A Role for the Proteasome in the Light Response of the Timeless Clock Protein

See allHide authors and affiliations

Science  10 Sep 1999:
Vol. 285, Issue 5434, pp. 1737-1741
DOI: 10.1126/science.285.5434.1737


The cyclic expression of the period (PER) and timeless (TIM) proteins is critical for the molecular circadian feedback loop inDrosophila. The entrainment by light of the circadian clock is mediated by a reduction in TIM levels. To elucidate the mechanism of this process, the sensitivity of TIM regulation by light was tested in an in vitro assay with inhibitors of candidate proteolytic pathways. The data suggested that TIM is degraded through a ubiquitin-proteasome mechanism. In addition, in cultures from third-instar larvae, TIM degradation was blocked specifically by inhibitors of proteasome activity. Degradation appeared to be preceded by tyrosine phosphorylation. Finally, TIM was ubiquitinated in response to light in cultured cells.

In organisms ranging from cyanobacteria to mammals, the endogenous circadian clock is a feedback loop composed of cycling gene products that control their own synthesis. Such autoregulation has been demonstrated for theperiod (per) and timeless(tim) genes in Drosophila melanogaster, thefrequency (frq) gene inNeurospora, the kai gene cluster in cyanobacteria, and appears to be true of the mammalian pergenes (1, 2). In all cases, precisely timed negative feedback by the proteins results in rhythmic expression of the RNA.

Rhythmic feedback is effected largely through circadian control of protein levels. Thus, when protein levels are high, negative feedback occurs, resulting in low levels of RNA (2). Degradation of the proteins eliminates feedback and allows RNA levels to rise once again. Early disappearance of the protein (in the case of the pers mutant) results in an earlier rise of RNA levels and a shortening of circadian period by ∼5 hours (3). In addition, the accumulation of both PER and TIM proteins lags behind the rise of their RNAs by ∼6 hours and, at least in the case of PER, is regulated at the level of protein stability rather than RNA translation (4, 5). Phosphorylation of PER by the casein kinase encoded by thedouble-time gene renders it unstable in the absence of TIM (6). Finally, levels of the TIM protein, but not of the RNA, are reduced by light, a response that shifts the phase of the clock and apparently mediates resetting of behavioral rhythms (7–12).

To investigate the nature of the TIM light response, we developed an in vitro assay. We entrained flies to a 12 hour light/12 hour dark cycle and prepared head extracts from flies collected at either ZT (zeitgeber time) 20 or immediately after a 1-hour light pulse delivered at ZT 19 (ZT0 = lights on; ZT12 = lights off). These extracts were incubated with TIM protein immunoprecipitated from fly heads. After a 1-hour incubation at room temperature, we assayed TIM levels by protein immunoblots. Addition of the pulsed extract reduced the TIM signal (Fig. 1A). Unpulsed head extract had no effect on the level of TIM, indicating that the reduction was light-specific. This light-induced reduction was also observed in tim0 flies (which lack TIM protein) and was, in fact, routinely higher in these flies, which may suggest some down-regulation by the clock in wild-type flies. An immunoprecipitated PER substrate was not degraded by addition of the pulsed extract (Fig. 1B).

Figure 1

Degradation of TIM by a light-induced activity. Dash indicates incubated without the addition of extract. Approximately 1 mg of head extract made fromyellow-white (yw) flies collected at ZT19 (36) was incubated either with an antibody to TIM (UPR8) (7) or with an antibody to PER (UPR1) (5) overnight at 4oC. The immune complex was bound to protein G beads by incubation at 4oC for 1 hour and then washed in PBS. Head extract (37) from yw ortim 0 flies collected at ZT20 either after a 1-hour pulse of light (light) or unpulsed (dark) was added to the immunoprecipitated TIM and incubated for 1 hour at 25oC. TIM levels were assayed by SDS–polyacrylamide gel electrophoresis and immunoblotting. We routinely detect two bands with this particular antibody to TIM. Based on data in the laboratory, we believe that they reflect products of alternatively spliced forms of TIM (13). They both respond to light. The variability of this assay is addressed in (16). (A) Degradation of TIM by a light-induced activity in yw andtim 0 flies. (B) A light-treatedyw extract that degraded TIM (upper panel) failed to degrade PER (lower panel). (C) Inhibition of degradation activity by inhibitors of the proteasome. These experiments were carried out as above, except for incubation of the light-treated extracts with the inhibitors MG115 (40 μM), MG132 (40 μM), and ALLN (40 μM) for 30 min at 4oC before their incubation with immunoprecipitated TIM. Light and dark control extracts had no inhibitors added.

In order to determine the nature of the proteolytic activity, we assayed several general classes of protease inhibitors. Inhibitors of serine proteases [phenylmethylsulfonyl fluoride (PMSF) and aprotinin] and aspartate proteases (pepstatin) were not very effective in blocking TIM degradation (13). However, degradation was inhibited by the proteasomal inhibitors acetyl-leu-leu-norleucinal (ALLN), cbz-leu-leu-norvalinal (ZL2NVaH or MG115) and cbz-leu-leu-leucinal (ZL3H or MG132) (Fig. 1C). These peptide aldehydes strongly inhibit the chymotryptic activity of the eukaryotic 26S proteasome (14). TIM degradation was also blocked by bestatin, a metalloprotease inhibitor, and by leupeptin, which inhibits cysteine proteases and has some effects on other proteolytic systems, including the proteasome (15). The precise mechanism of action in this case is not known. Consistent with a role for the proteasome, we found that depletion of ubiquitin from the extract blocked TIM degradation (13).

Although the in vitro assay indicated a mechanism for TIM's response to light, its usefulness was limited by its variability (16). To verify the findings of this assay, we sought to develop an in vivo system. Thus, we used a primary culture assay in which we investigated the effect of two proteasomal inhibitors, lactacystin and MG115 on the TIM light response. Lactacystin, a microbial metabolite, is the most specific known naturally occuring inhibitor of the proteasome (17). It spontaneously hydrolyzes into clastolactacystin B lactone, which is the active species that reacts with the proteasome, inhibiting its chymotryptic and tryptic peptidase–like activity (18). MG115 (Fig. 1C) is a potent synthetic peptide aldehyde inhibitor with an inhibition constant (K i) of 0.02 μM.

For our assay, we dissected the central nervous system (CNS) of third-instar larvae as described (19) and maintained them in culture medium for 1 hour. Some samples were exposed to a pulse of light for 20 min and were fixed at the end of the hour. Dark control samples were also incubated for an hour in the dark (Fig. 2). We then examined TIM expression in the lateral neurons (clock cells), which were located by costaining with an antibody to pigment-dispersing hormone (PDH) (Fig. 2). We observed strong TIM staining in lateral neurons of unpulsed tissue, but little to no TIM in CNS tissue that had received a light pulse (Fig. 2 and Table 1). The effect of inhibitors was tested by adding them to the culture medium at the start of the incubation. Tissue treated with lactacystin and MG115 before the light pulse revealed robust TIM staining in the lateral neurons. The strong inhibition by MG115 was consistent with a report that this is a much more effective inhibitor of proteolysis in intact cells than of in vitro hydrolysis of macromolecular substrates (14). The <100% block by lactacystin may reflect variable permeability or instability of the lactone metabolite. The dimethyl sulfoxide vehicle did not inhibit degradation. Treatment with PMSF, a serine protease inhibitor, also did not affect degradation (13).

Figure 2

TIM degradation in lateral neurons is mediated by the proteasome. (A) TIM degradation was assayed in larval lateral neurons in the presence or absence of specific inhibitors. TIM staining is shown on the right and PDH on the left. The lateral neurons are indicated by arrows. Third-instar larvae were collected at ZT20 on ice, washed in 70% ethanol, and held in Schneider's medium on ice. The CNS was dissected out and maintained in culture medium [Schneider's medium containing 100 penicillin (μg/ml), streptomycin (100 μg/ml), fungizone (0.25 μg/ml), gentamycin (20 μg/ml), insulin (500 ng/ml), and 20% fetal bovine serum]. (B) For the inhibitor studies, the inhibitors MG115 (200 μM) or lactacystin (50 μM) or the vehicle control (DMSO) were present in the culture medium during dissection and incubation. The tissue was incubated for approximately 45 min with the inhibitor before receiving a 20-min light pulse, followed by an additional 40-min incubation in low light. The tissue was fixed in 4% paraformaldehyde in PBS for 30 min before washing and staining with TIM and PDH antibodies (38).

Table 1

Light-induced TIM degradation was assayed in lateral neurons in the presence of protease or kinase inhibitors. Each sample corresponds to lateral neurons from a single larval brain hemisphere (in some larvae both hemispheres were retained, but this was not always the case). n indicates the total number of samples from at least two experiments for each inhibitor (except PMSF and PD 98059, which were only tested once). Each experiment assayed three to five samples. “Residual” staining indicates samples in which staining was reduced but still detectable.

View this table:

Proteasomes are multicatalytic, multisubunit proteolytic complexes with highly conserved structures, and they play a key role in a variety of cellular processes, including the cell cycle, transcriptional regulation, removal of abnormal proteins from the cell, antigen presentation (20), and even in the turnover of a mammalian circadian-regulated protein (21). The TIM response to light is blocked specifically, in two different assays, by several inhibitors of the proteasome, which is important given that lactacystin, which was thought to affect only the proteasome (17), was recently shown to also act on a second multisubunit enzyme (22). Because the newly identified enzyme is insensitive to ALLN, it cannot account for the TIM response (see Fig. 1C). For the ubiquitin-proteasome system, proline glutamate serine threonine (PEST) regions sometimes serve as putative degradation/phosphorylation signals in the target molecule (23). The TIM protein sequence revealed the presence of seven PEST regions concentrated near the NH2 and COOH termini.

Most cellular proteins that are degraded by the proteasome are ubiquitinated and then targeted to the proteasome (24). To determine whether TIM is ubiquitinated, which would also demonstrate that it is a direct target of the proteasome, we used a cell culture system. We expressed TIM and a hemagglutinin (HA)–tagged ubiquitin octamer (25) under heat shock control in Drosophila S2 cells. After a 30-min heat shock, cells were either maintained in the dark or treated with light for 2 hours, after which the cells were lysed and immunoprecipitates of TIM were probed with an antibody to HA. We found that TIM was ubiquitinated in response to light (Fig. 3). The effect was specific for TIM, because PER was not ubiquitinated with or without light treatment (Fig. 3). Extended light treatment also degraded TIM in these cells, and this degradation was inhibited by the proteasome inhibitor MG115 (13).

Figure 3

Light-induced ubiquitination of TIM in cultured cells. per, tim, and HA-tagged ubiquitin were cloned into the hsp70 promoter-driven expression vector pCasPer-hs, to form hs-per, hs-tim, and hs-Ub. Drosophila S2 cells were transfected by means of the calcium phosphate procedure, and expression was induced by heat shock (39). After 2 hours of light or dark treatment, the cells were harvested and lysed. TIM or PER was immunoprecipitated as described in Fig. 1. The protein immunoblots were probed with an antibody to HA (1:1000 dilution). The constructs used in each transfection are indicated above each lane. The number at the left of the panels indicates the position of the molecular weight marker.

Although our data implicate a ubiquitin-proteasomal mechanism, they do not preclude a role for other proteolyic systems. In particular, bestatin is an effective blocker in the in vitro assay, which may indicate a role for a metalloprotease, in particular an aminopeptidase. The inhibitory effects of bestatin on this process could reflect a recently proposed proteasomal mechanism that involves a dipeptide pump (26) or, alternatively, preprocessing by a metalloprotease. The ErbB4 tyrosine kinase receptor is cleaved by a metalloprotease, then ubiquitinated and degraded by the proteasome (27).

As mentioned above, some cell cycle proteins are degraded via the proteasome pathway. In many cases, this degradation is preceded by a phosphorylation event (28–30). Moreover, PER degradation requires its phosphorylation by the casein kinase encoded by the double-time gene (6). To investigate a possible role for phosphorylation in the degradation of TIM, we examined the effect of several kinase inhibitors in the in vivo primary culture assay. The tyrosine kinase inhibitor genistein blocked the degradation of TIM in the lateral neurons after a pulse of light, whereas the serine-threonine inhibitors staurosporin and calphostin C and the MEK inhibitor PD98059 did not (Fig. 4A and Table 1) (31–35). These results suggest that tyrosine kinase activity precedes degradation of TIM. The concentrations of genistein that were effective in this assay (>50 μm) suggest ac-src–like kinase activity (31), although the concentration dependence must be interpreted with caution, because it could be a measure of permeability or drug stability.

Figure 4

TIM degradation in response to light involves tyrosine kinase activity. (A) The experiment was carried out as described in Fig. 2, except that genistein (25 μg/ml) or 1 μM staurosporin (Biomol) was added before light treatment in the samples indicated. The dark control was fixed immediately in this experiment. (B) Head extracts (containing phosphatase inhibitors, okadaic acid, and bromotetramisole oxalate) were prepared from adult fly heads collected after a 20-min light pulse at ZT19 (light) or from unpulsed controls collected at the same time (dark). Extracts were incubated at 25oC for 40 min (foryw only) and then treated with an antibody to TIM. Aliquots of the TIM immunoprecipitates were run on two gels, and the protein immunoblots were probed either with an antibody to phosphotyrosine (Sigma, 1:200 dilution) (top) or with an antibody to TIM (bottom). The location of the TIM/phosphotyrosine band is indicated by an arrow. The numbers at the left indicate the position of the molecular weight marker. The 40-min 25oC incubation was done to promote the TIM degradation process, which would allow us to examine the effect of proteasome inhibitors on phosphorylation. However, the additional incubation did not lead to any TIM degradation, and so we observed the same level of phosphorylation with and without proteasome inhibitors (13). Please note that the tim0 sample was not subjected to the extra incubation before immunoprecipitation. It is included here to distinguish the TIM band from a cross-reacting lower band. In other experiments whereyw and tim0 flies were treated identically, the phosphotyrosine band in TIM immunoprecipitates was specific to yw.

To determine whether the tyrosine phosphorylation occured on TIM itself, we probed protein immunoblots of TIM immunoprecipitates with an antibody to phosphotyrosine. After 20 min of light treatment at ZT19, we were able to detect TIM with the antibody to phosphotyrosine (Fig. 4B). TIM in the “dark” samples was sometimes detected with this antibody but not consistently, which suggests that tyrosine phosphorylation of TIM is increased by light. The mobility of the TIM band in the light-treated sample was also reduced, presumably because of phosphorylation.

Together, our data indicate that the TIM response to light involves tyrosine phosphorylation and ubiquitination, followed by proteasomal degradation. An important question that arises from these studies is, what is the role of the proteasome pathway in free-running behavioral rhythms. Are the mechanisms that degrade TIM in response to light the same as those that degrade it in constant darkness, in which case light may serve only to further activate a process that is already under way. We propose that cyclic turnover of TIM under free-running conditions is mediated by phosphorylation, which targets it for degradation, perhaps by the proteasome. TIM is progressively phosphorylated throughout the night, and maximally phosphorylated forms are found just before the rapid decline of protein levels (10). From this point on, until the middle of the day, TIM levels remain low because of the low levels of RNA. As the repression of transcription is released, most likely because of the decrease in PER levels, RNA accumulates and protein also starts to accumulate, albeit slowly because it is still subject to phosphorylation and degradation. When the rate of TIM synthesis exceeds the rate of phosphorylation/degradation, higher levels of protein are observed, but as the phosphorylation program continues and RNA levels are reduced (because of negative feedback), levels of the protein drop off. Light could enhance TIM degradation by increasing TIM phosphorylation or increasing proteolytic activity in some manner (or both). This model would predict that the presence of light accelerates the falling phase of the protein and delays the rising phase, both through the same mechanism.

  • * Present address: Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA.

  • To whom correspondence should be addressed. E-mail: amita{at}


View Abstract

Stay Connected to Science

Navigate This Article