The ion channel narrow abdomen is critical for neural output of the Drosophila circadian pacemaker.

Abstract

Circadian clocks consist of transcriptional feedback loops housed in interdependent pacemaker neurons. Yet little is known about the neuronal output components essential for rhythmic behavior. Drosophila mutants of a putative ion channel, narrow abdomen (na), exhibit poor circadian rhythms and suppressed daylight activity. We find that NA is expressed in pacemaker neurons and induced expression within circadian neurons is sufficient to rescue these mutant phenotypes. Selective na rescue in distinct pacemaker neurons influences rhythmicity and timing of behavior. Oscillations of the clock protein PERIOD are intact in na mutants, indicating an output role. Pore residues are required for robust rescue consistent with NA action as an ion channel. In na mutants, expression of potassium currents and the key neuropeptide PDF are elevated, the latter consistent with reduced release. These data implicate NA and the pacemaker neural network in controlling phase and rhythmicity.

Full text

Page 1

Neuron, Vol. 48, 965–976, December 22, 2005, Copyright ª2005 by Elsevier Inc.DOI 10.1016/j.neuron.2005.10.030
The Ion Channel Narrow Abdomen Is
Critical for Neural Output of the
Drosophila Circadian Pacemaker
Bridget C. Lear,1Jui-Ming Lin,1J. Russel Keath,1
Jermaine J. McGill,1Indira M. Raman,1
and Ravi Allada1,*
1Department of Neurobiology and Physiology
Northwestern University
Evanston, Illinois 60208
Summary
Circadian clocks consist of transcriptional feedback
loops housed in interdependent pacemaker neurons.
Yet little is known about the neuronal output compo-
nents essential for rhythmic behavior. Drosophila mu-
tants of a putative ion channel, narrow abdomen (na),
exhibit poor circadian rhythms and suppressed day-
light activity. We find that NA is expressed in pace-
makerneuronsandinducedexpressionwithincircadian
neurons is sufficient to rescue these mutant pheno-
types. Selective na rescue in distinct pacemaker neu-
ronsinfluencesrhythmicityandtimingofbehavior.Os-
cillations of the clock protein PERIOD are intact in na
mutants, indicating an output role. Pore residues are
required for robust rescue consistent with NA action
as an ion channel. In na mutants, expression of potas-
sium currents and the key neuropeptide PDF are ele-
vated,thelatterconsistentwithreducedrelease.These
dataimplicateNAandthepacemakerneuralnetworkin
controlling phase and rhythmicity.
Introduction
Daily behaviors are governed by circadian clocks. Cen-
tral to these circadian pacemakers are transcriptional
feedback loops. In Drosophila, the transcription factor
CLOCK (CLK) and its partner CYCLE (CYC) activate
transcription of the circadian rhythm genes, period
(per) and timeless (tim) (Stanewsky, 2003). PER nega-
tively feeds back to repress this CLK-CYC activator.
This feedback loop is essential for both behavioral and
molecular rhythms. A second feedback loop has been
identified that regulates Clk transcription (Allada, 2003;
Cyran et al., 2003; Glossop et al., 2003). Mammalian
pacemakers consist of a conserved set of core compo-
nents essential for its circadian feedback loops and be-
havior (Stanewsky, 2003). Changes in the phase and
periodicity of these molecular oscillators are thought to
largely determine the phase and periodicity of circadian
behavior. Despite the progress in understanding these
coreoscillators,littleisknownaboutthecomponentses-
sential for these feedback loops to control neural func-
tions.
InDrosophila,theseclocksarehousedinneuronalclus-
ters essential for behavioral rhythms (Helfrich-Forster,
2003). PER is expressed in neurons of the lateral proto-
cerebrum (lateral neurons; LNs) subdivided into ventral
(LNv) and dorsal groups (LNd) (Hall, 1998; Stanewsky,
2003). LNv neurons are distinguished by expression of
Pigment-Dispersing Factor (PDF), a neuropeptide es-
sential for robust circadian behavior (Helfrich-Forster,
1995; Park and Hall, 1998; Renn et al., 1999). PER is
also found in three dorsal neuronal clusters (dorsal neu-
rons; DN) (Kaneko et al., 1997). During 12 hr light: 12 hr
dark (LD) cycling conditions, flies typically exhibit in-
creases in locomotor activity in anticipation of light tran-
sitions. These morning and evening activity bouts have
distinct anatomical requirements. Tissue-specific res-
cue and ablation experiments suggest that the LNv are
responsible for morning anticipation, whereas the LNd
and/or DN1 are responsible for evening anticipation
(Grima et al., 2004; Stoleru et al., 2004). In addition, per
expression in the LNv is necessary and sufficient to
maintain rhythmic locomotor activity in constant condi-
tions (Grima et al., 2004; Stoleru et al., 2004).
Additional cells likely contribute to these rhythmic be-
haviors. disco mutant flies lack both LNvand LNd yet re-
tain evening anticipation, suggesting a role for the DNs
(Hardin et al., 1992). Moreover, studies of Clock and cy-
cle mutants indicate a broader circadian requirement
(LNv and LNd at a minimum) to rescue DD rhythmicity
(Allada et al., 2003; Peng et al., 2003). These neural clus-
ters may form reciprocal connections between ipsilat-
eral small LNv and DN1 groups as well as contralateral
LNv (Kaneko and Hall, 2000; Klarsfeld et al., 2004; Veleri
et al., 2003). Despite the importance of the LNd and
DN groups, no molecules essential for neuronal outputs
have been shown to function in these cell groups.
Mutants of a conserved Drosophila ion channel, nar-
row abdomen (na)/Dma1U, display altered behavior in
both LD and DD conditions (Nash et al., 2002). na mu-
tants were discovered on the basis of their abdominal
phenotype as well as altered responses to the general
anesthetichalothane(KrishnanandNash,1990;Leibovitch
et al., 1995). The na mutant alleles na1and naharare
strongloss-of-functionalleleswithveryloworundetect-
able levels of NA (Krishnan and Nash, 1990; Nash et al.,
2002). na mutants fail to robustly increase activity in an-
ticipation of LD transitions and in response to lights on.
In addition, they exhibit weak locomotor rhythms during
DD, consistent with a disruption in circadian function
(Nash et al., 2002). Our previous study suggested that
core clock function in the head (largely because of the
eye) was intact (Nash et al., 2002). However, it failed to
address the role of NA in the LN and DN neurons.
naisadistinctivememberofthevoltage-gatedsodium
andcalciumchannelsuperfamily.Membersofthesuper-
family consist of four repeats (domains I–IV) of six trans-
membrane segments (S1–S6). NA homologs are unique
in their intracellular domains between the second and
third repeat and the C terminus. A single amino acid
within each of four pore loops between S5 and S6 is es-
pecially important for ion selectivity (Tang et al., 1993;
Yangetal.,1993).InallNAhomologs,thisporeselectivity
filter is distinct, consisting of the amino acids EEKE (E =
glutamate, K = lysine). The unique features of this puta-
tive channel are highly conserved with homologs in hu-
mans, rats, and C. elegans. na is broadly expressed in
*Correspondence: r-allada@northwestern.edu

Page 2

both the fly and mammalian CNS (Lee et al., 1999; Nash
etal.,2002).Intheadultfly,NAisexpressedinmajorneu-
ropil regions, suggesting a role at synapses (Nash et al.,
2002).Theseobservationssuggestafundamentalrolein
neural function. However, heterologous expression to
characterize its function has been unsuccessful (Lee
et al., 1999). Here, we demonstrate that this putative
channel functions to regulate the output of Drosophila
circadian neurons to control behavioral rhythmicity and
timing.
Results
naLikelyFunctionsinCircadianPacemaker Neurons
to Regulate Diurnal Behavior
To assess the tissue-specific requirements for na func-
tion in regulating diurnal and circadian behavior, we
usedthebinaryGAL4-UASsystem(BrandandPerrimon,
1993). We generated strains containing na cDNA under
controloftheUAS promoter(UAS-na).Toquantify antic-
ipation, we have developed an anticipation index that
measures sustained increases in activity in advance of
lights on and lights off (Supplemental Experimental Pro-
ceduresandTableS1).Ofnote,innamutantcontrols(na;
UAS-na/+), we do not observe significant morning or
evening anticipation if we restrict our analysis to the
hours just before the lights-on and lights-off transitions
(data not shown). However, if we analyze 12 hr before
each transition, we find a significant activity upswing
about 4 hr prior to lights on in na mutants (Figure 1A
and Table S1). We will refer to this as anticipation, al-
though a role for light cannot be excluded.
To determine if na functions in neurons, we expressed
na in all postmitotic neurons of na mutant flies with
the elav-GAL4 driver (Lin and Goodman, 1994). Pan-
neuronal expression of na provides strong rescue of
the LD phenotypes of na mutants, including evening an-
ticipation, timing of morning anticipation and response
to lights on (Figures 1A–1B and Table S1). Given the
lack of robust evening anticipation in na mutants, we ex-
pressed UAS-na in all major circadian cell groups (LN
and DN) with a tim-GAL4 driver (tim-GAL4-62) (Kaneko
and Hall, 2000). tim-GAL4-62 also promotes robust res-
cue of the na LD phenotypes (Figure 1C and Table S1),
as does expression with other broad circadian drivers
(tim-GAL4-27 and cryp-GAL4-16) (data not shown and
Figure S1) (Kaneko and Hall, 2000; Zhao et al., 2003). A
comparison of the published expression of these lines
outside the pacemaker neurons strongly suggests non-
overlapping expression elsewhere (Zhao et al., 2003).
Nonetheless, we also utilized a Clk promoter GAL4
(Clk8.0-GAL4)(Glossopetal.,2003).Inordertocharacter-
ize its expression, we examined brains from Clk-GAL4/
UAS-nuclear GFP adults. We observe GFP expression
in most or all LNv neurons, as well as a large subset of
LNd, DN1, and DN3 neurons (Figure S2A); additional
GFP-expressing cells are observed near these circadian
cell groups, but we observe limited expression in other
brain regions (Figure S2A and data not shown). This
line also robustly rescues the LD phenotypes of na mu-
tants (Figure 1D and Table S1). Interestingly, more re-
stricted expression of na in ventral lateral neurons with
pdf-GAL4 fails to rescue these phenotypes (Figure 1E
and Table S1). The GAL4 line gal1118, which drives ex-
pression heterozygously in the LNv and weakly in the
LNd (Blanchardon et al., 2001), also fails to rescue na
LDphenotypes(datanotshownandTableS1).Thesere-
sults suggest a broader na requirement within the pace-
maker network.
To further map neurons important for rescue of even-
ing anticipation, we made use of two GAL80 lines: cry-
GAL80 and pdf-GAL80. Novel complements of GAL4-
mediated expression can be generated by expressing
GAL80,aninhibitorofGAL4,insubsetsofGAL4express-
ing cells (Lee and Luo, 1999). cry-GAL80 inhibits GAL4-
mediated transcription in all LNs and some DN1s and
DN3s (Stoleru et al., 2004). Consistent with na function
inpacemakerneurons,weobserveanearcompleteblock
of elav-GAL4-mediated na rescue with cry-GAL80, in-
cluding blocking of the lights-on response (Figure 1F
and Table S1). pdf-GAL80 inhibits GAL4 in the LNv
(Stoleru et al., 2004). We find that pdf-GAL80 fails to
strongly affect the LD behavior of tim-GAL4-62, Clk-
GAL4 or elav-GAL4 lines (Figures 1G and 1H and data
not shown). However, in tim-GAL4-62/pdf-GAL80; UAS-
GFP flies, GFP expression is still observed in w10% of
LNv, and we cannot entirely exclude na function in
cells in which driven GFP is not observed (Stoleru et al.,
2004). Nonetheless, the finding that driven na expres-
sion can be substantially reduced in the LNv without
strong effects on morning anticipation combined with
the lack of LNv-specific rescue suggests that na func-
tion in non-PDF neurons regulates morning anticipation.
Interestingly, we observe more selective rescue of
evening anticipation with the crypi-GAL4-13 and Mai179-
GAL4 drivers (Figures 1I and 1J and Table S1). These
lines also fail to rescue the lights-on activity response.
Interestingly, Mai179-GAL4 did not alter the degree or
timing of morning behavior in na mutants, whereas crypi-
GAL4-13 suppressed the degree of morning anticipa-
tion. We find that inhibiting na expression in the LNv
withpdf-GAL80incrypi-GAL4-13fliesstillrescueseven-
ing anticipation, suggesting that na function in the LNd
and/or DN1 is responsible for this behavior (data not
shown and Table S1). There have been contradictory
data concerning the extent of expression in this line
(Emery et al., 2000; Zhao et al., 2003, Stoleru et al.,
2004). We therefore reassayed crypi-GAL4-13 with a nu-
clear GFP and find expression in LNv, LNd, and in a few
DN1s (Figure S2B). We believe this pattern, broader
than those previously described, provides an accurate
representation of crypi-GAL4-13 expression because
concentrating GFP in the nucleus may render the assay
moresensitive.Importantly,weobservelittleexpression
beyond the LNs and DNs and thus limited overlap with
the other circadian GAL4 lines described. We also as-
sessed Mai179-GAL4 expression and find expression
in LNs as previously reported (Grima et al., 2004) as
well as in several DN1s not previously described (Fig-
ureS2C).Thus,naexpressioninallLNgroupsandasub-
set of DN1 is sufficient to promote evening anticipation
but fails to fully rescue the morning anticipation pheno-
type, indicating that additional DN expression may be
requiredfor wild-type morningbehavior. Although varia-
tion in driven expression levels may account for differ-
ences in rescue, these data suggest that na expression
within lateral neurons and a broad group of dorsal neu-
rons is required for complete wild-type LD behavior.
Neuron
966

Page 3

NA Is Expressed in a Core Set of Pacemaker Neurons
Given the evidence for NA function in pacemaker neu-
rons, we assessed NA expression in pacemaker neu-
rons. Broad NA expression has been observed synaptic
neuropil of the adult brain (Nash et al., 2002). To deter-
mine NA expression in pacemaker neurons, we utilized
apdf-GAL4/UAS-tau-lacZstrain,whichlabelsLNvpace-
maker neurons and their projections. We performed
double-label immunofluorescence studies on sectioned
adult brains with anti-NA and anti-b-gal. Consistent with
ourrescuestudies,wefoundcolabelinginbothcellbod-
ies (Figure 2A) and terminals (Figure 2B) of these neu-
rons. Of note, staining was not uniform throughout the
cell with more prominent staining in portions of the
cell body and in the more distant terminal. We did not
note any diurnal variation in NA, although this technique
may not be sufficiently quantitative to detect low ampli-
tude rhythms (data not shown).
To address na expression more broadly in the pace-
maker neuron network, we constructed na promoter
GAL4 fusions. A transcriptional start site for na has not
been defined, and at least two start methionines have
been defined with bioinformatics (Nash et al., 2002).
Alignmentofthetranslatedsequenceaftertheupstream
methionine identifies 18 consecutive amino acid identity
with sequences in D. pseudoobscura and yakuba (data
not shown), suggesting that the transcription start lies
upstream of this methionine. We fused 1.8 kb of na
Figure 1. Expression of na within LNs and DNs Rescues LD Anticipation
Charts represent normalizedactivity during 12hr Light:12 hr darkconditions (LD). Light phase (ZT 0–12) is indicated bywhite bars, whereas dark
phase (ZT 12–24) is represented by black bars. Error bars represent standard error of the mean for daily activity levels. All flies assayed were
nahar;UAS-na with the following GAL4/80 drivers: (A) none (n = 100); (B) elav-GAL4 (all postmitotic neurons; n = 21); (C) tim-GAL4-62 (all circadian
neurons; n = 46); (D) Clk-GAL4 (some/all LNv, LNd, DN1, DN3; n = 20), see Figure S2A; (E) pdf-GAL4 (LNv; n = 40); (F) elav-GAL4;cry-GAL80 (n =
29); (G) elav-GAL4;pdf-GAL80; (n = 31); (H) tim-GAL4-62/pdf-GAL80 (n = 21); (I) crypi-GAL4-13 (all LNs, DN1 subset; n = 75), see Figure S2B; (J)
Mai179-GAL4 (all LNs, DN1 subset; n = 37), see Figure S2C.
Figure 2. na Is Endogenously Expressed in
Circadian Neurons
(A and B) Individual confocal sections of sec-
tionedadultpdf-GAL4/UAS-taulacZbrainsla-
beled with antibodies to b-GAL (green) and
NA (red). (A) Arrow indicates the region of
LNv cell bodies in which NA colabeling is ob-
served.Thisregionismagnifiedinthecaption
in the upper right corner of each panel. The
LNv contralateral projection is also labeled
with b-GAL but does not exhibit noticeable
NA expression. (B) Arrow indicates NA label-
ing in the terminal region of the LNv dorsal
projection.
(C–E) na-GAL4 UAS-nGFP adult brains la-
beled with antibodies to PER. (C) Maximum
projection of ventral brain region, with LN
subgroups indicated. Colabeling of GFP and
PERcanbeidentifiedinsubsetsofcellswithin
all LN subgroups. (D) Higher magnification of
(C) with sections encompassing only the sLNv
and LNd. (E) Maximum projection of sections
encompassing posterior/dorsal brain regions.
ColabelingofPERandGFPisobservedinsub-
setsoftheDN1andDN3groups.Thearrowin-
dicatesregion ofcolabeling inthe DN3group.
In this image, saturated GFP signal in the re-
gion of the DN2 group (unlabeled) prevents
an accurate assessment of GFP expression
in those cells.
Ion Channel Function in Circadian Pacemakers
967

Page 4

upstream sequence including this methionine to GAL4.
We isolated transgenic lines and crossed them to UAS-
GFP to assess their expression pattern. As expected,
we observe expression throughout the brain (Figure 2C),
including at least a subset of cells in several pacemaker
neuron groups, including sLNv and LNd (Figure 2D), as
well as DN1 and DN3 (Figure 2E). We verified this pattern
with a second independent insert (data not shown).
Taken together, these results are most consistent with
na expression in pacemaker neurons.
na Functions Downstream of the Core Oscillator to
Regulate LD Behavioral Output
Given thestriking behavioral phenotype of na mutantsin
LD and the evidence of na function within circadian neu-
rons,weexamined molecularclock functioninthesecell
groupsinnamutants.Wehavepreviouslydemonstrated
robust circadian oscillation of PER protein in whole
heads of na mutants (Nash et al., 2002), but such analy-
ses might not detect cycling alterations within subsets
of circadian neurons themselves. We have now exam-
ined PER protein levels within both LN and DN circadian
groups at four LD time points. We find that PER robustly
cycles in na mutants in each of these cell groups. More-
over, we fail to identify strong differences in the PER ex-
pression profile between wild-type and na mutants that
may explain the altered or absent anticipation behavior
(Figure 3). Thus, na likely acts outside of the core molec-
ular clock but within circadian neurons to promote LD
anticipation.
naLikelyFunctionsinCircadianPacemaker Neurons
to Regulate Circadian Behavior
To assess the circadian role of na in constant condi-
tions, we analyzed daily activity patterns of out-crossed
wild-type and naharmutant flies over 2 days of LD en-
trainment and 4 days constant darkness (DD) (Figure 4).
We observe strong decreases in the rhythmic amplitude
of na mutants upon release into DD (Figure 4A). A closer
examination reveals differences not only in amplitude
but also in activity distribution. Wild-type flies have
a prominent activity peak that approximates the phase
of the evening LD peak and a lesser peak that approxi-
mates the morning LD peak (Figure 4A). na flies exhibit
a single broad activity peak across the subjective day
(Figure 4A). We confirmed these changes by examining
single fly activity profiles, although we observe consid-
erable variability (data not shown). Importantly, these
data indicate a circadian function for na from the onset
of DD, regulating rhythmicity and timing of activity.
These data also indicate that na mutants often exhibit
residual rhythmicity that dampens over time in DD. Sim-
ilar observations of dampening rhythmicity have been
made in pdf01mutant flies, although the timing of activ-
ity and perhaps the degree of rhythmicity at the onset of
DD appear different in this mutant (Renn et al., 1999).
To address the tissue-specific requirements for na
function in DD, we again utilized the GAL4/UAS system.
Pan-neural expression of na with elav-GAL4 rescues
robust rhythmicity (Figure 4B and Table 1). We also ob-
serve significant rescue of rhythmicity by expressing
UAS-na broadly in circadian neurons (tim-GAL4-62, Clk-
GAL4, crypi-GAL4-13, Mai179-GAL4, and tim-GAL4-27,
p < 0.0001), whereas expression limited to the LNv
(pdf-GAL4) provides no detectable rescue (Table 1). na
is not merely important for rhythmic amplitude but also
forthetimingofactivityinDD.Rescuewithbroaddrivers
suchaselav-GAL4andClk-GAL4resultinrescueofboth
morning and evening peaks in DD (Figure 4B and data
not shown). Indeed, statistically significant differences
areobservedinthreetimedomainsofthenaactivitypro-
fileonday1.First,elav-GAL4rescuesthetimingofmorn-
ingbehaviorsuchthatthereisreducedactivityduringZT
18.5–23(p<0.001)andadelayedriseofmorningactivity.
Theseresultsareconsistentwithdelayedmorningantic-
ipation rescue observed in LD (Table S1). Second, elav-
GAL4 rescues results in reduced activity during the
middle of the day (CT2-4.5; p < 0.001). Third, this GAL4
line promotes a robust evening peak (CT7.5-11.5; p <
0.001). The result is that the behavioral profile of elav-
GAL4 rescue flies is comparable to that of wild-type.
RescuewithmorerestrictedGAL4linesresultinselec-
tive rescue ofna mutant phenotypes. Rescue with crypi-
GAL4-13yieldsonlyasinglebroadeveningpeak,consis-
tentwithitsLDbehavior(Figure4C).First,morningantic-
ipationbehaviorisnotobservedincrypi-GAL4-13rescue
Figure 3. PERIOD Protein Oscillates Ro-
bustly in na Mutants during LD
(A and B) Maximum projections of confocal
sections taken through representative w and
w naharadult Drosophila brains labeled with
PER antibody. Sections encompass either
lateral neurons (A) or dorsal neurons (B) at
times of peak (ZT 0) and trough (ZT 13) PER
expression. LN and DN subgroups are indi-
cated by arrows.
(C) Plots of average maximum intensity ver-
sus Zeitgeber time for each circadian cell
group (n = 13 to 25 hemispheres). See Exper-
imental Procedures for details of quantitation
method. Error bars represent standard error
of the mean.
Neuron
968

Page 5

flies or is very delayed, as seen by strong reductions of
activity from ZT 18.5–24 (Figure 4C, p < 0.001). There is
then an upswing of activity but delayed relative to elav-
GAL4 rescue. Interestingly, this altered behavior dissi-
pates over time consistent with the notion that it is a de-
layed morning peak (Figure 4C, left). crypi-GAL4-13 also
rescues an evening peak with increases in relative activ-
ity evident between CT11.5 and CT13.5. We do not ob-
serve reductions in midday activity as observed for
elav-GAL4. We also examined the activity pattern with
Mai179-GAL4,whichisexpressedintheLNsandalarger
subsetofDN1sthanobservedincrypi-GAL4-13(Figures
S2BandS2C).Here,weobserveathirdpatternofrescue.
Like with crypi-GAL4-13 and elav, we observed strong
rescue of the evening activity peak (ZT 11–12.5, p <
0.001). However, unlike with these lines, we did not
observe any change in morning anticipation behavior.
Moreover,unlikewithcrypi-GAL4-13,wedidobservere-
ductions in midday behavior (ZT 0.5–1.5, p < 0.001). Al-
thoughwecannotexcludethatdifferencesinexpression
levels may contribute to the observed differences in
phase, these data suggest that differential expression
of na in distinct pacemaker neuronal groups, especially
the DN1s, may have important consequences on behav-
ioral phase early in DD.
One line cryp-GAL4-16 that rescued LD behavior (Fig-
ure S1A) also rescues DD behavior (p < 0.0005), but this
rescue is relatively weak. Notably, cryp-GAL4-16 rescue
flies display a higher proportion of flies with stronger
rhythms (P-S > 20; cryp-GAL4-16 rescue = 57%, na con-
trol =37%).Moreover, these fliesexhibit clear persistent
morning and evening peaks in DD absent in na mutants
(Figures S1B and S1C). Thus, we favor the notion that
cryp-GAL4-16 partially rescues DD behavior. The vari-
abilityinrescueamong thebroadercircadian GAL4driv-
ers may reflect differential expression within LNd and/or
DN groups (e.g., gal1118, crypi-GAL4-13), inadequate
expression levels, or mis- or overexpression effects. Al-
though we cannot exclude the possibility that cryp-
GAL4-16drivesinadequateexpressioninkeypacemaker
cells, we favor the notion that ectopic expression may
lead to some degree of arrhythmicity evident only in an
na mutant background (Table 1).
GiventheobservationthatLNv-directedexpressionof
na (pdf-GAL4) is not sufficient to rescue DD rhythmicity,
we assayed whether LNv expression was required for
Figure 4. Distinct Behavioral Rhythms in na Mutant Rescue during DD
(A) Normalized average activity plots of outcrossed wild-type (red squares, n = 60) and nahar(blue diamonds, n = 100) adult males over 2 days LD
followed by 4 days DD. For LD, light phase is indicated by white bars, and dark phase by black bars. For DD, subjective day is indicated by gray
bars, and subjective night by black bars.
(B–D) Left: normalized average activity plots of GAL4 (elav, elav-GAL4 [n = 76]; crypi-13, crypi-GAL4-13 [n = 55]; mai179, Mai179-GAL4 [n = 37])
rescueflies (red)andna;;UAS-na/+mutants (n=74, bluediamonds).Right: amagnificationof behavioral traces during last12hrofLDand first12
hr of constant darkness is shown. Error bars indicate SEM. Black bars indicate where rescue and mutants strain activity is significantly different
(p < 0.001). Different GAL4s rescue different features of the na mutant phenotype.
Ion Channel Function in Circadian Pacemakers
969