Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Nuclear localization is required for function of the essential clock protein FRQ

Nuclear localization is required for function of the essential clock protein FRQ The EMBO Journal Vol.17 No.5 pp.1228–1235, 1998 Nuclear localization is required for function of the essential clock protein FRQ of frq mRNA from an inducible promoter rapidly depresses Chenghua Luo, Jennifer J.Loros and the level of the endogenous frq message (Aronson et al., Jay C.Dunlap 1994a; Merrow et al., 1997). These data place frq mRNA Department of Biochemistry, Dartmouth Medical School, Hanover, and FRQ protein in a negative feedback loop defining the NH 03755, USA circadian oscillator. frq mRNA accumulates very quickly Corresponding author (within 2 min) in response to short light pulses, and the e-mail: [email protected] threshold and kinetics of the response are similar to those of light-induced resetting (Crosthwaite et al., 1995). Two The frequency (frq) gene in Neurospora encodes central FRQ polypeptides arise from the frq transcript, resulting components of a circadian oscillator, a negative feed- from alternative initiation of FRQ translation from two back loop involving frq mRNA and two forms of FRQ in-frame initiation codons: codon #1 and codon #100 protein. Here we report that FRQ is a nuclear protein (Garceau et al., 1997). Both FRQ forms are immediately and nuclear localization is essential for its function. and progressively phosphorylated after being synthesized Deletion of the nuclear localization signal (NLS) (Garceau et al., 1997). Temperature regulates the ratio of renders FRQ unable to enter into the nucleus and the two different FRQ forms by favoring different transla- abolishes overt circadian rhythmicity, while reinsertion tion initiation sites at different temperatures and thus sets of the NLS at a novel site near the N-terminus of FRQ the physiological temperature limits for rhythmicity (Liu restores its function. Each form of FRQ enters the et al., 1997). nucleus soon after its synthesis in the early subjective The mechanism by which FRQ acts to depress the level day; there is no evidence for regulated sequestration of its own transcript or more generally contributes to in the cytoplasm prior to nuclear entry. The kinetics clock function is unknown. This negative action may be of the nuclear entry are consistent with previous achieved transcriptionally, post-transcriptionally or both. data showing rapid depression of frq transcript levels In the Drosophila system, a complex composed of the following the synthesis of FRQ, and suggest that early clock proteins PER and TIM is thought to regulate per in each circadian cycle, when FRQ is synthesized, it and tim transcription through a negative feedback loop enters the nucleus and depresses the level of its own (Gekakis et al., 1995; Myers et al., 1996; Zeng et al., transcript. 1996) to generate the circadian oscillation. Both PER and Keywords: circadian rhythm/frequency/Neurospora/ TIM are predominantly nuclear proteins in wild-type flies nuclear localization (Liu et al., 1992), and nuclear translocation is regulated by the dimerization between the two proteins (Saez and Young, 1995). Studies done in Drosophila cell cultures have shown that each protein contains sequences that Introduction confer cytoplasmic localization, and dimerization between them suppresses the cytoplasmic localization function Circadian rhythmicity is a general aspect of regulation (Saez and Young, 1995). Thus, both proteins accumulate existing in a wide variety of organisms ranging from in the cytoplasm following their translation, then enter prokaryotes to eukaryotes (Dunlap, 1993, 1996). The into the nucleus in a concerted manner after this delay to cellular machinery that generates this capacity is known execute their effect on transcription (Curtin et al., 1995; as the biological clock. Such clocks are intrinsic to the Saez and Young, 1995). On the other hand, in the silkmoth cell, are endogenous and self-sustaining under constant Antheraea pernyi, the PER protein is nuclear in eye conditions, and respond to environmental cues such as photoreceptor cells but is predominantly cytoplasmic in light and temperature (Crosthwaite et al., 1995; Hunter- the brain cells thought to drive behavioral rhythm (Sauman Ensor et al., 1996; Lee et al., 1996; Zeng et al., 1996; and Reppert, 1995). per mRNA and PER protein oscillate Liu et al., 1997). in both cell types, and the predominantly cytoplasmic The products of the frequency (frq) gene are central PER in the brain cells has suggested that the silkmoth components of the Neurospora oscillator (Dunlap, 1993, may use a different mechanism to construct the molecular 1996; Loros, 1995), and aspects of their regulation have oscillator. In Neurospora, FRQ has several characteristics pointed to roles for transcription and translation in the consistent with an involvement in transcription, including a clock. frq mRNA is expressed rhythmically with a period putative helix–turn–helix DNA-binding domain, a nuclear reflecting that of the overt rhythm. Constantly elevated localization signal (NLS) and highly charged regions expression of frq message from a heterologous inducible promoter eliminates overt rhythmicity in a wild-type (Aronson et al., 1994b; Merrow and Dunlap, 1994; Lewis background, and is unable to support rhythmicity in a et al., 1997). Both frq mRNA and FRQ oscillate, the peak loss-of-function strain (frq ). Step reduction of frq mRNA of FRQ lags the peak of frq mRNA by 4–6 h, and frq resets the clock to a predicable phase. Elevated expression mRNA levels begin to fall before the level of FRQ reaches 1228 © Oxford University Press Nuclear localization of FRQ It is known that two different forms of FRQ arise from two in-frame AUGs in the frq transcript generating ~135 and ~145 kDa proteins. Each form of FRQ is phosphoryl- ated progressively after being synthesized (Garceau et al., 1997). Biochemically isolating nuclei and detecting where FRQ is has the advantage of providing information on how these different FRQ forms behave. Figure 1 shows the results of Western analysis of FRQ, CYS-3 and CCG-1 from a mid-subjective morning sample (DD15, see Materials and methods) in the total cell lysate and nuclear fractions. CYS-3 was present in both lysate and Fig. 1. Both forms of FRQ are found inside nuclei in Neurospora. nuclear fractions, and its signal was enriched ~3- to 4-fold Western blots show FRQ and two control proteins (CYS-3 and in the nuclear fraction compared with that in the lysate CCG-1) in total cell lysate and nuclear fractions. The two arrows point fraction. This is consistent with incomplete nuclear localiz- to the two different forms of FRQ resulting from two in-frame translational initiation sites in the same message. Both FRQ forms are ation combined with breakage of nuclei upon isolation. detected in the nuclear fraction. There is no CCG-1 signal in the CCG-1 was present exclusively in the lysate fraction. The nuclear fraction, indicating that the isolated nuclei are free from results from the two protein markers confirm that the cytosolic contamination. isolated nuclear fraction has no cytoplasmic contamination. FRQ, like CYS-3, was found in the nuclear fraction (FRQ its peak (Garceau et al., 1997), suggesting that FRQ might nuclear panel), suggesting it is a nuclear protein. Its signal execute its role quickly, probably very soon after being is enriched in the nuclear fraction, and the fold enrichment synthesized. is similar to that of the nuclear marker protein CYS-3. In If FRQ functions at a transcriptional level in the addition, both large and small FRQ forms are found in feedback inhibition loop, it must be in the nucleus at least the nuclear fraction (as denoted by the two arrows in at certain times during a circadian cycle. To examine Figure 1). this, we adapted a cellular fractionation technique and determined the intracellular compartmentation of FRQ as Deletion of the NLS from FRQ eliminates nuclear a function of time of day. Our data demonstrate that FRQ localization and abolishes overt circadian is indeed a nuclear protein, and its nuclear location is rhythmicity essential for its function in the circadian clock. A single To determine whether the nuclear localization of FRQ is NLS located at amino acids 195–200 is necessary and required for its function, we identified and deleted the sufficient for FRQ nuclear entry. This NLS is still func- NLS of FRQ. Two putative SV40 large T antigen-type tional when moved to a different site near the N-terminus NLSs located at amino acids 194–199 and 564–568 are of FRQ; it allows nuclear localization of the large form predicted from computer analysis of FRQ’s amino acid of FRQ alone and is sufficient to restore clock function. sequence (Aronson et al., 1994b; Merrow and Dunlap, Both forms of FRQ move into the nucleus very soon after 1994; Lewis et al., 1997). The first one is phylogenetically being synthesized and well before levels of FRQ peak or more conserved among frq homologs (Lewis et al., 1997). phosphorylation of FRQ has progressed very far. The Constructs bearing deletions of each NLS separately and absence of a noticeable delay prior to nuclear entry a double deletion of both were made (Figure 2A), and suggests that delay prior to nuclear localization contributes targeted by transformation to the his-3 locus in the frq little to the long-term control of the circadian cycle, nor null strain (strain #93-4) (Aronson et al., 1994b). A is it required to generate the long time constant of the putative NLS (RKKRK) starting at amino acid 564 was circadian oscillator. deleted in construct pCL7; the frq transformants bearing this construct still had apparently normal nuclear localiz- Results ation of FRQ and retained the overt circadian conidial Both forms of FRQ are found inside the nucleus banding rhythm (data not shown). In construct pCL10, a In order to study the subcellular localization of FRQ, we putative NLS (PRRKKR) beginning at amino acid 194 was have utilized a biochemical method to isolate nuclei from deleted. Transformants bearing this construct produced a Neurospora, resolved the nuclear fractions by SDS–PAGE, normal amount of FRQ, but almost no FRQ was detected and detected FRQ by Western blot analysis. Two proteins in the nuclear fraction (Figure 2B and C), suggesting that were used as control markers in this method: CYS-3 and this region is the important NLS directing FRQ into the CCG-1. CYS-3 is used as the nuclear protein marker. It nucleus. To assess the functional significance of this is a transcription factor involved in regulating structural mutation on the action of FRQ in the oscillator, the proteins in the sulfur metabolism pathway in Neurospora overt circadian rhythm in these transformant strains was (Fu et al., 1989; Paietta, 1992; Kanaan and Marzluf, examined in race tubes (Figure 3A). If the nuclear localiz- 1993). CCG-1 is the product of clock-controlled gene-1 ation of FRQ is important for its function, we would (McNally and Free, 1988; Loros et al., 1989). It is an expect that this construct could not rescue the loss of abundant and exclusively cytoplasmic protein (Loros et al., circadian rhythmicity in the frq (null) strain. As expected, 1995; Garceau, 1996), and it is used as a cytoplasmic this NLS-deleted frq failed to support the overt conidiation protein marker in this study. The two protein markers rhythm (compare pCL10 with pKAJ120 race tubes). Aside were used to show the quality of the isolated nuclei, from the loss of rhythmicity, the growth rate and general specifically to verify the absence of cytoplasmic contamin- morphology were not altered in the pCL10 transformants, ation in the nuclear preparation (Figure 1). suggesting that no other major defects were associated 1229 C.Luo, J.J.Loros and J.C.Dunlap Fig. 3. Deletion of the NLS in FRQ abolishes overt circadian rhythmicity. (A) Race tube analysis showing the banding rhythm in different strains at 25°C: wild-type (frq ), a transformant bearing an intact frq locus (pKAJ 120), a transformant bearing the NLS deletion construct (pCL10) and the frq strain (null). No conidial banding rhythm is found in the pCL10 transformant. Cultures were inoculated at the left ends of race tubes, incubated in constant light for 1 day and transferred into darkness at 25°C. The growth front (vertical lines) was marked at 24 h intervals thereafter. (B) Western blot of FRQ from ΔNLS strain pCL10 over time points spanning a circadian cycle. Both the amount and the phosphorylation states of FRQ remain unchanged Fig. 2. An NLS located between amino acids 194 and 199 is required over time, indicating an absence of rhythmicity. for FRQ nuclear localization. (A) A schematic diagram of the ΔNLS construct. The open box represents the frq ORF, and arrows mark the positions of the methionines initiating the large and small FRQ into the N-terminal region of ΔNLS construct pCL10 translation products. The black box marked NLS corresponds to amino (Figure 4A). A putative casein kinase II phosphorylation acids 194–199. (B) Western blot of FRQ in total lysate and nuclear site (T LRD) was included in the 16 amino acid fractions of three different strains: bdA (wild-type, frq ), pCL10 10 10 transformant (ΔNLS construct) and frq (frq null). frq served as a sequence, since it has been reported that phosphorylation negative control. In the nuclear fraction in the pCL10 transformant, the at a CKII site close to an NLS can enhance the rate of amount of FRQ is reduced significantly. (C) Densitometry of the nuclear translocation (Jans and Jans, 1994). This construct amount of FRQ in different fractions in each strain shown in (B). (pCL11) was transformed into the frq null strain (strain Levels of FRQ are normalized against the FRQ level in the lysate fraction in the bdA (frq ) strain at DD15. #93-4) at the his-3 locus, and the transformants were examined for nuclear localization of FRQ and for overt rhythmicity. Theoretically, NLSs can be located anywhere with this mutation. In confirmation of the race tube data, in a protein’s primary structure, provided that after protein neither frq mRNA nor FRQ showed any rhythmicity in folding, the NLS is on the surface of the protein (reviewed pCL10 transformants, as shown in Figure 3B. Interestingly, by Garcia-Bustos et al., 1991; Jans and Hubner, 1996). both the amount and the phosphorylation states of FRQ The insertion was made into an SphI site near the 5 end remained relatively unchanged across different time points of frq’s open reading frame (ORF), just 11 amino acids spanning a circadian cycle in the pCL10 transformants. after AUG#1 (Garceau et al., 1997). We chose to insert The observation that the phosphorylation states remain the putative NLS near to the N-terminus of FRQ since the same over time is similar to the behavior of wild- structural predictions place this on the surface of the type FRQ under constant light conditions (N.Y.Garceau, protein and this insertion also provides a convenient J.J.Loros and J.C.Dunlap, in preparation). In both cases, internal control in the subsequent analysis. Specifically, there is no circadian oscillation in the organism. This insertion of the NLS at residue 11 should support nuclear suggests that the phosphorylation of FRQ is coordinated localization of the large form of FRQ but not the small with the circadian cycle. Together, these data show that it form which arises from alternative translation initiation at is necessary for FRQ to enter into the nucleus to generate amino acid 100. This insertion is not likely to destroy any both molecular and overt rhythmicity. essential functional domain(s) in FRQ since small FRQ alone has been shown to be sufficient for establishing Insertion of the NLS near to the N-terminal region rhythmicity (Liu et al., 1997). of FRQ restores its function in the clock Reinsertion of the NLS restored nuclear localization of To test if the nuclear localization of FRQ could be restored the large form of FRQ. Figure 4B shows the Western blot and be sufficient for its function in the clock, the NLS and results of total cell lysate and nuclear fractions of strains its surrounding region (amino acids 193–208, including the bdA (frq ), pCL11 (N-terminal NLS insertion construct) six amino acids extending from 194 to 199) was inserted and pCL10 (NLS deletion construct) at DD15. All three 1230 Nuclear localization of FRQ Fig. 5. Insertion of the NLS into the N-terminal region of ΔNLS FRQ pCL10 restores a functional clock. Race tube analysis showing the overt conidial banding rhythm of transformants bearing a wild-type frq locus (pKAJ120), the NLS deletion (pCL10) and the NLS insertion (pCL11) constructs. Race tubes were inoculated and incubated in constant light for a day before being transferred into constant darkness at different temperatures. The rhythm was restored in the pCL11 transformant; rhythmicity was apparent at 25°C and at a higher temperature (27°C), but the rhythm was lost at a lower temperature (22°C). Table I. Period length in different strains Strains Period (h  SD) bdA/25°C 21.2  1.3 pKAJ120/25°C 23.6  1.1 pCL10/25°C ND pCL11/25°C 24.8  1.5 pCL11/27°C 24.6  0.7 pCL11/22°C ND frq /25°C ND Fig. 4. Insertion of the NLS at a different position in ΔNLS FRQ directs the FRQ protein to enter the nucleus. (A) Schematic diagram of the constructs (see Figure 2 and text for details). (B) Western blot Figure 4C, both FRQ forms were in wild-type and pCL11 of FRQ in the total cell lysate and nuclear fractions of different lysate fractions, but only the large form of FRQ was found strains: bdA (wild-type, frq ), pCL10 (ΔNLS) transformant and pCL11 in the pCL11 nuclear fraction. The faint nuclear-enriched (NLS insertion) transformant. Both forms of FRQ shown by the two arrowheads were found in the total lysate in all three strains. In the band sandwiched between the two FRQ bands is non- nuclear fractions, both FRQ forms are detected in the wild-type strain specific, since it is recognized by the antiserum in the but only the large FRQ form is detected in the pCL11 transformant, frq (null) strain also. and no FRQ is detected in the pCL10 transformant. (C) Western blot Reinsertion of the NLS also restores the function of the of total cell lysate fractions and nuclear fractions together with controls, subjected to λ phosphatase treatment. JCC101 bears a large form of FRQ. Transformation of pCL11 into a frq disruption at the third AUG and YL15 bears a disruption at the first null strain rescued the clock null phenotype (Figure 5; AUG; they were used as controls for strains making large and small Table I), suggesting that once FRQ gets into nuclei it can FRQ, respectively. Both FRQ forms were detected in the wild-type function normally. It has been shown that the small FRQ and pCL11 transformants (wt, pCL11 lysate), while only large FRQ is form is important for the clock to function at the lower found in the nuclear fraction from the pCL11 transformant. end of the physiological temperature range (Liu et al., 1997). Thus, a reasonable prediction is that pCL11 should strains synthesized both FRQ forms (Figure 4B lysate support rhythmicity at higher temperatures but not at panel, as pointed out by the arrowheads). Both FRQ forms lower temperatures since only the large FRQ enters the were inside the nucleus in the bdA (frq ) strain. FRQ was nucleus in this strain. This prediction was proved to be found inside the nucleus of strains bearing construct correct: at 27°C the clock was functional in pCL11 pCL11 (pCL11 nuclear panel), confirming that the NLS transformants while at 22°C the clock no longer operated can direct FRQ into the nucleus despite the altered position. (Figure 5, pCL11 race tubes at 27°C versus 22°C). Data In addition, only the large FRQ form was found in the shown in Table I confirm that the period lengths of the nucleus; this is in agreement with our prediction, since rescued strains were comparable with those rescued by only the large FRQ has the reinserted NLS. The progressive wild-type frq, again confirming normal function of large phosphorylation of FRQ described above results in a series FRQ with a displaced NLS. of FRQ bands that tend to obscure the appearance of protein bands on the gels. In order to generate a visually The amount of FRQ in nuclei cycles with a peak clearer result, total cell lysate and the nuclear fraction amount at CT4 were treated with λ phosphatase (Garceau et al., 1997) to In order to address the behavior of FRQ in terms of convert all different FRQ forms (arising from different cellular localization during a whole circadian cycle, a levels of phosphorylation) into just the two bands corres- quantitative analysis of FRQ in the nuclear extract was ponding to large and small FRQ. The result is shown in performed. Liquid cultures were grown in constant light Figure 4C. JCC101 and YL15 are strains that make only at room temperature before being transferred into constant large and small FRQ, respectively, since they bear a darkness at 25°C. The cultures were harvested at 4 h genetically engineered knock-out of the third AUG intervals after the light to dark transfer. Nuclei were (JCC101) and first AUG (YL15) (Liu et al., 1997). In isolated from samples at different circadian time points 1231 C.Luo, J.J.Loros and J.C.Dunlap of the circadian cycle (CT0), significant amounts of FRQ were detected in the nuclear fraction (nuclear panel CT0). This suggests that once FRQ is synthesized, it rapidly enters the nucleus with essentially no delay. Consistent with this, the maximal amount of FRQ inside nuclei is at CT4, earlier than that in the total cell lysate, and, at CT0, when there is little FRQ detected in the total lysate, FRQ accumulates to a significant level inside the nucleus. Although the nuclear FRQ distribution profile is phase- advanced with respect to the total cell lysate, it is very similar to that of frq mRNA, which normally peaks at CT4 (Crosthwaite et al., 1995); this is apparent comparing Figure 6A, nuclear panel, with Figure 6B (see also Garceau et al., 1997), although the frq RNA level at CT0 was abberantly high in this series. Since it has been reported that different phosphorylation forms of the same protein may be extracted differently during nuclei isolation (Mittnacht and Weinberg, 1991; Templeton et al., 1991), we examined the effect of salt concentration on the apparent cellular localization of FRQ. Nuclei were isolated under different salt concentrations (0 and 200 mM NaCl), and the nuclear FRQ profiles were the same (data not shown), indicating that different FRQ phosphorylation forms are not artificially producing an apparent change in localization. Figure 6C plots the amount of FRQ normal- ized against the time-invariant protein CYS-3 and the amount of frq mRNA over a circadian cycle. It is clear that the amount of FRQ oscillates in a circadian manner in both lysate and nuclear fractions, although the phase Fig. 6. Nuclear cycling of the FRQ protein. (A) Western blots of FRQ, of each cycle is different: the peak amount of FRQ inside CYS-3 and CCG-1 during a circadian cycle. In the total cell lysate the nucleus occurs at CT4 and has a similar phase to that (left panel) and nuclear fractions (right panel), CYS-3 protein levels of the frq mRNA cycle, peaking ~4 h before the total are constant throughout the circadian cycle. CCG-1 is a clock- controlled protein, peaking in the subjective morning (CT0). No amount of FRQ peaks in the cell (CT8). CCG-1 is detected in the nuclear fraction at any time. FRQ cycles in both fractions, but the phases are different. The amount of FRQ in the nuclear fraction peaks at DD15 (CT4) while the amount of FRQ in the Discussion total cell lysate fraction peaks at DD18 (CT8). These data are representative of five experiments. (B) Northern blot of frq mRNA in It has been well established that frq mRNA and FRQ a wild-type (frq ) strain over a circadian cycle. (C) Densitometric protein, the central components of a negative feedback analysis plotting the amount of FRQ normalized against the time- loop, are defining elements of the Neurospora circadian invariant protein CYS-3 in total cell lysate (u), nuclear fraction (j) clock. When FRQ is induced from a heterologous inducible and frq mRNA (s) over a circadian cycle. The nuclear cycle is promoter, the level of endogenous frq mRNA is depressed phase-advanced with respect to that of the total cycle by ~4 h, and it has a similar phase to that of the mRNA cycle. The data are plotted as (Aronson et al., 1994a; Merrow et al., 1997). Nevertheless, mean  2 SEM (n5). many details of how this negative feedback loop works are poorly understood. Data presented here clearly demon- spanning a whole cycle, and the total cell lysate and strate that both the large and small forms of FRQ, arising nuclear fractions were subjected to Western blot analysis. from different translation initiation sites in the same Western blot results for FRQ, CYS-3 and CCG-1 are message, are nuclear proteins. This result is consistent shown in Figure 6A. CYS-3 is a constitutively expressed with FRQ acting transcriptionally to down-regulate its protein; the amount of CYS-3 at different circadian times own message. The same NLS exists for both FRQ forms remained unchanged in both lysate and nuclear fractions and, in the context of the rest of FRQ, this signal is (Figure 6A, CYS-3 lysate and nuclear panel). For the necessary and sufficient to direct the FRQ protein into the clock-controlled protein CCG-1, the amount of protein nuclei. Previous results showed that rhythmic expression cycled over the circadian cycle with a peak at subjective of frq, rather than simple constitutive expression, was morning, confirming that the cultures are normally rhyth- essential for a functional clock (Aronson et al., 1994a). mic (Figure 6A, CCG-1 lysate panel). There was no We have demonstrated further here that in order to generate CCG-1 signal detected in the nuclear fractions at any time the rhythmic expression of frq mRNA and protein, FRQ point (Figure 6A, CCG-1 nuclear panel). In the total lysate must be transferred into the nucleus. When the NLS is fractions, FRQ showed the characteristic mobility shifts deleted, FRQ does not enter the nucleus and both molecular resulting from progressive phosphorylation over the circa- and overt rhythms are abolished; when the NLS is put dian cycle, and the amount of the protein peaked at CT8 back at a different position in the protein sequence, the as previously reported (Garceau et al., 1997; Figure 6A, rhythm of the clock is restored. FRQ lysate panel). In the nuclear fractions, FRQ was At all the time points tested during a circadian cycle, present at all time points and, especially at early stages some FRQ is always found inside the nucleus. Particularly 1232 Nuclear localization of FRQ at early stages of the circadian cycle such as CT0 and Additionally, we have shown previously (Liu et al., 1997) CT4, a significant amount of FRQ has accumulated in the that strains expressing only one of two forms of FRQ nucleus, suggesting that FRQ enters the nucleus very exhibit rhythms of reduced quality such as that seen here. quickly after being synthesized. If FRQ executes its Although both FRQ forms are expressed here, only large negative effect on its own transcription while inside the FRQ enters the nucleus and hence it may be that only it nucleus, then it could repress transcription and reduce is active. transcript levels within just a few hours, consistent with Among the systems used for clock research, Neurospora previous data (Merrow et al., 1997). In this study, we and Drosophila are the best characterized. A common showed that the amount of FRQ not only cycles in the theme arising from years of molecular research is that total cell lysate but also cycles within the nucleus. The both oscillators are transcription–translation-based nega- nuclear cycling is phase-advanced with respect to that of tive feedback loops, i.e. the clock proteins (FRQ in the total cell lysate, but has a similar phase to that of frq Neurospora, PER and TIM in Drosophila) enter the mRNA. This is consistent with a model in which once nucleus to down-regulate their own transcript levels. FRQ is made, it enters the nucleus and represses its own However, the mechanisms through which these proteins transcription, resulting in a significant decline in frq achieve negative feedback have seemed somewhat differ- mRNA before maximum FRQ levels are reached. This ent: in Neurospora there appears to be little temporal lag early negative effect executed by FRQ would help to between the appearance of frq RNA, its translation and generate the phase lag between total FRQ and frq message the movement of FRQ into the nucleus in that total frq previously observed (Garceau et al., 1997). RNA and nuclear FRQ protein both peak around CT4. These data are also consistent with the rapid kinetics FRQ appears to execute its function in reducing frq mRNA of FRQ negative feedback as reported using a frq strain, levels rather quickly, beginning before peak nuclear FRQ that is the negative feedback itself occurs very rapidly levels are reached, with the result that frq mRNA levels (Merrow et al., 1997). In a frq background, frq mRNA decline significantly before total cellular FRQ levels peak fluctuates around a high level and (because of a frameshift at around CT8. Thus the lag between total frq RNA and mutation) no functional FRQ is made (Aronson et al., protein levels simply reflects the well-known lag that 1994a). If functional FRQ is induced from a heterologous accompanies the translation of proteins with half-lives of promoter, a substantial decrease in frq mRNA is observed even a few hours (Wuarin et al., 1992; Wood, 1995). within 3 h, and the repression is essentially complete by Later in the cycle, some FRQ remains in the nucleus to 6 h (Merrow et al., 1997). These rapid repression kinetics keep frq transcript levels reduced till the proteins turn are seen following induction of FRQ at or below normal over completely. In contrast to Neurospora,in Drosophila physiological levels (Merrow et al., 1997), so it is clear there is reported to be a controlled lag between the that amounts of FRQ lower than the peak are quite synthesis/accumulation of PER in the cytoplasm and its effective in establishing and maintaining depressed levels translocation to the nucleus (Curtin et al., 1995). In this of frq mRNA. We have confirmed here that at least modest view, the regulation of PER nuclear entry, as distinct from levels of FRQ remain in the nucleus throughout the PER accumulation, was seen as playing an essential role subjective day and into the early evening. Depressed levels in clock dynamics by providing the necessary temporal of frq mRNA are maintained until about the time when delay between PER synthesis and its effect on transcrip- FRQ is turned over (Garceau et al., 1997), when positive tion. However, per mRNA levels peak at CT (or ZT) 14 factor(s), possibly the factors encoded by the wc genes (Zeng et al., 1994) and have declined significantly before (Crosthwaite et al., 1997), activate frq transcription to ZT18 when sequestered PER was seen as entering the start the next cycle. It seems likely that all the different nucleus (Curtin et al., 1995); it thus seems likely that phosphorylation states of FRQ are effective in depressing PER has executed its negative function well before ZT18. frq transcript levels; all the different phosphorylation Hence, accumulation of nuclear PER–TIM after ZT14 is forms of FRQ are found inside the nucleus, frq mRNA probably not contributing to delays in RNA suppression. decreases significantly when hypo-phosphorylated FRQ Instead, delays in the Drosophila clock cycle probably accumulates in the nucleus (in the early time points arise earlier in the cycle perhaps as a result of the CT0, 4), and the message remains very low later in the requirement for TIM to stabilize PER (Vosshall et al., circadian cycle when only more highly phosphorylated 1994; Price et al., 1995). PER and TIM would be seen as forms of FRQ are found in the nucleus. Given the nuclear accumulating in the cytoplasm until they form a hetero- localization of FRQ, its apparent action in depressing the dimer, an action that both stabilizes PER and masks level of its own transcript, and the reported weak similarity cytoplasmic localization domains on each protein, thereby to a helix–turn–helix DNA-binding domain (Lewis et al., promoting translocation into the nucleus prior to the 1997), it will be of interest to test FRQ’s ability to bind proteins’ effect on transcription (Saez and Young, 1996). to DNA, especially to frq-specific sequences. In this view, rising per RNA keeps climbing because no It is noticeable that the overt rhythm rescued by construct PER monomers accumulate because TIM levels are not pCL11 (NLS insertion construct) is not as tight and clean yet high enough to exercise a protective function or to as that rescued by a wild-type frq construct (compare race take PER–TIM heterodimers to the nucleus. By CT14, low tubes pKAJ120 and pCL11 at 25°C in Figure 6). There levels of PER/TIM can accumulate, and they immediately are two possible reasons for this: first, this NLS originally enter the nucleus and are effective in turning down resides in a phylogenetically conserved region in the frq transcript levels. Here then, PER instability and a require- amino acid sequence (Merrow and Dunlap, 1994; Lewis ment for heterodimerization to initiate RNA suppression et al., 1997), so it is possible that this deletion may have collaborate to delay PER accumulation in the late sub- a general non-specific effect on FRQ function or stability. jective day and to give a system that accumulates enough 1233 C.Luo, J.J.Loros and J.C.Dunlap Nuclei isolation protein to ensure that once shut off, the genes stay off Nuclei were isolated by a modification of the method described by Baum until mid-morning. and Giles (1985). The mycelial pads (2–3 g, wet weight) were ground In contrast to the Drosophila story, the PER protein in with 2 g of glass beads and 6 ml of buffer A [1 M sorbitol, 7% (w/v) the giant silkmoth A.pernyi is a predominantly cytoplasmic ficoll, 20% (v/v) glycerol, 5 mM magnesium acetate, 5 mM EGTA, 3 mM CaCl , 3 mM dithiothreitol (DTT), 50 mM Tris–HCl, pH 7.5] for protein in brain neurons. Both per mRNA and PER cycle 2 2 min on ice. The crude homogenate was filtered through cheesecloth, in amount in the silkmoth brain, but there is no significant and 2 vols of buffer B [10% (v/v) glycerol, 5 mM magnesium acetate, phase difference between the two. A TIM homolog and a 5 mM EGTA, 25 mM Tris–HCl, pH 7.5] were slowly added to the per antisense RNA were also found in the silkmoth brain. supernatants with stirring. The diluted homogenate was layered over 15 ml of a solution consisting of a 1:1.7 mix of buffers A and B in a The fact that PER is a cytoplasmic protein in A.pernyi 50 ml tube, and centrifuged at 3000 g for 7 min at 4°C in a SW 28 suggests the possibility that it uses a different, possibly rotor to remove cell debris. The resulting supernatants were removed post-transcriptional mechanism to achieve the molecular (this is referred as the total cell lysate), layered onto 5 ml step gradients oscillation, maybe through the action of the antisense per [1 M sucrose, 10% (v/v) glycerol, 5 mM magnesium acetate, 1 mM RNA (Sauman and Reppert, 1996; reviewed by Hall, DTT, 25 mM Tris–HCl, pH 7.5] in a 50 ml tube and centrifuged at 9400 g for 15 min at 4°C in a SW 28 rotor to pellet the nuclei. The 1996). These results demonstrate that although similar resulting nuclear pellets were gently resuspended in ice-cold storage molecular components (PER, TIM and per mRNA) exist buffer [25% (v/v) glycerol, 5 mM magnesium acetate, 3 mM DTT, in insects, the mechanism used to generate rhythms 0.1 mM EDTA, 25 mM Tris–HCl, pH 7.5] and stored at –80°C. All the buffers contain 10 μg/ml leupeptin, 10 μg/ml pepstatin A, 5 mM may be very different. Taken together among different phenanthroline and 1 mM phenylmethylsulfonyl fluoride (PMSF). organisms, the data suggest that clock proteins may use different approaches to achieve a similar result: a negative Protein analysis feedback essential for generating a circadian rhythm in Western blot analysis was performed as previously described (Garceau, the organism. What has been found with Neurospora FRQ 1996; Garceau et al., 1997). X-ray films of Western blots were scanned in terms of cellular biochemistry is consistent with the and densitometry was performed using NIH Image 1.59. FRQ was normalized against time-invariant protein CYS-3. expected nuclear action of clock proteins while again The exclusively cytoplasmic localization of CCG-1 was demonstrated revealing another of the ways through which circadian previously by immunofluorescence (Garceau, 1996; N.Y.Garceau, oscillators operate. K.M.Lindgren, S.J.Free, J.C.Dunlap and J.J.Loros, in preparation). Phosphatase treatment Total cell lysate and nuclei fractions were prepared as described above. Materials and methods A portion of each fraction corresponding to 40 μg of total protein was taken and put into 1 phosphatase treatment buffer (New England Strains, growth conditions and race tube assay 10 Biolabs), 1000 U of λPPase were added, and the reaction was incubated Strains used were 30-7 (bd; frq A) and 93-4 (bd; frq A; his-3) at 30°C for 45 min. Sample buffer was added to the resulting reaction, (Aronson et al., 1994b). All frq constructs were introduced into the frq and the samples were subjected to SDS–PAGE and Western blot analysis. null strain 93-4 and were targeted to the his-3 locus during transformation (Sachs and Ebbole, 1990). Methods for growing rhythmic cultures have been published previously Acknowledgements (Aronson et al., 1994a; Crosthwaite et al., 1995). Briefly, mycelial pads were inoculated into 500 ml of liquid culture and shaken at 110 r.p.m. We thank Dr Zuoyu Xu for his constructive suggestions, stimulating in light at room temperature. The cultures were then transferred into discussions and critical reading of the manuscript, and Dr Mike Young darkness and harvested at the time indicated. The liquid culture contains and members of our laboratories for discussion of the manuscript. The 1 sulfate-free Vogel’s salts, 2% glucose, 50 ng/ml biotin and 0.25 mM anti-CYS-3 and anti-CCG-1 antibodies were generously provided by Dr methionine. DD15 represents cultures held for 15 h in the dark prior George Marzluf and Dr Stephen Free. This work was supported by to harvesting. grants from AFOSR (F49620-94-1-0260 to J.J.L.), the National Science Race tube assay medium contains 1 Vogel’s salts, 0.1% glucose, Foundation (MCB-9307299 to J.J.L.), the National Institute of Health 0.17% arginine and 50 ng/ml biotin. Calculations of period length were (GM 34984 and MH 01186 to J.C.D. and MH 44651 to J.C.D. and done using the Chrono II version 9.3 (Dr Till Roenneberg, Ludwigs- J.J.L.) and the Norris Cotton Cancer Center core grant at Dartmouth Maximillian University, Munich). Medical School. Plasmids and Neurospora transformation References pKAJ120, which contains the entire frq locus, was the parental plasmid for all the constructs. pCL10 was constructed using the Transformer Aronson,B.D., Johnson,K.A., Loros,J.J. and Dunlap,J.C. (1994a) Site-Directed Mutagenesis kit from Clontech Laboratories, Inc. (Palo Negative feedback defining a circadian clock: autoregulation of clock Alto, CA). The SphI–EcoRV fragment from pKAJ120 was cloned into gene frequency. Science, 263, 1533–1656. pUC19, and the resulting plasmid was used as the template for the Aronson,B.D., Johnson,K. and Dunlap,J.C. (1994b) Circadian clock in vitro mutagenesis. The mutagenic primer for making the NLS deletion locus frequency: protein encoded by a single open reading frame was 5-GAGCGTTGCCTCCAACTCTAAGCCATGTACCTTGATCTC- defines period length and temperature compensation. Proc. Natl Acad. 3. The Trans Oligo NdeI–NcoI from the kit was used as the selection Sci. USA, 91, 7683–7686. primer. The SphI–BssHII fragment from the resulting plasmid was then Baum,J. and Giles,N. (1985) Genetic control of chromatin structure 5 inserted into the SphI–BssHII-digested pKAJ120 to create pCL10. The to the qa-x and qa-2 genes of Neurospora. J. Mol. Biol., 182, 79–89. pCL11 construct was made by cloning the PCR-amplified fragment Crosthwaite,S.K., Loros,J.J. and Dunlap,J.C. (1995) Light-induced containing the NLS into the SphI site in pCL10. The primers used resetting of a circadian clock is mediated by a rapid increase in for PCR amplification are 5-ACATGCATGCGCTTACCTAGAAGA- frequency transcript. Cell, 81, 1003–1012. AAG-3 and 5-ACATGCATGCGAAAGTCGCGGAGCGTTG-3. The Crosthwaite,S.K., Dunlap,J.C. and Loros,J.J. (1997) Neurospora wc-1 resulting PCR fragment was digested with SphI and then cloned into the and wc-2: transcription, photoresponses, and the origins of circadian SphI-digested pCL10. This resulted in the insertion of 16 amino acids rhythmicity. Science, 276, 763–769. including the NLS into the N-terminal region of FRQ, 11 amino acids Curtin,K.D., Huang,Z.J. and Rosbash,M. (1995) Temporally regulated after the first methionine. Both constructs were confirmed by DNA nuclear entry of the Drosophila period protein contributes to the sequencing. circadian clock. Neuron, 14, 365–372. Both constructs were transformed into strain 93-4 and targeted at the Dunlap,J.C. (1993) Genetic analysis of circadian clocks. Annu. Rev. his-3 locus as described previously (Sachs and Ebbole, 1990) Physiol., 55, 683–728. 1234 Nuclear localization of FRQ Dunlap,J.C. (1996) Genetic and molecular analysis of circadian rhythms. Saez,L. and Young,M. (1996) Regulation of nuclear entry of Drosophila Annu. Rev. Genet., 30, 579–601. clock proteins Period and Timeless. Neuron, 17, 911–920. Fu,Y.H., Paietta,J., Mannix,D. and Marzluf,G. (1989) cys-3, the positive- Sauman,I. and Reppert,S. (1996) Circadian clock neurons in the silkmoth activating sulfur regulatory gene of Neurospora crassa, encodes a Antheraea pernyi: novel mechanisms of period protein regulation. protein with a putative leucine zipper DNA-binding element. Mol. Neuron, 17, 889–900. Cell. Biol., 9, 1120–1127. Templeton,D., Park,S.H., Lanier,L. and Weinberg,R. (1991) Garcia-Bustos,J., Heitman,J. and Hall,M. (1991) Nuclear protein Nonfunctional mutants of the retinoblastoma protein are characterized localization. Biochim. Biophys. Acta, 1071, 83–101. by defects in phosphorylation, viral oncoprotein association, and Garceau,N.Y. (1996) Identification and regulation of a circadian clock nuclear tethering. Proc. Natl Acad. Sci. USA, 88, 3033–3037. protein, FRQ, and a clock controlled protein, CCG-1, in Neurospora Vosshall,L.B., Price,J.L., Sehgal,A., Saez,L. and Young,M.W. (1994) crassa. Ph.D. Thesis. Block in nuclear localization of period protein by a second clock Garceau,N.Y., Liu,Y., Loros,J. and Dunlap,J.C. (1997) Alternative mutation, timeless. Science, 263, 1606–1609. initiation of transcription and time-specific phosphorylation yield Wood,K. (1995) Marker proteins for gene expression. Curr. Opin. multiple forms of essential clock protein FREQUENCY. Cell, 89, Biotechnol., 6, 50–58. 469–476. Wuarin,J., Falvey,E., Lavery,D., Talbot,D., Schmidt,E., Ossipow,V., Gekakis,N., Saez,L., Delahaye-Brown,A.-M., Myers,M.P., Sehgal,A., Fonjallaz,P. and Schibler,U. (1992) The role of the transcriptional Young,M.W. and Weitz,C.J. (1995) Isolation of timeless by PER activator protein DBP in circadian liver gene expression. J. Cell Sci., protein interactions: defective interaction between timeless protein Suppl., 16, 123–127. and long-period mutant PER . Science, 270, 811–815. Zeng,H., Qian,Z., Myers,M.P. and Rosbash,M. (1996) A light- Hall,J. (1996) Are cycling genes products as internal Zeitgebers no entrainment mechanism for the Drosophila circadian clock. Nature, longer the Zeitgeist of chronobiology? Neuron, 17, 799–802. 380, 129–135. Hunter-Ensor,M., Ousley,A. and Sehgal,A. (1996) Regulation of the Drosophila protein timeless suggests a mechanism for resetting the Received September 15, 1997; revised and accepted January 7, 1998 circadian clock by light. Cell, 84, 677–685. Jans,D.A. and Hubner,S. (1996) Regulation of protein transport to the nucleus: central role of phosphorylation. Physiol. Rev., 76, 652–675. Jans,D.A. and Jans,J. (1994) Negative charge at the casein kinase II site flanking the nuclear localization signal of SV 40 large T-antigen is mechanistically important for enhanced nuclear import. Oncogene, 9, 2961–2968. Kanaan,M. and Marzluf,G. (1993) The positive-activating sulfur regulatory protein CYS3 of Neurospora crassa: nuclear localization, autogenous control, and regions required for transcriptional activation. Mol. Gen. Genet., 239, 334–344. Lee,C., Parikh,V., Itsukaichi,T., Bae,K. and Edery,I. (1996) Resetting of the Drosophila clock by photic regulation of PER and PER–TIM complex. Science, 271, 1740–1744. Lewis,M.T., Morgan,L.W. and Feldman,J.F. (1997) Analysis of frequency (frq) clock gene homologs: evidence for a helix–turn–helix transcription factor. Mol. Gen. Genet., 253, 401–414. Liu,X., Zwiebel,L., Hinton,D., Benzer,S., Hall,J. and Rosbash,M. (1992) The per gene encodes a predominantly nuclear protein in adult Drosophila. J. Neurosci., 12, 2735–2744. Liu,Y., Garceau,N.Y., Loros,J.J. and Dunlap,J.C. (1997) Thermally regulated translational control of FRQ mediates aspects of temperature responses in the Neurospora circadian clock. Cell, 89, 477–486. Loros,J.J. (1995) The molecular basis of the Neurospora clock. Semin. Neurosci., 7, 3–13. Loros,J.J., Denome,S.A. and Dunlap,J.C. (1989) Molecular cloning of genes under the control of the circadian clock in Neurospora. Science, 243, 385–388. Loros,J.J., Free,S.J., Lindgren,K.M., Garceau,N. and Dunlap,J.C. (1995) The ccg-1 gene of Neurospora displays multiple levels of regulation including the clock, development, light and heat shock. Fungal Genet. Newslett., 42A, 22. McNally,M.T. and Free,S.J. (1988) Isolation and characterization of a Neurospora glucose-repressible gene. Curr. Genet., 14, 545–551. Merrow,M.W. and Dunlap,J.C. (1994) Intergeneric complementation of a circadian rhythmicity defect: phylogenetic conservation of structure and function of the clock gene frequency. EMBO J., 13, 2257–2266. Merrow,M.W., Garceau,N.Y. and Dunlap,J.C. (1997) Dissection of a circadian oscillation into discrete domains. Proc. Natl Acad. Sci. USA, 94, 3877–3882. Mittnacht,S. and Weinberg,R. (1991) G1/S phosphorylation of the retinoblastoma protein is associated with an altered affinity for the nuclear compartment. Cell, 65, 381–393. Myers,M.P., Wager-Smith,K., Rothenfluh-Hilfiker,A. and Young,M.W. (1996) Light-induced degradation of TIMELESS and entrainment of the Drosophila circadian clock. Science, 271, 1736–1740. Paietta,J. (1992) Production of the CYS3 regulator, a bZIP DNA-binding protein, is sufficient to induce sulfur gene expression in Neurospora crassa. Mol. Cell. Biol., 12, 1568–1577. Price,J.L., Dembinska,M.E, Young,M.W. and Rosbash,M. (1995) Suppression of PERIOD protein abundance and circadian cycling by the Drosophila clock mutation timeless. EMBO J., 14, 4044–4047. Sachs,M.S. and Ebbole,D. (1990) The use of lacZ gene fusions in Neurospora crassa. Fungal Genet. Newslett., 37, 35–37. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The EMBO Journal Springer Journals

Nuclear localization is required for function of the essential clock protein FRQ

Download PDF
8 pages

Loading next page...
 
/lp/springer-journals/nuclear-localization-is-required-for-function-of-the-essential-clock-nJOb1WuBan

References (47)

Publisher
Springer Journals
Copyright
Copyright © European Molecular Biology Organization 1998
ISSN
0261-4189
eISSN
1460-2075
DOI
10.1093/emboj/17.5.1228
Publisher site
See Article on Publisher Site

Abstract

The EMBO Journal Vol.17 No.5 pp.1228–1235, 1998 Nuclear localization is required for function of the essential clock protein FRQ of frq mRNA from an inducible promoter rapidly depresses Chenghua Luo, Jennifer J.Loros and the level of the endogenous frq message (Aronson et al., Jay C.Dunlap 1994a; Merrow et al., 1997). These data place frq mRNA Department of Biochemistry, Dartmouth Medical School, Hanover, and FRQ protein in a negative feedback loop defining the NH 03755, USA circadian oscillator. frq mRNA accumulates very quickly Corresponding author (within 2 min) in response to short light pulses, and the e-mail: [email protected] threshold and kinetics of the response are similar to those of light-induced resetting (Crosthwaite et al., 1995). Two The frequency (frq) gene in Neurospora encodes central FRQ polypeptides arise from the frq transcript, resulting components of a circadian oscillator, a negative feed- from alternative initiation of FRQ translation from two back loop involving frq mRNA and two forms of FRQ in-frame initiation codons: codon #1 and codon #100 protein. Here we report that FRQ is a nuclear protein (Garceau et al., 1997). Both FRQ forms are immediately and nuclear localization is essential for its function. and progressively phosphorylated after being synthesized Deletion of the nuclear localization signal (NLS) (Garceau et al., 1997). Temperature regulates the ratio of renders FRQ unable to enter into the nucleus and the two different FRQ forms by favoring different transla- abolishes overt circadian rhythmicity, while reinsertion tion initiation sites at different temperatures and thus sets of the NLS at a novel site near the N-terminus of FRQ the physiological temperature limits for rhythmicity (Liu restores its function. Each form of FRQ enters the et al., 1997). nucleus soon after its synthesis in the early subjective The mechanism by which FRQ acts to depress the level day; there is no evidence for regulated sequestration of its own transcript or more generally contributes to in the cytoplasm prior to nuclear entry. The kinetics clock function is unknown. This negative action may be of the nuclear entry are consistent with previous achieved transcriptionally, post-transcriptionally or both. data showing rapid depression of frq transcript levels In the Drosophila system, a complex composed of the following the synthesis of FRQ, and suggest that early clock proteins PER and TIM is thought to regulate per in each circadian cycle, when FRQ is synthesized, it and tim transcription through a negative feedback loop enters the nucleus and depresses the level of its own (Gekakis et al., 1995; Myers et al., 1996; Zeng et al., transcript. 1996) to generate the circadian oscillation. Both PER and Keywords: circadian rhythm/frequency/Neurospora/ TIM are predominantly nuclear proteins in wild-type flies nuclear localization (Liu et al., 1992), and nuclear translocation is regulated by the dimerization between the two proteins (Saez and Young, 1995). Studies done in Drosophila cell cultures have shown that each protein contains sequences that Introduction confer cytoplasmic localization, and dimerization between them suppresses the cytoplasmic localization function Circadian rhythmicity is a general aspect of regulation (Saez and Young, 1995). Thus, both proteins accumulate existing in a wide variety of organisms ranging from in the cytoplasm following their translation, then enter prokaryotes to eukaryotes (Dunlap, 1993, 1996). The into the nucleus in a concerted manner after this delay to cellular machinery that generates this capacity is known execute their effect on transcription (Curtin et al., 1995; as the biological clock. Such clocks are intrinsic to the Saez and Young, 1995). On the other hand, in the silkmoth cell, are endogenous and self-sustaining under constant Antheraea pernyi, the PER protein is nuclear in eye conditions, and respond to environmental cues such as photoreceptor cells but is predominantly cytoplasmic in light and temperature (Crosthwaite et al., 1995; Hunter- the brain cells thought to drive behavioral rhythm (Sauman Ensor et al., 1996; Lee et al., 1996; Zeng et al., 1996; and Reppert, 1995). per mRNA and PER protein oscillate Liu et al., 1997). in both cell types, and the predominantly cytoplasmic The products of the frequency (frq) gene are central PER in the brain cells has suggested that the silkmoth components of the Neurospora oscillator (Dunlap, 1993, may use a different mechanism to construct the molecular 1996; Loros, 1995), and aspects of their regulation have oscillator. In Neurospora, FRQ has several characteristics pointed to roles for transcription and translation in the consistent with an involvement in transcription, including a clock. frq mRNA is expressed rhythmically with a period putative helix–turn–helix DNA-binding domain, a nuclear reflecting that of the overt rhythm. Constantly elevated localization signal (NLS) and highly charged regions expression of frq message from a heterologous inducible promoter eliminates overt rhythmicity in a wild-type (Aronson et al., 1994b; Merrow and Dunlap, 1994; Lewis background, and is unable to support rhythmicity in a et al., 1997). Both frq mRNA and FRQ oscillate, the peak loss-of-function strain (frq ). Step reduction of frq mRNA of FRQ lags the peak of frq mRNA by 4–6 h, and frq resets the clock to a predicable phase. Elevated expression mRNA levels begin to fall before the level of FRQ reaches 1228 © Oxford University Press Nuclear localization of FRQ It is known that two different forms of FRQ arise from two in-frame AUGs in the frq transcript generating ~135 and ~145 kDa proteins. Each form of FRQ is phosphoryl- ated progressively after being synthesized (Garceau et al., 1997). Biochemically isolating nuclei and detecting where FRQ is has the advantage of providing information on how these different FRQ forms behave. Figure 1 shows the results of Western analysis of FRQ, CYS-3 and CCG-1 from a mid-subjective morning sample (DD15, see Materials and methods) in the total cell lysate and nuclear fractions. CYS-3 was present in both lysate and Fig. 1. Both forms of FRQ are found inside nuclei in Neurospora. nuclear fractions, and its signal was enriched ~3- to 4-fold Western blots show FRQ and two control proteins (CYS-3 and in the nuclear fraction compared with that in the lysate CCG-1) in total cell lysate and nuclear fractions. The two arrows point fraction. This is consistent with incomplete nuclear localiz- to the two different forms of FRQ resulting from two in-frame translational initiation sites in the same message. Both FRQ forms are ation combined with breakage of nuclei upon isolation. detected in the nuclear fraction. There is no CCG-1 signal in the CCG-1 was present exclusively in the lysate fraction. The nuclear fraction, indicating that the isolated nuclei are free from results from the two protein markers confirm that the cytosolic contamination. isolated nuclear fraction has no cytoplasmic contamination. FRQ, like CYS-3, was found in the nuclear fraction (FRQ its peak (Garceau et al., 1997), suggesting that FRQ might nuclear panel), suggesting it is a nuclear protein. Its signal execute its role quickly, probably very soon after being is enriched in the nuclear fraction, and the fold enrichment synthesized. is similar to that of the nuclear marker protein CYS-3. In If FRQ functions at a transcriptional level in the addition, both large and small FRQ forms are found in feedback inhibition loop, it must be in the nucleus at least the nuclear fraction (as denoted by the two arrows in at certain times during a circadian cycle. To examine Figure 1). this, we adapted a cellular fractionation technique and determined the intracellular compartmentation of FRQ as Deletion of the NLS from FRQ eliminates nuclear a function of time of day. Our data demonstrate that FRQ localization and abolishes overt circadian is indeed a nuclear protein, and its nuclear location is rhythmicity essential for its function in the circadian clock. A single To determine whether the nuclear localization of FRQ is NLS located at amino acids 195–200 is necessary and required for its function, we identified and deleted the sufficient for FRQ nuclear entry. This NLS is still func- NLS of FRQ. Two putative SV40 large T antigen-type tional when moved to a different site near the N-terminus NLSs located at amino acids 194–199 and 564–568 are of FRQ; it allows nuclear localization of the large form predicted from computer analysis of FRQ’s amino acid of FRQ alone and is sufficient to restore clock function. sequence (Aronson et al., 1994b; Merrow and Dunlap, Both forms of FRQ move into the nucleus very soon after 1994; Lewis et al., 1997). The first one is phylogenetically being synthesized and well before levels of FRQ peak or more conserved among frq homologs (Lewis et al., 1997). phosphorylation of FRQ has progressed very far. The Constructs bearing deletions of each NLS separately and absence of a noticeable delay prior to nuclear entry a double deletion of both were made (Figure 2A), and suggests that delay prior to nuclear localization contributes targeted by transformation to the his-3 locus in the frq little to the long-term control of the circadian cycle, nor null strain (strain #93-4) (Aronson et al., 1994b). A is it required to generate the long time constant of the putative NLS (RKKRK) starting at amino acid 564 was circadian oscillator. deleted in construct pCL7; the frq transformants bearing this construct still had apparently normal nuclear localiz- Results ation of FRQ and retained the overt circadian conidial Both forms of FRQ are found inside the nucleus banding rhythm (data not shown). In construct pCL10, a In order to study the subcellular localization of FRQ, we putative NLS (PRRKKR) beginning at amino acid 194 was have utilized a biochemical method to isolate nuclei from deleted. Transformants bearing this construct produced a Neurospora, resolved the nuclear fractions by SDS–PAGE, normal amount of FRQ, but almost no FRQ was detected and detected FRQ by Western blot analysis. Two proteins in the nuclear fraction (Figure 2B and C), suggesting that were used as control markers in this method: CYS-3 and this region is the important NLS directing FRQ into the CCG-1. CYS-3 is used as the nuclear protein marker. It nucleus. To assess the functional significance of this is a transcription factor involved in regulating structural mutation on the action of FRQ in the oscillator, the proteins in the sulfur metabolism pathway in Neurospora overt circadian rhythm in these transformant strains was (Fu et al., 1989; Paietta, 1992; Kanaan and Marzluf, examined in race tubes (Figure 3A). If the nuclear localiz- 1993). CCG-1 is the product of clock-controlled gene-1 ation of FRQ is important for its function, we would (McNally and Free, 1988; Loros et al., 1989). It is an expect that this construct could not rescue the loss of abundant and exclusively cytoplasmic protein (Loros et al., circadian rhythmicity in the frq (null) strain. As expected, 1995; Garceau, 1996), and it is used as a cytoplasmic this NLS-deleted frq failed to support the overt conidiation protein marker in this study. The two protein markers rhythm (compare pCL10 with pKAJ120 race tubes). Aside were used to show the quality of the isolated nuclei, from the loss of rhythmicity, the growth rate and general specifically to verify the absence of cytoplasmic contamin- morphology were not altered in the pCL10 transformants, ation in the nuclear preparation (Figure 1). suggesting that no other major defects were associated 1229 C.Luo, J.J.Loros and J.C.Dunlap Fig. 3. Deletion of the NLS in FRQ abolishes overt circadian rhythmicity. (A) Race tube analysis showing the banding rhythm in different strains at 25°C: wild-type (frq ), a transformant bearing an intact frq locus (pKAJ 120), a transformant bearing the NLS deletion construct (pCL10) and the frq strain (null). No conidial banding rhythm is found in the pCL10 transformant. Cultures were inoculated at the left ends of race tubes, incubated in constant light for 1 day and transferred into darkness at 25°C. The growth front (vertical lines) was marked at 24 h intervals thereafter. (B) Western blot of FRQ from ΔNLS strain pCL10 over time points spanning a circadian cycle. Both the amount and the phosphorylation states of FRQ remain unchanged Fig. 2. An NLS located between amino acids 194 and 199 is required over time, indicating an absence of rhythmicity. for FRQ nuclear localization. (A) A schematic diagram of the ΔNLS construct. The open box represents the frq ORF, and arrows mark the positions of the methionines initiating the large and small FRQ into the N-terminal region of ΔNLS construct pCL10 translation products. The black box marked NLS corresponds to amino (Figure 4A). A putative casein kinase II phosphorylation acids 194–199. (B) Western blot of FRQ in total lysate and nuclear site (T LRD) was included in the 16 amino acid fractions of three different strains: bdA (wild-type, frq ), pCL10 10 10 transformant (ΔNLS construct) and frq (frq null). frq served as a sequence, since it has been reported that phosphorylation negative control. In the nuclear fraction in the pCL10 transformant, the at a CKII site close to an NLS can enhance the rate of amount of FRQ is reduced significantly. (C) Densitometry of the nuclear translocation (Jans and Jans, 1994). This construct amount of FRQ in different fractions in each strain shown in (B). (pCL11) was transformed into the frq null strain (strain Levels of FRQ are normalized against the FRQ level in the lysate fraction in the bdA (frq ) strain at DD15. #93-4) at the his-3 locus, and the transformants were examined for nuclear localization of FRQ and for overt rhythmicity. Theoretically, NLSs can be located anywhere with this mutation. In confirmation of the race tube data, in a protein’s primary structure, provided that after protein neither frq mRNA nor FRQ showed any rhythmicity in folding, the NLS is on the surface of the protein (reviewed pCL10 transformants, as shown in Figure 3B. Interestingly, by Garcia-Bustos et al., 1991; Jans and Hubner, 1996). both the amount and the phosphorylation states of FRQ The insertion was made into an SphI site near the 5 end remained relatively unchanged across different time points of frq’s open reading frame (ORF), just 11 amino acids spanning a circadian cycle in the pCL10 transformants. after AUG#1 (Garceau et al., 1997). We chose to insert The observation that the phosphorylation states remain the putative NLS near to the N-terminus of FRQ since the same over time is similar to the behavior of wild- structural predictions place this on the surface of the type FRQ under constant light conditions (N.Y.Garceau, protein and this insertion also provides a convenient J.J.Loros and J.C.Dunlap, in preparation). In both cases, internal control in the subsequent analysis. Specifically, there is no circadian oscillation in the organism. This insertion of the NLS at residue 11 should support nuclear suggests that the phosphorylation of FRQ is coordinated localization of the large form of FRQ but not the small with the circadian cycle. Together, these data show that it form which arises from alternative translation initiation at is necessary for FRQ to enter into the nucleus to generate amino acid 100. This insertion is not likely to destroy any both molecular and overt rhythmicity. essential functional domain(s) in FRQ since small FRQ alone has been shown to be sufficient for establishing Insertion of the NLS near to the N-terminal region rhythmicity (Liu et al., 1997). of FRQ restores its function in the clock Reinsertion of the NLS restored nuclear localization of To test if the nuclear localization of FRQ could be restored the large form of FRQ. Figure 4B shows the Western blot and be sufficient for its function in the clock, the NLS and results of total cell lysate and nuclear fractions of strains its surrounding region (amino acids 193–208, including the bdA (frq ), pCL11 (N-terminal NLS insertion construct) six amino acids extending from 194 to 199) was inserted and pCL10 (NLS deletion construct) at DD15. All three 1230 Nuclear localization of FRQ Fig. 5. Insertion of the NLS into the N-terminal region of ΔNLS FRQ pCL10 restores a functional clock. Race tube analysis showing the overt conidial banding rhythm of transformants bearing a wild-type frq locus (pKAJ120), the NLS deletion (pCL10) and the NLS insertion (pCL11) constructs. Race tubes were inoculated and incubated in constant light for a day before being transferred into constant darkness at different temperatures. The rhythm was restored in the pCL11 transformant; rhythmicity was apparent at 25°C and at a higher temperature (27°C), but the rhythm was lost at a lower temperature (22°C). Table I. Period length in different strains Strains Period (h  SD) bdA/25°C 21.2  1.3 pKAJ120/25°C 23.6  1.1 pCL10/25°C ND pCL11/25°C 24.8  1.5 pCL11/27°C 24.6  0.7 pCL11/22°C ND frq /25°C ND Fig. 4. Insertion of the NLS at a different position in ΔNLS FRQ directs the FRQ protein to enter the nucleus. (A) Schematic diagram of the constructs (see Figure 2 and text for details). (B) Western blot Figure 4C, both FRQ forms were in wild-type and pCL11 of FRQ in the total cell lysate and nuclear fractions of different lysate fractions, but only the large form of FRQ was found strains: bdA (wild-type, frq ), pCL10 (ΔNLS) transformant and pCL11 in the pCL11 nuclear fraction. The faint nuclear-enriched (NLS insertion) transformant. Both forms of FRQ shown by the two arrowheads were found in the total lysate in all three strains. In the band sandwiched between the two FRQ bands is non- nuclear fractions, both FRQ forms are detected in the wild-type strain specific, since it is recognized by the antiserum in the but only the large FRQ form is detected in the pCL11 transformant, frq (null) strain also. and no FRQ is detected in the pCL10 transformant. (C) Western blot Reinsertion of the NLS also restores the function of the of total cell lysate fractions and nuclear fractions together with controls, subjected to λ phosphatase treatment. JCC101 bears a large form of FRQ. Transformation of pCL11 into a frq disruption at the third AUG and YL15 bears a disruption at the first null strain rescued the clock null phenotype (Figure 5; AUG; they were used as controls for strains making large and small Table I), suggesting that once FRQ gets into nuclei it can FRQ, respectively. Both FRQ forms were detected in the wild-type function normally. It has been shown that the small FRQ and pCL11 transformants (wt, pCL11 lysate), while only large FRQ is form is important for the clock to function at the lower found in the nuclear fraction from the pCL11 transformant. end of the physiological temperature range (Liu et al., 1997). Thus, a reasonable prediction is that pCL11 should strains synthesized both FRQ forms (Figure 4B lysate support rhythmicity at higher temperatures but not at panel, as pointed out by the arrowheads). Both FRQ forms lower temperatures since only the large FRQ enters the were inside the nucleus in the bdA (frq ) strain. FRQ was nucleus in this strain. This prediction was proved to be found inside the nucleus of strains bearing construct correct: at 27°C the clock was functional in pCL11 pCL11 (pCL11 nuclear panel), confirming that the NLS transformants while at 22°C the clock no longer operated can direct FRQ into the nucleus despite the altered position. (Figure 5, pCL11 race tubes at 27°C versus 22°C). Data In addition, only the large FRQ form was found in the shown in Table I confirm that the period lengths of the nucleus; this is in agreement with our prediction, since rescued strains were comparable with those rescued by only the large FRQ has the reinserted NLS. The progressive wild-type frq, again confirming normal function of large phosphorylation of FRQ described above results in a series FRQ with a displaced NLS. of FRQ bands that tend to obscure the appearance of protein bands on the gels. In order to generate a visually The amount of FRQ in nuclei cycles with a peak clearer result, total cell lysate and the nuclear fraction amount at CT4 were treated with λ phosphatase (Garceau et al., 1997) to In order to address the behavior of FRQ in terms of convert all different FRQ forms (arising from different cellular localization during a whole circadian cycle, a levels of phosphorylation) into just the two bands corres- quantitative analysis of FRQ in the nuclear extract was ponding to large and small FRQ. The result is shown in performed. Liquid cultures were grown in constant light Figure 4C. JCC101 and YL15 are strains that make only at room temperature before being transferred into constant large and small FRQ, respectively, since they bear a darkness at 25°C. The cultures were harvested at 4 h genetically engineered knock-out of the third AUG intervals after the light to dark transfer. Nuclei were (JCC101) and first AUG (YL15) (Liu et al., 1997). In isolated from samples at different circadian time points 1231 C.Luo, J.J.Loros and J.C.Dunlap of the circadian cycle (CT0), significant amounts of FRQ were detected in the nuclear fraction (nuclear panel CT0). This suggests that once FRQ is synthesized, it rapidly enters the nucleus with essentially no delay. Consistent with this, the maximal amount of FRQ inside nuclei is at CT4, earlier than that in the total cell lysate, and, at CT0, when there is little FRQ detected in the total lysate, FRQ accumulates to a significant level inside the nucleus. Although the nuclear FRQ distribution profile is phase- advanced with respect to the total cell lysate, it is very similar to that of frq mRNA, which normally peaks at CT4 (Crosthwaite et al., 1995); this is apparent comparing Figure 6A, nuclear panel, with Figure 6B (see also Garceau et al., 1997), although the frq RNA level at CT0 was abberantly high in this series. Since it has been reported that different phosphorylation forms of the same protein may be extracted differently during nuclei isolation (Mittnacht and Weinberg, 1991; Templeton et al., 1991), we examined the effect of salt concentration on the apparent cellular localization of FRQ. Nuclei were isolated under different salt concentrations (0 and 200 mM NaCl), and the nuclear FRQ profiles were the same (data not shown), indicating that different FRQ phosphorylation forms are not artificially producing an apparent change in localization. Figure 6C plots the amount of FRQ normal- ized against the time-invariant protein CYS-3 and the amount of frq mRNA over a circadian cycle. It is clear that the amount of FRQ oscillates in a circadian manner in both lysate and nuclear fractions, although the phase Fig. 6. Nuclear cycling of the FRQ protein. (A) Western blots of FRQ, of each cycle is different: the peak amount of FRQ inside CYS-3 and CCG-1 during a circadian cycle. In the total cell lysate the nucleus occurs at CT4 and has a similar phase to that (left panel) and nuclear fractions (right panel), CYS-3 protein levels of the frq mRNA cycle, peaking ~4 h before the total are constant throughout the circadian cycle. CCG-1 is a clock- controlled protein, peaking in the subjective morning (CT0). No amount of FRQ peaks in the cell (CT8). CCG-1 is detected in the nuclear fraction at any time. FRQ cycles in both fractions, but the phases are different. The amount of FRQ in the nuclear fraction peaks at DD15 (CT4) while the amount of FRQ in the Discussion total cell lysate fraction peaks at DD18 (CT8). These data are representative of five experiments. (B) Northern blot of frq mRNA in It has been well established that frq mRNA and FRQ a wild-type (frq ) strain over a circadian cycle. (C) Densitometric protein, the central components of a negative feedback analysis plotting the amount of FRQ normalized against the time- loop, are defining elements of the Neurospora circadian invariant protein CYS-3 in total cell lysate (u), nuclear fraction (j) clock. When FRQ is induced from a heterologous inducible and frq mRNA (s) over a circadian cycle. The nuclear cycle is promoter, the level of endogenous frq mRNA is depressed phase-advanced with respect to that of the total cycle by ~4 h, and it has a similar phase to that of the mRNA cycle. The data are plotted as (Aronson et al., 1994a; Merrow et al., 1997). Nevertheless, mean  2 SEM (n5). many details of how this negative feedback loop works are poorly understood. Data presented here clearly demon- spanning a whole cycle, and the total cell lysate and strate that both the large and small forms of FRQ, arising nuclear fractions were subjected to Western blot analysis. from different translation initiation sites in the same Western blot results for FRQ, CYS-3 and CCG-1 are message, are nuclear proteins. This result is consistent shown in Figure 6A. CYS-3 is a constitutively expressed with FRQ acting transcriptionally to down-regulate its protein; the amount of CYS-3 at different circadian times own message. The same NLS exists for both FRQ forms remained unchanged in both lysate and nuclear fractions and, in the context of the rest of FRQ, this signal is (Figure 6A, CYS-3 lysate and nuclear panel). For the necessary and sufficient to direct the FRQ protein into the clock-controlled protein CCG-1, the amount of protein nuclei. Previous results showed that rhythmic expression cycled over the circadian cycle with a peak at subjective of frq, rather than simple constitutive expression, was morning, confirming that the cultures are normally rhyth- essential for a functional clock (Aronson et al., 1994a). mic (Figure 6A, CCG-1 lysate panel). There was no We have demonstrated further here that in order to generate CCG-1 signal detected in the nuclear fractions at any time the rhythmic expression of frq mRNA and protein, FRQ point (Figure 6A, CCG-1 nuclear panel). In the total lysate must be transferred into the nucleus. When the NLS is fractions, FRQ showed the characteristic mobility shifts deleted, FRQ does not enter the nucleus and both molecular resulting from progressive phosphorylation over the circa- and overt rhythms are abolished; when the NLS is put dian cycle, and the amount of the protein peaked at CT8 back at a different position in the protein sequence, the as previously reported (Garceau et al., 1997; Figure 6A, rhythm of the clock is restored. FRQ lysate panel). In the nuclear fractions, FRQ was At all the time points tested during a circadian cycle, present at all time points and, especially at early stages some FRQ is always found inside the nucleus. Particularly 1232 Nuclear localization of FRQ at early stages of the circadian cycle such as CT0 and Additionally, we have shown previously (Liu et al., 1997) CT4, a significant amount of FRQ has accumulated in the that strains expressing only one of two forms of FRQ nucleus, suggesting that FRQ enters the nucleus very exhibit rhythms of reduced quality such as that seen here. quickly after being synthesized. If FRQ executes its Although both FRQ forms are expressed here, only large negative effect on its own transcription while inside the FRQ enters the nucleus and hence it may be that only it nucleus, then it could repress transcription and reduce is active. transcript levels within just a few hours, consistent with Among the systems used for clock research, Neurospora previous data (Merrow et al., 1997). In this study, we and Drosophila are the best characterized. A common showed that the amount of FRQ not only cycles in the theme arising from years of molecular research is that total cell lysate but also cycles within the nucleus. The both oscillators are transcription–translation-based nega- nuclear cycling is phase-advanced with respect to that of tive feedback loops, i.e. the clock proteins (FRQ in the total cell lysate, but has a similar phase to that of frq Neurospora, PER and TIM in Drosophila) enter the mRNA. This is consistent with a model in which once nucleus to down-regulate their own transcript levels. FRQ is made, it enters the nucleus and represses its own However, the mechanisms through which these proteins transcription, resulting in a significant decline in frq achieve negative feedback have seemed somewhat differ- mRNA before maximum FRQ levels are reached. This ent: in Neurospora there appears to be little temporal lag early negative effect executed by FRQ would help to between the appearance of frq RNA, its translation and generate the phase lag between total FRQ and frq message the movement of FRQ into the nucleus in that total frq previously observed (Garceau et al., 1997). RNA and nuclear FRQ protein both peak around CT4. These data are also consistent with the rapid kinetics FRQ appears to execute its function in reducing frq mRNA of FRQ negative feedback as reported using a frq strain, levels rather quickly, beginning before peak nuclear FRQ that is the negative feedback itself occurs very rapidly levels are reached, with the result that frq mRNA levels (Merrow et al., 1997). In a frq background, frq mRNA decline significantly before total cellular FRQ levels peak fluctuates around a high level and (because of a frameshift at around CT8. Thus the lag between total frq RNA and mutation) no functional FRQ is made (Aronson et al., protein levels simply reflects the well-known lag that 1994a). If functional FRQ is induced from a heterologous accompanies the translation of proteins with half-lives of promoter, a substantial decrease in frq mRNA is observed even a few hours (Wuarin et al., 1992; Wood, 1995). within 3 h, and the repression is essentially complete by Later in the cycle, some FRQ remains in the nucleus to 6 h (Merrow et al., 1997). These rapid repression kinetics keep frq transcript levels reduced till the proteins turn are seen following induction of FRQ at or below normal over completely. In contrast to Neurospora,in Drosophila physiological levels (Merrow et al., 1997), so it is clear there is reported to be a controlled lag between the that amounts of FRQ lower than the peak are quite synthesis/accumulation of PER in the cytoplasm and its effective in establishing and maintaining depressed levels translocation to the nucleus (Curtin et al., 1995). In this of frq mRNA. We have confirmed here that at least modest view, the regulation of PER nuclear entry, as distinct from levels of FRQ remain in the nucleus throughout the PER accumulation, was seen as playing an essential role subjective day and into the early evening. Depressed levels in clock dynamics by providing the necessary temporal of frq mRNA are maintained until about the time when delay between PER synthesis and its effect on transcrip- FRQ is turned over (Garceau et al., 1997), when positive tion. However, per mRNA levels peak at CT (or ZT) 14 factor(s), possibly the factors encoded by the wc genes (Zeng et al., 1994) and have declined significantly before (Crosthwaite et al., 1997), activate frq transcription to ZT18 when sequestered PER was seen as entering the start the next cycle. It seems likely that all the different nucleus (Curtin et al., 1995); it thus seems likely that phosphorylation states of FRQ are effective in depressing PER has executed its negative function well before ZT18. frq transcript levels; all the different phosphorylation Hence, accumulation of nuclear PER–TIM after ZT14 is forms of FRQ are found inside the nucleus, frq mRNA probably not contributing to delays in RNA suppression. decreases significantly when hypo-phosphorylated FRQ Instead, delays in the Drosophila clock cycle probably accumulates in the nucleus (in the early time points arise earlier in the cycle perhaps as a result of the CT0, 4), and the message remains very low later in the requirement for TIM to stabilize PER (Vosshall et al., circadian cycle when only more highly phosphorylated 1994; Price et al., 1995). PER and TIM would be seen as forms of FRQ are found in the nucleus. Given the nuclear accumulating in the cytoplasm until they form a hetero- localization of FRQ, its apparent action in depressing the dimer, an action that both stabilizes PER and masks level of its own transcript, and the reported weak similarity cytoplasmic localization domains on each protein, thereby to a helix–turn–helix DNA-binding domain (Lewis et al., promoting translocation into the nucleus prior to the 1997), it will be of interest to test FRQ’s ability to bind proteins’ effect on transcription (Saez and Young, 1996). to DNA, especially to frq-specific sequences. In this view, rising per RNA keeps climbing because no It is noticeable that the overt rhythm rescued by construct PER monomers accumulate because TIM levels are not pCL11 (NLS insertion construct) is not as tight and clean yet high enough to exercise a protective function or to as that rescued by a wild-type frq construct (compare race take PER–TIM heterodimers to the nucleus. By CT14, low tubes pKAJ120 and pCL11 at 25°C in Figure 6). There levels of PER/TIM can accumulate, and they immediately are two possible reasons for this: first, this NLS originally enter the nucleus and are effective in turning down resides in a phylogenetically conserved region in the frq transcript levels. Here then, PER instability and a require- amino acid sequence (Merrow and Dunlap, 1994; Lewis ment for heterodimerization to initiate RNA suppression et al., 1997), so it is possible that this deletion may have collaborate to delay PER accumulation in the late sub- a general non-specific effect on FRQ function or stability. jective day and to give a system that accumulates enough 1233 C.Luo, J.J.Loros and J.C.Dunlap Nuclei isolation protein to ensure that once shut off, the genes stay off Nuclei were isolated by a modification of the method described by Baum until mid-morning. and Giles (1985). The mycelial pads (2–3 g, wet weight) were ground In contrast to the Drosophila story, the PER protein in with 2 g of glass beads and 6 ml of buffer A [1 M sorbitol, 7% (w/v) the giant silkmoth A.pernyi is a predominantly cytoplasmic ficoll, 20% (v/v) glycerol, 5 mM magnesium acetate, 5 mM EGTA, 3 mM CaCl , 3 mM dithiothreitol (DTT), 50 mM Tris–HCl, pH 7.5] for protein in brain neurons. Both per mRNA and PER cycle 2 2 min on ice. The crude homogenate was filtered through cheesecloth, in amount in the silkmoth brain, but there is no significant and 2 vols of buffer B [10% (v/v) glycerol, 5 mM magnesium acetate, phase difference between the two. A TIM homolog and a 5 mM EGTA, 25 mM Tris–HCl, pH 7.5] were slowly added to the per antisense RNA were also found in the silkmoth brain. supernatants with stirring. The diluted homogenate was layered over 15 ml of a solution consisting of a 1:1.7 mix of buffers A and B in a The fact that PER is a cytoplasmic protein in A.pernyi 50 ml tube, and centrifuged at 3000 g for 7 min at 4°C in a SW 28 suggests the possibility that it uses a different, possibly rotor to remove cell debris. The resulting supernatants were removed post-transcriptional mechanism to achieve the molecular (this is referred as the total cell lysate), layered onto 5 ml step gradients oscillation, maybe through the action of the antisense per [1 M sucrose, 10% (v/v) glycerol, 5 mM magnesium acetate, 1 mM RNA (Sauman and Reppert, 1996; reviewed by Hall, DTT, 25 mM Tris–HCl, pH 7.5] in a 50 ml tube and centrifuged at 9400 g for 15 min at 4°C in a SW 28 rotor to pellet the nuclei. The 1996). These results demonstrate that although similar resulting nuclear pellets were gently resuspended in ice-cold storage molecular components (PER, TIM and per mRNA) exist buffer [25% (v/v) glycerol, 5 mM magnesium acetate, 3 mM DTT, in insects, the mechanism used to generate rhythms 0.1 mM EDTA, 25 mM Tris–HCl, pH 7.5] and stored at –80°C. All the buffers contain 10 μg/ml leupeptin, 10 μg/ml pepstatin A, 5 mM may be very different. Taken together among different phenanthroline and 1 mM phenylmethylsulfonyl fluoride (PMSF). organisms, the data suggest that clock proteins may use different approaches to achieve a similar result: a negative Protein analysis feedback essential for generating a circadian rhythm in Western blot analysis was performed as previously described (Garceau, the organism. What has been found with Neurospora FRQ 1996; Garceau et al., 1997). X-ray films of Western blots were scanned in terms of cellular biochemistry is consistent with the and densitometry was performed using NIH Image 1.59. FRQ was normalized against time-invariant protein CYS-3. expected nuclear action of clock proteins while again The exclusively cytoplasmic localization of CCG-1 was demonstrated revealing another of the ways through which circadian previously by immunofluorescence (Garceau, 1996; N.Y.Garceau, oscillators operate. K.M.Lindgren, S.J.Free, J.C.Dunlap and J.J.Loros, in preparation). Phosphatase treatment Total cell lysate and nuclei fractions were prepared as described above. Materials and methods A portion of each fraction corresponding to 40 μg of total protein was taken and put into 1 phosphatase treatment buffer (New England Strains, growth conditions and race tube assay 10 Biolabs), 1000 U of λPPase were added, and the reaction was incubated Strains used were 30-7 (bd; frq A) and 93-4 (bd; frq A; his-3) at 30°C for 45 min. Sample buffer was added to the resulting reaction, (Aronson et al., 1994b). All frq constructs were introduced into the frq and the samples were subjected to SDS–PAGE and Western blot analysis. null strain 93-4 and were targeted to the his-3 locus during transformation (Sachs and Ebbole, 1990). Methods for growing rhythmic cultures have been published previously Acknowledgements (Aronson et al., 1994a; Crosthwaite et al., 1995). Briefly, mycelial pads were inoculated into 500 ml of liquid culture and shaken at 110 r.p.m. We thank Dr Zuoyu Xu for his constructive suggestions, stimulating in light at room temperature. The cultures were then transferred into discussions and critical reading of the manuscript, and Dr Mike Young darkness and harvested at the time indicated. The liquid culture contains and members of our laboratories for discussion of the manuscript. The 1 sulfate-free Vogel’s salts, 2% glucose, 50 ng/ml biotin and 0.25 mM anti-CYS-3 and anti-CCG-1 antibodies were generously provided by Dr methionine. DD15 represents cultures held for 15 h in the dark prior George Marzluf and Dr Stephen Free. This work was supported by to harvesting. grants from AFOSR (F49620-94-1-0260 to J.J.L.), the National Science Race tube assay medium contains 1 Vogel’s salts, 0.1% glucose, Foundation (MCB-9307299 to J.J.L.), the National Institute of Health 0.17% arginine and 50 ng/ml biotin. Calculations of period length were (GM 34984 and MH 01186 to J.C.D. and MH 44651 to J.C.D. and done using the Chrono II version 9.3 (Dr Till Roenneberg, Ludwigs- J.J.L.) and the Norris Cotton Cancer Center core grant at Dartmouth Maximillian University, Munich). Medical School. Plasmids and Neurospora transformation References pKAJ120, which contains the entire frq locus, was the parental plasmid for all the constructs. pCL10 was constructed using the Transformer Aronson,B.D., Johnson,K.A., Loros,J.J. and Dunlap,J.C. (1994a) Site-Directed Mutagenesis kit from Clontech Laboratories, Inc. (Palo Negative feedback defining a circadian clock: autoregulation of clock Alto, CA). The SphI–EcoRV fragment from pKAJ120 was cloned into gene frequency. Science, 263, 1533–1656. pUC19, and the resulting plasmid was used as the template for the Aronson,B.D., Johnson,K. and Dunlap,J.C. (1994b) Circadian clock in vitro mutagenesis. The mutagenic primer for making the NLS deletion locus frequency: protein encoded by a single open reading frame was 5-GAGCGTTGCCTCCAACTCTAAGCCATGTACCTTGATCTC- defines period length and temperature compensation. Proc. Natl Acad. 3. The Trans Oligo NdeI–NcoI from the kit was used as the selection Sci. USA, 91, 7683–7686. primer. The SphI–BssHII fragment from the resulting plasmid was then Baum,J. and Giles,N. (1985) Genetic control of chromatin structure 5 inserted into the SphI–BssHII-digested pKAJ120 to create pCL10. The to the qa-x and qa-2 genes of Neurospora. J. Mol. Biol., 182, 79–89. pCL11 construct was made by cloning the PCR-amplified fragment Crosthwaite,S.K., Loros,J.J. and Dunlap,J.C. (1995) Light-induced containing the NLS into the SphI site in pCL10. The primers used resetting of a circadian clock is mediated by a rapid increase in for PCR amplification are 5-ACATGCATGCGCTTACCTAGAAGA- frequency transcript. Cell, 81, 1003–1012. AAG-3 and 5-ACATGCATGCGAAAGTCGCGGAGCGTTG-3. The Crosthwaite,S.K., Dunlap,J.C. and Loros,J.J. (1997) Neurospora wc-1 resulting PCR fragment was digested with SphI and then cloned into the and wc-2: transcription, photoresponses, and the origins of circadian SphI-digested pCL10. This resulted in the insertion of 16 amino acids rhythmicity. Science, 276, 763–769. including the NLS into the N-terminal region of FRQ, 11 amino acids Curtin,K.D., Huang,Z.J. and Rosbash,M. (1995) Temporally regulated after the first methionine. Both constructs were confirmed by DNA nuclear entry of the Drosophila period protein contributes to the sequencing. circadian clock. Neuron, 14, 365–372. Both constructs were transformed into strain 93-4 and targeted at the Dunlap,J.C. (1993) Genetic analysis of circadian clocks. Annu. Rev. his-3 locus as described previously (Sachs and Ebbole, 1990) Physiol., 55, 683–728. 1234 Nuclear localization of FRQ Dunlap,J.C. (1996) Genetic and molecular analysis of circadian rhythms. Saez,L. and Young,M. (1996) Regulation of nuclear entry of Drosophila Annu. Rev. Genet., 30, 579–601. clock proteins Period and Timeless. Neuron, 17, 911–920. Fu,Y.H., Paietta,J., Mannix,D. and Marzluf,G. (1989) cys-3, the positive- Sauman,I. and Reppert,S. (1996) Circadian clock neurons in the silkmoth activating sulfur regulatory gene of Neurospora crassa, encodes a Antheraea pernyi: novel mechanisms of period protein regulation. protein with a putative leucine zipper DNA-binding element. Mol. Neuron, 17, 889–900. Cell. Biol., 9, 1120–1127. Templeton,D., Park,S.H., Lanier,L. and Weinberg,R. (1991) Garcia-Bustos,J., Heitman,J. and Hall,M. (1991) Nuclear protein Nonfunctional mutants of the retinoblastoma protein are characterized localization. Biochim. Biophys. Acta, 1071, 83–101. by defects in phosphorylation, viral oncoprotein association, and Garceau,N.Y. (1996) Identification and regulation of a circadian clock nuclear tethering. Proc. Natl Acad. Sci. USA, 88, 3033–3037. protein, FRQ, and a clock controlled protein, CCG-1, in Neurospora Vosshall,L.B., Price,J.L., Sehgal,A., Saez,L. and Young,M.W. (1994) crassa. Ph.D. Thesis. Block in nuclear localization of period protein by a second clock Garceau,N.Y., Liu,Y., Loros,J. and Dunlap,J.C. (1997) Alternative mutation, timeless. Science, 263, 1606–1609. initiation of transcription and time-specific phosphorylation yield Wood,K. (1995) Marker proteins for gene expression. Curr. Opin. multiple forms of essential clock protein FREQUENCY. Cell, 89, Biotechnol., 6, 50–58. 469–476. Wuarin,J., Falvey,E., Lavery,D., Talbot,D., Schmidt,E., Ossipow,V., Gekakis,N., Saez,L., Delahaye-Brown,A.-M., Myers,M.P., Sehgal,A., Fonjallaz,P. and Schibler,U. (1992) The role of the transcriptional Young,M.W. and Weitz,C.J. (1995) Isolation of timeless by PER activator protein DBP in circadian liver gene expression. J. Cell Sci., protein interactions: defective interaction between timeless protein Suppl., 16, 123–127. and long-period mutant PER . Science, 270, 811–815. Zeng,H., Qian,Z., Myers,M.P. and Rosbash,M. (1996) A light- Hall,J. (1996) Are cycling genes products as internal Zeitgebers no entrainment mechanism for the Drosophila circadian clock. Nature, longer the Zeitgeist of chronobiology? Neuron, 17, 799–802. 380, 129–135. Hunter-Ensor,M., Ousley,A. and Sehgal,A. (1996) Regulation of the Drosophila protein timeless suggests a mechanism for resetting the Received September 15, 1997; revised and accepted January 7, 1998 circadian clock by light. Cell, 84, 677–685. Jans,D.A. and Hubner,S. (1996) Regulation of protein transport to the nucleus: central role of phosphorylation. Physiol. Rev., 76, 652–675. Jans,D.A. and Jans,J. (1994) Negative charge at the casein kinase II site flanking the nuclear localization signal of SV 40 large T-antigen is mechanistically important for enhanced nuclear import. Oncogene, 9, 2961–2968. Kanaan,M. and Marzluf,G. (1993) The positive-activating sulfur regulatory protein CYS3 of Neurospora crassa: nuclear localization, autogenous control, and regions required for transcriptional activation. Mol. Gen. Genet., 239, 334–344. Lee,C., Parikh,V., Itsukaichi,T., Bae,K. and Edery,I. (1996) Resetting of the Drosophila clock by photic regulation of PER and PER–TIM complex. Science, 271, 1740–1744. Lewis,M.T., Morgan,L.W. and Feldman,J.F. (1997) Analysis of frequency (frq) clock gene homologs: evidence for a helix–turn–helix transcription factor. Mol. Gen. Genet., 253, 401–414. Liu,X., Zwiebel,L., Hinton,D., Benzer,S., Hall,J. and Rosbash,M. (1992) The per gene encodes a predominantly nuclear protein in adult Drosophila. J. Neurosci., 12, 2735–2744. Liu,Y., Garceau,N.Y., Loros,J.J. and Dunlap,J.C. (1997) Thermally regulated translational control of FRQ mediates aspects of temperature responses in the Neurospora circadian clock. Cell, 89, 477–486. Loros,J.J. (1995) The molecular basis of the Neurospora clock. Semin. Neurosci., 7, 3–13. Loros,J.J., Denome,S.A. and Dunlap,J.C. (1989) Molecular cloning of genes under the control of the circadian clock in Neurospora. Science, 243, 385–388. Loros,J.J., Free,S.J., Lindgren,K.M., Garceau,N. and Dunlap,J.C. (1995) The ccg-1 gene of Neurospora displays multiple levels of regulation including the clock, development, light and heat shock. Fungal Genet. Newslett., 42A, 22. McNally,M.T. and Free,S.J. (1988) Isolation and characterization of a Neurospora glucose-repressible gene. Curr. Genet., 14, 545–551. Merrow,M.W. and Dunlap,J.C. (1994) Intergeneric complementation of a circadian rhythmicity defect: phylogenetic conservation of structure and function of the clock gene frequency. EMBO J., 13, 2257–2266. Merrow,M.W., Garceau,N.Y. and Dunlap,J.C. (1997) Dissection of a circadian oscillation into discrete domains. Proc. Natl Acad. Sci. USA, 94, 3877–3882. Mittnacht,S. and Weinberg,R. (1991) G1/S phosphorylation of the retinoblastoma protein is associated with an altered affinity for the nuclear compartment. Cell, 65, 381–393. Myers,M.P., Wager-Smith,K., Rothenfluh-Hilfiker,A. and Young,M.W. (1996) Light-induced degradation of TIMELESS and entrainment of the Drosophila circadian clock. Science, 271, 1736–1740. Paietta,J. (1992) Production of the CYS3 regulator, a bZIP DNA-binding protein, is sufficient to induce sulfur gene expression in Neurospora crassa. Mol. Cell. Biol., 12, 1568–1577. Price,J.L., Dembinska,M.E, Young,M.W. and Rosbash,M. (1995) Suppression of PERIOD protein abundance and circadian cycling by the Drosophila clock mutation timeless. EMBO J., 14, 4044–4047. Sachs,M.S. and Ebbole,D. (1990) The use of lacZ gene fusions in Neurospora crassa. Fungal Genet. Newslett., 37, 35–37.

Journal

The EMBO Journal – Springer Journals

Published: Mar 2, 1998

Keywords: circadian rhythm; frequency; Neurospora; nuclear localization

There are no references for this article.