THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 40, Issue of October 3, pp. 38149–38158, 2003 © 2003 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Constructing a Feedback Loop with Circadian Clock Molecules from the Silkmoth, pernyi*□S

Received for publication, June 30, 2003 Published, JBC Papers in Press, July 17, 2003, DOI 10.1074/jbc.M306937200

Dennis C. Chang,a,b,c,d Harriet G. McWatters,a,e Julie A. Williams,a,f Anthony L. Gotter,a,c,g Downloaded from Joel D. Levine,a,h and Steven M. Repperta,b,c,i From the aLaboratory of Developmental Chronobiology, MassGeneral Hospital for Children, Massachusetts General Hospital, Boston, Massachusetts 02114, the bProgram in Neuroscience, Harvard Medical School, Boston, Massachusetts 02115, and the cDepartment of Neurobiology, University of Massachusetts Medical Center, Worcester, Massachusetts 01605 www.jbc.org

Circadian clocks are important regulators of behavior Circadian rhythms are driven by cell autonomous pacemak- and physiology. The circadian clock of Drosophila de- ers consisting of molecular feedback loops (1). In the fruit fly pends on an autoinhibitory feedback loop involving Drosophila melanogaster, circadian feedback loops require the at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006 dCLOCK, CYCLE (also called dBMAL, for Drosophila transcription factors dCLOCK (dCLK) and CYCLE (CYC, also brain and muscle ARNT-like protein), dPERIOD, and called dBMAL) (2). dCLK and CYC heterodimerize to bind dTIMELESS. Recent studies suggest that the clock E-box enhancer elements, presumably through their basic he- mechanism in other may differ strikingly lix-loop-helix (bHLH)1 domains (3). In addition, dCLK and CYC from that of Drosophila. We cloned Clock, Bmal, and each possesses a PER-ARNT-SIM homology (PAS) domain (3– Timeless homologs (apClock, apBmal, and apTimeless) 5), which is thought to facilitate protein-protein interactions (6) from the silkmoth Antheraea pernyi, from which a Pe- and contains two PAS structural motifs (PAS-A and PAS-B) riod homolog (apPeriod) has already been cloned. In and a region downstream of PAS-B called PAC (7). Further Schneider 2 (S2) cell culture assays, apCLOCK:apBMAL downstream, dCLK possesses a glutamine-rich transactivation activates transcription through an E-box enhancer ele- domain (3, 4). One of the genes activated by dCLK:CYC, period ment found in the 5 region of the apPeriod gene. Fur- thermore, apPERIOD can robustly inhibit apCLOCK: (dper), encodes another PAS protein, dPER (2, 3). dPER inhib- apBMAL-mediated transactivation, and apTIMELESS its dCLK:CYC through a non-PAS, C-terminal domain, called can augment this inhibition. Thus, a complete feedback the dCLK:CYC inhibition domain (CCID), thus completing the loop, resembling that found in Drosophila, can be con- key circadian negative feedback loop (2, 3, 8). dCLK:CYC also structed from silkmoth CLOCK, BMAL, PERIOD, and activates transcription of the timeless (dtim) gene, whose pro- TIMELESS. Our results suggest that the circadian auto- tein product, dTIM, regulates dPER protein stability and nu- inhibitory feedback loop discovered in Drosophila is clear transport and may also contribute to the inhibition of likely to be widespread among . However, dCLK:CYC (2, 3, 8). dTIM also plays a role in light input to the whereas the transactivation domain in Drosophila lies clock (2). in the C terminus of dCLOCK, in A. pernyi, it lies in the The circadian feedback loops of mice depend upon mCLK: C terminus of apBMAL, which is highly conserved with mBMAL1, the elements of which are homologs of dCLK:CYC the C termini of BMALs in other insects (except Dro- (9). Cryptochromes (mCRYs) are the main inhibitors of mCLK: sophila) and in vertebrates. Our analysis sheds light on mBMAL1, whereas mouse PERs seem to regulate mCRY nu- the molecular function and of clock genes in clear entry (10). Mammals lack a true ortholog of dTIM (11, 12). the kingdom. The presence of CLOCK, BMAL, and PERIOD (and other mol- ecules) at the heart of circadian clocks in both Drosophila and * This work was supported in part by National Institutes of Health mice suggests that both insects and vertebrates inherited their Grant R01-NS39303. The costs of publication of this article were de- clocks from a common ancestor (1). However, the differences in frayed in part by the payment of page charges. This article must fly and mouse clock mechanisms indicate that animal clocks therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. have diversified during the course of evolution. □S The on-line version of this article (available at http://www.jbc.org) Indeed, Drosophila even differs from other insects in its clock contains Figs. S1 and S2. mechanism. The temporally regulated nuclear transport of The nucleotide sequence(s) reported in this paper has been submitted dPER and dTIM (cytoplasmic during the afternoon, nuclear at to the GenBankTM/EBI Data Bank with accession number(s) AY330486 and AY330487. night) in pacemaker neurons (13, 14) is thought to be critical d Supported by a Howard Hughes Medical Institute Predoctoral for Drosophila clock function (2). In contrast, PER and TIM in Fellowship. the Chinese oak silkmoth, Antheraea pernyi (apPER and ap- e Current address: Dept. of Plant Sciences, Oxford University, South TIM), are primarily cytoplasmic at all times of day in brain Parks Rd., Oxford OX1 3RB, United Kingdom. f Current address: Dept. of Neuroscience, Howard Hughes Medical neurons (15). PER in a damselfly (order Odonata) is also cyto- Institute, University of Pennsylvania, 232 Stemmler Hall/6074, Phila- plasmic throughout the day (16). In fact, in the early morning, delphia, PA 19104. when dPER is nuclear in Drosophila (14), PER is cytoplasmic g Current address: Division of Human Genetics and Molecular Biol- ogy, Children’s Hospital of Philadelphia, 3516 Civic Center Blvd., ARC 1002, Philadelphia, PA 19104. h Current address: Dept. of Biology, MS-008, Brandeis University, 1 The abbreviations used are: bHLH, basic helix-loop-helix; PAS, Waltham, MA 02454. PER-ARNT-SIM homology; CCID, dCLK:CYC inhibition domain; i To whom correspondence should be addressed: Dept. of Neurobiol- mCRY, cryptochrome; S2, Schneider 2; NLS, nuclear localization se- ogy, University of Massachusetts Medical School, 364 Plantation St., quence; SFM, serum-free medium; PBS, phosphate-buffered saline; Worcester, MA 01605. Tel.: 508-856-6148; Fax: 508-856-6233; E-mail: CLD, cytoplasmic localization domain; aa, amino acid(s); BCTR, BMAL [email protected]. C-terminal region; DAPI, 4Ј,6-diamidino-2-phenylindole.

Supplemental Material can be found at: This paper is available on line at http://www.jbc.org 38149 http://www.jbc.org/cgi/content/full/M306937200/DC1 38150 A Silkmoth Circadian Feedback Loop in representatives of many insect orders: Thysanura, Ephemer- 25 °C in Schneider’s Drosophila medium (Invitrogen) with 9% heat- optera, Orthoptera, Plecoptera, Hemiptera, Coleoptera, Hy- inactivated fetal bovine serum (Invitrogen), and Sf9 cells were main- menoptera, Trichoptera, and even Diptera (16). Thus the pace- tained at 27 °C in Sf-900 II serum-free medium (SFM, Invitrogen). Transient transfection of plasmids involved mixing DNA with 5–10 ␮l maker mechanism in A. pernyi () may be more of CellFECTIN reagent (Invitrogen), and 15 min later, applying the representative of insect clocks than the mechanism in mixture to S2 cells in Drosophila SFM (Invitrogen) supplemented with Drosophila. 2mML-glutamine, or to Sf9 cells in Sf-900 II SFM. Approximately 4 h To further our understanding of insect clock function and later, S2 cells were fed with an equal volume of 9% fetal bovine serum evolution, we cloned Clock, Bmal, and Timeless homologs from in Schneider’s medium, and Sf9 cells were given fresh SFM. Cells were Downloaded from ϳ A. pernyi (apClk, apBmal, and apTim). We characterized their incubated for 48 h before use. Transcription Assays—To test the ability of apCLK:apBMAL to ac- functions and interactions with apPer using Drosophila tivate transcription in S2 cells, we used a modified version of the Schneider 2 (S2) cells, because many aspects of both fly and transcription assay of Darlington and coworkers (3). In brief, 1 ng of Ϯ mammalian circadian biochemistry can be simulated in S2 apClk 1ngofapBmal was co-transfected with 10 ng of apPer 4Ep- cells. These include dCLK:CYC-mediated transactivation via hs-luc and 30 ng of ␤gal pAc5.1V5/HisA (␤gal-V5). For each assay, a

E-boxes, dPER inhibition of dCLK:CYC (3, 8), mCLK:mBMAL1 control transfection, including only the reporters apPer4Ep-hs-luc and www.jbc.org ␤ transactivation, and mCRY inhibition of that activation (17, gal-V5, was used to establish baseline reporter activity. In the apPer promoter assays, we substituted apPer genomic DNA constructs for 18). In S2 cells, apCLK:apBMAL can activate transcription apPer4Ep-hs-luc. To measure the inhibitory activity of apPer/apTim,we through an E-box in the apPer promoter, apPER inhibits ap- co-transfected 2–250 ng of apPer, apTim,orapPer mutant constructs

CLK:apBMAL-mediated transcription, and apTIM can aug- with apClk, apBmal, and the reporters. The total amount (nanograms) at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006 ment the inhibitory activity of apPER. Like dPER, apPER of DNA in each transfection was normalized using empty vector inhibits apCLK:apBMAL through a C-terminal domain that (pAc5.1V5). After transfection and a 48-h incubation at 25 °C, cells were also contains the primary nuclear localization sequence (NLS) harvested, washed in phosphate-buffered saline (Dulbecco’s PBS, In- vitrogen), and lysed in lysis buffer (Promega). Cell extracts were as- of apPER. Thus, despite initial appearances, a Drosophila-like sayed for ␤-galactosidase and luciferase activities using commercial negative feedback loop involving CLOCK, BMAL, PERIOD, assay kits (Tropix and Promega, respectively) and an MLX microtiter and TIMELESS may exist in the silkmoth A. pernyi, and per- plate luminometer (Dynex). Average ratios of luciferase activity to haps in other insects with similarly cytoplasmic PER. However, ␤-galactosidase activity were computed, and, for the apPER/apTIM the domain responsible for apCLK:apBMAL-mediated transac- inhibition assays, the ratios were normalized so that the relative acti- tivation differs strikingly from that found in Drosophila. Our vation by apCLK:apBMAL alone equaled 100%. The assays in Sf9 cells were similar but used ϳ10-fold higher plasmid doses. comparative analysis provides new insight into the molecular Immunocytochemistry—S2 (or Sf9) cells seeded onto glass coverslips evolution of circadian pacemakers in . in 6-well tissue culture plates were transfected with ϳ300 ng (or 1 ␮g for Sf9 cells) V5-tagged apPer, apTim,orapBmal constructs. 48 h after EXPERIMENTAL PROCEDURES transfection, the cells were washed in PBS and fixed onto the coverslips Cloning and Sequence Analysis—Fragments of apClock, apBmal, with 5% paraformaldehyde in PBS. The coverslips were then processed and apTimeless were cloned by degenerate PCR. cDNA templates for with blocking solution (10% normal goat serum and 0.2% Triton X-100 PCR were prepared from RNA purified from A. pernyi brains dissected in PBS), primary antibody (monoclonal mouse anti-V5 IgG (Invitrogen) 1–3 days after eclosion. The ends of the coding regions were obtained by 1:500 in blocking solution diluted 1:4 in PBS), three PBS washes, rapid amplification of cDNA ends (Clontech kits). Complete open read- secondary antibody (Cy3-conjugated goat anti-mouse IgG (Jackson) ing frames were obtained by using Pfu Turbo (Stratagene) PCR from 1:300 in diluted blocking solution), two PBS washes, DAPI stain (bis- cDNA. Clones were sequenced at core facilities at Massachusetts benzimide 1:500 in PBS), two brief PBS washes, and a rinse in ddH2O, General Hospital and the University of Massachusetts Medical School. before being mounted onto slides with mounting medium for fluores- Sequence analysis was facilitated by software from the Genetics cence microscopy (Kierkegaard and Perry). Slides were viewed under a Computing Group (GCG) and the NCBI website (www.ncbi.nlm. fluorescence microscope (Olympus IX70) at the Digital Imaging Core nih.gov/BLAST/). Facility at the University of Massachusetts Medical Center. For each Plasmids—S2 cell expression constructs were generated by subclon- slide, 30 cells with strong fluorescent signal were classified as having ing lacZ, apClock, apBmal, apTimeless, apPeriod, or fragments thereof staining in the nucleus, cytoplasm, or both nucleus and cytoplasm. into the pAc5.1V5/HisA vector (Invitrogen) or a modified pAc5.1 vector Scoring was done blind to which construct was being analyzed. Each with the V5 tag at the N terminus. Sf9 cell expression constructs were construct was analyzed using three or more separate transfections. generated by subcloning into a pIB/V5-His vector (Invitrogen) modified so that its multiple cloning site resembled that of pAc5.1V5/HisA. apPer RESULTS promoter constructs were obtained by sequential digests and subclon- ing of genomic DNA fragments into the promoter-less luciferase vector, apClock, apBmal, and apTimeless: Cloning and Sequence TM pGL3 Basic (Promega). To generate E-box hs-luc constructs, the hsp70 Analysis—apClock (GenBank AY330486) encodes a pre- promoter (from XhoItoNcoI) from the dper 4E-box construct provided dicted protein of 611 amino acids with clear homology with by Steve Kay (3) was first subcloned into pGL3 Basic to generate an other CLOCK proteins. For example, the sequence identity “hs-luc” construct. Overlapping PCR primers (with restriction sites at between apCLK and dCLK/mCLK bHLH domains was Ն67%, the ends) were used to generate fragments containing E-boxes (apPer PAS-A motifs were Ն42%, PAS-B was Ն80%, and the PAC was E , mutant E ,E ,E , dper E, and mutant E) and 10- to 11-bp flanking p p m d Ն sequences; these were subcloned into hs-luc. A construct with four 76% (Table I). The high N-terminal sequence similarity clearly identifies apCLK as a homolog of CLK and not of any tandem repeats of the apPer Ep enhancer (with 7- to 8-bp flanking sequences) upstream of hs-luc (“apPer 4Ep-hs-luc”) was also generated other bHLH-PAS transcription factor. In contrast, the C ter- by overlapping primers containing appropriate restriction sites. The mini (amino acids downstream of the PAS domain) in silkmoth, pAct-dClk construct was provided by Steve Kay (3). fly, and mouse CLK proteins are not well conserved (6–17%; Site-directed Mutagenesis—NLS mutant apPer M693–4 was gener- Table I). In particular, although both dCLK and mCLK possess ated using complementary primers containing the desired mutation, along with a diagnostic restriction site, to PCR-amplify apPer-V5 tem- polyglutamine repeats in their C termini, apCLK has none plate. The PCR reaction was treated with DpnI to remove template (Fig. 1A). This is surprising, because the glutamine enrichment DNA and used to transform cells (DH5␣). apPer of mCLK/dCLK C termini is thought to be indicative of their M707–9 and apPer M717–9 were generated by designing primers (both transactivation function (3, 4, 19). However, the absence of sense and antisense) containing the mutation and a restriction site for polyglutamine stretches in apCLK does not rule out the pres- subcloning. N-terminal and C-terminal fragments, overlapping at the ence of a transactivation domain in its C terminus. MOP4 (also mutation site, were generated by PCR, subcloned, and sequenced. The C-terminal fragment was subcloned downstream to the N-terminal called NPAS2), a CLK paralog found in mammals, lacks long piece to produce the correct, full-length mutant construct. polyglutamine repeats, yet is capable of transcriptional activa- Insect Cell Culture and Transfections—S2 cells were maintained at tion when partnered with mBMAL1 (20). A Silkmoth Circadian Feedback Loop 38151

TABLE I Interestingly, whereas CYC terminates at the end of the PAS Domain-by-domain comparison of amino acid identities among domain (3, 5), the silkmoth and mouse BMAL proteins possess CLOCK/BMAL/TIMELESS homologs a C-terminal region of a few hundred amino acids downstream Peptide sequences for apCLK, dCLK, and mCLK were aligned using ϳ the Pileup program in the Genetics Computing Group (GCG) software of PAS. Moreover, the very C-terminal part of this region ( 40 package. Peptide sequences for apBMAL, CYC, and mBMAL1, and amino acids) is very well conserved (70% identical) between sequences for apTIM and dTIM were aligned using the same software. silkmoth and mouse BMALs (Table I and Fig. 1B). This C- bHLH, PAS-A, and PAS-B regions were demarcated according to the terminal end of mBMAL1 was previously implicated in tran-

SMART protein domain analysis available on the NCBI website Downloaded from (www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). The PAS linker is the scriptional activation in both yeast and mammalian cell culture region between PAS-A and PAS-B. The C terminus of CLK or BMAL (22). The authors of that study suggested that the C terminus was defined as the entire peptide sequence downstream of the PAS of mBMAL1 is the primary transactivation domain of mCLK: ⌬ domain. The “ 19” region was defined according to the entire exon 19 of mBMAL1, and that the C terminus of mCLK may serve only to mCLOCK (19). The BCTR was defined as the final 40 amino acids of apBMAL (which correspond to the final 39 amino acids of mBMAL1). augment or stabilize the activity of mBMAL1 (22). The conser- The Arm/HEAT domains were defined according to the analysis of vation of this BMAL C-terminal region (“BCTR”) in apBMAL Vodovar et al. (24), and the “Arm linker” is the region between the two suggests that apBMAL may also possess a C-terminal transac- www.jbc.org Arm/HEAT domains. The C terminus of TIM was defined as the peptide tivation domain. sequence downstream of Arm/HEAT 2. The CLD was defined according TM to the deletion mutant mapping of dTIM by Saez and Young (25). apTim (GenBank number AF132032) encodes a predicted Sequence gaps were counted as amino acids in the calculation of % protein of 1233 amino acids (GenBankTM number AAF66996) identity. GenBank™ sequences used in this comparison were: apCLK: homologous to dTIM. TIMELESS belongs to a protein family at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006 AY330486, dCLK: AAD10630, mCLK: O08785, apBMAL: AY330487, that includes a second Drosophila protein, dTIMEOUT, or CYC: AAC39124, mBMAL1: BAA81898, apTIM: AAF66996, and dTIM: P49021. dTIM2, which is homologous to mouse (mTIM) and Caenorh- abditis elegans (ceTIM-1) proteins involved in development (11, apCLK versus apCLK versus dCLK versus dCLK mCLK mCLK 12, 23). However, apTIM is more closely related to dTIM than to dTIMEOUT (11, 12), indicating that it is an ortholog of % dTIM, rather than of dTIMEOUT. Whole protein 30 23 18 bHLH 76 67 59 Sequence analysis software has suggested that dTIM pos- PAS-A 45 42 41 sesses two domains of Armadillo (Arm)/HEAT motif repeats PAS linker 30 22 19 (24). The role of these structural features in dTIM are un- PAS-B 80 88 80 known, although functional studies in S2 cells have revealed PAC 76 84 78 C-terminus 17 6 8 that dTIM contains two dPER binding sites (25), one overlap- “⌬19” region 57 31 29 ping the second Arm/HEAT domain and the other lying within the second Arm/HEAT domain (24) (Fig. 1C). In S2 cells, dTIM apBMAL versus apBMAL versus dCYC versus dCYC mBMAL1 mBMAL1 was also shown to possess a nuclear localization sequence % (NLS) and a C-terminal cytoplasmic localization domain (CLD) Whole protein 34 38 33 (25) (Fig. 1C). apTIM and dTIM are 30% identical overall and bHLH 87 68 68 are especially homologous in the Arm/HEAT domains (46% PAS-A 64 70 70 identity in each, Table I). apTIM also has a homologous puta- PAS linker 36 33 41 tive NLS in the region of the dTIM NLS, and a region homol- PAS-B 54 57 59 PAC 30 39 45 ogous to the CLD in its C terminus (Fig. 1C), but the homology C terminus 25 in the C terminus is poor (Table I). BCTR (“T” in 70 apCLOCK:apBMAL Activates Transcription via an apPer Fig. 1B) E-box—To test whether apCLK:apBMAL can activate tran- apTIM versus scription from the apPer promoter, we analyzed a genomic dTIM DNA fragment containing the first two exons of apPer and ϳ15 Whole protein 36 kb of upstream sequence (26). Within this clone, there are three Arm/HEAT 1 46 putative E-boxes (i.e. CACGTG sequences), ϳ3.5 kb, 9.0 kb, Arm linker 30 and 13.4 kb upstream of exon 1 (Fig. 2A). These we designated Arm/HEAT 2 46 C terminus 19 Ep,Em, and Ed, respectively (for proximal, medial, and distal CLD 13 E-boxes). We subcloned genomic fragments of varying lengths upstream from (and including part of) exon 1 into a promoter- less luciferase vector. In S2 cells, we found that apCLK:apB- A subregion of the C terminus is relatively well conserved in MAL do not activate luciferase transcription from any apPer apCLK, dCLK, and mCLK (Fig. 1A). This region corresponds to promoter construct evaluated (data not shown). We performed the peptide sequence encoded by exon 19 of mClk, which is a similar assay in Sf9 cells, which are derived from the spliced out of the mRNA in Clock mutant mice (19). Because Spodoptera frugiperda, but again obtained no activation (data the mutant mCLK protein (mCLK⌬19) is defective for tran- not shown). scriptional activation (21), this region may be an important There are at least three explanations for this negative result. part of the transactivation domain, even though it does not First, the apPer promoter may not be functional in S2 or Sf9 contain polyglutamine stretches. Alternatively, the “⌬19” re- cells, perhaps because essential activators or co-activators are gion may be involved in stabilizing the CLK:BMAL het- absent in those cell lines. This possibility is supported by the erodimer (22), or function as a binding site for important reg- extreme spatial restriction of apPER expression in silkmoth ulatory proteins, such as PER (see below). brain to only ϳ8 cells (15). Second, the apPer genomic frag- apBmal (GenBankTM number AY330487) encodes a pre- ments may contain a binding site for transcriptional repressors dicted protein of 589 amino acids with high homology with endogenously expressed in S2 and Sf9 cells. The inhibitory other BMAL proteins. apBMAL was especially similar to Dro- action of these repressors could be blocking transcription from sophila CYC or mBMAL1 in the bHLH (Ն68% identity), PAS-A the reporter constructs even in the presence of apCLK:apB- (Ն64% identity), and PAS-B (Ն54% identity) regions (Table I). MAL. These first two possibilities are supported by the low 38152 A Silkmoth Circadian Feedback Loop Downloaded from www.jbc.org at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006

FIG.1.A, schematic comparison of apCLOCK, dCLOCK, and mCLOCK protein features. The gray bar represents primary amino acid sequence of each protein (to scale). White box with “bH” ϭ basic helix-loop-helix (bHLH) domain; remaining white boxes ϭ parts of the PAS domain labeled “A” (PAS-A), “B” (PAS-B), and “C” (PAC); gray box containing “⌬” ϭ the region homologous to exon 19 in mCLK, which is eliminated in Clock mutant mice (generating the mCLK⌬19 mutant protein) (19); “Q” ϭ a polyglutamine region, where Ն 60% of the amino acid residues are glutamines; Jrk ϭ the site of the dClkJrk nonsense mutation (4). In the sequence of the “⌬19” region of apCLK, dCLK, and mCLK, amino acids identical in all three species are marked black, those identical in two species are dark gray, and those similar in two or more species are light gray. B, schematic comparison of apBMAL, CYCLE (dBMAL), and mBMAL1 protein features. The gray bar represents primary amino acid sequence of each protein (to scale). White box with “bH” ϭ basic helix-loop-helix (bHLH) domain; remaining white boxes ϭ parts of the PAS domain labeled “A” (PAS-A), “B” (PAS-B), and “C” (PAC); gray box with “T” ϭ region homologous to the transactivation domain of mBMAL1 (22); called the BMAL C-terminal region (BCTR) in the text. In the sequence of the BCTR of apBMAL and mBMAL1, identical amino acids are marked black and similar amino acids are light gray. C, schematic comparison of apTIM and dTIM protein features. The gray bar represents primary amino acid sequence (to scale). Black boxes labeled “Arm” ϭ Arm/HEAT domains; gray circle ϭ nuclear localization sequence (NLS); white box labeled “CLD” ϭ region homologous to the cytoplasmic localization domain (CLD) in dTIM. dPER binding sites of dTIM (25) are underlined. baseline level of luciferase (in the absence of apClk/apBmal) This E-box, the CACGTG and its 4-bp flanking sequences, is generated by the apPer promoter compared with that produced extremely similar (12/14 bases identical) to the functional E- by the hsp70 promoter, which was used as a control (data not box described for the Drosophila clock gene, vrille (27). shown). The third possibility is that the apPer genomic con- To enhance the measurement of apCLK:apBMAL-mediated structs do not contain a viable E-box element to which apCLK: transcriptional activation, we followed the example of others (3, apBMAL can bind to activate transcription. 27, 28) and generated a reporter construct containing four

To test the third possibility, we generated new constructs in tandem repeats of the functional apPer E-box, apPer 4Ep-hs- which individual E-boxes (the dper E-box or apPer Ep,Em,or luc, for use in subsequent assays. apCLK:apBMAL activate Ed) with 10–11 bp of surrounding sequence were subcloned transcription robustly from the apPer 4Ep-hs-luc construct in upstream of the hsp70 promoter in a luciferase reporter con- both S2 cells (Fig. 3) and Sf9 cells (data not shown). struct (Fig. 2B). These constructs would not rely on a possibly apCLOCK:apBMAL Transactivation Depends on the C Ter- inactive promoter for transcription initiation, and (because minus of apBMAL—To determine the domain(s) responsible for minimal fragments of apPer genomic DNA are used) they are transcriptional activation, we generated deletion mutants of not likely to contain repressor elements found in larger frag- apClk and apBmal and tested them in the S2 cell transcription ments of the apPer gene 5Ј-region. We tested these constructs assay (Fig. 3). We found that the BCTR was essential for in S2 cell transcription assays and found that apCLK:apBMAL activation, because a truncated apBMAL (aa 1–552) possessed could activate transcription of the dper E-box and the most no transcriptional activity (Fig. 3). The C terminus of apCLK proximal apPer E-box, Ep, to comparable levels (Fig. 2B). ap- (everything downstream of PAS), on the other hand, was dis- CLK:apBMAL could not robustly activate transcription from pensable for apCLK:apBMAL-mediated transcriptional activa- the other two apPer E-boxes (Em and Ed), or from mutated tion, because a truncated apCLK (aa 1–364) paired with full- E-boxes reading CTGCAG, instead of CACGTG (Fig. 2B). Thus, length apBMAL was still capable of transcriptional activation apCLK:apBMAL specifically activate transcription from the (Fig. 3). The truncated proteins tagged with the V5 epitope most proximal E-box element in the apPer promoter region. were robustly expressed in S2 cells, as assayed by anti-V5 A Silkmoth Circadian Feedback Loop 38153 Downloaded from www.jbc.org at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006

FIG.2.A, apPer genomic clone containing the 5Ј-regulatory region. The gray bar represents an ϳ18 kb genomic DNA clone containing exons 1 and2ofapPer and three putative E-box elements (CACGTG sequences) upstream of exon 1. Proximal, medial, and distal E-boxes are labeled Ep, ϭ ϭ Em, and Ed. Black boxes apPer exons, labeled 1 and 2; light gray boxes labeled “E” CACGTG, putative E-boxes. B, apCLK:apBMAL activates transcription via the most proximal E-box 5Ј of the apPer gene. Presence (ϩ) or absence (Ϫ) of co-transfected apClk-V5 and apBmal (2 ng of each plasmid) is indicated to the immediate left of the graph. Luciferase reporter constructs used (described above) are shown further to the left. White ϭ ϭ ϭ box labeled “luc” luciferase cDNA; white box labeled “hs” hsp70 minimal promoter; gray boxes E-boxes, labeled “E” for the dper E-box, Ep for the proximal apPer E-box, Em for the medial apPer E-box, and Ed for the distal apPer E-box, and with a bold “X” for a mutated E-box. The sequence of each E-box and the flanking regions included in the luciferase construct is shown to the left of the schematics. The luciferase activity relative to the ␤-galactosidase activity was computed and normalized such that the mean value for the hs-luc construct in the absence of apCLK:apBMAL was 1. Each value is mean Ϯ S.E. of three replicates.

Western blot (data not shown), and both full-length and C- scriptional activation mediated by the C-terminal truncation terminal truncated apBMAL were predominantly nuclear by mutant apCLK 1–364 paired with apBMAL (Fig. 4B). It is immunocytochemistry (Table II). Thus, the effect of the dele- therefore possible that the apCLK C terminus contains a bind- tion mutations on transcriptional activity could not be ascribed ing site important for apPER-mediated inhibition of apCLK: to protein instability or failure of nuclear transport. apBMAL. The “⌬19” region is a candidate for this binding site, apPER Inhibits apCLOCK:apBMAL, and apTIM Augments because it is relatively well conserved in dCLK (Fig. 1A and This Inhibition—In the Drosophila circadian clock, dPER is the Table I), and apPER can also robustly inhibit dCLK:CYC- primary inhibitor of dCLK:CYC-mediated transcription, al- mediated transcriptional activation (Supplementary Fig. S2). though dTIM has some inhibitory activity as well (8). In mice, apPER Possesses a C-terminal Inhibition Domain and a although mPERs can modestly inhibit mCLK:BMAL1-medi- Classic Bipartite NLS—apPER inhibition of apCLK:apBMAL- ated transcription, they are not the primary inhibitors (29). To mediated transcription is somewhat surprising in the context examine the function of silkmoth PER and TIM, we expressed of its cytoplasmic localization in brain neurons in vivo.Itis different concentrations of apPer and apTim in S2 cells along possible that apPER inhibits apCLK:apBMAL by sequestering with apClk and apBmal and assessed the ability of apPER Ϯ one or both of the transcription factors in the cytoplasm. This apTIM to inhibit apCLK:apBMAL-mediated transactivation inhibitory mechanism has been demonstrated for other pro- (Fig. 4A). teins, such as the aryl hydrocarbon receptor, which is also a We found that apTIM could not inhibit apCLK:apBMAL on bHLH-PAS protein (30). However, because the activity of its own (Fig. 4A), even though it was robustly nuclear in S2 apPER in S2 cells resembled that of dPER, we investigated cells (90% of cells, Table II). apPER alone could inhibit apCLK: another possibility: that apPER contains a C-terminal inhibi- apBMAL-mediated transcription in a dose-dependent manner, tion domain, homologous to the CCID of dPER (8), and, because although low doses of apTIM could augment the inhibitory the dPER CCID requires nuclear entry (8), a functional nuclear activity of apPER. Thus, in this assay, apPER functionally localization sequence (NLS). resembles its Drosophila homolog and is likely to be the pri- By testing deletion mutants of apPer in the transcription mary inhibitor of apCLK:apBMAL. It is therefore likely that assay, we found that a C-terminal CCID-like domain of apPER the silkmoth homologs of Drosophila circadian pacemaker ele- was responsible for its inhibition of apCLK:apBMAL (Fig. 5A), ments are capable of producing a complete transcriptional feed- similar to what was discovered in Drosophila (8). We next back loop. examined the subcellular location of apPER and apPER dele- Interestingly, apPER is not effective at inhibiting the tran- tion mutants in S2 cells via immunocytochemistry. In contrast 38154 A Silkmoth Circadian Feedback Loop

FIG.3.An essential transactivation domain is found at the C terminus of apBMAL. Presence (ϩ) or absence (Ϫ)of apClk-V5 or apClk 1–364 V5 (1 ng of plas- mid) is indicated to the immediate left of the graph. Co-transfection (protein illus- tration) or omittance (Ϫ)of1ngofapB- mal or apBmal deletion mutants is indi- cated at the far left. The relative Downloaded from luciferase activity (from a luciferase re- porter driven by 4 apPer E-boxes, as shown to the far left) was computed rela- tive to the ␤-galactosidase activity (driven by an actin promoter that lacks E-boxes) and normalized such that the mean value for the apPer 4Ep-hs-luc construct in the www.jbc.org absence of apCLK/apBMAL was 1. Each value is the mean Ϯ S.E. of three repli- cates. The numbers associated with each mutant represent the amino acids of full-

length apBMAL (589 aa) or apCLK (611 at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006 aa) contained in each mutant protein. The white box with “b” ϭ bHLH domain; other white boxes ϭ parts of the PAS domain, labeled “A” (PAS-A), “B” (PAS-B), and “C” (PAC); gray box labeled “⌬” ϭ “⌬19” ho- mology region; gray box labeled “T” ϭ pu- tative transactivation domain (BMAL C- terminal region, BCTR). The results shown are representative of three inde- pendent experiments.

TABLE II inhibition of apCLK:apBMAL by apPER is a nuclear activity, Subcellular localization of V5-tagged apPER constructs similar to dPER (8). The NLS of apPER mapped to a C-terminal in S2 and Sf9 cells region of the protein (aa 684–738; Fig. 5B and Table II). By S2 cells seeded onto glass coverslips were transiently transfected with Ϸ300 ng of plasmid for the expression of the indicated C-terminal mutating clusters of basic amino acids, which are typical of V5-tagged apPER or apTIM constructs or N-terminal V5-tagged apB- classic NLSs (31), we found two clusters in this region (aa MAL. Sf9 cells (last three rows) were transfected with Ϸ1 ␮g of plasmid. 693–4 and aa 707–9) essential for robust nuclear entry (Fig. 5B Numerical intervals in the leftmost column indicate which amino acids and Table II). The spacing of these clusters is consistent with in the 849-aa apPER or the 589-aa apBMAL are present in the ex- pressed protein. Two days after transfection, the cells were processed that of a classic bipartite NLS (31). Thus, like dPER (8), apPER for immunocytochemistry using a monoclonal anti-V5 primary anti- nuclear entry in S2 cells seems to be mediated by a bipartite body, and a Cy3-conjugated secondary antibody. The cells were also NLS. Moreover, the NLSs of apPER and dPER are homologous DAPI-stained to visualize the nuclei. Slides were viewed under a fluo- to each other, and to putative NLSs in other insect PERs rescence microscope (Olympus IX70), and 30 cells per slide were cate- (Supplementary Fig. S1). gorized as having primarily nuclear staining (N), cytoplasmic staining (C), or staining in both nucleus and cytoplasm (B). Cells were scored DISCUSSION blind to which construct was transfected into them. The scores were converted into percentages (rounded to the nearest integer), and the Evolution of CLOCK:BMAL Transactivation Domains—In mean Ϯ S.E. of three or more replicates are shown. contrast to mouse and Drosophila CLK, apCLK does not pos-

Both nuclear and sess a C-terminal transactivation domain (Fig. 3). Further- S2/Sf9 cell expression Nuclear Cytoplasmic cytoplasmic more, in the Anopheles gambiae genome (32) the closest hom- constructs staining, N staining, C staining, B olog of dClk (agClk ϭ “ebiP4355,” GenBankTM number % cells Ϯ S.E. EAA11642) resembles apCLK in its lack of polyglutamine re- apPER 1-849 V5 96 Ϯ 20Ϯ 03Ϯ 2 peats. Thus, even among the Diptera, a glutamine-rich CLK apPER 1-430 V5 34 Ϯ 54Ϯ 262Ϯ 4 protein is not universal. It seems likely that, in the common apPER 1-237 V5 1 Ϯ 122Ϯ 677Ϯ 5 apPER 238-430 V5 1 Ϯ 111Ϯ 488Ϯ 4 ancestor of flies and , the CLK protein resembled that of apPER 430-849 V5 98 Ϯ 10Ϯ 01Ϯ 1 apCLK and agCLK more than dCLK. Indeed, the putative apPER 642-849 V5 98 Ϯ 10Ϯ 02Ϯ 1 transactivation domains, including the positions of the polyglu- apPER 430-684 V5 0 Ϯ 088Ϯ 612Ϯ 6 tamine stretches, are not well conserved between vertebrates Ϯ Ϯ Ϯ apPER 430-738 V5 94 20062 and flies (Table I and Fig. 1A). This suggests that the glu- apPER 539-849 V5 96 Ϯ 30Ϯ 04Ϯ 3 apPER 527-794 V5 96 Ϯ 20Ϯ 04Ϯ 2 tamine-rich transactivation domains found in Drosophila and apPER M693-4 V5 0 Ϯ 097Ϯ 13Ϯ 1 mouse CLOCKs are not conserved from an ancestral CLOCK apPER M707-9 V5 2 Ϯ 174Ϯ 324Ϯ 3 protein. We propose that CLOCK originally lacked polyglu- Ϯ Ϯ Ϯ apPER M717-9 V5 87 13 13 13 0 0 tamine stretches and acquired them independently in the Dro- apTIM V5 90 Ϯ 41Ϯ 19Ϯ 3 V5 apBMAL 1-589 68 Ϯ 20Ϯ 032Ϯ 2 sophila and vertebrate lineages. V5 apBMAL 1-552 55 Ϯ 22Ϯ 243Ϯ 0 In contrast to the C termini of CLOCK proteins, the C ter- Sf9: apPER V5 97 Ϯ 33Ϯ 31Ϯ 1 mini of BMALs are highly conserved in both sequence and (at Ϯ Ϯ Ϯ Sf9: apTIM V5 60 23 38 25 2 2 least for apBMAL and mBMAL1) transactivation function. Sf9: dTIM V5 0 Ϯ 097Ϯ 23Ϯ 2 Thus, we conclude that the BCTR is evolutionarily ancient and likely to be widespread in the animal kingdom. A search of the to its cytoplasmic localization in brain neurons, apPER was GenBankTM data base supports this assertion. The only BMAL robustly nuclear in S2 cells (96% of cells; Fig. 5B and Table II) homolog lacking the BCTR is Drosophila CYCLE (Fig. 6A). The and Sf9 cells (97% of cells; Table II). This indicates that the Bmal homolog recently cloned from the domestic silkmoth, A Silkmoth Circadian Feedback Loop 38155 Downloaded from

FIG.4. A, apPER inhibits apCLOCK: apBMAL, and low doses of apTIM aug- ment that inhibition. Presence (ϩ)orab- sence (Ϫ)ofapClk-V5 and apBmal (1 ng of each plasmid) is indicated to the imme-

diate left of the graph. Presence (ng of www.jbc.org plasmid) or absence (Ϫ)ofapPer-V5 or apTim-V5 is indicated to the far left. The luciferase activity (from the apPer 4Ep-hs- luc construct) relative to the ␤-galactosid- ase activity was computed and normal- at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006 ized such that the mean value in the presence of apCLK:apBMAL alone was 100%. Each value is mean Ϯ S.E. of three replicates. B, the C terminus of apCLOCK is required for apPER-mediated inhibi- tion of apCLOCK:apBMAL. Presence (ϩ) or absence (Ϫ)ofapClk-V5, apClk 1–364 V5,orapBmal (1 ng of plasmid) is indi- cated in the three columns to the immedi- ate left of the graph. Presence (“20”, for 20 ng of plasmid) or absence (Ϫ)ofapPer-V5 or apTim-V5 is indicated in the two left- most columns. The luciferase activity

(from the apPer 4Ep-hs-luc construct) rel- ative to the ␤-galactosidase activity was computed and normalized such that the mean value in the presence of apCLK: apBMAL alone was 100%. Each value is mean Ϯ S.E. of three replicates.

Bombyx mori (33), contains an extremely well conserved BCTR domain made the BCTR redundant. More species must be (87% identical to apBMAL). The A. gambiae genome (32) also examined to verify this model of CLOCK-BMAL evolution. contains a Bmal homolog (“ebiP7299,” GenBankTM number Evolution of TIMELESS—In contrast to the widespread con- EAA11829) containing a bHLH domain, PAS, and the BCTR. servation of Clock and Bmal, the central clock gene timeless Among vertebrates, the BCTR is universally found in close has not yet been found outside of insects. It is absent from BMAL homologs, such as BMAL1 (also called MOP3 or C. elegans, humans, puffer fish (Fugu rubripes), and sea squirt ARNT3) and BMAL2 (also called MOP9 or CLIF) in mammals (Ciona intestinalis), four animals whose genomes have been (34, 35), and BMAL1 through BMAL3 in the zebrafish, Danio almost completely sequenced (39–43). timeout-like genes, on rerio (36, 37). It is not found in ARNT proteins, although ARNT the other hand, are found universally in animals, suggesting is more closely related to BMAL than most other bHLH-PAS that timeless evolved from a gene duplication of timeout, per- proteins. haps in the lineage (12). timeless orthologs have been Thus, in the currently most parsimonious model for CLOCK: cloned from a number of dipterans, including Drosophila spe- BMAL evolution (Fig. 6A), the common ancestor of insects and cies (44, 45), the drosophilid fruit fly Chymomyza costata (Gen- vertebrates possessed a BMAL homolog possessing a C-termi- BankTM number BAB91179), and the non-drosophilid fly Sar- nal transactivation domain and a CLOCK homolog lacking a cophaga crassipalpis (46). A timeless homolog is also present in glutamine-rich domain. CLOCK acquired polyglutamine re- the genome (32) of the mosquito A. gambiae (GenBankTM num- peats very early in the vertebrate lineage, because similar ber EAA12266). apTim is the first timeless ortholog identified repeats are found in zebrafish and mammals. CLOCK inde- in a lepidopteran insect, or indeed in any non-dipteran. The pendently acquired polyglutamine repeats in the Drosophila cloning of apTim establishes that the origin of timeless oc- lineage sometime after brachyceran flies diverged from mos- curred before the evolutionary divergence of Diptera and Lep- quitoes ϳ250 million years ago (38). Meanwhile, BMAL re- idoptera ϳ330 million years ago (38). tained its ancient transactivation domain in vertebrates and in The sequence homology in the Arm/HEAT domains between most insects, but lost this domain in the Drosophila lineage, dTIM and apTIM suggests that they might share similar func- presumably because the newly acquired dCLK transactivation tions. Indeed, both apTIM and dTIM augment PER-mediated 38156 A Silkmoth Circadian Feedback Loop Downloaded from www.jbc.org at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006

FIG.6.A, model of CLOCK and BMAL evolution. The tree represents the known phylogenetic relationships among the species indicated at the far right. Next to each species name are symbolic representations of the CLOCK and BMAL homologs present in that animal. CLOCK and FIG.5.A, a C-terminal domain in apPER is responsible for its inhib- BMAL symbols at other points on the tree represent the homologs itory action on apCLOCK:apBMAL. Presence (ϩ) or absence (Ϫ)of believed to be present in the ancestral animal at that place in the apClk-V5 and apBmal (1 ng of each plasmid) is indicated to the imme- phylogenetic scheme. Black boxes ϭ transactivators; circle labeled diate left of the graph. Co-transfection of 100 ng of apPer full-length or “ϩ” ϭ transactivation domain; white boxes ϭ binding partner lacking a deletion mutant constructs is indicated by the protein illustrations and transactivation domain. B, model of circadian clock function in A. per- labels at the far left. The luciferase activity relative to the ␤-galacto- nyi. There are three E-boxes (black bars) in the 5Ј-regulatory region of sidase activity was computed and normalized such that the mean value the apPer gene (horizontal black line). apCLK (white box labeled “CLK”) in the presence of apCLK:apBMAL alone was 100%. Each value is and apBMAL (black box labeled “BMAL” with transactivation domain mean Ϯ S.E. of three replicates. The numbers associated with each shown as a circle labeled “ϩ”) bind and activate transcription specifi- mutant represent the amino acids of full-length apPER (849 aa) cally from the most proximal E-box. The apPer mRNA (wavy line) contained in each mutant protein. White boxes ϭ parts of the PAS translocates to the cytoplasm (above the curved line), where it is trans- domain, labeled “A” (PAS-A), “B” (PAS-B), and “C” (PAC); black box ϭ lated into apPER protein (gray boxes labeled “PER” with inhibitory apCLK:apBMAL inhibition domain. The results shown are represent- domain shown as a circle labeled “Ϫ”). Some apPER enters the nucleus ative of three independent experiments. B, apPER nuclear transport (below the curved line), where it can inhibit apCLK:apBMAL, thus is dependent on a bipartite nuclear localization sequence (NLS) C- shutting off its own transcription. Our inability to detect apPER in the terminal to PAS. V5-tagged apPer constructs (ϳ300 ng) were trans- nucleus of A. pernyi brain neurons may be due to high rates of nucleus- fected into S2 cells, and the cellular location of their protein products specific degradation and/or nuclear export (white arrows) and/or cyto- was assayed by immunocytochemistry using a monoclonal anti-V5 plasmic sequestration of apPER. primary antibody, and a Cy3-conjugated secondary antibody. The cells were also DAPI-stained to visualize the nuclei. For each cell examined, the V5-tagged protein was classified as having one of three inhibition of CLK:BMAL-driven transcription in S2 cells (3, 8) staining patterns: nuclear (N), cytoplasmic (C), or both nuclear and (Fig. 4). dTIM may augment the activity of dPER by facilitating cytoplasmic (B). For each construct, the proportion of cells in each dPER nuclear entry (25, 47), but this cannot be how apTIM category (N, C,orB) relative to the total number of cells examined was calculated as a percentage; the immunostaining pattern (N, C,or enhances the inhibitory activity of apPER, because apPER is B)ofϾ50% of cells examined is shown in the right-hand column, and already nuclear in S2 cells (Fig. 5B). the percentage of cells with that staining pattern is shown in paren- apTIM differs from dTIM in at least two ways. First, in S2 theses (complete data are given in Table II). The vertical dashed lines cells, apTIM shows little ability to inhibit and at high doses indicate the NLS-containing region (aa 684–738) determined using may even enhance apCLK:apBMAL-activated transcription deletion mutants. The sequence of this region is shown below the constructs. The basic amino acids (lysine, K; arginine, R; histidine, H) (Fig. 4A), whereas dTIM clearly inhibits dCLK:CYC to a mod- are underlined, mutations to alanine are indicated above the se- erate level (3, 8, 48). Second, apTIM is nuclear in S2 cells and quence, and functional components of the bipartite NLSs, as shown Sf9 cells (Table II), whereas dTIM is cytoplasmic in the same ϭ by mutagenesis, are indicated below the sequence. White boxes cell lines (25) (Table II). The “CLD” in apTIM is thus unable to parts of the PAS domain, labeled “A” (PAS-A), “B” (PAS-B), and “C” (PAC); black box ϭ apCLK:apBMAL inhibitory domain; “X” ϭ muta- function as a CLD in S2 or Sf9 cells. Although sequences genized putative NLS. similar to the CLD are present in most TIM proteins, such a A Silkmoth Circadian Feedback Loop 38157 region is absent in Drosophila hydei TIM (45). Furthermore, 3. Darlington, T. K., Wager-Smith, K., Ceriani, M. F., Staknis, D., Gekakis, N., Steeves, T. D. L., Weitz, C. J., Takahashi, J. S., and Kay, S. A. (1998) the homology in this CLD region is very weak outside of the Science 280, 1599–1603 Drosophilidae, with only 14% identity to dTIM in A. gambiae 4. Allada, R., White, N. E., So, W. V., Hall, J. C., and Rosbash, M. (1998) Cell 93, TIM. Thus, the generality of the TIM CLD for cytoplasmic 791–804 5. Rutila, J. E., Suri, V., Le, M., So, W. V., Rosbash, M., and Hall, J. C. (1998) Cell localization is unclear. 93, 805–914 PERIOD Evolution and Inhibitory Mechanism—Functional 6. Huang, Z. J., Edery, I., and Rosbash, M. (1993) Nature 364, 259–262 7. Ponting, C. P., and Aravind, L. (1997) Curr. Biol. 7, R674-R677 similarity of apPER to dPER was predicted based on the ability 8. Chang, D. C., and Reppert, S. M. (2003) Curr. Biol. 13, 758–762 of an apPer transgene to rescue behavioral rhythms in dper 9. Reppert, S. M., and Weaver, D. R. (2001) Annu. Rev. Physiol. 63, 647–676 Downloaded from null mutant flies (49). Although apPER is only two-thirds the 10. Lee, C., Etchegaray, J.-P., Cagampang, F. R., Loudon, A. S., and Reppert, S. M. (2001) Cell 107, 855–867 size of dPER, its function is well conserved in our S2 cell 11. Gotter, A. L., Manganaro, T., Weaver, D. R., Kolakowski, Jr., L. F., Possidente, assays. Both PERs possess a homologous C-terminal domain B., Sriram, S., MacLaughlin, D. T., and Reppert, S. M. (2000) Nat. Neurosci. 3, 755–756 that inhibits CLK:BMAL-mediated transcription, and a con- 12. Benna, C., Scannapieco, P., Piccin, A., Sandrelli, F., Zordan, M., Rosato, E., served bipartite NLS that is functional in S2 cells (8) (Fig. 5). Kyriacou, C. P., Valle, G., and Costa, R. (2000) Curr. Biol. 10, R512–R513 The PAS domain of apPER also resembles that of dPER (50); 13. Curtin, K. D., Huang, Z. J., and Rosbash, M. (1995) Neuron 14, 365–372 www.jbc.org 14. Shafer, O., Rosbash, M., and Truman, J. W. (2002) J. Neurosci. 22, 5946–5954 however, it remains unknown whether the apPER PAS domain 15. Sauman, I., and Reppert, S. M. (1996) Neuron 17, 889–900 can bind apTIM as the dPER PAS domain binds dTIM. Inter- 16. Za´vodska´, R., Sauman, I., and Sehnal, F. (2003) J. Biol. Rhythms 18, 106–122 17. Shearman, L. P., Sriram, S., Weaver, D. R., Maywood, E. S., Chaves, I., Zheng, estingly, the PAC of apPER, which is homologous to the CLD of B., Kume, K., Lee, C. C., van der Horst, G. T., Hastings, M. H., and Reppert, dPER, does not override the NLS to produce predominantly S. M. (2000) Science 288, 1013–1019 at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006 cytoplasmic immunostaining in S2 cells (Fig. 5 and Table II) or 18. Froy, O., Chang, D. C., and Reppert, S. M. (2002) Curr. Biol. 12, 147–152 19. King, D. P., Zhao, Y., Sangoram, A. M., Wilsbacher, L. D., Tanaka, M., Antoch, Sf9 cells (Table II). M. P., Steeves, T. D., Vitaterna, M. H., Kornhauser, J. M., Lowrey, P. L., In S2 cells, apPER robustly inhibits not only apCLK:apB- Turek, F. W., and Takahashi, J. S. (1997) Cell 89, 641–653 20. Hogenesch, J. B., Gu, Y.-Z., Jain, S., and Bradfield, C. A. (1998) Proc. Natl. MAL (Fig. 4A) but also dCLK:CYC (Supplementary Fig. S2), Acad. Sci. U. S. A. 95, 5474–5479 even though the mechanism of transcriptional activation 21. Gekakis, N., Staknis, D., Nguyen, H. B., Davis, F. C., Wilsbacher, L. D., King (BCTR of apBMAL versus the glutamine-rich C terminus of D. P., Takahashi, J. S., and Weitz, C. J. (1998) Science 280, 1564–1569 22. Takahata, S., Ozaki, T., Mimura, J., Kikuchi, Y., Sogawa, K., and Fujii- dCLK) is different. The versatility of the inhibitory activity of Kuriyama, Y. (2000) Genes Cells 5, 739–747 apPER suggests that its CCID does not negate a specific trans- 23. Jeon, M., Gardner, H. F., Miller, E. A., Deshler, J., and Rougvie, A. E. (1999) Science 286, 1141–1146 activation domain. Instead, the C-terminal inhibitory domain 24. Vodovar, N., Clayton, J. D., Costa, R., Odell, M., and Kyriacou, C. P. (2002) of apPER may be a general transcriptional repressor domain, Curr. Biol. 12, R610–R611 or it may act on the DNA binding and/or dimerization of CLK: 25. Saez, L., and Young, M. W. (1996) Neuron 17, 979–990 26. Gotter, A. L., Levine, J. D., and Reppert, S. M. (1999) Neuron 24, 953–965 BMAL, which is mediated by the bHLH and PAS domains and 27. Blau, J., and Young, M. W. (1999) Cell 99, 661–671 is therefore likely to be independent of the transactivation 28. Jin, X., Shearman, L. P., Weaver, D. R., Zylka, M. J., de Vries, G. J., and Reppert, S. M. (1999) Cell 96, 57–68 domain(s). In vitro gel-shift assays have shown that dPER 29. Kume, K., Zylka, M. J., Sriram, S., Shearman, L. P., Weaver, D. R., Jin, X., binding to dCLK:CYC can inhibit DNA binding without dis- Maywood, E. S., Hastings, M. H., and Reppert, S. M. (1999) Cell 98, rupting dCLK:CYC heterodimer formation (48), suggesting 198–205 30. Gu, Y.-Z., Hogenesch, J. B., and Bradfield, C. A. (2000) Annu. Rev. Pharmacol. that dPER can interfere with the bHLH domains of dCLK:CYC. Toxicol. 40, 519–561 However, the inhibitory activity of apPER is dependent on the 31. Go¨rlich, D., and Kutay, U. (1999) Annu. Rev. Cell Dev. Biol. 15, 607–660 C terminus of apCLK (Fig. 4B). Thus apPER cannot act only 32. Holt, R. A., Subramanian, G. M., Halpern, A., Sutton, G. G., Charlab, R., Nusskern, D. R., Wincker, P., Clark, A. G., Ribeiro, J. M., Wides, R., through the bHLH domains of apCLK:apBMAL. Salzberg, S. L., Loftus, B., Yandell, M., Majoros, W. H., Rusch, D. B., Lai, Z., Toward a General Insect Circadian Clock Model—Our data Kraft, C. L., Abril, J. F., Anthouard, V., Arensburger, P., Atkinson, P. W., Baden, H., de, Berardinis, V., Baldwin, D., Benes, V., Biedler, J., Blass, C., in S2 cells indicate that a transcriptional feedback loop resem- Bolanos, R., Boscus, D., Barnstead, M., Cai, S., Center, A., Chaturverdi, K., bling that found in the circadian clock of Drosophila can be Christophides, G. K., Chrystal, M. A., Clamp, M., Cravchik, A., Curwen, V., constructed from A. pernyi CLOCK, BMAL, PERIOD, and Dana, A., Delcher, A., Dew, I., Evans, C. A., Flanigan, M., Grundschober- Freimoser, A., Friedli, L., Gu, Z., Guan, P., Guigo, R., Hillenmeyer, M. E., TIMELESS proteins, and the apPeriod gene (Fig. 6B). Such a Hladun, S. L., Hogan, J. R., Hong, Y. S., Hoover, J., Jaillon, O., Ke, Z., feedback loop could explain the previously detected in vivo Kodira, C., Kokoza, E., Koutsos, A., Letunic, I., Levitsky, A., Liang, Y., Lin, J. J., Lobo, N. F., Lopez, J. R., Malek, J. A., McIntosh, T. C., Meister, S., rhythms in apPer RNA and protein levels (15, 26). However, it Miller, J., Mobarry, C., Mongin, E., Murphy, S. D., O’Brochta, D. A., remains unclear why apPER is always primarily cytoplasmic in Pfannkoch, C., Qi, R., Regier, M. A., Remington, K., Shao, H., Sharakhova, A. pernyi brain neurons, when it possesses an evolutionarily M. V., Sitter, C. D., Shetty, J., Smith, T. J., Strong, R., Sun, J., Thomasova, D., Ton, L. Q., Topalis, P., Tu, Z., Unger, M. F., Walenz, B., Wang, A., Wang, ancient, classic bipartite NLS (Fig. 5B). We propose that ap- J., Wang, M., Wang, X., Woodford, K. J., Wortman, J. R., Wu, M., Yao, A., PER may indeed enter the nucleus, if only transiently or at low Zdobnov, E. M., Zhang, H., Zhao, Q., Zhao, S., Zhu, S. C., Zhimulev, I., Coluzzi, M., della Torre, A., Roth, C. W., Louis, C., Kalush, F., Mural, R. J., levels. The conservation of the NLS in insect PER proteins Myers, E. W., Adams, M. D., Smith, H. O., Broder, S., Gardner, M. J., (Supplementary Fig. S1) suggests that, although PER has been Fraser, C. M., Birney, E., Bork, P., Brey, P. T., Venter, J. C., Weissenbach, detected only in the cytoplasm in most insects, PER nuclear J., Kafatos, F. C., Collins, F. H., and Hoffman, S. L. (2002) Science 298, 129–149 entry may be widely true. The regulation of PER subcellular 33. Markova, E. P., Ueda, H., Sakamoto, K., Oishi, K., Shimada, T., and Takeda, localization in Drosophila differs dramatically from that of M. (2003) Comp. Biochem. Physiol. B 134, 535–542 34. Hogenesch, J. B., Gu, Y. Z., Moran, S. M., Shimomura, K., Radcliffe, L. A., A. pernyi and most other insects (15, 16). The tobacco hawk- Takahashi, J. S., and Bradfield, C. A. (2000) J. Neurosci. 20, RC83 moth, Manduca sexta, is also atypical in that PER nuclear 35. Maemura, K., de la Monte, S. M., Chin, M. T., Layne, M. D., Hsieh, C. M., Yet, immunostaining (but not PER rhythmicity) is readily detecta- S. F., Perrella, M. A., and Lee, M. E. (2000) J. Biol. Chem. 275, 36847–36851 ble in brain neurons (51). However, regardless of how PER 36. Cermakian, N., Whitmore, D., Foulkes, N. S., and Sassone-Corsi, P. (2000) subcellular transport is controlled, a Drosophila-like feedback Proc. Natl. Acad. Sci. U. S. A. 97, 4339–4344 37. Ishikawa, T., Hirayama, J., Kobayashi, Y., and Todo, T. (2002) Genes Cells 7, loop may form the core of the circadian clockwork in insects. 1073–1086 38. Gaunt, M. W., and Miles, M. A. (2002) Mol. Biol. Evol. 19, 748–761 Acknowledgments—We thank Sriram Sathyanarayanan for help 39. The Caenorhabditis elegans Sequencing Consortium. (1998) Science 282, generating some of the constructs. We thank Aditi V. Chavda, Kurtis N. 2012–2018 Gray, and Kasia Macko for expert technical assistance. 40. International Human Genome Sequencing Consortium (2001) Nature 409, 860–921 41. Venter, J. C., Adams, M. D., Myers, E. W., Li, P. W., Mural, R. J., Sutton, G. G., REFERENCES Smith, H. O., Yandell, M., Evans, C. A., Holt, R. A., Gocayne, J. D., 1. Young, M. W., and Kay, S. A. (2001) Nat. Rev. Genet. 2, 702–715 Amanatides, P., Ballew, R. M., Huson, D. H., Wortman, J. R., Zhang, Q., 2. Williams, J. A., and Sehgal, A. (2001) Annu. Rev. Physiol. 63, 729–755 Kodira, C. D., Zheng, X. H., Chen, L., Skupski, M., Subramanian, G., 38158 A Silkmoth Circadian Feedback Loop

Thomas, P. D., Zhang, J., Gabor Miklos, G. L., Nelson, C., Broder, S., Clark, E., Wang, M., Wen, M., Wu, D., Wu, M., Xia, A., Zandieh, A., and Zhu, X. A. G., Nadeau, J., McKusick, V. A., Zinder, N., Levine, A. J., Roberts, R. J., (2001) Science 291, 1304–1351 Simon, M., Slayman, C., Hunkapiller, M., Bolanos, R., Delcher, A., Dew, I., 42. Aparicio, S., Chapman, J., Stupka, E., Putnam, N., Chia, J. M., Dehal, P., Fasulo, D., Flanigan, M., Florea, L., Halpern, A., Hannenhalli, S., Kravitz, Christoffels, A., Rash, S., Hoon, S., Smit, A., Gelpke, M. D., Roach, J., Oh, S., Levy, S., Mobarry, C., Reinert, K., Remington, K., Abu-Threideh, J., T., Ho, I. Y., Wong, M., Detter, C., Verhoef, F., Predki, P., Tay, A., Lucas, S., Beasley, E., Biddick, K., Bonazzi, V., Brandon, R., Cargill, M., Richardson, P., Smith, S. F., Clark, M. S., Edwards, Y. J., Doggett, N., Chandramouliswaran, I., Charlab, R., Chaturvedi, K., Deng, Z., Di, Zharkikh, A., Tavtigian, S. V., Pruss, D., Barnstead, M., Evans, C., Baden, Francesco, V., Dunn, P., Eilbeck, K., Evangelista, C., Gabrielian, A. E., Gan, H., Powell, J., Glusman, G., Rowen, L., Hood, L., Tan, Y. H., Elgar, G., W., Ge, W., Gong, F., Gu, Z., Guan, P., Heiman, T. J., Higgins, M. E., Ji, Hawkins, T., Venkatesh, B., Rokhsar, D., and Brenner, S. (2002) Science R. R., Ke, Z., Ketchum, K. A., Lai, Z., Lei, Y., Li, Z., Li, J., Liang, Y., Lin, X., 297, 1301–1310 Lu, F., Merkulov, G. V., Milshina, N., Moore, H. M., Naik, A. K., Narayan, 43. Dehal, P., Satou, Y., Campbell, R. K., Chapman, J., Degnan, B., De Tomaso, A., Downloaded from V. A., Neelam, B., Nusskern, D., Rusch, D. B., Salzberg, S., Shao, W., Shue, Davidson, B., Di Gregorio, A., Gelpke, M., Goodstein, D. M., Harafuji, N., B., Sun, J., Wang, Z., Wang, A., Wang, X., Wang, J., Wei, M., Wides, R., Hastings, K. E., Ho, I., Hotta, K., Huang, W., Kawashima, T., Lemaire, P., Xiao, C., Yan, C., Yao, A., Ye, J., Zhan, M., Zhang, W., Zhang, H., Zhao, Q., Martinez, D., Meinertzhagen, I. A., Necula, S., Nonaka, M., Putnam, N., Zheng, L., Zhong, F., Zhong, W., Zhu, S., Zhao, S., Gilbert, D., Baumhueter, Rash, S., Saiga, H., Satake, M., Terry, A., Yamada, L., Wang, H. G., Awazu, S., Spier, G., Carter, C., Cravchik, A., Woodage, T., Ali, F., An, H., Awe, A., S., Azumi, K., Boore, J., Branno, M., Chin-Bow, S., DeSantis, R., Doyle, S., Baldwin, D., Baden, H., Barnstead, M., Barrow, I., Beeson, K., Busam, D., Francino, P., Keys, D. N., Haga, S., Hayashi, H., Hino, K., Imai, K. S., Carver, A., Center, A., Cheng, M. L., Curry, L., Danaher, S., Davenport, L., Inaba, K., Kano, S., Kobayashi, K., Kobayashi, M., Lee, B. I., Makabe,

Desilets, R., Dietz, S., Dodson, K., Doup, L., Ferriera, S., Garg, N., Gluecks- K. W., Manohar, C., Matassi, G., Medina, M., Mochizuki, Y., Mount, S., www.jbc.org mann, A., Hart, B., Haynes, J., Haynes, C., Heiner, C., Hladun, S., Hostin, Morishita, T., Miura, S., Nakayama, A., Nishizaka, S., Nomoto, H., Ohta, D., Houck, J., Howland, T., Ibegwam, C., Johnson, J., Kalush, F., Kline, L., F., Oishi, K., Rigoutsos, I., Sano, M., Sasaki, A., Sasakura, Y., Shoguchi, E., Koduru, S., Love, A., Mann, F., May, D., McCawley, S., McIntosh, T., Shin-i, T., Spagnuolo, A., Stainier, D., Suzuki, M. M., Tassy, O., Takatori, McMullen, I., Moy, M., Moy, L., Murphy, B., Nelson, K., Pfannkoch, C., N., Tokuoka, M., Yagi, K., Yoshizaki, F., Wada, S., Zhang, C., Hyatt, P. D., Pratts, E., Puri, V., Qureshi, H., Reardon, M., Rodriguez, R., Rogers, Y. H., Larimer, F., Detter, C., Doggett, N., Glavina, T., Hawkins, T., Richardson,

Romblad, D., Ruhfel, B., Scott, R., Sitter, C., Smallwood, M., Stewart, E., P., Lucas, S., Kohara, Y., Levine, M., Satoh, N., and Rokhsar, D. S. (2002) at UNAM. DIRECCION GENERAL DE BIBLIOTECAS. DEPARTAMENTO SUSCRIPCIONES on August 8, 2006 Strong, R., Suh, E., Thomas, R., Tint, N. N., Tse, S., Vech, C., Wang, G., Science 298, 2157–2167 Wetter, J., Williams, S., Williams, M., Windsor, S., Winn-Deen, E., Wolfe, 44. Myers, M. P., Wager-Smith, K., Wesley, C. S., Young, M. W., and Sehgal, A. K., Zaveri, J., Zaveri, K., Abril, J. F., Guigo, R., Campbell, M. J., Sjolander, (1995) Science 270, 805–808 K. V., Karlak, B., Kejariwal, A., Mi, H., Lazareva, B., Hatton, T., 45. Ousley, A., Zafarullah, K., Chen, Y., Emerson, M., Hickman, L., and Sehgal, A. Narechania, A., Diemer, K., Muruganujan, A., Guo, N., Sato, S., Bafna, V., (1998) Genetics 148, 815–825 Istrail, S., Lippert, R., Schwartz, R., Walenz, B., Yooseph, S., Allen, D., 46. Goto, S. G., and Denlinger, D. L. (2003) J. Insect Physiol. 48, 803–816 Basu, A., Baxendale, J., Blick, L., Caminha, M., Carnes-Stine, J., Caulk, P., 47. Vosshall, L. B., Price, J. L., Sehgal, A., Saez, L., and Young, M. W. (1994) Chiang, Y. H., Coyne, M., Dahlke, C., Mays, A., Dombroski, M., Donnelly, Science 263, 1606–1609 M., Ely, D., Esparham, S., Fosler, C., Gire, H., Glanowski, S., Glasser, K., 48. Lee, C., Bae, K., and Edery, I. (1999) Mol. Cell. Biol. 19, 5316–5325 Glodek, A., Gorokhov, M., Graham, K., Gropman, B., Harris, M., Heil, J., 49. Levine, J. D., Sauman, I., Imbalzano, M., Reppert, S. M., and Jackson, F. R. Henderson, S., Hoover, J., Jennings, D., Jordan, C., Jordan, J., Kasha, J., (1995) Neuron 15, 147–157 Kagan, L., Kraft, C., Levitsky, A., Lewis, M., Liu, X., Lopez, J., Ma, D., 50. Reppert, S. M., Tsai, T., Roca, A. L., and Sauman, I. (1994) Neuron 13, Majoros, W., McDaniel, J., Murphy, S., Newman, M., Nguyen, T., Nguyen, 1167–1176 N., Nodell, M., Pan, S., Peck, J., Peterson, M., Rowe, W., Sanders, R., Scott, 51. Wise, S., Davis, N. T., Tyndale, E., Noveral, J., Folwell, M. G., Bedian, V., J., Simpson, M., Smith, T., Sprague, A., Stockwell, T., Turner, R., Venter, Emery, I. F., and Siwicki, K. K. (2002) J. Comp. Neurol. 447, 366–380