Molecular Cloning and Characterization of Hemolymph 3-Dehydroecdysone 3β-Reductase from the Cotton Leafworm,Spodoptera littoralis

The primary product of the prothoracic glands of last instar larvae of Spodoptera littoralis is 3-dehydroecdysone (3DE). After secretion, 3DE is reduced to ecdysone by 3DE 3β-reductase in the hemolymph. We have previously purified and characterized 3DE 3β-reductase from the hemolymph of S. littoralis. In this study, cDNA clones encoding the enzyme were obtained by reverse transcription-polymerase chain reaction, employing primers based on the amino acid sequences, in conjunction with 5′- and 3′-rapid amplification of cDNA ends. Multiple polyadenylation signals and AT-rich elements were found in the 3′-untranslated region, suggesting that this region may have a role in regulation of expression of the gene. Conceptual translation and amino acid sequence analysis suggest that 3DE 3β-reductase from S. littoralis is a new member of the third superfamily of oxidoreductases. Northern analysis shows that 3DE 3β-reductase mRNA transcripts are widely distributed, but are differentially expressed, in some tissues. The developmental profile of the mRNA revealed that the gene encoding 3DE 3β-reductase is only transcribed in the second half of the last larval instar and that this fluctuation in expression accounts for the change in the enzyme activity during the instar. Southern analysis indicates that the 3DE 3β-reductase is encoded by a single gene, which probably contains at least one intron.

order, the ratio of ecdysteroids from the prothoracic glands are markedly different. The physiological significance of these differences remains to be explained. Previously, we demonstrated that the prothoracic glands of last instar larvae of Spodoptera littoralis primarily secreted 3DE (ϳ82%), with lesser amounts of ecdysone (ϳ18%) (6). The fact that interconversion of ecdysone and 3DE by prothoracic glands was not detectable suggested that 3DE is more likely an independent product of pathways of ecdysteroid biosynthesis in the glands.
After secretion from the prothoracic glands, the 3DE is reduced to ecdysone by an NAD(P)H-linked 3DE 3␤-reductase in the hemolymph (3,4,6,7). Ecdysone undergoes C-20 hydroxylation in certain peripheral tissues yielding 20-hydroxyecdysone, which is considered to be the true molting hormone in most insects (8). Thus, 3DE 3␤-reductase-catalyzed reduction of 3DE to ecdysone is viewed as an important regulatory step in the production of the molting hormone in Lepidopteran species. 3DE 3␤-reductase was demonstrated in the hemolymph and the activity of the enzyme during the last larval instar reached a peak just preceding that of the hemolymph ecdysteroid titer, supporting a role of the enzyme in production of ecdysteroids. It was revealed that the enzyme exhibited maximum activity at low 3DE substrate concentrations, with a drastic inhibition of activity at higher concentrations (Ͼ5 M), suggesting the existence of an inhibition site for 3DE (6).
The 3␤-reductase enzyme was purified from the hemolymph and was shown to be a monomer with molecular mass of approximately 36 kDa (6). Amino acid sequences of the NH 2 terminus as well as of interior tryptic peptides of the purified enzyme have been determined. Here, we report the molecular cloning and characterization of the cDNA encoding 3DE 3␤reductase of the cotton leafworm, S. littoralis. Conceptual translation and amino acid sequence analysis indicates that 3DE 3␤-reductase is closely related to those of the third superfamily of oxidoreductases.

EXPERIMENTAL PROCEDURES
3DE 3␤-Reductase Purification and Protein Sequencing-3DE 3␤reductase from hemolymph of S. littoralis was purified to homogeneity using a combination of polyethylene glycol 6000 precipitation and successive FPLC fractionation on Mono Q, phenyl Superose (twice), and hydroxyapatite columns, as described previously (6). Purified protein was subjected to SDS-PAGE on 10% gels, electrotransferred to ProBlott TM membrane, and visualized by Coomassie staining (9). A single band was observed, excised and sequenced by an automated pulsed liquid-phase sequencer (Applied Biosystems 471A). The NH 2 -terminal amino acid sequence thus determined was as follows, ATIDVPMLKM-LNDREMPAIALGTYLGFDKG. To obtain sequences of the interior region of the enzyme, the purified protein was resolved by SDS-PAGE as described above, visualized by Coomassie staining, and the corresponding band was excised and cleaved with trypsin. The resulting tryptic peptides were purified by high-performance liquid chromatography and * This work was supported by a grant from The Leverhulme Trust. 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 GenBank TM /EBI Data Bank with accession number(s) AJ131966.
sequenced. The sequences of four of the resulting peptides were determined. These peptides had the following sequences: peptide 1, HFDTA-AIYNTEAEVGEAIR; peptide 2, FGMDLPGPK; peptide 3, LKEEEIEK;  peptide 4, INQFNSNTR. cDNA Cloning and Sequencing-A PCR-based cloning strategy was used to clone a cDNA fragment encoding the region between the NH 2 terminus and an internal peptide sequenced as described above. Three degenerate primers were synthesized. Primer N was designed on the basis of a part of the NH 2 -terminal amino acid sequence (5Ј-GCGAAT-TCYTNAAYGAYMGNGARATGCC, where Y represents T/C, R is A/G, M is A/C, N is A/T/C/G); reverse primers A and B were made according to the sequences of peptide 1 and peptide 2 (primer A, 5Ј-CAGGATCC-ACYTCNGCYTCNGTRTTRTA; primer B, 5Ј-ATGGATCCCCNGGNA-RRTCCATNCCRAA). An EcoRI restriction site was built into primer N and a BamHI site into primer A and primer B for use in cloning; these are underlined in the primer sequences.
Since the site of synthesis of the reductase enzyme was unknown, total RNA was extracted using TRIzol (Life Technologies, Inc.) from total insect tissues (including hemolymph, but with the cuticle, head, and gut contents removed) dissected from larvae 90 h into the last larval instar. mRNA from total RNA was isolated using Dynabeads mRNA Purification Kit (Dynal) (UK) Ltd. First strand cDNA was reverse transcribed from the mRNA using a 1 st Strand cDNA Synthesis Kit from Boehringer Mannheim with either random primer p(dN) 6 supplied with the kit or Q T adapter primer, 5Ј-CCATCAGTGCTAGA-CAGCTAAGCTTGAGCTCGGATCC(T) 17 (modified from Ref. 10). cDNA synthesized with random primer served as template for the PCR in which the above degenerate primers were used. PCR reactions were carried out as follows: 1 cycle of 94°C for 4 min and 4 cycles of 94°C for 30 s, 37°C for 45 s, 72°C for 2 min; followed by 31 cycles of 94°C for 30 s, 58°C for 45 s, 72°C for 2 min. PCR products were analyzed by electrophoresis on a 1% agarose gel. This revealed that PCR with primer N and primer B yielded a product of approximately 660 bp, while PCR with primer N and primer A yielded two products of approximately 380 and 180 bp, respectively. The 660-bp PCR product was gel-purified using Hybaid TM DNA Purification Kit (Hybaid) and used as a template along with primer N and primer A in a second PCR reaction using the same cycling conditions as described above. This second PCR only yielded the 180-bp product, implying that peptide 1 is located in between the NH 2 terminus and peptide 2.
PCR products were purified, digested with EcoRI and BamHI, and subsequently cloned into EcoRI/BamHI-digested pBluescript vector. Transformants were screened by colony PCR using primers T 3 and T 7 (5Ј-ATTAACCCTCACTAAAG and 5Ј-AATACGACTCACTATAG, respectively), and those showing the correct size of inserts were propagated in LB broth containing 100 g/ml ampicillin and plasmid DNA was purified after 16 h incubation at 37°C. Double-stranded DNA sequencing was performed by the dideoxy termination method using Sequenase Version 2.0 (USB TM , Amersham Pharmacia Biotech).
Rapid Amplification of cDNA 5Ј-and 3Ј-Ends (5Ј-and 3Ј-RACE)-5Јand 3Ј-RACE were carried out to obtain the 5Ј-and 3Ј-ends of the cDNA. A 5Ј-RACE System (Life Technologies, Inc.) was used to amplify the 5Ј terminus of the message for sequencing. Briefly, a gene-specific primer 5Ј-1 (5Ј-TTCAGCCTCCGTGTTGTAG) was hybridized to the mRNA prepared as described above and cDNA was synthesized using Superscript II reverse transcriptase. The RNA was then degraded with RNase mix (RNase H and RNase T1), and the cDNA was purified by GlassMax spin cartridge supplied with the kit. A poly(dC) tail was added to the 3Ј terminus of the purified cDNA using dCTP and terminal deoxynucleotidyl transferase, and the cDNA region corresponding to the 5Ј-end of the mRNA was amplified by two successive rounds of PCR using additional gene-specific primers 5Ј-2 and 5Ј-3 (5Ј-TCGAATTCGTGTCTGTAC-CCGAGGTC and 5Ј-GCGAATTCTTGCATCACTACATTGCG, respectively, which incorporated an EcoRI site (underlined)), together with the anchor primers supplied by the manufacturer. The second round of PCR yielded a single amplified product of approximately 240 bp. To amplify the 3Ј-end, another two gene-specific primers C 1 and C 2 (5Ј-GAAGACCTGAT-CACGTACG and 5Ј-GCGAATTCAGGCACCATTGTCATGGGC, respectively, incorporating an EcoRI site (underlined)) in combination with the adapter primers Q 0 and Q 1 (5Ј-CCATCAGTGCTAGACAGCT and 5Ј-TA-AGCTTGAGCTCGGATCC, respectively, modified from Ref. 10) were used for the PCR in which Q T primed cDNA served as template (see earlier). The nested PCR with primers C 2 and Q 1 yielded a product of approximately 750 bp. The 5Ј-and 3Ј-RACE products were digested with the appropriate restriction enzymes, cloned into pBluescript vector, and sequenced across both strands.
Construction of a cDNA Containing an Open Reading Frame and 3Ј-Noncoding Region-The DNA sequence containing the entire open reading frame and 3Ј-noncoding region was amplified by PCR from the cDNA synthesized with primer Q T as described above using the following gene-specific primers: ␤-5Ј (5Ј-GCGAATTCATGTTTCGCGC-CAGTTTT) and ␤-3Ј (5Ј-CGGGATCCTGCAGAGATTGATTTCACATA-TT). These primers were designed to incorporate EcoRI, and BamHI and PstI sites (underlined in the primer sequences) into the 5Ј-and 3Ј-ends of the sense strand, respectively. PCR was conducted as follows: 1 cycle of 94°C for 4 min, and 4 cycles of 94°C for 30 s, 37°C for 45 s, 72°C for 2.5 min, followed by 31 cycles of 94°C for 30 s, 45°C for 45 s, 72°C for 2.5 min. The resulting PCR product was gel-purified, digested with EcoRI and PstI, ligated into pBluescript vector that had been previously digested with EcoRI and PstI, transformed into Escherichia coli strain DH5␣, selected, and sequenced. The sequence of three independent cDNA clones were compared to detect errors that could have occurred during the reverse-transcription and the PCR amplification.
Northern and Southern Blot Analysis-Total RNA from various tissues or whole animals with the gut content removed was isolated using TRIzol reagent (Life Technologies, Inc.). 10 g of RNA was fractionated on a formaldehyde/agarose gel, transferred to Electran® nylon membrane (BDH), and hybridized with a probe corresponding to the open reading frame and 3Ј-noncoding region of the 3␤-reductase cDNA. Probes were radiolabeled by random priming (Boehringer Mannheim), and loading was normalized by probing with a PstI fragment of the A3 actin gene from B. mori (a gift from Dr. A. Mange, Université Claude Bernard Lyon 1) (11) and an 18 S rRNA fragment from mouse (a gift from Ms Yi-Ping Bao, Liverpool University). Blot signals were quantified using a Molecular Dynamics densitometer.
Total genomic DNA was prepared from whole animals using Qiagen Genomic-Tip 500/G (Qiagen), following the manufacturer's recommendations. 10-g aliquots of DNA were digested with EcoRI, SalI, HindIII, PvuII, NcoI as well as EcoRI ϩ SalI, fractionated on a 1% agarose gel, transferred to a nylon membrane, and hybridized using radiolabeled probes corresponding to different regions of the 3␤-reductase cDNA. Prehybridization and hybridization were carried out using QuikHyb (Stratagene) under the conditions recommended by the manufacturer. The blots were washed at high stringency (0.1 ϫ SSC, 0.1% SDS) and labeled bands visualized by autoradiography.
Preparation of Antibodies and Western Immunoblot Analysis-Polyclonal antibodies against 3DE 3␤-reductase were raised in chickens using the purified protein as antigen. The purified protein was first resolved by SDS-PAGE on a 10% gel, visualized with 0.25 M KCl. The resulting white band was cut out, extracted with 50 mM Tris-HCl buffer, pH 7.2, containing 0.1 mM EDTA, 150 mM NaCl, and used as antigen. About 50 g of the antigen was emulsified with complete Freund's adjuvant and injected into a chicken. Booster injections were made 12 and 20 days later with 100 g of the antigen. Extraction of antibodies from the yolk of the immunized chicken eggs with polyethylene glycol was carried out following the procedures described by Gassmann et al. (12).
Western blotting was carried out essentially as described previously (13). Briefly, hemolymph samples taken at various times during the last larval instar were separated by SDS-PAGE on a 10% gel in accordance with the method of Laemmli (14) and electrophoretically transferred to a nitrocellulose membrane (Schleicher and Schuell) (15). Immunoblotting was performed with the antibodies at a final dilution of 1:500. Blots were developed by incubation with an alkaline phosphatase-conjugated goat anti-chicken IgG at a final dilution of 1:3000, followed by incubation in 15 ml of alkaline phosphatase color-development buffer (100 mM NaHCO 3 , pH 9.8) to which was added 150 l each of Bio-Rad AP color reagents A and B. Immunoreactive protein was quantified using a Molecular Dynamics densitometer.

RESULTS
Cloning of the cDNA Encoding 3DE 3␤-Reductase-Using a combination of polyethylene glycol 6000 precipitation and column chromatography on Mono Q, phenyl Superose, and hydroxyapatite, we purified 3DE 3␤-reductase to homogeneity from Spodoptera hemolymph. A silver-stained SDS gel revealed that the purified 3DE 3␤-reductase consisted of a single protein band of apparent molecular mass 36 kDa (see Fig. 4. in Ref 6). NH 2 -terminal and tryptic peptide amino acid sequences were obtained from this 36-kDa protein as detailed under "Experimental Procedures." At the time when these sequence data were obtained, data base searches using FASTA revealed that these sequences were novel.
A PCR-based cloning strategy, as detailed under "Experi-mental Procedures," allowed us to obtain a clone corresponding to the sequence between nucleotides 142-785 in Fig. 1. Genespecific primers derived from this sequence were synthesized and used for 5Ј-and 3Ј-RACE to obtain the 5Ј-and 3Ј-ends of the clone. 5Ј-RACE produced, homogeneously, a cDNA clone of 237 bp that contains two putative translation start sites at nucleotides 23-25 and 68 -70, respectively. On the other hand, 3Ј-RACE resulted in heterogeneous cDNA clones. Sequencing experiments revealed that this resulted from multiple polyadenylation signals (AATAAA) occurring in the 3Ј-noncoding region which directs a poly(A) tail to be added to different sites. The longest cDNA clone obtained with 3Ј-RACE was 746 bp in length and contained a stop codon and three putative polyadenylation signals. Taken together, all the overlapping cDNAs span a total of 1425 bp. This sequence was further confirmed by sequencing of a cDNA clone, which was constructed to contain the open reading frame and 3Ј-noncoding region, as detailed under "Experimental Procedures." As shown in Fig. 1, using the first ATG as the start codon, the full-length 3DE 3␤-reductase cDNA encodes a protein of 345 amino acids with a predicted molecular weight of 39,591. The predicted molecular weight of mature protein is 37,689, which is similar to its apparent M r observed in our SDS-PAGE analysis. The protein is mildly acidic with an estimated pI of 6.23.
Similarity of the Deduced Amino Acid Sequence to Those of the Third Superfamily of Oxidoreductases-The deduced amino acid sequence for the cDNA coding region was compared with all sequences in the data base Swiss-Prot using FASTA in the GCG package. The cDNA gene product was found to be most similar to proteins which belong to the third superfamily of oxidoreductases. 3DE 3␤-reductase particularly resembles mammalian aldose reductases. Among the 20 proteins most closely matched, six are mammalian aldose reductases, others include a mammalian alcohol dehydrogenase, a probable mammalian trans-1,2-dihydrobenzene-1,2-diol dehydrogenase, two mammalian aldose reductase-related proteins, an amphibian lens protein, a putative reductase from Leishmania, a mammalian 3␣-hydroxysteroid dehydrogenase, a mammalian chlordecone reductase, a mammalian trans-1,2-dihydrobenzene-1,2diol dehydrogenase, a yeast xylose reductase, a mammalian 3-oxo-5-␤-steroid 4-dehydrogenase, a plant D-sorbitol-6-phosphate dehydrogenase, a hypothetical 37.1-kDa protein in hxt5-cdc12 intergenic region from yeast and a plant 6-deoxychalcone synthase. The amino acid similarity extends across the entire length of the proteins compared. Fig. 2 shows the comparison of the deduced amino acid sequence of mature 3DE 3␤-reductase to some proteins corresponding to different subgroups of the third superfamily of oxidoreductases. It demonstrates that there is 42.1, 41.3, 39.2, 37.6, and 35.4% identity, respectively, between 3DE 3␤-reductase and human alcohol dehydrogenase, rabbit aldose reductase, rat 3-oxo-5-␤-steroid 4-dehydrogenase, rat 3␣-hydroxysteroid dehydrogenase, and frog crystallin.
Tissue Distribution and Developmental Expression of 3DE 3␤-Reductase mRNA-As demonstrated in Fig. 3A, a cDNA probe representing the protein coding and 3Ј-noncoding regions of 3DE 3␤-reductase detected a major transcript of approximately 1.4 kilobases in a variety of tissues prepared from larvae 78 h into the last larval instar. The size of this major transcript is consistent with that of the cDNA obtained. The highest expression of 3DE 3␤-reductase is detected in Malpighian tubules, followed by midgut and fat body, with low levels expressed in hemocyte. No detectable expression was found in the central nervous system (data not shown). Less expression is detected in Malpighian tubules and hemocytes dissected at 46 h into the last larval instar, but no expression was found in all four tissues from early instar larvae. It is noticeable that RNAs at approximately 2.6 and 3.6 kb are also recognized in fat body by this probe but at lower intensity (Fig.  3A). It remains to be determined whether these larger RNAs represent pre-mRNAs for 3DE 3␤-reductase or other mRNAs with significant sequence similarity to 3␤-reductase, although the former seems most probable due to the high stringency used. The signals detected by the actin probe used for normalization are substantially lower in the fat body than those in other tissues (Fig. 3B). rRNA probe was therefore used to further verify the equal loading (Fig. 3C). Some apparent degradation of rRNA is observed in midgut, Malpighian tubules, and hemocytes, which may be due to the fact that 28 S ribosomal RNA in insects frequently occurs as two equal-sized fragments due to RNase activity within certain cells (16). This activity seems least in the fat body and the total amount of hybridization to the rRNA probe seems comparable with the other samples.
Northern analysis of total RNA isolated from whole animals from different developmental stages of the last larval instar revealed that the 3DE 3␤-reductase mRNA is only expressed in the second half of the instar (Fig. 4). The mRNA starts to be detected at 66 h into the last larval instar and increases in intensity, as normalized to actin and rRNA, to 90 h with an apparent decrease at 96 h before increasing to its highest levels observed at 114 h.
Southern Blot Analysis of Genomic DNA-Genomic DNA was prepared from whole animals and digested with various restriction enzymes, of which SalI, HindIII, and NcoI sites are found in the 3␤-reductase cDNA, whereas EcoRI and PvuII sites are not found in the cDNA. The resultant DNA fragments were analyzed by Southern blot using probes representing different regions of the cDNA. As shown in Fig. 5A, usually two bands were observed in all of the digested samples, when the blot was subjected to hybridization with a probe representing the coding and 3Ј-noncoding regions of the 3␤-reductase cDNA. However, when the same blot was re-probed with a cDNA fragment that corresponded to the NH 2 -terminal region to the first HindIII site in the cDNA, essentially only one band was detected except for the restriction digestion with PvuII, where two bands were observed (Fig. 5B). Again, a probe that represents the region from the SalI site to the poly(A) site detected only one band in each case (Fig. 5C). The analysis displayed in Fig. 5 is most simply interpreted if the 3␤-reductase is encoded by a single copy gene in the Spodoptera genome and if this gene has at least one intron that contains EcoRI, PvuII, and SalI sites.

DISCUSSION
Using a reverse transcription-PCR-based cloning strategy, employing degenerate primers designed on the basis of the partial amino acid sequences of fragments of 3DE 3␤-reductase, together with 3Ј-and 5Ј-RACE, the complete cDNA encoding the enzyme was isolated and sequenced. The predicted amino acid sequence of the cDNA contained the NH 2 -terminal sequence of the enzyme as well as the four internal peptides obtained from tryptic digestion of the band corresponding to the M r 36,000 polypeptide, confirming that we had cloned the 3DE 3␤-reductase cDNA. The presence of 3DE 3␤-reductase enzyme in hemolymph is due to secretion from tissues that synthesize it. The first 17 predicted amino acids appear to constitute a signal peptide, since the sequence shares significant characteristics with the signal sequences of some eukaryotic secretory proteins (17): (i) the 17 amino acids are within the typical range of 13-36 residues of signal peptides; (ii) a positively charged residue, arginine, is found in the aminoterminal part of the signal; (iii) it contains a highly hydrophobic stretch of 11 residues; (iv) a serine residue at the end of the signal has a small neutral side chain that is recognized by signal peptidase for cleavage.
Three putative polyadenylation signals were found in the 3Ј-noncoding region of the cDNA, suggesting alternative polyadenylation sites may exist, as the cleavage and polyadenylation specificity factor may bind to any one of these three sequences (18). Indeed, during sequencing following 3Ј-RACE, we found that one cDNA clone showed a shorter 3Ј-noncoding region, with the poly(A) tail attaching 14 nucleotides downstream of the second AATAAA sequence (Fig. 1). Furthermore, three ATTTA sequences were found in the 3Ј-untranslated region (3Ј-UTR) of the 3␤-reductase cDNA, all located in between the second and the third polyadenylation signals, with two occurring in the sequence which is missing in the shorter 3␤-reductase cDNA (Fig. 1). In fact, the 3Ј-untranslated region has been increasingly recognized as one of the important sequences which may contain distinct elements involved in regulating mRNA degradation (19). A most well studied example is the AU-rich element (ARE) found on many unstable mRNAs. The sequence consensus for the AREs is loosely defined as AUUUA repeated once or several times within the 3Ј-UTR. This pentanucleotide is often found within an AU-rich region of the mRNA. It seems to be clear that (i) both nuclear and cytoplasmic proteins can specifically bind to the ARE, (ii) the binding activity of some of those proteins appears to increase or decrease in response to the changes in cellular metabolism that lead to alterations in the stability of ARE-containing mRNA, and (iii) ARE activity can be mediated by a complex of proteins as in the case of a 20 S complex (19). We were interested in addressing the question as to whether the alternative polyadenylation, using AATAAA signals at different locations in the 3␤-reductase cDNA, is related to the regulation of mRNA stability and what the role of the AREs in the 3Ј UTR is likely to be. For this, total RNA from fat body, midgut, and Malpighian tubules from 78 h last larval instar larvae was subjected to Northern analysis using probes corresponding, respectively, to nucleotides 1072-1176, 1177-1281, and 1282-1386 of the cDNA (see Fig. 1). Because all three probes gave identical signals, the AREs in the 3Ј-UTR do not seem to contribute to the regulation of the mRNA stability in the tissues examined (data not shown).
It has long been recognized that oxidoreductases, dependent upon nicotinamide coenzymes, can be divided into two extended superfamilies, according to their structural type and catalytic mechanism (20). The enzymes of one superfamily contain catalytically important zinc, many utilizing primary alcohols as substrates, though use of secondary alcohols and other activities are also noticed (21)(22)(23)(24). A structurally noteworthy feature of these proteins is the highly conserved cysteine residues, involved in zinc binding, scattered within the first half of the amino acid sequences. Enzymes of the other superfamily generally do not contain any catalytically active metal atom, have shorter subunits, and frequently but not always use large secondary alcohols as substrates (25)(26)(27)(28). Recently, a third superfamily has been proposed (29,30). Enzymes of this superfamily are not homologous to those of the two superfamilies mentioned above and their polypeptide chain lengths generally lie between those of the two superfamilies. They contain no metal atom and their substrates include primary and secondary alcohols. Sequence data base searching revealed that the predicted amino acid sequence of 3DE 3␤reductase is most similar to the proteins that belong to the third superfamily of oxidoreductases (30). The existence of only one cysteine residue (in position 192, see Fig. 1) in the predicted amino acid sequence of the mature protein indicated that the 3␤-reductase of S. littoralis contains no zinc. Motif searching (MOTIF at http://www.motif.genome.ad.jp/) with the predicted amino acid sequence of 3␤-reductase uncovered the three consensus patterns present in aldo/keto reductases (residues 68 -85, 163-180, 277-292, Fig. 1), which are specific to this family of proteins. In addition, its molecular weight (37,689/subunit) is in the same range as found for these proteins. Hence, we conclude that 3DE 3␤-reductase is a new member of the third superfamily of oxidoreductases.
Northern analysis has revealed that the mRNA for 3␤-reductase has a wide tissue distribution with considerable variability in the abundance of the mRNA among the tissues examined. It is not clear at present which tissue is the major source for the 3␤-reductase activity in the hemolymph or whether all the tissues containing 3␤-reductase mRNA contribute to the enzyme activity in hemolymph. The cytosolic enzymes reversibly interconverting ecdysone and 3DE (ecdysone oxidase and 3DE 3␤-reductase) have been reported to be widely distributed, even in tissues that do not reduce 3DE to 3-epiecdysone (31,32). If such tissue 3DE 3␤-reductase is closely related to the hemolymph 3␤-reductase, the wide distribution of transcripts in the Northern blots could be explained by hybridization of the probe to mRNA corresponding to such tissue 3␤-reductase. Sakurai and co-workers (33) examined the tissue distribution of the 3␤-reductase in B. mori at the protein level, as well as the level of enzyme activity, and they too found a wide tissue distribution. The absence of the 3␤-reductase protein in central nervous system in B. mori is consistent with our observation that 3␤-reductase mRNA is not detectable in that tissue from S. littoralis. However, the fact that Malpighian tubules and midgut from B. mori do not contain detectable enzyme protein seems to be surprising, since we found that the mRNA for 3␤-reductase is abundantly transcribed in these tissues from S. littoralis. Preliminary results of Western blotting with tissues dissected from S. littoralis at 90 h into the last instar revealed a similar distribution pattern as that of B. mori, with 3␤reductase being detected in hemolymph as well as hemocytes, but little in midgut and none in Malpighian tubules (data not shown). It seems, therefore, that although the mRNA is abundantly transcribed in Malpighian tubules and midgut, none or only little is translated into the protein in these tissues. This observation may indicate that selective translation of mRNA is indeed an important step in regulation of expression of the gene encoding 3DE 3␤-reductase.
Northern analysis also revealed that the mRNA for the enzyme is only expressed in the second half of the last larval instar. The mRNA starts to be detected 48 -66 h into the last larval instar, at which time the actual enzyme activity is just about to appear (6). The increase in expression can be seen from 72 to 90 h, which coincides well with the sharp rise in the enzyme activity. The small decline in the expression between 90 and 96 h is also coincident with the decrease in the enzyme activity at this time. The second increase in the mRNA level observed between 102 to 114 h is presumably responsible for the small rise in the enzyme activity at the end of the instar. It appears evident, therefore, that the change in the activity of 3␤-reductase is, at least in part, regulated by gene expression at the transcriptional level, although, as mentioned above, post-transcriptional regulation may also be as important in the expression of 3DE 3␤-reductase in certain tissues.
Southern analysis suggest that 3␤-reductase is encoded by a single copy gene in the genome of S. littoralis. Although the details of the organization of the gene are far from clear, it is predicted from the results of Southern analysis that the gene encoding 3␤-reductase has at least one intron that contains EcoRI, SalI, and PvuII sites. It remains to be determined whether the existence of two larger RNAs detected in fat body by a probe representing the entire coding region and 3Ј-noncoding region of the cDNA is due to the possible occurrence of intermediates in the process of mRNA splicing.
The availability of the cDNA clones encoding 3DE 3␤-reductase should facilitate studies of genomic organization, regulation of expression of the gene, structure-function relationships of the enzyme, and its involvement in regulation of the production of ecdysteroids.