Behavioral consequences of dopamine deficiency in the Drosophila central nervous system.

Abstract

The neuromodulatory function of dopamine (DA) is an inherent feature of nervous systems of all animals. To learn more about the function of neural DA in Drosophila, we generated mutant flies that lack tyrosine hydroxylase, and thus DA biosynthesis, selectively in the nervous system. We found that DA is absent or below detection limits in the adult brain of these flies. Despite this, they have a lifespan similar to WT flies. These mutants show reduced activity, extended sleep time, locomotor deficits that increase with age, and they are hypophagic. Whereas odor and electrical shock avoidance are not affected, aversive olfactory learning is abolished. Instead, DA-deficient flies have an apparently "masochistic" tendency to prefer the shock-associated odor 2 h after conditioning. Similarly, sugar preference is absent, whereas sugar stimulation of foreleg taste neurons induces normal proboscis extension. Feeding the DA precursor L-DOPA to adults substantially rescues the learning deficit as well as other impaired behaviors that were tested. DA-deficient flies are also defective in positive phototaxis, without alteration in visual perception and optomotor response. Surprisingly, visual tracking is largely maintained, and these mutants still possess an efficient spatial orientation memory. Our findings show that flies can perform complex brain functions in the absence of neural DA, whereas specific behaviors involving, in particular, arousal and choice require normal levels of this neuromodulator.

Full text

Page 1

Behavioral consequences of dopamine deficiency in
the Drosophila central nervous system
Thomas Riemenspergera,b, Guillaume Isabelc, Hélène Coulomb, Kirsa Neuserd, Laurent Seugnetc, Kazuhiko Kumee,
Magali Iché-Torresb, Marlène Cassara, Roland Straussd, Thomas Preatc, Jay Hirshf, and Serge Birmana,b,1
aGenetics and Physiopathology of Neurotransmission andcGenes and Dynamics of Memory Systems, Laboratoire de Neurobiologie, Centre National de la
Recherche Scientifique, Ecole Supérieure de Physique et de Chimie Industrielles ParisTech, 75005 Paris, France;bInstitut de Biologie du Développement de
Marseille-Luminy, Centre National de la Recherche Scientifique, Université de la Méditerranée, Campus de Luminy, 13009 Marseille, France;dInstitut für
Zoologie III, Johannes Gutenberg Universität, 55099 Mainz, Germany;eInstitute of Molecular Embryology and Genetics, Kumamoto University, Kumamoto
860-0811, Japan; andfDepartment of Biology, University of Virginia, Charlottesville, VA 22903
Edited by Howard Nash, National Institutes of Health, Bethesda, MD, and approved November 30, 2010 (received for review August 3, 2010)
The neuromodulatory function of dopamine (DA) is an inherent
feature of nervous systems of all animals. To learn more about the
functionofneuralDAinDrosophila,wegeneratedmutantfliesthat
lack tyrosine hydroxylase, and thus DA biosynthesis, selectively in
thenervoussystem.WefoundthatDAisabsentorbelowdetection
limits in the adult brain of these flies. Despite this, they have a life-
span similar to WT flies. These mutants show reduced activity, ex-
tended sleep time, locomotor deficits that increase with age, and
they are hypophagic. Whereas odor and electrical shock avoidance
are not affected, aversive olfactory learning is abolished. Instead,
DA-deficient flies have an apparently “masochistic” tendency to
prefer the shock-associated odor 2 h after conditioning. Similarly,
sugar preference is absent, whereas sugar stimulation of foreleg
taste neurons induces normal proboscis extension. Feeding the DA
precursor L-DOPAtoadultssubstantiallyrescuesthelearningdeficit
as well as other impaired behaviors that were tested. DA-deficient
flies are also defective in positive phototaxis, without alteration in
visual perception and optomotor response. Surprisingly, visual
tracking is largely maintained, and these mutants still possess an
efficient spatial orientation memory. Our findings show that flies
can perform complex brain functions in the absence of neural DA,
whereas specific behaviors involving, in particular, arousal and
choice require normal levels of this neuromodulator.
neurotransmitters|locomotor activity|memory formation|choice
behavior|feeding behavior
A
stasis and functioning. Dopamine (DA), a biogenic amine bio-
synthesized from tyrosine, is an essential neuromodulator in the
mammalian central nervous system that is involved in attention,
movement control, motivation, and cognition. Studies in Dro-
sophila melanogaster indicate that DA also plays central regula-
tory roles in insects, specifically in the neural networks controlling
locomotor activity and stereotypical behaviors (1–3), sleep and
arousal (4–7), registration of salient stimuli (4, 8, 9), and asso-
ciative olfactory learning (10–15). Some of these studies were
based on genetic inactivation or overactivation of dopaminergic
neurons. Dopaminergic neurons can corelease other neuroactive
agents, such as neuropeptides, however. Therefore, one must
ensure that the behavioral phenotypes observed specifically result
from the lack of DA release to draw firm conclusions on brain
DA function.
Nearly all neuropil regions of the insect CNS receive dense
dopaminergic innervation. In particular, the Drosophila adult
brain contains six paired clusters of dopaminergic neurons, some
of which specifically project to higher brain centers, such as the
central complex and the mushroom bodies (1, 10, 12, 13, 16–18).
Tyrosine hydroxylase (TH) catalyzes the first and rate-limiting
step in DA biosynthesis (Fig. S1A). Because DA is also required
in Drosophila as a precursor substrate for cuticle sclerotization
andmelanization, inactivating mutations ofthe genomic TH,alias
n important challenge in neuroscience is to understand the
roles of specific neurotransmitter systems on brain homeo-
pale (ple) locus, results in unpigmented cuticle and late embryonic
lethality (19).
Alternative splicing of Drosophila TH (DTH) produces two
enzyme isoforms, DTH1 and DTH2 (20, 21) (Fig. S1B). DTH1 is
selectively expressed in DA neurons within the CNS, whereas
DTH2 is expressed in peripheral nonnervous tissues, which in-
clude the hypodermal cells that secrete the cuticle matrix (20,
22). Here, we take advantage of this tissue-specific alternative
splicing to construct a mutated transgene, DTHgFS±, that only
expresses an active TH enzyme in nonneural cells. Homozygous
ple mutants are rescued by this transgene to adult stage, gener-
ating Drosophila essentially devoid of DA in the adult brain. We
then studied the consequences of this neural-specific DA de-
pletion on adult survival and behavior.
Results
Generation of Viable DA-Deficient Drosophila. We first established
the conditions for full rescue of DTH deficiency with the GAL4-
upstream activation sequence (UAS) binary expression system by
expressing genomic DTH (DTHg) driven by TH-GAL4 and Ddc-
GAL4, each of which contains regulatory sequences from genes
involved in DA biosynthesis (Fig. S1A). The combined action of
the Ddc-GAL4 and TH-GAL4 drivers is required for full rescue,
with neither GAL4 driver alone being sufficient for full rescue to
adult stage (Table S1).
We then used in vitro mutagenesis to introduce frameshift
mutationsinDTHg(Fig.S1CandDandSIMaterialsandMethods).
DTHgFS+contains one additional base in a hypoderm-specific
exon, and DTHgFS±contains the same mutation plus another
compensating mutation that removes one base in an adjacent
common exon. When transcribed, DTHgFS+only expresses active
TH in neural tissues and, conversely, DTHgFS±only expresses ac-
tiveTHinnonneuraltissues.Noplerescuecouldbeobservedwith
DTHgFS+(Table S2), confirming that an active DTH2 isoform
producing cuticular DA is required for Drosophila development
(22). In contrast, expression of DTHgFS±yielded full rescue of ple
toadulthood(TableS2),indicatingthatbiosynthesisofneuralDA
is not essential for fly development and survival.
Immunohistochemistry confirmed that TH and DA are not de-
tectable in adult brain of the DTHgFS±-rescued ple mutants (Fig.
1A3and6), whereasthey arepresent in WTflies (Fig.1A1and4)
or ple mutants rescued by the native DTHg construct (Fig. 1A 2
Author contributions: S.B. designed research; T.R., G.I., H.C., K.N., L.S., K.K., M.I.-T., J.H.,
and S.B. performed research; T.R., H.C., L.S., K.K., M.C., R.S., T.P., J.H., and S.B. contributed
new reagents/analytic tools; T.R., G.I., H.C., K.N., L.S., K.K., R.S., T.P., J.H., and S.B. analyzed
data; and T.R., J.H., and S.B. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed. E-mail: serge.birman@espci.fr.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1010930108/-/DCSupplemental.
834–839
| PNAS
| January 11, 2011
| vol. 108
| no. 2 www.pnas.org/cgi/doi/10.1073/pnas.1010930108

Page 2

and 5). Accordingly, RT-PCR shows that the neural-specific
DTH1 mRNA is absent in head and brain extracts of DTHgFS±;
ple (Fig. 1B). Similarly, no TH is present in the larval CNS of
DTHgFS±; ple, whereas it is expressed at this stage in DTHg; ple
(Fig. S2A). In contrast, DA is detected in larval neurons of both
strains, although fewer positive cells could be seen in larval brain
hemispheres of DTHgFS±; ple versus DTHg; ple (Fig. S2 A3 and
A4). This suggests that systemic DA produced by peripheral TH
can partially supply the mutant larval CNS with the missing bio-
genic amine.
DA was then assayed by HPLC to determine the absolute
magnitude of reduction of DA level in adult brain (Fig. 1C and
Tables S3 and S4). Brain extracts from control flies and DTHg;
ple showed equivalent levels of DA per brain (52.0 and 48.9 pg
per brain, respectively), indicating that expression of the DTHg
transgenequantitativelyrestoredbrainDA.Incontrast,DTHgFS±;
plefliesapparently had7.5pgofDAper brain,suggestinga strong
(∼85%) but incomplete reduction of DA level in these mutants.
HPLC was then repeated using a decreased detector potential of
+420 mV, near the half-maximal oxidation potential of DA.
Whereas the DA peak in an extract from a control brain behaves
as expected (relative DA peak area at +420 vs. 500 mV: 0.43), the
residual peak in the DTHgFS±; ple strain was unaffected by the
reduced detector potential (relative “DA” peak area: 1.05) (Fig.
1C and Tables S3 and S4). We conclude that the residual peak in
this strain is largely, if not totally, a non-DA contaminant and that
brainsofDTHgFS±;pleflies lack detectable DA,withan estimated
sensitivity limit of 2% WT levels.
Nevertheless, adult brain of DA-deficient flies contains differ-
entiated dopa decarboxylase (Ddc)-positive neurons that do not
express serotonin and so are most likely “dopaminergic” neurons
devoid of TH and DA. These cells indeed distribute in clusters
identical to the WT pattern of dopaminergic neurons (Fig. S2B1–
3). Moreover, characteristic dopaminergic innervation of the
mushroom bodies appears normal in the mutant (Fig. S2B4–6).
Therefore, maintenance of the adult brain dopaminergic system
does not seem to depend on the activity of TH or the presence of
DA in Drosophila.
DA Deficiency Reduces Activity and Arousal. Unexpectedly, we ob-
served that both DTHg; ple and DTHgFS±; ple lines show similar
and normal adult life expectancy (Fig. S3A), demonstrating that
the lack of neural DA has no detrimental effect on Drosophila
lifespan. This enabled us to compare the behaviors of these lines.
First, we studied the effect of neural DA deficiency on walking
behavior of young adult flies. DTHg
walking speed (median = 7.8 mm/s) compared with DTHg; ple
(median = 10.8 mm/s) and WT (median = 15 mm/s) (Fig. 2A,
Left). Similarly, the distance covered for 15 min is much shorter
in DTHgFS±; ple (median = 193 cm) than in DTHg; ple (me-
dian = 425 cm) and WT (median = 474 cm) (Fig. 2A, Right).
Therefore, flies deficient in neural DA show markedly decreased
locomotor behavior.
Startle-induced negative geotaxis monitors climbing ability
and excitability in Drosophila, and DA modulates this behavior
(1, 23). WT flies placed in a vial or a column respond to a gentle
FS±; ple show reduced
Fig. 1.
sophila (1) and DTHg (2) or DTHgFS±-rescued (3) ple mutants. No TH immunoreactivity was detected in the brain of DTHgFS±; ple. Anti-DA immunostaining. DA-
positive cells can be observed in the brains of WT (4) and DTHg; ple (5) flies (arrows) but not in the brains of DTHgFS±; ple flies (6). (B) Detection of the DTH1- and
DTH2-specific spliced mRNAs by RT-PCR in heads (Left) and dissected brains (Right) of control or DTHgFS±; ple flies. No DTH transcript is present in the brain of
DTHgFS±-rescued flies. (C) Representative graphs of DA assay by HPLC in brain extracts of control w1118(a and b) and DTHgFS±; ple (c and d) detected with an
optimalpotentialof+500mV(aandc)orreducedto+420mV(bandd).(d)Reducedpotentialdidnotdecreasetheamplitudeoftheresidual“DA”peakofneural
TH-deficient flies, indicating that this peak is not DA. The small additional peak on the right corresponds to a contaminant from eye pigments.
Lack of TH expression and DA in the brain of DTHgFS±-rescued ple Drosophila. (A) Anti-TH immunostaining on whole-mount adult brains of WT Dro-
Riemensperger et al. PNAS
| January 11, 2011
| vol. 108
| no. 2
| 835
NEUROSCIENCE

Page 3

mechanical shock by fastclimbing. This behavior isstable overthe
first 3 wk of adult life and then progressively declines with age,
such that by day ∼55, WT flies show no response (Fig. 2B). The
DTHgFS±; ple flies do not perform well in this test, although they
are definitely able to climb. Just after eclosion, their performance
index (PI) was ∼0.4, and at 15–21 d of age, their PI was only ∼0.2,
comparable to the PI of 50-d-old WT flies (Fig. 2B, dotted line).
DTHg; ple show PIs between those of WT and DA-deficient flies
(Fig. 2B).
Monitoring activity over the light/dark cycle indicates that
DTHgFS±; ple are overall less active and possibly sleep more
frequently than DTHg; ple during the day (Fig. S3B). Because
both strains showed frequent periods of relative inactivity, we
also monitored sleep by direct observation. The results demon-
strate that DTHgFS±; ple do sleep more than DTHg; ple during
both day and night (Fig. S3C). Furthermore, we found that sleep-
ing DTHgFS±; ple flies respond less frequently to mild and mod-
erate mechanical stimuli (Fig. S3D), whereas response to strong
stimuli was comparable to that of DTHg; ple flies. Overall, these
results argue for a decreased arousal state in the absence of
brain DA.
Treatment with 3-iodotyrosine (3IY), a TH inhibitor, de-
creases activity and increases resting periods in WT flies (4),
whereas caffeine has the opposite effects (24). We observed
that the neural TH-deficient DTHgFS±; ple, as expected, are not
responsive to 3IY, whereas DTHg; ple flies reacted like WT
(Fig. S5). In contrast, caffeine was found to decrease resting
periods of both DTHg; ple and DTHgFS±; ple (Fig. S4), sug-
gesting that caffeine does not act by modulating DA release
in Drosophila.
Impaired Aversive Olfactory Learning and Feeding Behavior.We used
classical olfactory conditioning to evaluate the ability of DTHgFS±;
ple flies to associate an aversive stimulus (electrical shock) with an
odor. Either immediately (t = 0) or 2 h after training, these flies
did not show any avoidance of the shock-associated odor (Fig. 3A).
This is in contrast to the DTHg-rescued controls that demonstrate
associative memory (PI = ∼0.6) (Fig. 3A), similar to WT (25).
Interestingly, the DA-deficient flies showed an inverse tendency to
choose the shock-associated odor during testing, an effect that was
significant 2 h after training (PI = −0.13 ± 0.05, n = 9). We
checked that shock avoidance (Fig. S5A) and odor perception (Fig.
S5B) are preserved in DA-deficient as well as DTHg-rescued flies.
Pharmacological rescue experiments were performed in adult
flies to distinguish between physiological or developmental
effects of brain DA deficiency. Remarkably, treatment of adult
DTHgFS±; ple with the DA precursor L-DOPA significantly im-
proved their negative geotaxis behavior (Fig. S6A). This indicates
that progressive locomotor impairments directly result from DA
deficiency and not from developmental abnormalities. Similarly,
L-DOPA feeding of adult DA-deficient flies substantially in-
creased their learning performance index towards a positive
value (PI = 0.26 ± 0.08 with L-DOPA compared to -0.11 ± 0.07
without L-DOPA at t = 0, n = 6) (Fig. S6B), confirming that the
lack of neural DA induces an absolute defect in aversive olfac-
tory memory.
Preparatory to appetitive conditioning trials, we tested the flies
for sugar preference after a 21-h period of starvation. Whereas
DTHg-rescued controls respond positively and comparably to
WT, DTHgFS±; ple flies showed no significant response to sucrose
(Fig. 3B). The proboscis extension reflex appears normal in these
mutants (Movie S1), however, indicating that sugar perception
and reflex appetitive reactions are preserved. This suggests that
neural DA is specifically required for behavioral attraction to
sugar. Quantification of food intake revealed that DTHgFS±; ple
eat one-third as much as DTHg; ple (Fig. S6C), demonstrating
that they are markedly hypophagic. Again, L-DOPA significantly
improves this phenotype, whereas the drug has no effect on food
intake of DTHg; ple flies. Interestingly, L-DOPA also has a posi-
tive effect on food intake of heterozygous ple flies (Fig. S6C),
suggesting a dominant effect of ple on feeding behavior.
Lack of Phototaxis but Preserved Visual Orientation and Spatial
Memory. Positive phototaxis is a characteristic behavior of Dro-
sophila and many other insect species. In countercurrent assay,
DTHgFS±; ple flies were mainly recovered in the first tube of the
apparatus, indicating that they are not attracted by light, whereas
DTHg; ple showed partial rescue of phototactic behavior (Fig.
3C). The mutants have normal electroretinograms (Fig. S5C) and
optomotor response (Fig. S5D), indicating that visual functions
arepreserved.Wealsoperformed individual assaystoruleoutthe
possibility that the lack of positive response in the countercurrent
test was attributable to reduced activity. A single fly was placed at
the entrance of a T-maze and allowed to choose freely between
a dark tube and an illuminated tube until it reached one of the
vials. Similar to WT, neurally rescued DTHg; ple show a signifi-
cantpreferencetowardthelightedtube.Incontrast,DTHgFS±;ple
single flies show no preference, choosing either tube with equal
probability (Fig. 3D).
We further checked the flies for visual tracking and orienta-
tion in the Buridan’s paradigm, a cylindrical virtual reality arena
in which two dark vertical stripes are presented opposite to each
other. Under standard conditions, WT flies patrol between two
given visual targets for a considerable length of time. Whole
orientation curves indicate that DTHg; ple and DTHgFS±; ple are
comparably able to target the two stripes, with a fixation ability
Fig. 2.
whisker plots showing the distribution of walking speed (Left) and cov-
ered distance (Right) during spontaneous locomotor behavior of 5-d-old
adult flies monitored in an open arena. The DA-deficient DTHgFS±; ple
flies show a statistically significant reduction compared with rescued
DTHg; ple in both walking speed (P = 0.015) and covered distance (P =
0.001). DTHg; ple show a difference compared with WT flies for walking
speed (P = 0.0003) but not walking distance (P = 0.130). Plots represent
the median (horizontal line), mean (square), 25% and 75% quartiles
(box), 10% and 90% quantiles (whiskers), and extreme values (crosses)
(**P < 0.02 and ***P < 0.005). (B) Climbing abilities of WT flies, DTHg; ple,
and DTHgFS±; ple monitored by startle-induced negative geotaxis as
a function of adult age. The robust negative geotactic behavior of young
WT Drosophila declines steadily over time after 3 wk. DTHgFS±; ple show
from the eclosion a strong impairment in this behavior; after 2 wk, their
mean PI was only 0.2 (horizontal dotted line). DTHgFS±; ple values were
compared for significance with DTHg; ple values (*P < 0.05 and **P <
0.02, Student’s t test).
Neural DA-deficient flies show locomotor deficits. (A) Box-and-
836
| www.pnas.org/cgi/doi/10.1073/pnas.1010930108Riemensperger et al.

Page 4

different from random behavior, although less efficiently than
WT flies (Fig. 4A). A closer look at the curves shows that both
mutants differ in their orientation ability at small deviation
angles (Fig. 4B), suggesting that the lack of DA makes visual
orientation slightly less accurate. Therefore, in Drosophila, the
presence of neural DA improves visual tracking but, neverthe-
less, is dispensable for this behavior.
Orientation in a changing environment requires the ability to
retain and recall the position of a target. Drosophila possesses
sucha spatial orientation memory that canbe testedby the detour
paradigm. Inthis test, a fly walks toward a landmark that suddenly
disappears while, simultaneously, a vertical distracter stripe
appears laterally to the fly. After the fly approaches this second
landmark, the distracter stripe also disappears within 1 s. In this
situation, WT flies will turn back to their previous, still invisible,
landmark, demonstrating a spatial orientation memory for their
former target with a median frequency of ∼80% (26). We ob-
served that both DTHg; ple and DTHgFS±; ple do recall the po-
sition of the first landmark with a median frequency of 90% and
73%, respectively (Fig. 4C). This demonstrates that the lack of
brain DA has only a minor impact on spatial orientation memory
in Drosophila.
Fig. 4.
absence of brain DA. (A) Whole orientation curves showing frequency of
angle deviation between the fly trajectory and approached target. Mutant
and rescue flies show a frequency distribution significantly different from
the calculated random behavior (DTHg; ple: P = 0.007 and DTHgFS±: ple: P =
0.005, F test), whereas both strains do not differ from each other (P =
0.455; mean of 13 flies of each genotype). This confirms that DTHg; ple and
DTHgFS±; ple flies are able to visualize and target the landmarks, although
less consistently than WT flies (DTHg; ple: P = 0.028 and DTHgFS±; ple: P =
0.036, F test). (B) Differences between angles to the approached target
and calculated random behavior for small deviations (derived from shaded
area of curves in B are shown. Trajectories of DTHgFS±; ple target the
landmarks but are not accurate, making angle frequency not significantly
different from random behavior for small deviations [*P < 0.05, not sig-
nificant (n.s.)]. This defect is rescued in DTHg; ple flies that contain neural
DA (***P < 0.005). The Wilcoxon signed-rank test was used against 0 as
random behavior. (C) Test of spatial orientation memory in the detour
paradigm (details are provided in Materials and Methods). Ten flies per
genotype and 10 tests per fly were recorded. The frequency of positive
choices is represented by box-and-whisker plots showing the median
(horizontal line), mean (square), 25% and 75% quartiles (box), 10% and
90% quantiles (whiskers), and extreme values (crosses). Random behavior
would result in a reference value of 58% (horizontal dotted line). The
DTHgFS±; ple flies still retain spatial orientation memory (**P = 0.0015,
one-sample sign test against the random value), although their perfor-
mance is reduced compared with the DTHg; ple flies (*P = 0.034, Mann–
Whitney U test).
Visual fixation and spatial memory are largely preserved in the
Fig. 3.
sive olfactory learning. The PI of flies tested either immediately (t = 0) or
2 h after electrical shock training is shown. Whereas DTHg; ple show
normal conditioned avoidance of the aversive odor, the neural DA-
deficient DTHgFS±; ple do not learn. They rather show an abnormal pref-
erence for the shock-associated odor that is statistically significant 2 h
after training (P = 0.312 at t = 0 and **P = 0.012 at t = 2 h, Wilcoxon signed-
rank test against 0 as random behavior). (B) Sugar preference. After a 21-h
starvation period, flies were tested in a T-maze for their response to su-
crose. WT and DTHg-rescued Drosophila show positive responses to sugar,
whereas DTHgFS±; ple flies appear not to be attracted (*P < 0.05, Student’s
t test). The sugar response of the DTHgFS±; ple flies was not different from
0 as random behavior (Wilcoxon signed-rank test). (C) Phototaxis assayed
by countercurrent distribution. WT Drosophila display strong phototactic
behavior and mostly distributes toward the light source in tubes 5 and 6. In
contrast, most DTHgFS±; ple flies remain in tube 1. This behavioral im-
pairment is partially rescued in DTHg; ple flies [***P < 0.0001, one-way
ANOVA (Bonferroni-corrected) compared with WT values]. (D) Single-fly
phototaxis assay. Flies were allowed to distribute freely between an illu-
minated tube and a dark tube. WT and DTHg-rescued Drosophila showed
a strong preference for the illuminated tube. In contrast, the neural DA-
deficient DTHgFS±; ple distributed equally [***P = 6.1 × 10−5, **P = 0.012,
and P = 0.499 (not significant [n.s.]) for WT, DTHg; ple, and DTHgFS±; ple,
respectively, Wilcoxon signed-rank test; mean tested against reference
constant (50% for equal distribution)].
Behaviors disrupted in neural DA-deficient Drosophila. (A) Aver-
Riemensperger et al.PNAS
| January 11, 2011
| vol. 108
| no. 2
| 837
NEUROSCIENCE

Page 5

Discussion
Here, we describe specific genetic tools to study DA as a signaling
molecule in the Drosophila brain. We show that the embryonic
lethality of ple is fully rescued by a combination of dopaminergic
drivers (Ddc-GAL4andTH-GAL4)expressingaframeshiftmutant
oftheTHgene(DTHgFS±).TheDTHgFS±;plemutantsonlyexpress
TH in nonneural tissues, principally the cuticle-producing cells of
the hypoderm. This leads to the absence of DA in adult brain
without apparent morphological disturbance of the dopaminergic
system, as attested by L-DOPA rescue of defective behaviors. The
resulting flies are fully viable with a normal lifespan and they show
a number of distinctive phenotypes.
Mutants without neural DA show reduced locomotor activity
and arousal, and they sleep significantly more than control flies
during both day and night, consistent with previous reports on the
role of DA in these behaviors (4–6). The lack of further effect of
a TH inhibitor (3IY) demonstrates that spontaneous activity is
not influenced by trace levels of neural DA or by cuticular DA
in these flies. Other reports indicate that blockade of dopami-
nergicsynaptictransmissionincreasesactivity,particularlystartle-
induced activity (1, 7, 27). This apparent discrepancy cannot be
fully explained at present butmight be relatedto the fact that only
part of brain DA signaling is inhibited with the use of GAL4
drivers that do not express in the entire dopaminergic system.
The role of dopaminergic neurons in Drosophila aversive ol-
factory learning has been widely studied before (10–15). The
present results show conclusively that DA is the critical signaling
molecule released by these neurons that is required for this form
of associative memory. Interestingly, although electrical shock
avoidance is not altered, we observed that the DA-deficient flies
have a striking tendency to choose the shock-associated odor 2 h
after training, indicative of an inversion of the evaluation system.
Such an apparently “masochistic” behavior could result from an
inability of these flies to attribute a negative value to a reinforc-
ing stimulus, leading to a bias toward an appetitive behavior. In
a possibly related observation, Krashes et al. (12) recently re-
ported that specific DA neurons can inhibit the expression of an
appetitive memory performance.
Our study reveals the requirement for neural DA in sugar
preference and normal food intake in Drosophila. The fact that
serotonin is also involved in feeding in flies (28) suggests that
these two biogenic amines directly interact in the control of this
behavior. The lack of sugar preference, despite the intact pro-
boscis extension reflex, may suggest that DA is required for re-
ward and motivation in flies, as is the case in mammals, but this
would need further work to be established. Remarkably, several
of the deficits we observe in DA-deficient Drosophila, (e.g.,
hypoactivity and hypophagy readily rescued by L-DOPA) are
quite similar to phenotypes observed in DA-deficient mice (29).
The same is true for the preservation of dopaminergic neuron
circuits. This argues for partial conservation of brain DA func-
tions between flies and mammals.
The lack of phototaxis does not result from blindness because
flies without neural DA have normal electroretinograms and
optomotor response. Circadian entrainment to dim light also
requires neural DA (30). We suspect that this is not related to the
phototaxis defect, which was tested under high levels of illumi-
nation. An unexpected result of this study is that visual tracking
and spatial orientation memory are largely retained in Drosophila
lacking neural DA. Although neural DA improves visual fixation
and orientation toward a landmark, possibly by increasing arousal
level and attention (8), this neuromodulator appears to be dis-
pensable for these behaviors. Our results, in agreement with
previous findings (26, 31), suggest that the neural circuits involved
in the formation of spatial memory do not depend on neural DA.
In conclusion, we find that activity, feeding, and certain choice
behaviors are markedly altered or abolished in neural DA-deficient
flies, whereas other complex behaviors are surprisingly well main-
tained and might be similar to WT without the decrease in arousal.
This suggests that DA is implicated in many but not all aspects of
brain functioning in Drosophila. The specific behaviors that abso-
lutely require DA could be related to the primordial and evolu-
tionarily conserved functions of this essential neuromodulator in
primitive nervous systems.
Materials and Methods
Further details and references of the procedures are provided in SI Materials
and Methods.
Drosophila Strains. DTHg- and DTHgFS±-rescued homozygous ple Drosophila
were generated before each experiment by crossing a line containing the
combined TH-GAL4 and Ddc-GAL4 drivers with the respective UAS trans-
gene, each in a heterozygous ple mutant background. The Canton S strain
was used as WT flies for comparison in behavior tests.
Immunohistochemistry. Adult brains were dissected in ice-cold Drosophila
Ringer’s solution and processed for whole-mount immunostaining by
standard procedure. The following primary antibodies were used: rabbit
anti-DA (1:100; ImmunoStar), rabbit anti-serotonin (1:500; Sigma-Aldrich),
mouse monoclonal anti-TH (1:50; ImmunoStar), rat anti-Ddc (1:200; pre-
pared in the laboratory of J.H.), and rabbit anti-GFP (1:500; Invitrogen
Molecular Probes).
DA Assay. DA levels in Drosophila brain extracts were determined by HPLC
coupled to electrochemical detection essentially using a mobile phase con-
taining 4 mM decanesulfonic acid, 50 mM citrate/acetate (pH 4.5), and 20%
(vol/vol) acetonitrile. Standard detection of DA was performed at a detector
potential of +500 mV relative to an Ag/AgCl electrode. Electrochemical
confirmation of the identity of the DA peaks was performed at a reduced
potential of +420 mV.
Locomotor Behavior. Two methods were used to test for spontaneous loco-
motor behavior of 2- to 5-d-old adult Drosophila. First, walking speed and
covered distance were computed from video-based recordings of individual
flight-disabled flies walking freely for 15 min in an open arena. Second,
activity was monitored by recording infrared beam crossings in glass tubes
(6.5-cm length, 3-mm inside diameter) using a Drosophila activity monitor-
ing system (TriKinetics).
OlfactoryConditioning.Totestforaversiveolfactorylearning,groupsofabout
35 flies were conditioned in a barrel-type machine by sequential exposure to
two odors, octanol and methylcyclohexanol, for 60 s with 45-s rest intervals
between presentation of odors (25). Exposure to the first odor was paired
with electrical shocks.
Phototaxis. Phototactic behavior was tested by mass and single-fly assays. The
mass assay was carried out in a dark room with the countercurrent procedure.
The results were the mean of the scores from five trials. Data were statistically
analyzed with a one-way ANOVA (Bonferroni-corrected). The single-fly pho-
totaxis assay was performed using a T-maze system in which flies choose freely
between an illuminated vial and a dark vial. The test was repeated five times
with each fly (n = 16), and the percentage of photopositive choices was scored.
Data were analyzed with the Wilcoxon signed-rank test.
Visual Fixation and Orientation. Orientation toward a landmark was analyzed
in the Buridan’s paradigm on freely walking 5-d-old Drosophila with short-
ened wings. The angular deviation between fly trajectory and the
approached target was measured every 0.2 s for 15 min. Spatial orientation
memory was tested in the detour paradigm (26). Ten consecutive trials of at
least 10 flies per genotype were recorded. Data were analyzed with the
Kruskal–Wallis ANOVA test.
ACKNOWLEDGMENTS. We thank Erich Buchner for help with the electro-
retinogram and optomotor assays, Isabelle Rivals for advice on statistical
analyses, and Scott Waddell and Burkhard Poeck for helpful discussions. This
work was supported by grants from the Centre National de la Recherche
Scientifique and the Fondation de France (to S.B.), the National Institutes of
Health (to J.H.), and Agence Nationale pour la Recherche (to T.P.), and by
Deutsche Forschungsgemeinschaft Grant STR590-4 (to R.S.). Fellowships
were provided by the Fondation pour la Recherche Médicale (to H.C. and
T.R.) and Fondation Pierre-Gilles de Gennes (to L.S. and T.R.).
838
| www.pnas.org/cgi/doi/10.1073/pnas.1010930108 Riemensperger et al.