified enzyme was estimated to be 0.66 min−1 with apparent K m values of 43 and 50 μmfor 17β-estradiol and NAD+, respectively. The catalytic efficiency of this enzyme for the oxidation of 17β-estradiol was comparable with that of peroxisomal 17β-hydroxysteroid dehydrogenase type 4. As a result, the human SCHAD gene product, a single-domain multifunctional enzyme, appears to function in two different pathways of lipid metabolism. Because the catalytic functions of human brain short chain l-3-hydroxyacyl-CoA dehydrogenase could weaken the protective effects of estrogen and generate aldehydes in neurons, it is proposed that a high concentration of this enzyme in brain is a potential risk factor for Alzheimer’s disease.

Transiently expressed ERAB was reportedly localized at the endoplasmic reticulum of the cultured cells (2). However, this finding does not agree with the presumed function of human SCHAD/ERAB in fatty acid ␤-oxidation (1). Therefore, the intracellular distribution of this dehydrogenase needed to be re-examined. Determination of the role(s) that human SCHAD plays in normal cells is of paramount importance for elucidating its role(s) in the pathogenesis of Alzheimer's disease. Because some members of the short chain dehydrogenase family display activities toward several structurally distinct substrates (7), we set out to determine whether human SCHAD harbors more than one enzymatic activity. Human SCHAD is clearly homologous to the amino-terminal domain of the peroxisomal multifunctional protein-2 (MFP-2), which exhibits D-3-hydroxyacyl-CoA dehydrogenase (D-3-HAD) activity of opposite stereospecificity (8). Interestingly, this domain also displays some 17␤-hydroxysteroid dehydrogenase (17␤-HSD) activity (9). MFP-2 was found to be identical to the previously described 17␤-hydroxysteroid dehydrogenase type 4 (8,9), although the maximal velocity (V max ) of the 17␤-HSD was only about 1/8000 of that of D-3-HAD (9). These clues prompted us to further characterize the catalytic properties of human brain SCHAD.
In this report, we provide evidence that human brain SCHAD is a new 17␤-hydroxysteroid dehydrogenase. Moreover, it also harbors alcohol dehydrogenase activity. This single domain multifunctional enzyme not only plays a part in the mitochondrial fatty acid ␤-oxidation, but also functions in steroid hormone metabolism.

EXPERIMENTAL PROCEDURES
Plasmid Constructs and Transient Transfections-The cDNA insert (BamHI-EcoRI) containing the whole coding region of human brain SCHAD was removed from the recombinant plasmid pGEM-T-HBHAD (1) by digestion with proper restriction enzymes and subcloned into the BamHI-EcoRI site of the vector pcDNA3.1 (Invitrogen) to yield a mammalian expression plasmid designated pcDNA3.1/SCHAD. The cDNA of human brain SCHAD was also amplified from pGEM-T-HBHAD by polymerase chain reaction using a pair of primers: 5Ј-AAGCTTCCGC-CACCATGGCAGCAGCGTGTCGGA-3Ј and 5Ј-CTGCAGAAGGCTG-CATACGAATGGCC-3Ј. The resulting polymerase chain reaction product was cloned into the pcDNA3.1/CT-GFP-TOPO vector (Invitrogen) according to the procedures of the manufacturer to generate a SCHAD-GFP (green fluorescent protein) fusion protein expression plasmid designated pcDNA3.1/SCHAD-CT-GFP. COS-7 and PC-12 cells were grown according to the instructions of the supplier (ATCC). All culture media were purchased from Life Technologies, Inc. Subconfluent cells were plated in 35-mm plates. The next day, cells were washed with fresh medium and transfected with 1 g of pcDNA3.1/CT-GFP (Invitrogen), pcDNA3.1/SCHAD-CT-GFP, pcDNA3.1/SCHAD, and pcDNA3.1, respectively, using LipofectAMINE PLUS reagent (Life Technologies, Inc.) in accordance with the protocol of the manufacturer. From 24 to 48 h post-transfection, the cells were examined for GFP expression or were used for immunocytochemical staining.
Antibody Preparation and Immunocytochemistry-Anti-SCHAD antibodies were generated by immunizing a rabbit with the peptide RLDGAIRMQP coupled to keyhole limpet hemocyanin with glutaraldehyde. Cells transfected with pcDNA3.1/SCHAD or pcDNA3.1 and growing on coverslips were incubated with primary anti-SCHAD antibody. The bound primary antibody was then detected with fluorescein isothiocyanate-labeled goat anti-rabbit IgG (Sigma) according to a published protocol (10). Cells were imaged using a laser scanning confocal microscope as described below.
Mitochondria Staining and Confocal Microscopy-Living cells were incubated in culture medium containing 1 M MitoTracker TM Red CMXRos (Molecular Probes) for 15 min at 37°C. The live cells were then either mounted in fresh culture medium and examined directly or were immunostained as described above. Fluorescense-labeled cells were examined using a Nikon Eclipse E800 microscope coupled to a Nikon PCM 2000 dual laser scanning confocal microscopy system. Images were analyzed with a C-Imaging-SIMPLE32 TM image analysis system (Compix Inc.).
Overexpression and Purification of Human Brain L-3-Hydroxyacyl-CoA Dehydrogenase-IPTG induction of Escherichia coli BL21 (DE3)pLysS/pSBET-HBHAD transformants, and the purification of human brain SCHAD from the transformants containing the overexpressed enzyme, were performed as described previously (1).
Protein Analysis and Enzyme Assays-Protein concentrations were determined by the method of Bradford (11). Proteins were separated on a 4 -20% gradient gel at pH 8.3, as described previously (12). The activity of L-3-hydroxyacyl-CoA dehydrogenase was measured with acetoacetyl-CoA as substrate according to the published method (13). Alcohol dehydrogenase and hydroxysteroid dehydrogenase activities were assayed at 25°C by spectrophotometrically measuring the absorbance change at 340 nm as a function of time (14,15). The molar extinction coefficient used for calculating rates is 6220 M Ϫ1 cm Ϫ1 . A standard assay mixture for dehydrogenation reactions contained 0.1 M potassium phosphate (pH 8.0), fatty acid-free bovine serum albumin (0.1 mg/ml), 100 M steroid substrate (added in 8 l of ethanol) or 160 mM alcohol substrate, 1 mM NAD ϩ (or NADP ϩ ), and appropriate quantities of enzyme. When the reduction of substrate was measured, the pH of phosphate buffer was adjusted to 6.6, and 100 M NADH (or NADPH), and keto/aldehyde substrate were substituted for NAD ϩ (or NADP ϩ ) and alcohol substrate, respectively. Confirmation of steroid products, after extraction with diethyl ether, was achieved by thin-layer chromatography. Kinetic parameters for the different substrates were estimated by analysis of the kinetic data with the computer program Leonora (16). A unit of activity is defined as the amount of enzyme that catalyzes the conversion of 1 mol of substrate to product/min.

Subcellular Location of Human Short Chain L-3-Hydroxyacyl-
CoA Dehydrogenase-The GFP and GFP-tagged human brain short chain L-3-hydroxyacyl-CoA dehydrogenase (SCHAD-GFP) were expressed in COS-7 cells transiently transfected with pcDNA3.1/CT-GFP and pcDNA3.1/SCHAD-CT-GFP, respectively. Because GFP and its fusion protein fluoresce under UV light, GFP is a convenient molecular reporter to monitor patterns of protein localization in living cells (17). As shown in Fig. 1a, GFP was distributed uniformly throughout the cytoplasm and nucleus. In contrast, the resulting SCHAD-GFP fusion protein displayed a granulated pattern of localization (see Fig. 1d). Once mitochondria were specifically stained by a mitochondrion-selective dye with red fluorescence (Fig. 1, b and e), the punctate organelles where the GFP-tagged SCHAD localized were unequivocally identified as mitochondria (see Fig. 1f). Because GFP itself cannot enter mitochondria (Fig. 1c), these results indicate that mitochondria are the proper location for this dehydrogenase.
The native SCHAD expressed in COS-7 cells transiently transfected with pcDNA3.1/SCHAD was located by immunocytochemistry (see Fig. 1g). The green fluorescence of the SCHAD/antibody complex co-localized with the red fluorescence of mitochondria (see Fig. 1, h and i). When preimmune rabbit serum was used instead of the anti-SCHAD antibodies, such green fluorescence was not observed (data not shown). Cells transfected with pcDNA3.1 vector emitted little green fluorescence after treatment with anti-SCHAD and FITC-labeled secondary antibody (data not shown). We obtained similar results when using PC-12 instead of COS-7 cells (data not shown).
The human SCHAD gene product was shown to be identical to the A␤-binding-protein, ERAB (1). This finding may cause some confusion about whether human SCHAD participates in mitochondrial fatty acid ␤-oxidation, particularly in light of a recent report localizing ERAB to the endoplasmic reticulum in transiently transfected cells (2). Our study, employing confocal microscopy and using either the immunocytochemistry or GFP method, resulted in fluorescence images differing from those published previously (2). We show that mitochondria are a major location for human SCHAD in living cells (see Fig. 1). These results strongly support our proposition (1) that this human dehydrogenase could act as a mitochondrial fatty acid ␤-oxidation enzyme.
Capability of Human Short Chain L-3-Hydroxyacyl-CoA Dehydrogenase to Oxidize Alcohols-Ethanol is usually a good substrate for alcohol dehydrogenases (14); however, its rate of oxidation catalyzed by human brain SCHAD is not detectable (data not shown). Because we considered SCHAD to be a new member of the short chain dehydrogenase superfamily (1), we were interested in determining whether this mitochondrial ␤-oxidation enzyme is capable of converting alcohols to aldehydes or ketones. Our experiments revealed that propanol and longer chain primary alcohols could be slowly oxidized. For example, the specific activity of the purified enzyme was estimated to be 2.0 Ϯ 0.3 nmol ⅐ min Ϫ1 ⅐ mg Ϫ1 for the dehydrogenation of 1-propanol. However, secondary alcohols were found to be better substrates. For example, 2-propanol was oxidized approximately 9-fold faster than was 1-propanol in a standard assay. Nevertheless, the rate of the backward reaction, i.e. the reduction of acetone to produce 2-propanol, was not detected. These observations suggest that human brain SCHAD, in contrast to the classic L-3-hydroxyacyl-CoA dehydrogenase, is a multifunctional enzyme.
Characterization of the Intrinsic Dehydrogenase Activities of Human Short Chain L-3-Hydroxyacyl-CoA Dehydrogenase-A variety of steroid substrates were tested for the ability of human brain SCHAD to oxidize or to reduce them. We found that SCHAD could oxidize the most potent estrogen 17␤-estradiol to a very weak one, estrone. This dehydrogenase can also oxidize and thereby inactivate androsterone to androstanedione. In addition, SCHAD can use NADH as coenzyme to reduce the potent androgen, 5␣-dihydrotestosterone, to a weak one, dihydroandrosterone. SCHAD also displayed low activities toward C 21 steroid hormones ( Table I). Details of this investigation will be described in a subsequent report. Our results reveal that human brain SCHAD is a new type of oxidative 17␤-hydroxysteroid dehydrogenase, which possesses some 3␣-hydroxysteroid dehydrogenase activity as well. Its primary catalytic properties were determined by steady-state kinetic measurements (16,18) and are summarized in Table II. To our knowledge, porcine 17␤-hydroxysteroid dehydrogenase type 4 was the only known oxidative 17␤-HSD that had been purified to near homogeneity (9). It is noteworthy that the V max and the K m for 17␤-estradiol of SCHAD were both 2 orders of magnitude higher than those reported for porcine 17␤-HSD type 4 (9). Therefore, the actual catalytic efficiencies of these two different oxidative 17␤-HSDs are comparable.
The effects of pH on the 17␤-HSD and L-3-HAD activities of human brain short chain L-3-hydroxyacyl-CoA dehydrogenase were determined. The reduction of acetoacetyl-CoA to L-3-hydroxybutyryl-CoA exhibited a pH optimum of 6.5-7.0 (Fig. 2), which is quite distinct from that reported for the bovine liver enzyme (19). In the latter case, reaction rates continuously increase with decreasing pH (19). For the oxidation of 17␤estradiol to estrone, human brain SCHAD showed a pH optimum of 9.5-10.0 (Fig. 2), which is similar to that reported for human 17␤-HSD type 2 (20). SCHAD is different from other 17␤-HSDs because of its inability to catalyze the reduction of either estrone or androstenedione by NADH or NADPH at an appreciable rate (Table I). However, its pH-activity profile for the reduction of acetoacetyl-CoA appears to be similar to the patterns determined for the reduction of estrone or androstenedione catalyzed by human 17␤-HSD type 2 (20). Human Short Chain L-3-Hydroxyacyl-CoA Dehydrogenase Is a Novel 17␤-Hydroxysteroid Dehydrogenase-Thus far, five different types of human 17␤-HSD have been identified (20 -24).
However, only type 1, which catalyzes the reduction of estrone to 17␤-estradiol, has been purified to near homogeneity (25). Recently, rat 17␤-HSD type 6 (26) and mouse 17␤-HSD type 7 (27) and Ke 6 (28) were identified and reported to be new isoenzymes. However, their corresponding enzymes from human tissues have not yet been identified. The sequence identity between human 17␤-HSD isoenzymes is less than 22% except that human brain SCHAD and the amino-terminal domain of peroxisomal 17␤-HSD type 4 show 29% identity. Nevertheless, all of them have eight conserved residues that represent "fingerprints" typical of the short chain dehydrogenase family (29) (see Fig. 3). By comparing the structural and functional features of SCHAD with the known 17␤-HSDs (see Table III), we conclude that human SCHAD represents a new type of 17␤-HSD.
A homolog of the human SCHAD gene was found on Drosophila chromosome X. A conceptual translation from the open reading frame of the Drosophila scully gene revealed a 68%  green (a, d, and g). c, f, and i show merged images of a and b, d and e, and g and h, respectively. The results shown are representative of three independent experiments. identity with the primary structure of human brain SCHAD, even though the scully gene product has not yet been purified and characterized (30). The putative Drosophila short chain L-3-hydroxyacyl-CoA dehydrogenase was thought to function in fatty acid ␤-oxidation because such a role agreed with the presence of aberrant mitochondria and cytoplasmic lipid inclusions in scully mutants described previously (30). However, other important scully mutant phenotypes, such as hypogenitalism and germinal cell aplasia, cannot be explained without the knowledge that human brain SCHAD is a multifunctional enzyme with intrinsic 17␤-hydroxysteroid dehydrogenase activity. Combining these observations, it seems likely that the invertebrate counterpart of human SCHAD, the scully gene product, retains enough hydroxysteroid dehydrogenase activity to play a significant part in sex steroid hormone metabolism as well as in fatty acid oxidation. DISCUSSION The results of our present study provide new evidence for a close relationship between fatty acid metabolism and sterol metabolism. The recent observation that the sterol regulatory element-binding protein (SREBP) stimulates the biosynthesis of cholesterol and fatty acids-particularly unsaturated fatty acids-established interdependent regulatory control at the transcriptional level for these two lipid pathways (31). Our study, which shows that the human SCHAD gene product, a mitochondrial ␤-oxidation enzyme, functions as an oxidative 17␤-hydroxysteroid dehydrogenase (17␤-HSD), establishes interdependence of metabolites in these two pathways upon a single-domain multifunctional enzyme. Together, these observations suggest that both the biosynthesis and the catabolism of intermediates in these two lipid pathways are interrelated at both the level of gene expression and enzyme function. These important observations could have wide ranging implications.
17␤-Hydroxysteroid dehydrogenase is essential for sex steroid metabolism. The reductive 17␤-HSDs catalyze the final step in the synthesis of potent androgens and estrogens, whereas the oxidative 17␤-HSDs inactivate these androgens and estrogens. The various 17␤-HSD isoenzymes display different tissue and cellular distributions and substrate specificities (Table III). Expression of reductive 17␤-HSD isoenzymes across tissue types is directly related to the formation of the respective sex steroid in each tissue type. For example, 17␤-HSD types 3 and 5 catalyze the formation of testosterone in the testis and peripheral tissues, respectively (22,24). In contrast, type 1 17␤-HSD converts estrone to the potent estrogen 17␤estradiol in both gonadal and nongonadal tissues, including brain (32). Reportedly, 30 -50% of total androgens in males and 75% of total estrogens in females prior to menopause are synthesized in peripheral tissues (33). Oxidative 17␤-HSD, on the other hand, play a significant role in maintaining the steadystate level of each sex steroid within individual cells. 17␤-HSD type 2, expressed in placenta, endometrium, and liver, inactivates androgens as well as estrogens (20). 17␤-HSD type 4, expressed in organs that are rich in peroxisomes, catalyzes the oxidation of 17␤-estradiol (8,9). We found that human SCHAD, expressed in a variety of tissues including brain, catalyzes the conversion of 17␤-estradiol and dihydrotestosterone to inactive steroids (Table I). This is the first mitochondrial enzyme found to function in the catabolism of steroid hormones although several reactions essential for the conversion of cholesterol into sex steroids and bile acids have been previously shown to occur in mitochondria. Moreover, the oxidation of androsterone catalyzed by human SCHAD would yield androstanedione, a potent inhibitor of the aromatase. Therefore, overexpression of this dehydrogenase may interfere with the generation of 17␤estradiol from androgens. After menopause, all estrogens in women are synthesized in peripheral tissues (33). In this situation, human SCHAD and other oxidative 17␤-HSDs would have an exceedingly profound impact on the control of intra- a When NAD ϩ and NADH were replaced by NADP ϩ and NADPH, respectively, no enzyme activity was detected.
b The data from triplicate assays are given as a mean Ϯ S.D. c ND, not detected, implies activity of Ͻ0.0005 unit/mg. The kinetic data are based on measurements of human brain short chain L-3-hydroxyacyl-CoA dehydrogenase at pH 8 with steroids as variable substrates at several fixed concentrations of NAD ϩ . Recent clinical studies have suggested that estrogen replacement therapy can delay or prevent Alzheimer's disease (AD) (34). A number of putative mechanisms whereby estrogens could interfere with the progression of AD have been proposed, including organizational and activational effects on the central nervous system (35), increase of dendritic spines and synapse formation (36), and reduction of neuronal generation of amyloid ␤-peptide (37), among others. Brain extracts of AD patients contain significantly more human brain SCHAD/ERAB than those of age-matched controls (2). 2 A high concentration of human brain SCHAD/ERAB would likely cause an estrogendeficient state in neurons. The identification of human brain SCHAD as a new 17␤-HSD leads us to propose that high concentrations of human brain SCHAD are a potential risk factor for AD.
Estrogens reportedly attenuate excitotoxicity, oxidative injury, and amyloid ␤-peptide (A␤) toxicity in hippocampal neurons (38). Overexpression of human brain SCHAD in cultured cells would deplete 17␤-estradiol, thus accounting for the high sensitivity of such cells to A␤-induced stress and apoptosis described previously (2). Moreover, it was reported that aldehyde reductase confers protection upon PC-12 cells to glyoxal cytotoxicity (39), so the alcohol dehydrogenase activity of human brain SCHAD may play the opposite role of enhancing the toxicity of aldehydes. Because the human brain SCHAD/A␤ complex displays considerable enzymatic activities, 2 the binding of A␤ to human brain SCHAD might exacerbate the adverse effects of this multifunctional enzyme if it would cause a relocation of SCHAD. On the basis of these observations, we believe that tuning down SCHAD expression to restore the protective effects of estrogen may ameliorate the neuronal dysfunction associated with AD. Additionally, because the human SCHAD gene product is responsible for estrogen inactivation, it would be important to assess if its expression in malignant cells of estrogen-responsive tissues such as breast cancer and endometrium cancer differs from that in normal controls. In our effort to expand insight into this important multifunctional enzyme, we are currently studying the regulatory mechanism of human SCHAD gene expression.   (21). Human 17␤-HSD type 5 is not included because its primary structure has not been published (24). Standard 1-letter amino acid abbreviations are used. Conserved residues that represent the fingerprint typical of the short chain dehydrogenase family (29) are marked by asterisks (*), whereas those conserved in three sequences or more are shown by letters in the bottom line. Such letters are capitalized where SCHAD and 17␤-HSD type 4 have the same residue.