Abstract
The Drosophila PAR domain protein 1 (Pdp1) gene encodes a transcription factor with multiple functions. One isoform, PDP1ε, was proposed to be an essential activator of the core clock gene, Clock (Clk). However, a central clock function for PDP1ε was recently disputed, and genetic analysis has been difficult due to developmental lethality of Pdp1-null mutants. Here we report the discovery of a mutation that specifically disrupts the Pdp1ε isoform. Homozygous Pdp1ε mutants are viable and exhibit arrhythmic circadian behavior in constant darkness and also in the presence of light:dark cycles. Importantly, the mutants show diminished expression of CLK and PERIOD (PER) in the central clock cells. In addition, expression of PDF (pigment-dispersing factor) is reduced in a subset of the central clock cells. Loss of Pdp1ε also alters the phosphorylation status of the CLK protein and disrupts cyclic expression of a per-luciferase reporter in peripheral clocks under free-running conditions. Transgenic expression of PDP1ε in clock neurons of Pdp1ε mutants can restore rhythmic circadian behavior. However, transgenic expression of CLK in these mutants rescues the expression of PER in the central clock, but fails to restore behavioral rhythms, suggesting that PDP1ε has effects outside the core molecular clock. Together, these data support a model in which PDP1ε functions in the central circadian oscillator as well as in the output pathway.
Introduction
The molecular circadian clock consists of interlocking feedback loops in which clock gene products regulate their own expression (Hardin, 2005). In one feedback loop in Drosophila, the circadian clock genes period (per) and timeless (tim) are activated by CLOCK-CYCLE (CLK-CYC) heterodimers during the late day, and their transcripts peak in the early night. Later in the night, the PER and TIM proteins translocate into the nucleus and inhibit CLK-CYC activity, thus repressing per and tim transcription. In a second feedback loop, CLK-CYC activate the basic leucine zipper transcription factors Pdp1ε and vrille (vri). Both PDP1ε and VRI proteins can bind to E4BP4 sites in the Clk promoter and activate or repress Clk transcription, respectively. PDP1ε expression lags behind that of VRI, and the sequential repression and activation may be the driving force behind rhythmic Clk expression (Cyran et al., 2003). This feedback loop also regulates overall Clk levels, so that overexpression of vri (Glossop et al., 2003) or reduced expression of Pdp1 results in reduced Clk expression (Cyran et al., 2003).
Pdp1 has seven isoforms, one of which, Pdp1ε, exhibits rhythmic expression (Cyran et al., 2003) and is thought to function in the circadian clock. A null mutation (P205) that removes all Pdp1 isoforms causes developmental lethality (Reddy et al., 2006) but shows dramatically reduced Clk levels in larvae (Cyran et al., 2003), consistent with cell culture data demonstrating that PDP1 can activate Clk transcription. However, two recent studies challenged the requirement of PDP1 in the molecular circadian oscillator (Benito et al., 2007). Overexpression of Pdp1ε or RNAi knock down of Pdp1 in clock cells disrupted circadian locomotor rhythms without affecting expression of Clk or other clock genes (Benito et al., 2007). Similar results were obtained when PDP1 activity was inhibited by transgenic expression of a dominant-negative form of PDP1 (Lim et al., 2007). Based on these results, it was suggested that PDP1 functions in the circadian output pathway rather than as an essential component of the circadian oscillator. The reduction of Clk expression in Pdp1-null mutant larvae was attributed to developmental defects. However, RNAi is unlikely to completely eliminate PDP1 activity [∼70% reduction by RNAi was reported in Benito et al. (2007)], and the residual PDP1 activity may be sufficient for normal molecular cycling, but not for regulation of downstream output genes. In addition, both RNAi and dominant-negative PDP1 indiscriminately affect all isoforms of Pdp1. Thus it remains unclear whether PDP1ε has a role in the circadian feedback loop of the adult fly as an essential positive regulator of Clk transcription.
Through a forward-genetic screen of EMS-mutagenized lines for abnormal circadian sleep:wake behavior (Wu et al., 2008), we identified a mutation that specifically disrupts the Pdp1ε isoform. Loss of Pdp1ε is not associated with any lethality, but causes arrhythmic circadian locomotor behavior. More importantly, we show here that expression of circadian clock proteins CLK and PER is substantially reduced in the clock neurons of Pdp1ε mutants. In addition, expression of pigment-dispersing factor (PDF) in some clock cells is reduced in Pdp1ε mutants. Transgenic expression of CLK in Pdp1ε mutants rescues the cyclic expression of PER in clock cells, but fails to restore behavioral rhythms. These data demonstrate that PDP1ε is required as an activator of Clk as well as a regulator of circadian output.
Materials and Methods
Drosophila genetics.
An arrhythmic line (Z2-3135) was identified in a screen of EMS-mutagenized lines from the Zuker collection (Koundakjian et al., 2004; Wu et al., 2008). Initial mapping revealed that the mutation causing arrhythmia was on the third chromosome even though the line was balanced on the second chromosome. For meiotic recombination mapping, we crossed Roughened (R) Dichaete (D)/3135 females to Z2-3135 homozygous males. Of 53 R not-D progeny, 27 were arrhythmic, and of 76 not-R, D progeny, 29 were arrhythmic. This placed the mutation approximately at cytological location 66. Complementation tests with deficiencies and candidate genes suggested that the mutation lies in the Pdp1 locus. Through sequencing, we identified a 4 bp deletion in an exon common to the Pdp1 RD and RJ isoforms. Next we outcrossed the mutant chromosome to an isogenic w 1118 background (iso31) for five generations, allowing recombination, and the molecular lesion was followed by PCR and sequencing. This mutation is referred to as Pdp1 3135 in this study.
A full-length Pdp1-RD cDNA clone (GH27708) was obtained from DGRC, and the coding region was amplified by PCR using the following primers: 5′-GTT GCG GCC GCC AGT GAT CAT GTC GTC-3′ and 5′-TGT CTC GAG ATC ATT ACA CAT CTT GGA A-3′. The PCR fragment was digested with NotI and XhoI, and inserted into the pUAST transformation vector. Transgenic lines carrying the UAS-Pdp1-RD construct were generated using standard techniques (Rubin and Spradling, 1982) in an isogenic w1118 (iso31) background (Rainbow Transgenic Flies), and two independent lines were used for rescue and overexpression experiments.
Circadian behavioral assays.
Stocks were raised on standard food containing cornmeal, molasses, and yeast. Three- to five-day-old adult male flies were entrained to a 12 h:12 h light:dark (LD) schedule at 25°C for 3 d, and then loaded into locomotor assay tubes containing 5% sucrose and 2% agarose. Activity records collected under LD and constant darkness (DD) conditions were analyzed by using Clocklab software (Actimetrics). Circadian periodicity was evaluated using χ2 periodogram analysis. In the aging experiment, 2- to 3-d-old adult males were collected and maintained on standard food for 22 d at room temperature (23°C) with change of food every 4 d, and then entrained to an LD cycle at 25°C for 3 d and analyzed as described above.
Whole-fly per-luciferase reporter assay.
Three- to five-day-old adult males were collected and entrained to a 12 h:12 h LD cycle at 25°C for 3 d on standard food, then loaded into Optiplate-96F plates (PerkinElmer) containing 100 μm firefly luciferin (Biosynth), 5% sucrose, and 2% agar. Data were collected in DD over 5 d in a Topcount NXT luminometer (Packard). Each well was read twice, for 6 s each time, in a 30 min interval. Raw data were imported and analyzed in Excel.
Western blot analysis.
Three- to five-day-old adult flies were entrained to 12 h:12 h LD cycles for 3 d, and heads were collected at indicated time points for protein extraction. For DD samples, flies were collected on the second day in constant darkness, and Western blot analysis was performed as described previously (Sathyanarayanan et al., 2004). Guinea pig anti-PER, guinea pig anti-VRI (Glossop et al., 2003), guinea pig anti-CLK (Lee et al., 1998), and rabbit anti-PDP1 (Cyran et al., 2003) were used at 1:3000 dilutions. Following enhanced chemiluminescence, images were taken in a Kodak image station or exposed to film, and images were processed using Adobe Photoshop.
Quantitative real-time PCR.
Three- to five-day-old adult flies were maintained at a 12 h:12 h light:dark cycle at 25°C for 3 d and then collected on dry ice at indicated time points on the third day of LD or the second day of DD. Total RNA was isolated using an Ultraspec RNA isolation system (Biotecx), and cDNAs were synthesized by using a high-capacity cDNA Archive kit (Applied Biosystems). Quantitative real-time PCR was performed in an ABI prism 7100 using a SYBR Green kit (Applied Biosystems).
Immunohistochemistry and microscopy.
Fly brains were dissected in 4% PFA, washed for 1 h in PBS buffer, and incubated with primary antibody (in PBS buffer with 3% normal donkey serum and 0.3% Triton X-100) overnight at 4°C. Brain samples were extensively washed with PBS buffer and incubated with Cy3 donkey anti-rat (or guinea pig) and FITC or Cy3 donkey anti rabbit secondary antibodies (Jackson ImmunoResearch Laboratories) for 1.5 h at room temperature, followed by extensive washes in PBS. Fluorescent images were taken with a Leica TCS SP5 confocal micro scope. Primary antibody dilutions were as follows: PER, 1:1500; PDF, 1:1000; and CLK (Houl et al., 2006), 1:500. Secondary antibody dilutions were 1:1000.
Results
A new allele of Pdp1 has arrhythmic circadian behavior
To isolate novel circadian mutants, we screened the Zuker collection of EMS-induced mutations (Koundakjian et al., 2004) for failure to show anticipatory locomotor behavior before light–dark transitions, as previously described (Wu et al., 2008). One of these lines (Z2-3135) exhibits arrhythmic behavior in both LD and DD conditions (Fig. 1). Through meiotic recombination and deficiency mapping, we found that this line carries a mutation that lies at the cytological position 66 on the third chromosome (see Materials and Methods). Since this region contains circadian clock genes Clk and Pdp1, we suspected that one of them is mutated in Z2-3135. Complementation tests with Clk and Pdp1 alleles indicated that the new mutation (hereafter referred to as Pdp1 3135) maps to the Pdp1 locus (Fig. 1, Table 1).
Molecular characterization of the Pdp1ε mutation
Multiple promoters and alternative splicing events produce several Pdp1 isoforms (Reddy et al., 2000): there are seven annotated isoforms in Flybase (www.flybase.org) (Fig. 2 A). To identify the molecular lesion of the new Pdp1 3135 allele, we sequenced the coding region of all Pdp1 isoforms. A 4 bp deletion was found in the second exon of the RD and RJ isoforms (Fig. 2 A; supplemental Fig. 1A, available at www.jneurosci.org as supplemental material). While the Pdp1ε sequence reported in Cyran et al. (2003) matches that of the RD isoform, the sequence provided by Reddy et al. (2000) is that of the RJ isoform. RD and RJ isoforms differ by one small exon coding for 14 aa (supplemental Fig. 1B, available at www.jneurosci.org as supplemental material). The two isoforms share the same promoter region and transcriptional start site, and thus are likely to be regulated in a similar manner. Indeed, we find that transcript and protein levels of both isoforms are regulated in a circadian manner (Fig. 2 B,C). We refer to RJ and RD collectively as the ε isoform in this study.
The 4 bp deletion in Pdp1 3135 mutants affects both the RD and RJ isoforms but none of the other isoforms. The deletion results in a frame shift starting from residue 262 and a predicted truncated polypeptide of 374 aa (supplemental Fig. 1A, available at www.jneurosci.org as supplemental material), which is expected to be nonfunctional. Western analysis confirmed that PDP1ε is undetectable in Pdp1 3135 mutants (Fig. 2 C). Therefore, the 3135 mutation is presumed to abolish the ε isoform, but leave other isoforms intact.
Pdp1ε is the dominant but not the only isoform that regulates circadian rhythms
Pdp1 P205-null mutants are lethal, and flies expressing a dominant-negative form of Pdp1 (Pdp1 DN) and Pdp1 RNAi flies show anatomical abnormalities in the clock neurons (Lim et al., 2007), underscoring the important role of PDP1 in development. It is likely, however, that spatial and temporal expression patterns of different isoforms of Pdp1 are differentially regulated. Pdp1 3135 mutants are viable, which suggests that the PDP1ε isoform does not play a significant developmental role. In addition, only the Pdp1ε isoform displays robust cycling at the mRNA and protein levels (Cyran et al., 2003). An antibody raised against the common region of the PDP1 isoforms detected robust cycling in the central clock cells in the larval brain (Cyran et al., 2003), suggesting that PDP1ε is the major isoform in these clock cells. Using the same antibody, we found that PDP1 immunoreactivity in the central pacemaker cells of the adult brain is markedly reduced in Pdp1 3135 mutants (Fig. 2 D), confirming that PDP1ε is indeed the predominant isoform expressed in the central circadian pacemaker.
We also identified a second mutation in Pdp1, d11071, which is caused by a P-element insertion in an intron common to all isoforms of Pdp1 (Fig. 2 A). The P-element carries a splicing acceptor, and thus is predicted to disrupt splicing of all Pdp1 isoforms. PDP1ε levels are markedly reduced, but detectable, in Pdp1 d11071 homozygous mutants, indicating that d11071 is a hypomorphic allele (supplemental Fig. 1C, available at www.jneurosci.org as supplemental material). Most (78%) Pdp1 d11071 homozygous flies exhibit arrhythmic behavior (Table 1). Notably, Pdp1 d11071/3135 transheterozygous flies show a milder phenotype, which becomes more severe with aging; ∼26% of the 7-d-old and ∼41% of the 25-d-old transheterozygotes are arrhythmic. Our finding that Pdp1 d11071/3135 transheterozygous flies have less PDP1ε than Pdp1 d11071 flies (supplemental Fig. 1D, available at www.jneurosci.org as supplemental material), but have a milder phenotype, suggests that PDP1ε is not the only PDP1 isoform contributing to behavioral rhythms.
Transgenic expression of Pdp1ε restores behavioral rhythms
To confirm that the arrhythmic phenotype of Pdp1 3135 mutants maps to Pdp1ε, we generated transgenic fly lines carrying a UAS-Pdp1-RD construct. Expression of wild-type Pdp1-RD in the clock cells under the control of tim-UAS-Gal4 (TUG) or cry-Gal4 fully rescues the mutant phenotype (Table 1, Fig. 3 A). Notably, the RJ isoform is more abundant than the RD isoform in wild-type flies (Fig. 2 C), and thus our rescue result suggests that the RD and RJ isoforms are functionally interchangeable. On the other hand, expression of Pdp1-RD in PDF-expressing cells using the Pdf-Gal4 driver fails to rescue the arrhythmic behavior, probably due to its restricted expression pattern. We note that expression of UAS-Clk or UAS-cyc by a Pdf-Gal4 driver is also ineffective in rescuing Clk ar or cyc 01 mutants (Allada et al., 2003; Peng et al., 2003). Overexpression of Pdp1ε in a wild-type background does not affect rhythmic behavior or free-running period (Table 1). Western blot analysis of head extracts confirmed that the UAS-Pdp1-RD transgene produces protein of the expected size whose abundance exhibits circadian cycling even when overexpressed by the TUG driver (Fig. 3 B). However, when overexpressed with another, possibly stronger driver (tim-Gal4), PDP1 protein cycling is abolished (Fig. 3 C). Interestingly, the behavioral rhythms of flies overexpressing PDP1 with tim-Gal4 are still unaffected (Table 1). These data suggest that cycling of PDP1 protein levels is not required for behavioral rhythms.
PDP1ε regulates expression of clock proteins in the central clock
Isolation of an isoform-specific mutant that specifically abolishes the ε isoform of Pdp1 provides us with a unique opportunity to ascertain the role of PDP1ε in adult clock neurons. We first asked whether Pdp1ε plays a role in regulating CLK expression in the adult central clock cells. While CLK expression in the brain of the Pdp1ε mutant is variable in the presence of LD cycles (that is, some cells have reduced expression of CLK, but others have relatively normal expression, data not shown), it is substantially reduced in the ventral lateral neurons in DD (Fig. 4 A), consistent with previous findings in Pdp1-null larvae (Cyran et al., 2003). CLK expression is also reduced in other clock neurons such as dorsal lateral neurons (LNds) and dorsal neurons (DN1) (Fig. 4 A; supplemental Fig. 2, available at www.jneurosci.org as supplemental material). In addition, expression of PER, a direct target of CLK, is also dramatically diminished in DD, and the cycling of PER is not discernable in mutants (Fig. 4 B). However, as noted for the Pdp1-null larvae (Blau et al., 2007), there is only marginal effect of the Pdp1 3135 mutation on PER oscillations in LD (data not shown). The effect of loss of Pdp1ε on the free-running molecular clock is in contrast to results obtained with Pdp1 RNAi or Pdp1 DN flies (Benito et al., 2007; Lim et al., 2007). The difference may stem from residual PDP1 activity in Pdp1 RNAi and Pdp1 DN flies.
Interestingly, downregulation of another CLK-regulated protein, PDF, is seen in the small ventral lateral neurons but not in the large ventral lateral neurons (Fig. 5). This is reminiscent of findings that PDF expression is severely affected only in the small ventral lateral neurons in Clk and cyc mutants (Park et al., 2000; Allada et al., 2003). Together, these results suggest that loss of Pdp1ε results in molecular phenotypes in the central clock similar to those caused by Clk or cyc mutation, and provide evidence that PDP1ε is indeed a major positive regulator of CLK expression in the central pacemaker cells.
Loss of Pdp1ε alters the expression of clock genes in peripheral clocks
To examine whether Clk transcription is affected in Pdp1ε mutants, we measured Clk mRNA levels in whole fly heads by quantitative real-time PCR. Most circadian gene transcripts and protein products in whole heads come from the eyes, and thus fly head assays are presumed to reflect the function of peripheral clocks. Clk mRNA levels are slightly reduced in LD and DD (Fig. 6 A,B). We also assayed CLK protein levels in fly heads by Western blot analysis. In the presence of LD cycles, there is a small reduction of CLK protein levels in Pdp1ε mutants, but the CLK protein still displays cyclic mobility shifts over the course of the day (Fig. 6 C). However, a more severe phenotype is observed in DD. Cyclic mobility shifts of CLK are dampened in DD in both wild type and Pdp1 3135 mutants, but the mutant, in addition, shows a reduction in the low mobility forms of CLK (Fig. 6 D). Decreased mobility of CLK is thought to arise from phosphorylation, which affects its transcriptional activity (Kim and Edery, 2006; Yu et al., 2006). An effect of the Pdp1 3135 mutation on the levels and the phosphorylation of CLK suggest that CLK activity is altered in multiple ways in Pdp1 3135 mutants.
Since altered activity of CLK is predicted to affect oscillations of its targets, we also assayed the expression of genes regulated by CLK. mRNA levels of per are reduced in Pdp1ε mutants (Fig. 6 A). In addition, Western blot analysis revealed that peak levels of PER protein are reduced in Pdp1 3135 mutants at late night (ZT20 and CT20), although the cyclic expression and phosphorylation pattern of PER are not dramatically changed in LD or DD (Fig. 6 B,D). Moreover, cyclic expression of a per-luciferase reporter in whole flies dampens rapidly in Pdp1ε mutant flies (Fig. 6 E). Although these data support the idea that CLK activity is affected, it is possible that PDP1 directly affects per expression. Regardless of whether or not the effects on per are CLK dependent, PDP1ε clearly has a role in peripheral circadian oscillators, in particular under free-running conditions.
Another CLK target, VRI, is also affected in Pdp1 mutants. As in the case of all the other proteins discussed above, the effect on VRI is more obvious in DD, with a broader peak and reduced peak levels at CT14 (supplemental Fig. 3, available at www.jneurosci.org as supplemental material). We also examined a target of VRI to determine whether PDP1ε has opposing effects on VRI targets. This was based upon mammalian and Drosophila studies showing that E4BP4 and PAR proteins antagonize each other in regulating circadian clock genes (Mitsui et al., 2001; Cyran et al., 2003). Since cry is repressed by overexpression of VRI (Glossop et al., 2003), we speculated that cry expression would be downregulated in Pdp1 3135 mutants. Indeed, peak expression of cry mRNA at ZT02 is reduced in Pdp1 3135 mutants, despite the decreased VRI levels, indicating that PDP1ε and VRI antagonize each other in regulating cry expression (Fig. 6 A).
Overexpression of Clk rescues PER expression but not PDF expression and behavioral rhythms
If reduced CLK activity is the underlying cause of the reduced expression of PER and PDF, then increased expression of Clk in the Pdp1ε mutant background might rescue these defects. Indeed, transgenic expression of Clk under the control of cry-Gal4 restores PER expression in the LNvs (Fig. 7). However, PDF expression is not restored, nor are behavioral rhythms (Fig. 7, Table 2). Transgenic expression of per or Pdf was also not able to rescue the behavioral rhythm defects of Pdp1ε mutants (Table 2), consistent with the idea that PDP1ε regulates Clk-dependent expression of per, but that it has additional roles in the circadian clock (see Discussion).
Discussion
As noted in the Introduction, the role of Pdp1 in circadian rhythms has been controversial. Recent reports on the subject concluded that Pdp1 does not function in the central clock and only affects circadian output (Benito et al., 2007) or lateral neuron morphology (Lim et al., 2007). However, these studies did not specifically address the role of the ε isoform, instead using methods that would knock down all isoforms. In addition, the methods used for knockdown (RNAi and DN) did not eliminate gene expression. We have identified an ε-specific mutation which allows us to unequivocally establish PDP1ε as an important component of the central clock.
PDP1ε affects CLK and PER expression in central and peripheral clocks
The mutation we discovered is a 4 bp deletion in the second exon of Pdp1ε that is predicted to introduce a frame shift and result in no functional PDP1ε protein. Loss of Pdp1ε has a strong effect on clock protein expression in the central pacemaker cells. We found that CLK levels are reduced in the lateral neurons of Pdp1 3135 mutants. Expression of PER is also dramatically reduced, demonstrating that PDP1ε is an important positive regulator of the central circadian clock.
Analysis of adult heads, which are representative of peripheral clocks, demonstrates that although the effect of the Pdp1 3135 mutation on CLK levels is small, the mutation decreases the phosphorylation of CLK. Surprisingly, there is a stronger effect on per levels, which may be indicative of a CLK-independent effect of PDP1 on per. In fact, the effect on CLK phosphorylation may be secondary to the effect on PER. PER is known to promote CLK phosphorylation by “delivering” the kinase DBT (Kim and Edery, 2006; Yu et al., 2006). While CLK is constantly present in the nucleus (Houl et al., 2006), PER only enters the nucleus during the late night; thus a small reduction of PER expression at that stage might account for the reduced PER-dependent phosphorylation of CLK (Kim and Edery, 2006; Yu et al., 2006) in Pdp1 3135 mutants. Alternatively, PDP1 may regulate the expression of a kinase that phosphorylates CLK.
In general, the phenotype in peripheral clocks is weaker than that observed in central clock cells, with the exception of the defect in CLK phosphorylation which would not be detected through the immunofluorescence assays used to study the central clock. We assume that this difference is due to the presence of other isoforms of PDP1 in peripheral clocks. While we find that Pdp1ε is the major isoform expressed in the central clock cells, other isoforms of Pdp1 may be more abundant in the head and body, although we were not able to reliably detect other isoforms in Western blot assays. Transcripts of other isoforms do not cycle (Cyran et al., 2003), but they may still be able to activate Clk expression, and cycling of other clock components such as VRI may be sufficient to drive rhythmic Clk expression. Our finding that the behavioral phenotype of the pan-isoform, hypomorphic Pdp1 d11071 homozygotes is more severe than that of transheterozygous Pdp1 d11071/3135 flies is consistent with the possibility that non-ε isoforms of Pdp1 function in the circadian oscillator (Table 1).
Differential effects of PDP1 on molecular and behavioral rhythms
All molecular effects of the Pdp1 3135 mutant are more pronounced in DD than in LD, suggesting that factors other than Pdp1 account for the molecular cycling in LD conditions. Thus, PDP1 likely functions in the clock predominantly under free-running conditions. However, the mutant flies are arrhythmic in LD cycles. While this is similar to, although perhaps stronger than, the LD phenotype reported for Clk and cyc mutants, it is not a general feature of clock mutants. For example, per and tim mutants are rhythmic in LD cycles. Rhythms in LD arise either from entrainment of the clock to the light:dark cycle or, in the absence of a functional clock, from a process known as “masking.” The arrhythmic LD behavior of Pdp1 mutants suggests that in addition to affecting clock function, they have defects in masking.
Effect of PDP1 on PDF expression
CLK-CYC heterodimers activate Pdf expression through an unknown mechanism (Park et al., 2000). Both Pdf RNA and PDF protein levels are extremely low in Clk and cyc mutants, and this reduction is more pronounced in the small lateral neurons relative to the large lateral neurons (Park et al., 2000; Allada et al., 2003). This is consistent with our finding that in Pdp1 3135 mutants, the reduction of CLK expression is associated with reduced PDF expression in the small lateral neurons of Pdp1 3135 mutants. These data suggest that different clock cells have different sensitivity to loss of PDP1 and CLK.
Interestingly, expression of CLK with a heterologous promoter does not rescue PDF expression in Pdp1 mutants. This may indicate a circadian clock-independent effect of PDP1 on PDF, but we cannot rule out an alternative explanation involving levels of VRI. VRI overexpressing flies also show reduced PDF, most likely due to a posttranscriptional effect of VRI on PDF (Blau and Young, 1999). Since CLK activates vri expression, forced expression of CLK in the Pdp13135 mutant background may increase VRI, leading to a further reduction of PDF. Most likely, it is this effect on PDF that accounts for the failure of CLK to restore behavioral rhythms in Pdp1 mutants. In a wild-type background, CLK overexpression does not cause arrhythmia, probably because both VRI and PDP are activated simultaneously and have opposing effects. Indeed, VRI overexpression (Blau and Young, 1999) and PDP1ε loss of function have the same effect on PDF expression.
Based upon the discussion above, it appears that the effect of PDP1 on PDF is not strictly through CLK. Thus, in this sense PDP1 has a more direct role in clock output (if PDF may be considered a clock output). In addition, it clearly has an essential role in the central clock. However, while these data establish a clock function for PDP1, they do not address the importance of the second feedback loop which drives the cyclic expression of Clk. PDP1 is required because it activates the expression of an essential clock gene, Clk, and not necessarily because it drives Clk cycling. An investigation of the importance of the second feedback loop must await analysis of mutations in the negative regulator of that loop, vri.
Footnotes
-
This work was supported by National Institute of Neurological Disorders and Stroke (NINDS) Grant R01NS048471 (A.S.), a National Alliance for Research on Schizophrenia and Depression Young Investigator award (K.K.), a Career Award for Medical Scientists Award from the Burroughs-Wellcome Foundation (M.N.W.), and NINDS Grant K08NS059671 (M.N.W.). A.S. is an investigator of the Howard Hughes Medical Institute. We thank Isaac Edery (Rutgers University), Paul Hardin (University of Houston), Ravi Allada (Northwestern University), Paul Taghert (Washington University in St. Louis), and Justin Blau (New York University) for providing fly stocks and antibodies. We are especially thankful to Andrea Lougheed, Robert Hardy, and Charles Zuker for providing Zuker collection mutants for screening.
- Correspondence should be addressed to Amita Sehgal, 232 Stemmler Hall, Department of Neuroscience, University of Pennsylvania Medical School, Philadelphia, PA 19104. amita{at}mail.med.upenn.edu