Extent and character of circadian gene expression in
Drosophila melanogaster : ident ification of twenty
oscillating mRNAs in the fly head
Russell N. Van Celder*+, Helen Bae*, Michael J.Palazzolo*
and Mark A. Krasnow*
Extent and character of circadian gene expression in
Drosophila melanogasfeu: identification of twenty
oscillating mRNAs in the fly head
Russell N. Van Gelder"), Helen Bae*, Michael J.Palazzolo*
and Mark A. Krasnow*
*Department of Biochemistry, Stanford University School of Medicine, Stanford, California 94305, USA. +Basic Sleep Research Laboratory,
Stanford University School of Medicine, Stanford, California 94305, USA. *Human Genome Project, Lawrence Berkeley Laboratories,
University of California, Berkeley, Berkeley, California 94720, USA.
Background: Although mRNAs expressed with a is also maximally expressed (but when the flies are
circadian rhythm have been isolated from many species, inactive). Further analysis of the three 'morning' cDNAs
the extent and character of circadianly regulated gene showed that each has a unique dependence on the pres-
expression is unknown for any animal. In Drosophila ence of a light-dark cycle, on timed feeding, and on the
nzelanogaster, only the period (per) gene, an essential com- function of the per gene for its oscillation. These depen-
ponent of the circadian pacemaker, is known to show dencies were different from those determined for per and
rhythmic niRNA expression. Recent work suggests that for a novel 'evening' gene. Sequence analysis indicated
the encoded Per protein controls its own transcription by that all but one of the 20 cDNAs identified previously
an autoregulatory feedback loop. Per might also control uncloned genes.
the rhythmic expression of other genes to generate Conclusions: Diurnal control of gene expression is a
circadian behavior and physiology. The goals of this work significant but limited phenomenon in the fly head,
were to evaluate the extent and character of circadian which involves many uncharacterized genes. Diurnal
control of gene expression in Drosoplzila, and to identify control is mediated by multiple endogenous and
genes dependent on per for circadian expression. exogenous mechanisms, even at the level of individual
Results: A large collection of anonymous, independent genes. A subset of circadianly expressed genes are
cDNA clones was used to screen for transcripts that are predominantly or exclusively dependent on per for their
rhythmically expressed in the fly head. 20 of the 261 rhythmic expression. The per gene can therefore in-
clones tested detected mRNAs with a greater than two- fluence the expression of genes other than itself, but for
fold daily change in abundance. Three mRNAs were many rhythmically expressed genes, per functions in con-
maximally expressed in the morning, whereas 17 mRNAs junction with external inputs to control their daily
were most abundant in the evening - when pev mRNA expression patterns.
Current Biology 1995, 5:1424-1436
Background without disrupting the actual behaviors themselves. Ani-
mals mutant in these genes are fully viable and without
Much of the behavior and physiology of the Drosophilidae any obvious morphological defects.
fruit flies is temporally organized around a 24-hour day.
Eclosion from the pupal case, for example, occurs exclu- Recent studies have begun to elucidate how single genes
sively in a brief window of time in the early morning [l] . can so profoundly affect the temporal organization of
Locomotor activity also has a strong diurnal rhythm, behavior. The per gene encodes a predominantly nuclear
with substantially greater levels of activity occurring dur- protein (Per) that is expressed in the fly brain ,with
ing the day than at: night . These circadian rhythms are homology in a dimerization domain (the PAS domain) to
endogenously generated, and persist in the absence of all several transcription factors . Both per m R N A and
external timing cues (see 1 1 for review).
3 protein oscillate with a free-running circadian rhythm
[9-111. The per m R N A reaches peak levels in the
The temporal organization of physiology and behavior evening, and Per protein reaches its peak 8-10 hours later
has a discrete genetic basis. Specific genes appear to func- 112,131. The oscillation of per m R N A depends o n the
tion primarily in organizing behavior around an approxi- intact function of Per protein [ l l ] - a feedback loop
mately 24-hour clock. Mutations in the period (per; ) which has been suggested to comprise the essential core
and timeless [5,6] genes can completely disrupt the daily of the circadian pacemaker [11,14-161. Heat-shock
organization of adult locomotion and pupal eclosion induction of per m R N A in a wild-type background shifts
Correspondence to: Russell N.Van Gelder. Present address: Department of Ophthalmology and Visual Sciences, Washington University
Medical School, 660 South Euclid Avenue, St. Louis, Missouri 631 10, USA. E-mail address: email@example.com
1424 0 Current Biology 1995, Vol 5 No 12
Circadian gene expression in Drosophila Van Gelder et a / . RESEARCH PAPER 1425
the phase of the circadian clock, with the phase of the Results
resultant rhythm being dependent on the phase of heat-
shock induction . This result strongly suggests that A screen for diurnally expressed mRNAs in the fly head
the instantaneous level of per m R N A in part determines By performing a subtractive hybridization between
the phase of the circadian clock. Thus, the circadian cDNA libraries prepared from adult Drosophila heads and
expression of the per gene is apparently essential to the 0-1 hour-old embryos, Palazzolo et al.  isolated and
function of the endogenous circadian pacemaker. characterized 436 independent cDNAs that were
expressed in the adult fly head but not in the early
Although per is the only gene in Drosophila rnelanogaster embryo. This collection of anonymous cDNA clones is
(D. rnelanogaster) that is known to undergo circadian well suited for screening for diurnally expressed genes,
rhythms of expression, genes expressed with diurnal and The collection contains cDNAs that correspond to a
circadian rhythms have been identified in many other wide range of transcript abundances, from transcripts
organisms from cyanobacteria to mice (reviewed in expressed at nearly 1 % of fly head m R N A to those
). These genes function in many physiological pro- expressed at only several parts per million of fly head
cesses, including light transduction [ 19,201, endocrine mRNA. Because the 0-1 hour embryo does not have a
function [21-231 and metabolism . Most of these nervous system, the collection is enriched for genes spe-
oscillating genes were identified by the testing of various cific to the nervous system, and the representation of
known genes for rhythmic expression, or by the ubiquitously expressed genes is reduced. Importantly, the
fortuitous observation of diurnal oscillation of a known procedure used in characterizing these transcripts elimi-
gene product. nated redundancy in the collection, as cross-hybridizing
clones were identified and removed. O f the 436 cDNAs
The extent and importance of circadian control of gene in the collection, 280 show ‘simple’ expression patterns
expression are not well known. The only previous large- manifest on Northern blots by one or more transcripts
scale survey was reported recently for a prokaryotic that are coordinately regulated throughout development.
cyanobacterium, Syneclzococczrs . By monitoring luci- The remaining cDNAs identift. multiple transcripts that
ferase expression in individual Synechococcus colonies are not coordinately regulated during development.
carrying random insertions of a luciferase reporter
gene, the authors found that almost all of the 800 To determine the diurnal expression patterns of these
colonies analyzed showed rhythmic expression. T h e only genes, adult D. rnelanogaster Canton-S (wild-type) flies
other systematic search for circadianly expressed genes were collected from large population cages at Zeitgeber
used a two-timepoint subtractive hybridization approach times (ZTs) 2, 8, 14 and 20, in a 12-hour light-12-hour
in the mold Neurospora c~assa; from a pool of total dark (LD 12:12) cycle (by convention, Z T 0 occurs at
Neurospora m R N A , the authors found two cDNAs the dark-light transition). Total R N A was prepared from
undergoing circadian rhythms of expression . the heads of these flies. A simple, rapid procedure was
Although this study did not identift. all genes under developed for the synthesis of single-stranded, radio-
circadian control in 2L’eurosyora (see ), the results labeled R N A probes from polymerase chain reaction
nevertheless suggest that circadian control of gene (PCR) amplification products of each of the ‘simple’
expression is a quite limited phenomenon in Neurospora. expression pattern cDNA phage clones (see Materials
N o comparable study has been reported for Duosophila or and methods). Each of these cRNAs was used to probe
any other animal. individual northern blots of R N A from each of the four
Zeitgeber times. A per probe was used in the experi-
The present study was initiated to answer two questions. ments as a positive control for rhythmic m R N A expres-
First, what is the extent of diurnal control of gene sion, and probes for the r p 4 9 m R N A , which encodes a
expression in D. rnelatzogaster? Second, to what extent is ribosomal protein, and for the ninaE m R N A , which
diurnal gene expression dependent o n the function of encodes an abundant opsin, served as non-oscillating
the per gene, and to what extent is it dependent o n negative controls.
external time cues? To begin to answer these questions,
we undertook a screen through a large collection of The vast majority (> 90 %) of the 261 cDNAs detected
independent cDNAs known to be expressed in the mRNAs that were expressed constitutively or that
adult Drosophila head but not in the early embryo, search- cycled very weakly (such as the 7C12 cDNA probe; Fig.
ing for mRNAs that demonstrate diurnal variations in 1). 20 cDNA probes, however, identified oscillating
abundance. In addition to answering the questions mRNAs with greater than two-fold peak-to-trough dif-
of extent and mechanism of circadian control of gene ferences in expression over the 24-hour period (Fig. 1).
expression, we also hoped to identifi- additional circadi- Each of the probes that detected cycling mRNAs was
anly controlled genes that might contribute to clock retested on northern blots of R N A from a separate pop-
function and to its control of physiology and behavior. ulation of flies to confirm the oscillation of expression.
Here, we describe the results of the screen, and report o n We named this collection of cycling mRNAs the Dregs,
further circadian and molecular characterization of the for Drosophila rhythmically expressed genes, and assigned
subset of the genes that were expressed at their highest each a number according to its overall level of expression
levels in the morning. in the adult fly head (Dreg-2 was the most abundant).
1426 Current Biology 1995, Vol 5 No 1 2
The 20 diurnally oscillating mRNAs display two general rhythmicity. In other instances (such as Dreg-9, Dreg- 12
temporal expression patterns and Di.eg-21), one transcript cycled while the other was
The Dregs displayed two general temporal patterns of constitutively expressed.
expression. All showed a single daily peak and a single
trough of expression. Dreg-I, Dreg-2 and Dreg-3 showed The amplitude of oscillation of the Dregs was comparable
highest levels of expression in the early morning or late at to, and in some cases greater than, the five-fold peak-to-
night just before day break (Fig. 1); these are referred to trough amplitude that was observed for per mRNA.
as the ‘morning’ Dregs hereafter. Dreg-I and Dreg-3 were Dreg-2 showed the lowest amplitude of oscillation, of
expressed with peak levels at ZT 2 and troughs at Z T 14, about 2.5-fold, while several of the Dregs that were
in the opposite phase from per niRNA. Dreg-2 followed a expressed in phase with per m R N A (such as Dreg-9)
similar temporal expression pattern, but demonstrated showed amplitudes as high as 50-fold. These values are
minimal transcript levels at ZT 20, rather than ZT14. not absolutes, however, as the amplitude of m R N A
The other 17 Dregs all showed highest expression in the oscillation for many of the Dregs, as well as for per, is
early night, at Z T 14, and lowest expression at Z T 2, in influenced by environmental conditions (see below).
the sanie phase as expression of per mRNA. These are
referred to as the ‘evening’ Dregs hereafter. Several Dregs Both Dreg-1 and Dreg-2 were expressed in the body of
showed multiple poly-adenylated transcripts o n northern the fly as well as in the head, whereas Dreg-3 m R N A was
blots (Fig. 1; Table 1). In some instances (such as Dreg-10 found exclusively in the head (Table 1). Among the
and Dreg- 16), the multiple transcripts all showed diurnal evening Dregs, only Dreg-6 and Dreg-9 transcripts were
2 8 14 20 Poly-A’
Dreg- 1 Dreg-2 Dreg-3 Dreg-5
Dreg-6 Dreg-7 Dreg-8 Dreg-9
Fig. 1. Diurnal expression of the Dreg
mRNAs in fly heads. Each panel shows
Dreg- 7 0 Dreg- 1 1 Dreg- 12 an individual northern blot of total head
RNA (20 pg per lane) from flies kept
under LD 12:12 lighting conditions and
sacrificed at (lane 1) ZT 2, (lane 2) ZT8,
(lane 3) ZT14 or (lane 4) ZT20. The lane
on the far right of each panel is poly-A+
mRNA (-1 pg) prepared from all time
Dreg- 14 Dreg- 15 Dreg- 16 Dreg- 17
points. Blots were hybridized with
probes from Dreg-1 to Dreg-21, cDNA
clone 7 0 2 , ninaE, rp49 and per, as
indicated. Blots were exposed for differ-
ent times to obtain the autoradiograms
shown; the estimated abundance (based
on exposure time and probe lengths)
and sizes of the poly-A+ transcripts are
Dreg- 18 Dreg- 19 Dreg-20 Dreg-2 1 given for each Dreg in Table 1. 7C12 is
an example of one of the many non-
cycling genes identified in the screen;
the sizes of its transcripts are 2.2 kb and
6.2 kb. The per, ninaE and rp49 genes
a were used throughout the screen as
cycling (per) and non-cycling (ninaE,
rp49) control mRNAs. Dreg-1 to Dreg-3
show highest expression in pre-dawn or
7C12 ninaE rp4 9 Per
early morning; Dreg-5 to Dreg-21 show
highest expression in the evening.
Circadian gene expression in Drosophila Van Gelder et a / . RESEARCH PAPER 1427
detected (albeit weakly) in polyAf R N A from fly bodies. and the endogenous circadian pacemaker. We tested the
All of the Dregs were expressed in the heads of eyes absent three morning Dregs for their dependence o n each of
( c p ) mutant flies , indicating that none of these genes these potential influences. Surprisingly, the dependencies
are expressed exclusively in the compound eye. of each of the morning Dregs were different and com-
plex; these dependencies were also different from the
There was an interesting and unexpected correlation dependencies of per and of one of the evening genes (see
between the absolute levels of expression of the Dregs and Discussion). The results are summarized in Table 2 and
their teniporal expression patterns. The three morning are described in detail below.
genes were of relatively high abundance (- 0.05-0.1 %
head mRNA; Table 1). In contrast, the evening Dregs
were universally of low abundance - approximately Dreg-1 and Dreg-3 mRNAs continue to oscillate in the
100-1000 times less abundant than the morning Dregs. absence of a light-dark cycle
The majority of evening niRNAs were also found to be We first tested whether Dveg-1, Dreg-2, Dreg-3 and per
large transcripts, with all except Dveg-5 and Dreg-15 mRNAs would oscillate in abundance in the absence of a
being 6 kb or greater in length. light-dark cycle. Dveg- 1, Dveg-3 and per mRNAs contin-
ued to oscillate with constant period and slightly reduced
Oscillations of the morning Dregs show distinct, complex amplitude in the transition from light-dark cycle to total
dependencies on light, feeding time and the per gene darkness (Fig. 2); these genes therefore do not have an
Because the original screen for cycling genes used flies absolute requirement for light to drive their circadian
maintained in a light-dark cycle, we expected to identie cycling. Dveg- 6 to Dreg- 10 and Dveg- 15 also continued to
mRNAs that are responsive to light as well as those that cycle in the transition from the light-dark cycle to total
fluctuate with an endogenous circadian rhythm, indepen- darkness (data not shown). (We have not, however, elimi-
dent of external time cues (Zeitgebers). The flies that nated the possibility that the cycling observed in these
were used in the initial screening experiments were experiments is due to ‘after effects’ of the lighting cycle
subject to three identifiable sources for their daily ). In contrast, the diurnal oscillation of Dreg-2
rhythmicity: the lighting cycle, the daily feeding time m R N A was dependent on the light-dark cycle, in an
Table 1. Size, abundance and distribution of Dreg mRNAs.
Gene cDNA size Size of poly-A+ Estimated fraction Expressed in body* Expressed in 24 h Sequence#
(bp)* mRNAs (kb) of head mRNA (“/o)+ embryos
Dreg- 7 350 1 .o 0.2 Yes No Adh
Dreg-2 465 1 .o 0.1 Yes No Novel
Dreg-3 31 0 2.7 0.05 No Yes Nove I
Dreg-5 385 1.8 0.001 No Yes Novel
Dreg-6 31 0 7.0 0.0005 Weakly No New
Dreg-7 3 60 8.2 0.0005 No Yes New
Dreg-8 540 7.4 0.0005 No Yes New
Dreg-9 800 7.0 + 0.0005 Weakly No New
Dreg- 7 0 460 6.9 + 0.0001 No No New
Dreg- 1 1 480 6.6 + 0.0005 No Yes New
Dreg- 12 290 9.0 + 0.00005 No Yes New
Dreg- 13 640 10.0 0.00005 No Yes New
Dreg- 14 375 6.2 0.00005 No Yes New
Dreg- I5 650 3.5 0.00005 No Yes New
Dreg- 16 350 7.3 0.00005 No Yes New
Dreg- 7 7 900 10.0 0.00005 No Yes New
Dreg- 7 8 100 14.0 + 0.00005 No Yes New
Dreg- 7 9 960 10.0 0.00005 No Yes New
Dreg-20 560 6.0 0.00005 No Yes New
Dreg-2 1 340 6.9 + 0.00005 No No New
*Size of Xbal-EcoRI cDNA insert in the original clones from the lambda-SWAJcDNA library. +Abundance was estimated from
exposure times required to obtain a signal approximately equal to ninaE, taking into account differing probe lengths. ninaE
represents -1 o/o of fly head mRNA [471. *Body RNA assessments were made on 2 vg poly-A+ body mRNA prepared from the same
flies as used for the fly head Northern blots, pooled for all time points. §See [281.#DNA sequence (>200 bp) was obtained from
each c D N A insert and w i l l be deposited in GeneBank. ’New’ indicates no matches were found in GeneBank for the partial cDNA.
‘Novel’ indicates no closely related proteins were found for the coding sequence of a complete cDNA.
1428 Current Biology 1995, Vol 5 No 12
I Table 2. Comparison of the cycling behavior of the
Dregs under different environmental and genetic conditions.
Gene Cycles in LD Cycles in DD Changes phase with Cycles in LD Cycles in LD Cycles in D D
in caged flies in caged flies altered feeding time in bottled flies in bottled in bottled
in caged flies pep flies p e p flies
Dreg- 7 +
LD: a lighting regimen o repeated 12-hours light followed by 12-hoursdark. DD: constant darkness.
interesting manner. This m R N A displayed increased of all three morning Dregs shifted to later in the day
expression from Z T 20-26 in the first cycle of total dark- ( Z T S), and the trough of expression shifted to Z T 2
ness. However, the transcript levels of Dreg-2 remained (Fig. 3). In contrast, this change had no effect on the
high and without fluctuation throughout the remainder phase or amplitude ofper m R N A cycling (Fig. 3) or on
of the cycle. This suggests that morning light is required the phase of expression of Dreg-5 to Dreg-9 ( and data
to effect the evening decline in Dreg-2 transcript levels not shown). These data demonstrate that the timing of
under these conditions, and that Dreg-2 expression is not the placement of the food trays is sufficient to alter the
driven directly by the endogenous circadian pacemaker. phase of expression o f all three morning Dregs, and to
dissociate their relative phase from that of the per m R N A
The phase of morning Dreg expression is altered by the and several other cycling genes.
time of feeding
Flies kept in large cages require regular replenishment of To determine whether the timed introduction of fresh
food. In our screen, a fresh food tray was added each day food was required for the cyclic expression of the
at Z T 17, during the subjective night (although some morning Dregs, we raised flies in small bottles with abun-
food remained available to the flies at all times). To deter- dant food and allowed them to feed ad libitum. Dreg-1,
mine whether the addition of fresh food acted as a Dreg-3 and per continued to demonstrate diurnal cycling
Zeitgeber for the expression of the Dregs, we added fresh in these conditions (Fig. 4a,c,d), but the amplitude of
food trays at ZT 1 instead of Z T 17. The peak expression oscillation was reduced by about half. Thus, although the
timing of food availability can alter the phase of Dreg-1
and Dreg-3 rhythmicity, it is not necessary for manifest
rhythmicity. Dreg-2 did not show significant oscillations
Dreg- 1 in flies housed in small bottles (Fig. 4b), and therefore
requires both a light-dark cycle and timed food availabil-
ity for its rhythmicity in wild-type flies.
In a light-dark cycle, the function of the per gene is
required for the oscillation of Dreg-1 but not Dreg-3
O n e of the goals of our screen was to identi@ genes that
are controlled by the per locus. We therefore tested
whether the morning Dregs would oscillate in per null
mutants. T h e pero allele used here, perD', is a point
Per mutation that eliminates per function and produces flies
with no apparent circadian rhythms . The temporal
expression patterns of the morning Dregs and per in
rp4 9 wild-type Canton-S and mutant y pero1 strains in LD
12:12 are shown in Figure 4. T h e per m R N A was
expressed in the mutants but remained at a constant
Poly-A level throughout the circadian cycle (Fig. 4d), confirm-
2 8 14 2 0 2 6 3 2 35 44
ing the results of Hardin et al. [ l 11. Similarly, the oscil-
Time after ZT 0 (hours)
lation of Dreg-1 m R N A ceased in the pero flies,
indicating that circadian oscillation of Dreg- 1 also
Fig. 2. Role of light in the oscillating expression o the morning requires wild-type per function (Fig. 4a). In contrast,
Dregs and per f
mRNA. Northern blots are shown o total head Dreg-3 oscillated in a similar manner in the presence or
RNA prepared from ilies housed in large population cages as in
Figure 1 , and sacrificed at the indicated times during the absence ofper function (Fig. 4c). This transcript thus has
transition from LD12:12 lighting conditions into a cycle of no dependence o n the per gene for its oscillation under
constant darkness. these conditions.
Circadian gene expression in Drosophila Van Gelder et a / . RESEARCH PAPER 1429
Fig. 3. Effect of altered feeding time on
the cycling mRNA expression of the
morning Dregs and per. Northern blots
of total head RNA from flies maintained
in LD 12:12 as in Figure 1, except that
for half of the cages food trays were
swapped at ZT 1 (right panels) instead
of the standard time of ZT 17 (left pan- Dreg-2
els). Arrowheads indicate the time of
food-tray swapping. Quantification of
the expression of each gene (as
described in the legend to Fig. 4)
showed that signal amplitudes did not Dreg-3
differ significantly under the t w o feeding
2 8 14 20 2 8 14 20
Zeitgeber time (hours)
The results with Dreg-2 in this experiment were more circadian cycling of these cDNAs in per' flies kept in con-
complex. In wild-type flies maintained in small bottles stant darkness (DD). Neither Dreg-2 nor Dreg-3 expres-
and in a LD 12:12 lighting cycle, this gene did not show sion oscillated under these conditions (data not shown).
significant oscillation (Fig. 4b). Surprisingly, Dreg-2 did The cycling of these genes in per' mutants is therefore
show weak rhythmicity in per' flies kept in the same con- attributable to an effect of light on their expression that is
ditions, with increased expression in the early day (Fig. independent of the effect of the per gene.
4b). Taken together with the previous results for Dveg-2,
this indicates that the diurnal control of Dreg-2 expression Sequence analysis of the Dreg genes
is a complex function of the interplay between light-dark As an initial step towards understanding the functions of
cycle, per gene function and feeding conditions. the Dveg genes, we obtained partial sequences of the
entire set of Dregs and complete coding sequences for the
In addition to its effects on the oscillatory behavior of the three morning Dregs. All of the original Dreg cDNA
Dregs, per also affected the overall level of expression of clones were relatively short (-200-1000 bp; see Table 1)
some of the Dregs. Significantly more Dreg- 1 transcript and probably represent the 3' ends of their respective
was found in wild-type flies than in per' flies at all times transcripts, as each is terminated by a poly-A sequence.
of the day, and particularly at those times when Dreg-1 We sequenced the ends of each Dreg cDNA clone to
was maximally expressed (Fig. 4a). Hence, per appears to establish identities with previously cloned Drosophila
potentiate the expression of Dreg- 1. In contrast, wild- genes. Only one of the 20 Dreg genes corresponded to a
type per suppressed Dveg-5 m R N A expression at all times known gene: the Dreg-1 sequence precisely matched the
of the day , and it potentiated and suppressed per 3' portion of the Dvosoplzila alcohol dehydrogenase (Adlz)
m R N A expression at different times of the day (Fig. 4d). gene, establishing that Dreg-1 is Adh (Table 1). The re-
N o significant effects of per genotype on the overall levels maining 19 Dregs did not show sequence identity with
of expression of Dreg-2 or Dveg-3 were detected. Thus, any cDNA or genomic sequences in the GeneBank or
per. affects the expression of several cycling genes, but it EMBL databases, indicating that all of these Dregs
appears to do so by modulating their patterns of expres- represent newly identified Drosophila genes (Table 1).
sion in different ways. This conclusion should be consid-
ered provisional, however, as the pero and wild-type flies To obtain the complete coding sequence of the two other
used in this experiment were not fully isogenic and we morning genes, full-length Dreg-2 and Dveg-3 cDNA
cannot exclude the possibility that other genetic clones were isolated from adult Drosophila cDNA libraries
differences between the strains may have contributed to and the sequences of the inserts were determined. The
differences in expression levels. sequence of Dveg-2 is shown in Figure 5a. The Dreg-2
cDNA encodes a predicted 260-residue protein of novel
To determine whether the observed oscillation of Dreg-2 sequence. Searches of GeneBank and EMBL databases
and Dreg-3 expression in per' flies was driven by the LD with the BLAST, Blaze and FastDB programs revealed
12:12 lighting conditions or reflected the presence of a no significant homologies, except for a weak homology
per-independent endogenous oscillator, we tested the to a conserved region of phenylalanine-ammonia lyase
1430 Current Biology 1995, Vol 5 No 12
(Fig. 5b). The Dreg-3 cDNA encodes a predicted 626-
residue protein (Fig. 6a). Residues 548-591 of Dreg-3
$ 1.0- show two Cys-X-X-Cys-X-X-Cys-X-X-X-Cys motifs
? found in the 4Fe-4S family of iron-sulfur-binding pro-
$ 0.8- teins [33,34]. Dreg-3 also contains a short region (resi-
.E 0.6- dues 378-394) with significant homology to the putative
m flavin adenine dinucleotide (FAD)-binding domain of
E 0.4- dihydroorotate dehydrogenase  (Fig. 6b).
go.2! 1 I I I I I I I I
LL 2 8 14 20 2 8 14 20
Zeitgeber time (hours)
In this study, we have screened a large, heterogeneous
collection of anonymous cDNA clones for those that
show daily rhythmic expression. We identified 20 genes
that showed significant cycling, only one of which
(Dreg-2)corresponds to a gene that has been cloned pre-
viously. We further characterized the three genes (Dreg-1,
Dreg-2 and Dreg-3) that are maximally expressed at pre-
1 1 I I 1 I I I I
dawn or early morning. Elsewhere, we examine in detail
U 2 8 14 20 2 8 14 20 one of the 17 evening genes identified in the screen,
Zeitgeber time (hours) Dreg-5, which is expressed strictly in phase with per
Dreg-3 m R N A . Comparison of the oscillatory behavior of
these four genes, as well as per, demonstrates a surprising
variety of diurnal regulatory controls: each of these five
cycling genes shows a unique dependence on the
endogenous circadian oscillator and on external time
0.2 I I I I I I I I I
cues, ranging from complete dependence on the per gene
to predominant dependence on external time cues.
The extent of oscillating gene expression in Drosophila
2 2 8 1 4 20 2 8 14 20 O u r screening experiment allows us to estimate the
Zeitgeber time (hours) prevalence of oscillating transcripts in the fly head. The
results indicate that a small but significant proportion of
transcripts specific to the adult fly undergo diurnal
rhythms of expression: 8 % of the genes in the collection
demonstrated greater than two-fold oscillations of
abundance. This estimate is based on flies housed in large
population cages and kept in LD 12:12 lighting con-
ditions, and it does not include genes that cycle in only a
C subset of the head tissues in which they are expressed -
such cycling would not have been detected in our screen.
Zeitgeber time (hours) O u r experiments further suggest that the number of
cycling genes in Drosophila is not fixed but rather differs
0 Wild-type flies
under different environmental conditions. Several genes
that we identified altered their cycling behavior or ceased
to cycle under conditions that were different from those
Fig. 4. Effects of a per null mutation on the cycling expression of used in the screen. For example, Dreg-2 m R N A levels
the morning Dregs. Canton-S flies and y p e P flies were entrained cycled in flies kept in LD 12:12 in large population cages,
for 4 days under LD 12:12 conditions in small bottles on corn- but did not oscillate in flies housed in small bottles under
meal-molasses agar covered with yeast. In the next cycle, head the same lighting regimen. Furthermore, initial experi-
RNA was prepared from the wild-type and mutant flies sacrificed
at the indicated circadian times. Individual northern blots of the
ments with the evening Dregs (Dreg-5 to Dreg-21) have
RNA were probed for (a) Dreg-I, (b) Dreg-2, ( c ) Dreg-3 and (d) shown that cycling in at least a subset of these genes is
per, and the level of each mRNA was quantified on a phospho- also highly dependent on environmental conditions.
rimager. Values shown are the mean and standard error of six Several evening Dregs showed robust cycling in large
independent experiments. Each point was normalized first to population cages, but oscillated with diminished ampli-
rp49 expression (to control for any differences in RNA loading)
and then to the maximum expression for that gene using the
tude or failed to cycle when the flies were kept in small
method of [1 11. The same data are plotted twice over t w o cycles bottles (R.N.V.G. and H.B., unpublished observations).
to facilitate comparison of waveforms. It also seems likely that some of the genes that did not
Circadian gene expression in Drosophila Van G e l d e r et a / . RESEARCH PAPER 1431
show cycling behavior in the population cages might ccg-1 and ccg-2, were identified [26,36], indicating that
oscillate in flies kept in other environmental conditions. circadian control of gene expression is a limited phenom-
Thus, our estimate of the extent of circadian gene enon in Neurospora, as in Drosophila. Another Neurospora
expression in Drosophila heads should not be interpreted gene, j e q u e n c y (frq), which has many functional similari-
as an absolute value but rather as a representative value ties and weak sequence similarity to per, has recently
for a standard set of environmental conditions. been shown to undergo autoregulated circadian oscilla-
tion in a manner analogous to that observed for per .
O u r estimate of the prevalence of rhythmically expressed Unlike per, however, j q m R N A is maximally expressed
genes in Drosophila heads differs dramatically from the during daytime. Both ccg-1 and ccg-2 also show maxi-
result of the only other large-scale survey of circadian mum expression during daytime, and both are at least
gene expression that has been conducted. Liu et al.  partially dependent on f r q for their circadian rhythmicity
reported recently that nearly every random insert of a . Thus, in both the work of Loros et al.  on
reporter construct in the cyanobacterium Synechococcus Neurospora and our study on Drosophila, the extent of cir-
genome manifests significant circadian rhythmicity in cadian control of gene expression appears to be limited to
expression. O u r results show that the Drosophila genome a relatively small proportion of the genome, and the
is not subject to such global circadian control, and majority of identified cycling transcripts are expressed in
that the result from Synechococcirs is unlikely to be general. the same phase as the transcripts of a gene known to be
Some of the difference between the estimates for the two essential to the function of the circadian clock.
organisms may be attributable to biases in the sets of
genes surveyed, to the different assays of gene expression The diverse character of cycling gene expression in
used, or to specific environmental conditions used in the Drosophila
experiments. But it is possible that much of the differ- Diurnal control of gene expression in Drosophila is not a
ence is real and reflects extreme differences in the circa- unitary phenomenon. Diurnal control of the five genes
dian control of gene expression, behavior and physiology that have been analyzed in detail is remarkably varied
between the two organisms. It should be noted that our (Table 2). O u r screen initially identified two major classes
data do not exclude the possibility that specific cells in of rhythmically expressed genes - those expressed at
Drosophila (such as the central circadian pacemaker cells) peak levels in the pre-dawn or early morning (Dreg-1 to
show a more widespread circadian regulation of gene Dreg-3), and a much larger set (Dreg-5 to Dreg-21)
expression, as in Syrzechococctis. expressed at peak levels in the evening like per mRNA.
Subsequent circadian characterization of the three
The only other systematic search for circadianly regulated morning genes revealed, however, that their oscillatory
transcripts used a subtractive hybridization approach in behavior was heterogeneous and complex, and that each
the mold iVeurospora c~assa.Two cycling transcripts, called was differentially dependent both on the endogenous,
M R S L 4
S R F R L I T F D V T N T L L Q F R T T 24
P G K Q Y G E I G A L F G A R C D N N E 44
L A K N F K A N W Y K M N R D Y P N F G 64
R D T N P Q M E W Q Q W W R K L I A G T 84
F A E S G A A I P D E K L H N F S N H L 104
I E L Y K T S I C W Q P C N G S V E L L 124
Q Q L R K E L K P E K C K j
L G V I A N F 144
D P R L P T L L Q N T K L D Q Y L D F A 164
I N S Y E V Q A E K P D P Q I F Q K A M 184
E K S 0 L K N L K P E E C L H I G D G P 204
Fig. 5. (a) Sequence of Dreg-2. The T T D Y L A A K E L G W H S A L V H E K 224
nucleotide sequence of a full-length 781 AGCTACGCATATCTGGTCAAGAAATACGGCGAGGACATCGATCGAGATCATGTCTTCCCC
S Y A Y L V K K Y G E D I D R D H V F P 244
Dreg-2 cDNA is shown along with the 841 AGTCTCTACGACTTCCACAAAAAGATCTCCGACGGCGCAGTTGTCTGGtgattgattgta
deduced amino-acid sequence of Dreg-2 S L Y D F H K K I S D G A V V W * 260
protein. The initiation codon was taken 901 cacacattaaattaataacaccaaaaaaaaaaaa
as the first methionine in the long open-
reading frame. The region of homology (b)
to phenylalanine-ammonia lyase is 156 K L D Q Y L - D F A - I N 166 179 I F Q K - A M E K S G L K N L K P E E 1 9 6 Dreg-2
shown in a grey box. (b) Comparison of KLRQVLVDHALVN IFQKIAIFEEELKNLLPKE Soybean FA lyase
the Dreg-2 protein sequence to KLRQVLVDHALVN IFQKIATFEDELKTLLPKE Pea FA lyase
sequences of phenylalanine-ammonia KLRQVLVDHALIN IFQKIATFEEELKTILPKE Human kidney FA lyase
lyase (FA lyase) in various organisms.
1432 Current Biology 1995, Vol 5 No 1 2
per-dependent circadian oscillator and on external time also investigated the oscillatory behavior of one of the
cues of feeding time and light-dark cycle. O f the three evening genes identified in the screen (Dreg-5), whose
morning genes, only Dreg- 1 was strictly dependent on m R N A is expressed precisely in phase with pev m R N A
the per gene for its rhythmic expression. However, its . Dveg-5 oscillation is dependent on per gene function,
phase of expression could be dissociated from that of per and its phase of expression could not be dissociated from
m R N A by altering the feeding time. In contrast, Dreg-3 that of pev m R N A under altered environmental condi-
m R N A was capable of robust oscillation in the absence of tions (see Table 2). Dreg-5 protein, however, oscillates
per gene function in flies maintained in a light-dark cycle, with a phase markedly different from that reported for
but was also capable of weak oscillation in the absence of Per protein .
external time-cues in the presence of wild-type per gene
function. Dreg-2 oscillation was the most complex The five cycling Dvosophilu genes that have been studied
observed, being dependent in a complex way on the in detail therefore demonstrate five different rhythms
presence of a light-dark cycle, on timed feeding and, to a and/or dependencies on internal and external oscillatory
lesser extent, on the function of the per gene. We have cues. Although the very low abundance and large size of
32 1 acgcaaagtgatcgtcaaggatggcctgatcacagccATGGAATTCTGTCGCACGGAGCAGAACGAAAACGATGAATGGG
M E F C R T E Q N E N D E W V 15
E D E E Q T Q R L K A N F V I S A F G S G L E D Q D 41
V K A A L A P L Q F R G E L P V V D R V T M Q S S V K 68
Q V F L G G D L A G V A N T T V E S V N D G K V A A W 95
S I H C Q L Q G L P L D T P A A L P L F Y T D I D A 121
72 1 GTGGACATATCGGTGGAGATGTGCGGCATCCGGTTCGAGAATCCCTTTGGCCTGGCCTCCGCACCGCCCACCACCAGCAC
V D I S V E M C G I R F E N P F G L A S A P P T T S T 148
A M I R R A F E Q G W G F V V T K T F G L D K D L V T 175
N V S P R I V R G T T S G Y K Y G P Q Q G C F L N I 201
E L I S E K R A E Y W L K S I G E L K R D F P E K I V 228
I A S I M C S F N E E D W T E L A I K A E Q S G A D A 255
L E L N L S C P H G M G E R G M G L A C G Q D P E L 281
V E Q I S R W V R K A V K L P F F I K L T P N I T D I 308
V S I A A A A K R E E P M R S A I N T V Q G L M G L K 335
A D S T A W P A I G K E Q R T T Y G G V S G N A T R 361
P M A L K A I S D I A N R V P G F P I L G I G G I D S 388
G E V A L Q F I H A G A T V L Q I C S S V Q N Q D F T 415
V I E D Y C T A L K A L L Y L K R I R H Q S M V P S 441
G M A S H H P R R S I R R Q A R C P F D R R G K A T L 468
G F F G P Y Q R Q R D I K M A E L R S Q K G A L S W D 495
A E Q V K A T P P A S N G A P N P A P R I K D V I G 521
A A L D K I G S Y N K L D N K Q Q K V A L I D D D M C - 548
I N C G K C Y M T C A D S G Y Q A I E F D K D T H I P 575
H V N D D C T G C T L C V S V C P I I D C I T M V P 601
K K I P H V I K R G V E E K I F Y T H A L S Q C Q * 626
P I L G I GG I D SGE VAL QF I H AGATV L Q Dreg-3
P I I GVG GV S SGQ N A L E K I R A G A S L VQ Human D H O D
P I I GVG G I D SV IA A R E K I AAGA SL V Q €.coli D H O D
. Fig. 6. (a) Sequence of Dreg-3. The nucleotide sequence of Dreg-3 derived from clone pDREG3BS is shown along with the deduced
amino-acid sequence of Dreg-3 protein. The initiation codon was selected as the first methionine codon in the long open reading
frame. The region of homology to the FAD-binding domain of dihydroorotate dehydrogenase is shown in a grey box. The t w o cysteine-
rich regions of homology to 4Fe-4S iron sulfur binding domains are underlined. (b) Comparison of the Dreg-3 protein sequence to the
sequences of dihydroorotate dehydrogenase ( D H O D ) in various organisms.
Circadian gene expression in Drosophila Van Gelder et a / . RESEARCH PAPER 1433
the other evening Dregs have hampered our efforts at diurnally active, and feed on decaying fruit during the
similar characterization of these genes, preliminary day. In both flies and mammals, it may be beneficial to
experiments suggest that they too are diverse, but that a trigger production of the detoxification enzyme prior to
subset may behave as a more homogeneous group. the actual feeding event.
Dreg-6 to Dreg-10 and Dreg-15 behaved like Dreg-5 in
that they continued to cycle in constant darkness, and As flies are active during the day, we were surprised to
oscillated independently of feeding time cues. Prelimi- find that the vast majority (85 %) of cycling transcripts
nary experiments suggest that the cycling of some of were maximally expressed in the evening. Why so many
these genes is dependent on per function (R.N.V.G. and of the cycling genes are maximally expressed when the
H.B., unpublished observations). A subset of the evening flies are behaviorally'inactive is unclear. It is also intrigu-
Dregs, including Dreg-6, Dreg- 7 and Dreg-9, also appeared ing that almost all of the cycling genes we identified are
to be very sensitive to caging conditions in a manner large transcripts of low abundance (see Table 1).
analogous to Dreg- 1.
Given that there is some advantage to restricting gene
O u r experiments establish that per influences the expression to particular times of the day, why do such
circadian expression of genes other than itself. We have varied and multiple mechanisms exist to achieve this end?
also found, however, that only a subset of rhythmically For example, the per gene appears to be essential for the
expressed genes are tightly coupled to per. For each of diurnal expression of Dreg-1 and Dreg-5, but is dispens-
the morning genes it appears that per functions in con- able for diurnal Dreg-3 expression. A similar phenome-
junction with extrinsic cues to determine the circadian non is seen at the behavioral level. Diurnal rhythms of
expression patterns. In contrast, cycling of the evening activity are observed both in wild-type flies kept in con-
gene Dreg-5 is tightly coupled to per . It will be of stant darkness and in pero flies kept in a light-dark cycle
interest to dissect the regulation of some of the Dreg [2,39]; two independent mechanisms clearly exist to gen-
genes in biochemical detail to determine whether there erate diurnal rhythmicity of behavior in the fly. Although
is a direct influence of Per on the expression of these the present set of experiments do not answer this ques-
genes, and how control by per and by various extrinsic tion, two hypotheses can be suggested. Firstly, although
cues are integrated to generate different patterns of exogenous cues may be adequate to drive many aspects
rhythmic gene expression. It will also be important to of circadian gene expression, certain genes may benefit
further characterize the cycling profiles of the other from regulation within a window of time during the day
evening genes, to learn how extrinsic cues and per influ- that is effectively constant with respect to environmental
ence their expression, and to obtain a refined estimate of conditions, or for which the environmental cues are not
the number of 'clock-controlled genes' whose cycling is highly dependable. Such genes may rely more on the
tightly linked to per. function of an endogenous circadian pacemaker than o n
exogenous cues. A second hypothesis is that certain genes
Roles of cycling gene expression may rely o n both per function and external cues to mod-
Why are certain genes expressed with diurnal rhythms? ulate their expression in a circannual fashion. The over-
Some, like per, may be part of the internal time-keeping lapping influences of the circadian pacemaker and photic
apparatus. For others, there is presumably some selective stimulation in modulating individual gene expression
advantage to restricting expression to particular times of could therefore be the mechanism of photoperiodic
day. Many plant genes, including the chlorophyll-binding induction of such circannual events as ovarian diapause in
proteins, catalase and others, have been shown to certain species of Drosophila . Left unexplained by
undergo diurnal regulation of expression [ 191; in these either hypothesis, however, are the genes like Dreg-2 that
cases, the peak levels of expression are during the day show rhythmic expression patterns that are dependent o n
when sunlight is available for photosynthesis. If a selec- complex interactions between per and the various
tive advantage is created by restricting expression of a environmental controls.
gene to a particularly advantageous time, one would
expect diurnally gated gene expression of particular The functions of most of the Dreg genes are not apparent
genes to be conserved across phylogeny. In this regard, it from their sequences. That 19 of the 20 oscillating
is interesting that Dreg-1 was determined to be the alco- cDNAs appear to represent new Drosophila genes suggests
hol dehydrogenase gene (Adz), alcohol dehydrogenase
as that these genetic functions, which may be either part of
activity is expressed with circadian rhythmicity in mam- the clock or functions that are advantageously seques-
nials [24,38]. This is the first gene identified with a circa- tered to particular times of day, have not been addressed
dian oscillation in expression that is conserved across at a molecular level. The biochemical roles of Dreg-2 and
phylogenic classes. However, it should be noted that in Dreg-3 are not revealed by their complete coding sequen-
the nocturnal mammals, where circadian variation of the ces as neither of these gene products shows substantial
enzyme activity has been described, alcohol dehydroge- similarity to those of previously cloned genes, and only
nase activity is maximal a t the end of the subjective day Dreg-3 protein has any clearly recognizable functional
and the beginning of the subjective night. In contrast, motifs. Dreg-3 contains two iron-sulfur-binding motifs
Drq-1 (Adh) m R N A was found to be maximally and a potential FAD-binding region. Such domains are
expressed during the early day. Flies living in the wild are found in a number of reductases. including ferredoxin
1434 Current Biology 1995, Vol 5 No 12
and cytochrome systems, as well as in the transferrin entire cage followed by quick-freezing in liquid nitrogen. Two
mRNA-binding protein cis-aconitase, suggesting that cages of flies were collected for each timepoint.
Dreg-3 may have a role in either metabolism or gene
regulation [413, For smaller scale experiments, 250-300 mg of flies were
maintained at 24 "C and 55 % relative humidity in each 200 ml
The collection of Dreg genes will be valuable in further plastic bottle containing cornmeal-molasses agar medium over-
laid with yeast. LD 12:12 was provided by either incandescent
studies of the mechanisms of circadian rhythms in or fluorescent lighting alternating with complete darkness in
Drosophila They can provide molecular genetic markers light-tight incubators. Flies were transferred to fresh bottles
of events that lead from the central circadian pacemaker three to four days prior to collection, and were collected seri-
to the expression of behavior. Analysis of their oscillatory ally from bottles kept in the same incubator. 10-15 bottles of
behavior in each of the different Drosophila mutants that each genotype were collected at each time point under 10 W
affect circadian rhythms and behavior should allow red-filtered lighting. Flies were collected directly onto dry ice.
ordering of the Dregs with respect to these genes and
elucidation of the genetic pathway for circadian rhythms RNA purification and northern blotting
in Drosophila. Establishing the functions of the Dreg genes Frozen fly heads were separated from antennae and other body
by mutational analysis and by characterization of their parts by vigorous vortexing and then purified by sieving over
protein products will also be invaluable in determining #25 and #40 US Standard brass sieves. Any remaining body
parts were removed manually. Total RNA was prepared from
the role of the circadian expression of these genes and
2 g fly heads per timepoint using chaotropic salt homogeniza-
their functions in generating circadian behavior and
tion and LiCl precipitation . Poly A+-containing RNA was
physiology. purified by oligo(dT) cellulose chromatography . RNA
concentrations were determined by A,, spectrometry.
Conclusions For northern blots, 15 or 20 pg total RNA was loaded per lane
on a 0.8 % agarose gel containing 2.2 M formaldehyde and run
O u r large scale survey of the daily expression patterns of in MOPS buffer . RNA was capillary blotted overnight to
individual Drosophila genes identified 20 genes that show Nytran membrane and bound to the membrane by UV
significant daily rhythms of gene expression. Circadian crosslinking at 1200 pJ per cm2, Hybridization conditions
control of gene expression is therefore a significant but were: 50 % formamide, 6X SSC, 5~ Denhardt's reagent, 0.2 %
limited phenomenon in the fly head. Circadian gene SDS, denatured salmon sperm DNA at 100 pg ml-l and cRNA
probe at 2 X 10, cpm d - l . Hybridizations were carried out at
expression is not a unitary phenomenon but rather is
65 "C in a rolling-bottle incubator for at least 12 h. The final
mediated by multiple mechanisms, which include the wash was in 0 . 1 ~ SSC, 0.2 % SDS at 65 O C . Blots were
timing of the daily light-dark cycle, the timed availability exposed to Kodak XAR film with intensifjring screen or
of food and the function of the per gene and endogenous quantified on a Molecular Dynamics Phosphorimager using
oscillator. Although a subset of rhythmically expressed ImageQuant v. 3.22 software.
genes is primarily or exclusively dependent on per gene
function, many cycling genes appear to depend on all Production of cRNA probes
three types of input for their daily expression patterns. PCR amplification of the cDNA insert in each phage stock
Al but one of the 20 cDNAs appear to identifjr new from the head-not-embryo cDNA collection  was per-
genes. These cycling genes provide a valuable new mol- formed using -3 x lo6 pfu phage and 30 pmol of the primers
ecular tool for elucidating the genetic pathway from per SWAJ T7 (5' TCG AAA TTA ATA CGA CTC ACT ATA
and the central time-keeping apparatus to the circadian GGG) and SWAJ SP6 (5' ACA CAT ACG ATT TAG GTG
behavior and physiology that it controls. ACA CTA TAG). SWAJ T7 contains a bacteriophage T7 pro-
moter and SWAJ SP6 contains an SP6 promoter. Standard
PCR conditions  were used for 30 cycles of 1 min denatu-
ration at 94 "C, 2 min annealing at 55 OC, and 3 min extension
Materials and methods at 72 "C. Products were separated on a 1.5 % low-melting-
temperature agarose gel (SeaPlaque GTG, FMC Biochemicals)
Fly strains, maintenance and collection in Tris-acetate buffer. The PCR product was isolated as an
D. rnelanogastev strains used in these experiments were either excised gel fragment. The cRNA probe was synthesized in a
Canton-S, or y pep' derived from an outcross of y per0', vy506 20 pl reaction containing 7 pl melted gel slice and T7 RNA
to Canton-S. For the large-scale screen, 60 g (- 6 x lo4) wild- CTP
polymerase with C X - [ ~ ~ P ] ~under transcription conditions,
type Canton-S flies were seeded at one week of age into each as described . The resultant cRNA probe (-1 X lo9 cpm
-20 1 plexiglass population cage, and maintained in LD 12:12 per pg cRNA) was purified on a Sephadex G50 spin column.
with overhead 60 W tungsten incandescent lighting during
light phase, and 15 W tungsten light with a red (Wratten 1A) To generate probes for ~ p 4 9 mRNA, the 640 bp
filter during dark phase. Cages were kept at 25 "C and 55 % EcoRI-Hind111 fragment extending from the 5' end of the rp49
humidity in a double-door isolation room. Flies were fed transcript  was subcloned into pBluescript SK+ (Strata-
baker's yeast overlaid on apple-juice agar plates. Excess yeast gene). The resulting clone, pRP49BSSK, was cleaved with
was used such that some food remained on the plates for the EcoRI and transcribed in vitvo with T7 RNA polymerase as
full 24 h. Food trays were changed at Z T 17 under 15 W red- above. To generate probes for per mRNA, the full-length
filtered light. Following five days of entrainment under LD cDNA of the 'A' form of per, pCDperF+ , was digested
12:12 lighting, flies were collected by CO, anesthesia of the with PstI and EcoRI, and the resulting 0.6 kb fragment (base
Circadian gene expression in Drosophila Van Gelder et a / . RESEARCH PAPER 1435
pairs 3965-4523) was cloned into pBluescript SK+. The resul- 8. Huang Z, Edery I, Rosbash M: PAS is a dimerization domain com-
tant plasmid, pPER0.5SK, was cleaved with PstI and tran- mon to Drosophila period and several transcription factors. Nature
scribed as above. The rzinaE gene  probe was produced 9. Siwicki K, Eastman C, Petersen C, Rosbash M, Hall J: Antibodies to
from phage clone 1F9 of the head-not-embryo collection using the period gene product of Drosophila reveal diverse tissue distrib-
the conditions described above. ution and rhythmic changes in the visual system. Neuron 1988,
Cloning of full-length cDNAs and DNA sequencing IO. Zerr DM, Hall JC, Rosbash MI Siwicki KK: Circadian fluctuations of
period protein immunoreactivity in the CNS and the visual system
PCR products from the original phage cDNA clones of the of Drosophila. J Neurosci 1990, 10:2749-2762.
Dreg genes were subcloned into the XbnI and EcoRI sites of 1 1 . Hardin PE, Hall JC, Rosbash M: Feedback of the Drosophila period
pBluescript SK+, and sequenced by chain termination sequenc- gene product on circadian cycling of i t s messenger RNA levels.
ing using T7 and T3 primers. Full-length clones of Dveg-2 were Nature 1990, 343:536-540.
12. Edery I, Zwiebel L, Dernbinska MI Rosbash M: Temporal phospho-
obtained by screening an adult fly head cDNA library in the rylation of the Drosophila period protein. Proc Natl Acad Sci USA
lambda-EXLX vector  using the origmal cDNA clone as 1994,91:2260-2264.
probe. The full length cDNA was then in vivo subcloned into 13. Zwiebel LJ, Hardin PE, Liu X I Hall IC, Rosbash M: A post-transcrip-
pEXLX yielding p2H12EXLX. An ApI-Sac1 fragment con- tional mechanism contributes to circadian cycling of a per-B-
taining the entire cDNA was subcloned into pMOB to generate galactosidase fusion protein. Proc Natl Acad Sci USA 1991, 88:
p2H12MOB. The insert was sequenced on both strands using 14. Page T: Time is the essence: molecular analysis of the biological
Tn3 transposon-facilitated DNA sequencing . clock. Science 1994, 263:1570-I 572.
15. Takahashi J: Circadian clock genes are ticking. Science 1992,
A partial cDNA of Dveg-3 (pDREG3EXLX) was obtained 258:2 38-240.
from the same adult-head lambda EXLX cDNA library, using 16. Hall JC, Rosbash M : Oscillating molecules and how they move cir-
cadian clocks across evolutionary boundaries. Proc Natl Acad Sci
the Dreg-3 cRNA as probe. A full-length clone (pDREG3BS) USA 1993, 90:5382-5383.
was obtained from another adult-head cDNA library in the 17. Edery I, Rutilla JE, Rosbash M : Phase shifting of the circadian clock
vector lambda-ZAP (provided by T. Schwarz, Stanford by induction of the Drosophila period protein. Science 1994,
University) by high-density screening of the library using the 263 :2 37-240.
18. Takahashi J: Circadian-clock regulation of gene expression. Curr
partial cDNA as probe, followed by PCR from the 5’ end of Opin Genet Dev 1993, 3:301-309.
the initial cDNA into the vector polylinker to identifjl full- 19. Kay SA, Millar AJ: Circadian-regulated cab gene transcription in
length clones . Both cDNAs were sequenced completely higher plants. In Molecular Genetics of Biological Rhythms. Edited
on both strands, using a combination of transposon-facilitated by Young MW. New York: Marcel Dekkar; 1993:73-89.
DNA sequencing and directed sequencing with custom oligo- 20. Danon A, Mayfield S: Light regulated translational activators: iden-
tification of chloroplast gene specific mRNA binding proteins.
nucleotide primers. EMBO J 1991, 10:3993-4001.
21. Yang J, Cagarnpang F, Nakayarna Y, lnouye S: Vasoactive intestinal
Database searches were performed using the IntelliGenetics polypeptide precursor mRNA exhibits diurnal variation in the rat
Suite programs with the GeneBank and EMBL databases. suprachiasmaticnuclei. Brain Res Mol Brain Res 1993, 20:259-262.
2 2 . Uhl G, Reppert S: Suprachiasmatic nucleus vasopressin messenger
ilckrzowle~gerrieric~:We thank Bruce Hanlilton and Elliot RNA: circadian variation in normal and Brattleboro rats. Science
Meyerowitz for generously providing cDNA clones and libraries as 23. Schwartz W: A clinician’s primer on the circadian clock: its local-
well as technical advice. We also thank Jeff Hall, Michael Rosbash, ization, function and resetting. Adv lnt Medicine 1993, 38:81-106.
Toni Schwarz and Michael Young for providing clones, libraries 24. Sturtevant R, Garber S : Circadian rhythms of alcohol dehydrogenase
and Drosophih strains. We thank the members of the Krasnow and and MEOS in the rat. Proc Soc Exp Biol Med 1984, 175:299-303.
Dement laboratories for discussions and comments on this nianu- 25. Liu Y, Tsinorernas N, Johnson C, Lebedeva N, Golden S, lshiura M,
script. R.N.V.G. also thanks Dale Edgar for his support and et a/.: Circadian orchestration of gene expression in cyanobacteria.
Gene Dev 1995, 9:1469-1478.
Suzanne Dintzis for encouragement. R.N.V.G. was supported by 26. Loros J, Denorne SI Dunlap J:Molecular cloning of genes under the
a Medical Scientist Training Program grant of the National control of the circadian clock in Neurospora. Science 1989,
Institutes of General Medical Sciences. M.A.K. is supported by a 243:385-388.
Presidential Young Investigator Award. This work was supported 27. Aronson B, Johnson K, Loros I, Dunlap I: Negative feedback defin-
by a grant from the SmithKlineBeecham foundation. ing a circadian clock: autoregulation of the clock gene frequency.
Science 1994, 263:1578-1584.
28. Palazzolo MJ,Hyde DR, VijayRaghavan K, Mecklenburg K, Benzer
S, Meyerowitz E: Use of a new strategy to isolate and characterize
References 436 Drosophila cDNA clones corresponding to RNAs detected in
adult heads but not early embryos. Neuron 1989, 3:527-539.
1. Pittendrigh C: Circadian systems, 1. The driving oscillation and i t s 29. Bonini N, Leiserson W, Benzer S: The eyes absent gene: genetic
assay in Drosophila pseudoobscura. Proc Natl Acad Sci USA 1967, control of cell survival and differentiation in the developing
58~1762-I767. Drosophila eye. Cell 1993, 72:379-395.
2. Harnblen-Coyle M, Wheeler D, Rutila J, Rosbash M, Hall J: Behavior 30. Pittendrigh C: Circadian oscillations in cells and the circadian orga-
of period-altered circadian rhythm mutants of Drosophila in nization of multicellular systems. In The Neurosciences Third Study
1ight:dark cycles (Diptera: Drosophilidae). J Insect Behav 1 992, Program. Edited by Schmitt F, Worden F. Cambridge, MA: The MIT
5:417-446. Press; 1974: 437-458.
3. Pittendrigh C: Temporal organization: reflections of a Darwinian 31. Van Gelder RN, Krasnow MA: A novel circadianly expressed
clock-watcher. Annu Rev Physioll993, 55:16-54. Drosophila melanogaster gene dependent on the period gene for i t s
4. Konopka R, Benzer S: Clock mutants of Drosophila melanogasfer. rhythmic expression. EMBO 1, in press.
Proc Natl Acad Sci USA 1971, 68:2112-2116. 32. Smith R, Konopka R: Circadian clock phenotypes of chromosome
5. Sehgal A, Price J, Man B, Young M: Loss of circadian behavioral aberrations with a breakpoint at the per locus. Mol Gen Genet
rhythms and per RNA oscillations in the Drosophila mutant fime- 1981,183~243-251.
less. Science 1994, 263:1603-I 606. 33. Beinert H: Recent developments in the field of iron-sulfur proteins.
6. Vosshall L, Price J, Sehgal A, Saez L, Young M : Block in nuclear FASEB J 1990,4:2483-2491.
localization of period protein by a second clock mutation, timeless. 34. Knaff D: The photosystem I reaction centre. Trends Biochern Sci
Science 1994, 263:1606-1609. 1988, 13:460-461.
7. Liu X, Zwiebel L, Hinton D, Benzer S, Hall J, Rosbash M: The period 35. Rawls J, Kirkpatrick R, Yang J, Lacy L: The dhod gene and deduced
gene encodes a predominantly nuclear protein in adult Drosophila. structure of mitochondrial dihydroorotate dehydrogenase in
) Neurosci 1992, 12:2735-2744. Drosophila melanogaster. Gene 1993, 124:191-197.
1436 Current Biology 1995, Vol 5 No 1 2
36. Loros JJ, Lichens-Park A, Lindgren KM, Dunlap JC: Molecular genet- 44. Van Gelder R, von Zastrow M, Yo01 A, Dement W, Barchas 1, Eber-
ics of genes under circadian temporal control in Neurospora. In wine J: Amplified RNA synthesized from limited quantities of het-
Molecular Genetics of Biological Rhythms. Edited by Young MW. erogeneous cDNA. Proc Natl Acad Sci USA 1990, 87:1663-1667.
New York: Marcel Dekkar; 1993:91-121. 45. O’Connell P, Rosbash M: Sequence, structure, and codon prefer-
37. Arpaia G, Loros J, Dunlap J, Morelli G, Macino G: The interplay of ence of the Drosophila ribosomal protein 49 gene. Nucleic Acids
light and the circadian clock. Independent dual regulation of clock- Res 1984,12:5495-5513.
controlled gene ccg-2 (eas). Plant Physiol1993, 102:1299-I 305. 46. Citri Y, Colot H, JacquierA, Yu Q, Hall I,Baltimore D, et a/.: A fam-
38. North C, Feurers R, Scheuring L, Pauly I,Tsai T, Casciano D: Circa- ily of unusually spliced biologically active transcripts encoded by a
dian organization of thirteen liver and six brain enzymes of the Drosophila clock gene. Nature 1987, 326:42-47.
mouse. Am 1 Anat 1981,182:183-199. 47. O’Tousa J, Baehr W, Martin R, Hirsh 1, Pak W, Applebury M: The
39. Wheeler D, Hamblen-Coyle M, Dushay M, Hall J: Behavior in light- Drosophila ninaE gene encodes an opsin. Cell 1985, 40:839-850.
dark cycles of Drosophila mutants that are arrhythmic, blind, or 48. Palazzolo M, Hamilton B, Ding D, Martin C, Mead D, Mierendorf R,
both. J Biol Rhyth 1993, 8:67-94. et al.: Phage lambda cDNA cloning vectors for subtractive
40. Pittendrigh C: Circadian organization and the photoperiodic phe- hybridization, fusion-protein synthesis and Cre-loxP automatic
nomena. In Biological Clocks in Seasonal Reproductive Cycles plasmid subcloning. Gene 1990, 88:25-36.
Edited by Follett D, Follett B. Bristol: Wright; 1981 :1-35. 49. Strathman M, Hamilton B, Mayeda C, Simon M, Meyerowitz E,
41. Kaptain S, Downey W, Tang C, Philpott C, Haile D, Orloff D, et a/.: Palazzolo M: Transposon-facilitated DNA sequencing. Proc Natl
A regulated RNA binding protein also possesses aconitase activity. Acad Sci USA 1991, 88:1247-I 250.
Proc Natl Acad Sci USA 1991, 88:10109-10113. 50. Hamilton B, Palazzolo M, Meyerowitz E: Rapid isolation of long
42. Cathala G, Savouret J-F, Mendez B, West B, Karin M, Martial J, et cDNA clones from existing libraries. Nucleic Acids Res 1991, 19:
a/.: A method for isolation of intact, translationally active ribonu- 1951-1 952.
cleic acid. DNA 1983, 2:329-335.
43. Sambrook 1, Fritsch E, Maniatis T: Molecular Cloning: a Laboratory Received: 9 August 1995; revised: 9 October 1995.
Manual. Cold Spring Harbor, NY: Cold Spring Harbor Press; 1989. Accepted: 18 October 1995.