Slimb是調(diào)節(jié)蛋白質(zhì)降解的蛋白質(zhì) Period and Timeless是兩個(gè)控制果蠅晝夜節(jié)律的蛋白質(zhì) 其中Period基因是1972年世界上第一個(gè)發(fā)現(xiàn)的控制晝夜節(jié)律的基因,，現(xiàn)在已經(jīng)知道高等動(dòng)物也是用同樣的基因控制晝夜節(jié)律 最早發(fā)現(xiàn)者是加州理工大學(xué)的Seymour Benzer和博后Konopka,，基因克隆是八十年代的事,，主要是Brandeis大學(xué)Michael Roshbash和Jeff Hall實(shí)驗(yàn)室,，中國(guó)學(xué)生俞強(qiáng)，劉欣,，黃佐石,，曾紅葵等做了一些重要實(shí)驗(yàn) 俞強(qiáng)現(xiàn)在波士頓大學(xué)副教授（妻子是哈佛的細(xì)胞生物教授袁鈞英），劉欣現(xiàn)在UCLA藥理系助理教授（妻子是UCLA的吳虹）,，黃佐石（現(xiàn)在冷泉港助理教授）,，曾紅葵（現(xiàn)在西雅圖，丈夫許文清是西雅圖華盛頓大學(xué)結(jié)構(gòu)生物學(xué)助理教授） 14 November 2002 Nature 420, 178 - 182 (2002); doi:10.1038/nature01122 The F-box protein Slimb controls the levels of clock proteins Period and Timeles s BRIGITTE GRIMA*, ANNIE LAMOUROUX*, ELISABETH CHÉLOT*, CHRISTIAN PAPIN*, B ERNADETTE LIMBOURG-BOUCHON† & FRANÇOIS ROUYER* Institut de Neurobiologie Alfred Fessard (NGI, CNRS UPR 2216) and Centre de Gén étique Moléculaire (CNRS UPR 2167), Centre National de la Recherche Scientifiq ue, av. de la terrasse, 91198 Gif-sur-Yvette, France Correspondence and requests for materials should be addressed to F.R. (e-mail: r ouyer@iaf.cnrs-gif.fr). The Drosophila circadian clock is driven by daily fluctuations of the proteins P eriod and Timeless, which associate in a complex and negatively regulate the tra nscription of their own genes1, 2. Protein phosphorylation has a central role in this feedback loop, by controlling Per stability in both cytoplasmic and nuclea r compartments3-6 as well as Per/Tim nuclear transfer7, 8. However, the pathways regulating degradation of phosphorylated Per and Tim are unknown. Here we show that the product of the slimb (slmb) gene9—a member of the F-box/WD40 protein f amily of the ubiquitin ligase SCF complex that targets phosphorylated proteins f or degradation10-13—is an essential component of the Drosophila circadian clock . slmb mutants are behaviourally arrhythmic, and can be rescued by targeted expr ession of Slmb in the clock neurons. In constant darkness, highly phosphorylated forms of the Per and Tim proteins are constitutively present in the mutants, in dicating that the control of their cyclic degradation is impaired. Because level s of Per and Tim oscillate in slmb mutants maintained in light:dark conditions, light- and clock-controlled degradation of Per and Tim do not rely on the same m echanisms. To test whether the SCF-mediated ubiquitin proteasome pathway is involved in the control of Per and Tim oscillations, we chose to analyse circadian rhythms of f lies defective for genes encoding F-box proteins that were known to target phosp horylated substrates for degradation. We started with the slimb (slmb) gene, whi ch encodes an F-box/WD40 protein regulating transcription factors' levels in the wingless and hedgehog signalling pathways9. slmb8 mutants that normally die as early larvae14 were brought to adulthood by providing the slmb gene product thro ughout development under the control of a heat-shock promoter. The rescued HS-sl mb slmb8 adult flies, hereafter referred to as slmbm mutants, were then tested f or their locomotor activity rhythms in both light:dark (LD) and constant darknes s (DD) conditions. slmbm mutants were completely arrhythmic in DD, whereas the h eterozygous genotype displayed wild-type rhythms (Table 1). The absence of anato mical defects of the PDF-expressing ventral lateral neurons (LNvs, data not show n), which control the behavioural rhythms15, 16, strongly argues against a devel opmental origin of the mutants' rhythm defect. Furthermore, targeted slmb expres sion using well characterized LNvs-specific gal4 drivers15, 16 restored near wil d-type activity rhythms, whereas similarly targeted overexpression in a wild-typ e background lengthened the circadian period (Table 1), indicating a cell-autono mous role of the slmb gene in circadian rhythmicity. In LD conditions, slmbm mut ants did not display the light-off anticipation of activity that characterizes a functional clock, whereas it was observed in the flies expressing slmb under th e LNvs-specific gal1118 driver (Fig. 1). These experiments identify the F-box/WD 40 protein Slmb as an essential component of the Drosophila brain clock. Figure 1 Locomotor activity in LD cycles. Full legend High resolution image and legend (44k) To understand how Slmb might affect the circadian oscillator, slmbm mutants were analysed for Per and Tim oscillations in the head. In wild-type flies maintaine d in LD cycles, Per and Tim proteins accumulate and are progressively phosphoryl ated during night time, with Tim disappearing at the end of the night whereas hy per-phosphorylated Per persists for a few hours in the morning17-19. A similar t emporal pattern persists in DD17-19, and is required to sustain behavioural rhyt hmicity16, 20, 21. In contrast, highly phosphorylated Per and Tim were present a t all circadian times in slmbm mutants kept in DD (Fig. 2a, and Supplementary In formation), although low-amplitude oscillations of the hypo-phosphorylated forms indicated a weak residual activity of the molecular clock (Fig. 2a). In agreeme nt with the persistence of weak protein cycling in slmbm heads, levels of per an d tim transcripts displayed low-amplitude oscillations (Fig. 2b). We then looked at Per immunoreactivity in the LNvs that control behavioural rhythms (Fig. 2c). At circadian time (CT) 0 and CT 12, which correspond to the peak and trough of Per labelling in w flies at 20 °C (Fig. 3e), slmbm mutants showed low levels of Per immunoreactivity, indicating that the oscillations of the proteins levels a re also abolished in the clock cells. To determine whether Slmb acts at the prot ein level or through a transcriptional control, per was constitutively overexpre ssed through a transgene. High-molecular-mass Per proteins were observed to accu mulate in head extracts of slmbm but not of wild-type flies carrying GMR-gal4 an d UAS-per transgenes that drive strong Per expression in the eye (Fig. 2d). Alto gether, these data indicate that Slmb is involved in the control of phosphorylat ed Per levels. Figure 2 Per and Tim proteins and RNAs in slmbm mutants. Full legend High resolution image and legend (77k) Figure 3 Slmb protein expression and interactions. Full legend High resolution image and legend (67k) In LD conditions, Per and Tim degradation in the morning is driven by both the c ircadian cycle and by light18, 19, 22, 23. Light-induced Tim degradation involve s ubiquitinylation of the protein, and is blocked by proteasome inhibitors24. To test whether Slmb is involved in the light-induced degradation pathway of the c lock proteins, Per and Tim levels were assayed in slmbm flies kept in LD conditi ons. In contrast to constant darkness, robust oscillations of Per and Tim amount s were observed in LD, with both proteins accumulating during the night and show ing a strong day-time decrease (Fig. 2e). This shows that light-induced Per and Tim degradation does not occur through the same slmb-dependent mechanism as thei r circadian-cycle-controlled degradation in constant darkness. In addition, the absence of light-off anticipation in the slmbm activity profiles (Fig. 1) sugges ts that the mutants' altered temporal regulation of phosphorylated Per and Tim d oes not allow rhythmic outputs to be driven, although protein levels clearly cyc le. Clock-dependent Per and Tim degradation occurs at the end of the circadian cycle , and relieves the transcriptional repression that the proteins exert on their o wn genes1, 2. Per degradation has also been proposed to take place during the ri sing phase of the protein levels in the early night, and to be responsible for t he shift (of 5 h) between per messenger RNA and Per protein peaks3, 4. In order to determine whether Slmb levels vary during a circadian cycle and may therefore affect Per and Tim only during a limited time window, anti-Slmb antibodies were raised and the Slmb protein was followed in head extracts at different circadia n times. A strongly reacting protein, as well as a faintly reacting one slighly above, were detected at a relative molecular mass of 45,000 (Mr 45K) in wild-typ e flies, and did not show any oscillations of their levels over a 24-h time cour se (Fig. 3a). Similarly, slmb mRNA did not show any cycling (data not shown). Sl mb therefore appears not to be circadianly regulated, and could act on different steps of the cycle. Both early- and late-night Per degradation steps appear to depend upon Per phosp horylation, which requires the casein kinase I encoded by the double-time (dbt) gene3-6. To final out how Slmb could affect Per and Tim phosphorylation, we firs t tested whether Dbt, and Shaggy (Sgg) that has been recently shown to phosphory late Tim7, were affected in slmbm mutants. No alterations of the level or the mo bility of these kinases were detected in slmbm head extracts (Fig. 3b). We then investigated whether Slmb could associate with the Per protein, by searching for Per–Slmb interactions in co-immunoprecipitation experiments on head extracts. We found that the Slmb protein was co-precipitated by anti-Per antibodies, and that anti-Slmb could precipitate Per in wild-type flies collected at CT 0 (Fig. 3c). Similar results were obtained with pooled extracts (CT 15-18-21-0-3, not shown). In addition, Slmb co-precipitated with Dbt (Fig. 3c). Because Per, but not Dbt, was profoundly affected in slmbm mutants, our results support Per rather than Dbt as a Slmb target for ubiquitinylation, and suggest that the three proteins constitute a complex. Slmb was co-immunoprecipitated by anti-Per antibodies in tim0 flies (Fig. 3c), indicating that Per?CSlmb complexes can form in the absence of Tim. Although twice as much extract w as used for tim0 flies to compensate for the low Per levels in this genotype, th e amount of immunoprecipitated Slmb suggests that the absence of Tim may favour Per–Slmb complexes. These results fit well with Slmb being involved in the control of unbound Per, either during its cytoplasmic accumulation at the beginning of the protein cycle or during its nuclear degradation at the end6. To test whether the formation of Per?CSlmb complexes is circadianly controlled, we performed co-immunoprecipitations at the beginning of the night when Per is mostly hypo-phosphorylated, or at the end of the night when Per is highly phosphorylated (Fig. 3d). All time points sh owed comparable levels of Per–Slmb complexes, and several forms of Per were immunoprecipitated by the anti-Slmb antibodies (compare CT 1 and CT 13). This indicates that differently phosphorylated Per molecules can be committed to Per?CSlmb complexes. Possible explanations for the accumulation of highly phosphorylated Per in slmbm mutants would be that partially phosphorylated Per is the relevant Slmb substra te for degradation, or that Slmb targets some Per kinase that is bound to Per. T he presence of highly phosphorylated Per in slmbm as early as zeitgeber time (ZT )15 in LD (Fig. 2e) indicates that Slmb is required for the control of phosphory lated Per accumulation in the early night. Moreover, Slmb overexpression in the LNvs resulted in a lengthening of the circadian period (Table 1). In agreement w ith the behavioural data, Slmb overexpression slowed down the oscillations of Pe r immunoreactivity in these cells, which showed a 6-h delay compared to wild-typ e controls after two days (Fig. 3e). These data can be explained by high levels of cytoplasmic Slmb inducing too much degradation of cytoplasmic Per, thus furth er delaying the night accumulation of the protein, whereas high levels of nuclea r Slmb would rather precipitate the fall of the Per protein and shorten the circ adian period. We therefore think that Slmb is, at least, involved in the control of cytoplasmic Per accumulation in the early night. The presence of low-mobility Tim proteins at all circadian times in slmbm mutant s (compare with control in Fig. 2a) indicates that the accumulation of phosphory lated Tim is also Slmb-dependent. Remarkably, the Tim kinase Sgg controls the Sl mb-dependent proteolysis of Cubitus Interruptus9, 25, 26 and degradation of Arma dillo9, 27. Our results suggest that phosphorylated Tim could be a Slmb target o r that Tim is phosphorylated by a Slmb-dependent kinase. Because Tim is hypo-pho sphorylated in per0 flies7, 19, it is also possible that the accumulation of hyp er-phosphorylated Per in slmbm influences Tim phosphorylation. Although protein degradation is commonly believed to have a major role in the co ntrol of the oscillations of clock proteins, the present work is (to our knowled ge) the first to implicate a characterized component of the ubiquitin proteasome pathway. Because cycling of phosphorylated Per proteins also occurs in the mamm alian clock1, 2, it would be interesting to determine whether the Slmb mammalian homologue -Trcp11 is involved in the control of phosphorylated Per levels. F-bo x proteins have been shown to be important at the G1/S transition of the cell cy cle, by targeting phosphorylated cyclins and inhibitors of cyclin kinases for de gradation by the proteasome10. Our study therefore suggests that the cell-cycle and the circadian-clock machineries share mechanisms to control the oscillations of phosphorylated proteins. Methods Fly strains Homozygous slmb8 adults were obtained in the progeny of slmb8 HS-slm bL/TM3Sb flies14 that were grown at 25 °C and heat-shocked daily at 37 °C for 1 h, from first-instar larvae until the end of pupariation. Other genotypes homo zygous for slmb8 were obtained with the same protocol after the appropriate cros sings with the w;;gal111816, yw;pdf-gal415 and w;GMR-gal4;UAS-per16 lines. We ca nnot exclude that residual Slmb is present in the slmbm (w;;slmb8 HS-slmb and w HS-slmb;;slmb8) adults. The UAS-slmb construct was made with the T slmb compleme ntary DNA14 inserted in the pUAST vector and injected into embryos to generate t he w;;UAS-slmb line. Behavioural analysis Experiments were carried out with 1–7-d-old males at 20 °C in Drosophila activity monitors (Trikinetics) as described before16. For DD analysis, flies were first kept in 12 h:12 h LD cycles over 2?C4 d, and recorded in DD over 6 d. Data analysis was done on a Macintosh compute r with the Faas software (available upon request), which is derived from the Bra ndeis Rhythm Package (http://hawk.bcm.tmc.edu/). Rhythmic flies were defined by 2 periodogram analysis with the following criteria (filter on): power 20, width 5, number of peaks 3. Western blotting Head total protein extracts and immunoblottings were made as pr eviously described16, except that anti-phosphatases (20 mM -glycerophosphate, 20 0 µM Na3VO4) were added to the extraction buffer. Anti-Slmb antibodies wer e generated in rats against the C-terminal peptide ENKTGRTPSPALMEH (Neosystem), and used at dilution 1/2,000. The rat anti-Tim antibodies were raised against a GST–Tim (residues 222–557) fusion protein18 and used at dilution 1/2,000. (GST , glutathione S-transferase) The rabbit anti-Per28 (1/10,000), rat anti-Dbt6 (1/ 500) and Mab2G2C5 anti-Sgg29 antibodies have been already reported. Immunocytochemistry Labellings were done on whole-mounted adult brains as previo usly described16 with the following modifications: anti-Per serum was applied ov ernight, Alexa Fluor 594 (Molecular Probes) anti-rabbit IgG goat antibody was us ed at 1:10,000 as a secondary antibody. Fluorescence intensity was quantified fr om digital images using the formula: I = 100 (S - B)/B, that gives the fluoresc ence percentage above background (S, fluorescence intensity; B, average intensit y of the region adjacent to the positive cell). Immunoprecipitations Extractions and immunoprecipitations were performed as desc ribed6, with the following modifications: 2 mg of head total proteins (4 mg for tim0 flies) were extracted in HE buffer supplemented with 20 mM -glycerophosphat e, 200 µM Na3VO4, 0.1% gelatin and 1 mM dithiothreitol. 2 µl of immu ne (or pre-immune) serum and 50 µl of a slurry containing 50% protein G se pharose (Pharmacia) were used for each immunoprecipitation. The UT31 rabbit anti -human CKI antibody30 was used to immunoprecipitate Dbt. For anti-Slmb immunopre cipitations, rabbit antibodies directed against the same peptide were used inste ad of rat antibodies. All immunoprecipitations were done at least twice, with si milar results. Quantitative RT–PCR cDNA were synthesized from 1 µg of head total RNA (Promega SV total RNA isolation system). Quantitative PCR was performed with a Lightcycler using the SYBR green detection protocol (Roche). cDNA samples were mixed with Faststart DNA master SYBR green I mix, 3 mM MgCl2, 500 nM of each primers in 20 µl and submitted to 40 cycles of PCR (95 ?鉉, 15 s; 60 °C, 10 sec; 72 °C, 20 s). RNA levels were normalized to the levels of tubulin RNA and converted to a multiple of the minimal level for each gene (minimal levels were the less variable between experiments). slmbm mutants are in a w genetic background but have red eyes owing to the HS-slmb transgene. We therefore used both w and Canton S (CS) flies as controls, because white-eyed flies usually show RNA oscillations of higher amplitude. The primers were 5'TCCTTGTCGCGTGTGAAACA (exon 1) and 3'CCGAACGAGTGGAAGATGAG (exon 2) for tubulin (464 bp), 5'CGCCGGACACTGAAAAGG (exon 7) and 3'ATATATAAACCTTAGGGCT (exon 8) for per (620 bp), and 5'AGTTGGTCATGCGCAGCAAATG (exon 12) and 3'TCCTTTTCGTACACAGATGCCA (exon 13) for tim (447 bp). Supplementary information accompanies this paper. Received 11 July 2002;accepted 23 August 2002 References 1. Allada, R., Emery, P., Takahashi, J. S. & Rosbash, M. Stopping tim e: the genetics of fly and mouse circadian clocks. Annu. Rev. Neurosci. 24, 1091 -1119 (2001) | Article | PubMed | 2. Young, M. W. & Kay, S. A. Time zones: a comparative genetics of circadian clo cks. Nature Rev. Genet. 2, 702-715 (2001) | Article | PubMed | 3. Kloss, B. et al. The Drosophila clock gene double-time encodes a protein clos ely related to human casein kinase I&epsis;. Cell 94, 97-107 (1998) | PubMed | 4. Price, J. L. et al. double-time is a novel Drosophila clock gene that regulat es PERIOD protein accumulation. Cell 94, 83-95 (1998) | PubMed | 5. Suri, V., Hall, J. C. & Rosbash, M. Two novel doubletime mutants alter circad ian properties and eliminate the delay between RNA and protein in Drosophila. J. Neurosci. 20, 7547-7555 (2000) | PubMed | 6. Kloss, B., Rothenfluh, A., Young, M. W. & Saez, L. Phosphorylation of period is influenced by cycling physical associations of double-time, period, and timel ess in the Drosophila clock. Neuron 30, 699-706 (2001) | Article | PubMed | 7. Martinek, S., Inonog, S., Manoukian, A. S. & Young, M. W. A Role for the segm ent polarity gene shaggy/GSK-3 in the Drosophila circadian clock. Cell 105, 769- 779 (2001) | Article | PubMed | 8. Curtin, K. D., Huang, Z. J. & Rosbash, M. Temporally regulated nuclear entry of the Drosophila period protein contributes to the circadian clock. Neuron 14, 365-372 (1995) | PubMed | 9. Jiang, J. & Struhl, G. Regulation of the Hedgehog and Wingless signalling pat hways by the F-box/WD40-repeat protein Slimb. Nature 391, 493-496 (1998) | Artic le | PubMed | 10. Skowyra, D., Craig, K. L., Tyres, M., Elledge, S. J. & Harper, J. W. F-box p roteins are receptors that recruit phosphorylated substrates to the SCF ubiquiti n-ligase complex. Cell 91, 209-219 (1997) | PubMed | 11. Margottin, F. et al. A novel human WD protein, h- TrCp, that interacts with HIV-1 Vpu connects CD4 to the ER degradation pathway through an F-box motif. Mol . Cell 1, 565-574 (1998) | PubMed | 12. Spencer, E., Jiang, J. & Chen, Z. J. Signal-induced ubiquitination of IB by the F-box protein Slimb/-TrCP. Genes Dev. 13, 284-294 (1999) | PubMed | 13. Winston, J. T. et al. The SCF-TRCP-ubiquitin ligase complex associates speci fically with phosphorylated destruction motifs in IB and -catenin and stimulates IB ubiquitination in vitro. Genes Dev. 13, 270-283 (1999) | Article | PubMed | 14. Miletich, I. & Limbourg-Bouchon, B. Drosophila null slimb clones transiently deregulate Hedgehog-independent transcription of wingless in all limb discs, an d induce decapentaplegic transcription linked to imaginal disc regeneration. Mec h. Dev. 93, 15-26 (2000) | Article | PubMed | 15. Renn, S. C., Park, J. H., Rosbash, M., Hall, J. C. & Taghert, P. H. A pdf ne uropeptide gene mutation and ablation of PDF neurons each cause severe abnormali ties of behavioral circadian rhythms in Drosophila. Cell 99, 791-802 (1999) | Pu bMed | 16. Blanchardon, E. et al. Defining the role of Drosophila lateral neurons in th e control of circadian activity and eclosion rhythms by targeted genetic ablatio n and PERIOD protein overexpression. Eur. J. Neurosci. 13, 871-888 (2001) | Arti cle | PubMed | 17. Edery, I., Zwiebel, L. J., Dembinska, M. E. & Rosbash, M. Temporal phosphory lation of the Drosophila period protein. Proc. Natl Acad. Sci. USA 91, 2260-2264 (1994) | PubMed | 18. Myers, M. P., Wager-Smith, K., Rothenfluh-Hilfiker, A. & Young, M. W. Light- induced degradation of TIMELESS and entrainment of the Drosophila circadian cloc k. Science 271, 1736-1740 (1996) | PubMed | 19. Zeng, H. K., Qian, Z. W., Myers, M. P. & Rosbash, M. A light-entrainment mec hanism for the Drosophila circadian clock. Nature 380, 129-135 (1996) | PubMed | 20. Kaneko, M., Park, J. H., Cheng, Y., Hardin, P. E. & Hall, J. C. Disruption o f synaptic transmission or clock-gene-product oscillations in circadian pacemake r cells of Drosophila cause abnormal behavioral rhythms. J. Neurobiol. 43, 207-2 33 (2000) | Article | PubMed | 21. Yang, Z. & Sehgal, A. Role of molecular oscillations in generating behaviora l rhythms in Drosophila. Neuron 29, 453-467 (2001) | PubMed | 22. Hunter-Ensor, M., Ousley, A. & Sehgal, A. Regulation of the Drosophila prote in timeless suggests a mechanism for resetting the circadian clock by light. Cel l 84, 677-685 (1996) | PubMed | 23. Lee, C. G., Parikh, V., Itsukaichi, T., Bae, K. & Edery, I. Resetting the Dr osophila clock by photic regulation of PER and a PER-TIM complex. Science 271, 1 740-1744 (1996) | PubMed | 24. Naidoo, N., Song, W., Hunter-Ensor, M. & Sehgal, A. A role for the proteasom e in the light response of the Timeless clock protein. Science 285, 1737-1741 (1 999) | Article | PubMed | 25. Price, M. A. & Kalderon, D. Proteolysis of the Hedgehog signaling effector C ubitus interruptus requires phosphorylation by glycogen synthase kinase 3 and ca sein kinase 1. Cell 108, 823-835 (2002) | PubMed | 26. Jia, J. et al. Shaggy/GSK3 antagonizes Hedgehog signalling by regulating Cub itus interruptus. Nature 416, 548-552 (2002) | Article | PubMed | 27. Pai, L. M., Orsulic, S., Bejsovec, A. & Peifer, M. Negative regulation of Ar madillo, a Wingless effector in Drosophila. Development 124, 2255-2266 (1997) | PubMed | 28. Stanewsky, R. et al. Temporal and spatial expression patterns of transgenes containing increasing amounts of the Drosophila clock gene period and a lacZ rep orter: Mapping elements of the PER protein involved in circadian cycling. J. Neu rosci. 17, 676-696 (1997) | PubMed | 29. Ruel, L., Pantesco, V., Lutz, Y., Simpson, P. & Bourouis, M. Functional sign ificance of a family of protein kinases encoded at the shaggy locus in Drosophil a. EMBO J. 12, 1657-1669 (1993) | PubMed | 30. Cegielska, A., Gietzen, K. F., Rivers, A. & Virshup, D. M. Autoinhibition of casein kinase I &epsis; (CKI &epsis;) is relieved by protein phosphatases and l imited proteolysis. J. Biol. Chem. 273, 1357-1364 (1998) | Article | PubMed | Acknowledgements. We thank M. Boudinot for the Faas software, M. Serrier and L. Collet for help with the figures, M. Rosbash, P. Emery, A. Klarsfeld, J.-F. Juli en and E. Petrochilo for their comments and suggestions on the manuscript, as we ll as J. Champagnat and J.-D. Vincent for their continuous support. We thank I. Miletich for the unpublished UAS-slmb line, and R. Myers, L. Saez, R. Stanewsky and D. Virshup for providing antibodies or constructs. This work was supported b y CNRS (ATIPE "Développement" and appel d'offres "Biologie cellulaire") and Fon dation pour la Recherche Médicale. F.R. is supp -- ※ 來(lái)源:．生命玄機(jī)站 bbs.cst.sh.cn.