cdc42-dependent forgetting regulates repetition effect in ...€¦ · cell reports article...
TRANSCRIPT
Article
Cdc42-Dependent Forgett
ing Regulates RepetitionEffect in Prolonging Memory RetentionGraphical Abstract
Highlights
d Single-session training leads to ARM formation and Cdc42
activation
d Repeated learning prolongs ARM through inhibition of
Cdc42-mediated forgetting
d Cdc42 inhibition mimics repeated learning sessions in
prolonging ARM retention
d Increased Cdc42 activity abolishes repetition-induced ARM
extension
Zhang et al., 2016, Cell Reports 16, 817–825July 19, 2016 ª 2016 The Author(s).http://dx.doi.org/10.1016/j.celrep.2016.06.041
Authors
Xuchen Zhang, Qian Li, Lianzhang Wang,
Zhong-Jian Liu, Yi Zhong
In Brief
Forgetting is mediated through multiple
biological mechanisms. Zhang et al. find
that Cdc42 specifically regulates
forgetting of anesthesia-resistant
memory (ARM) without affecting other
known memory components in
Drosophila. Repeated learning prolongs
ARM retention through suppression of
Cdc42-dependent forgetting.
Cell Reports
Article
Cdc42-Dependent Forgetting Regulates RepetitionEffect in Prolonging Memory RetentionXuchen Zhang,1,4 Qian Li,1,4 Lianzhang Wang,1 Zhong-Jian Liu,2,3 and Yi Zhong1,*1Tsinghua-Peking Center for Life Sciences, IDG/McGovern Institute for Brain Research, MOE Key Laboratory of Protein Sciences, School ofLife Sciences, Tsinghua University, Beijing 100084, China2Shenzhen Key Laboratory for Orchid Conservation and Utilization, The National Orchid Conservation Center of China and The Orchid
Conservation & Research Center of Shenzhen, Shenzhen 518114, China3The Center for Biotechnology and Biomedicine, Graduate School at Shenzhen, Tsinghua University, Shenzhen 518055, China4Co-first author
*Correspondence: [email protected]
http://dx.doi.org/10.1016/j.celrep.2016.06.041
SUMMARY
Repeated learning is used daily and is a powerful wayto improve memory. A fundamental question is howmultiple learning trials add up to improve memory.While the major studies so far of such a repetition ef-fect have emphasized the strengthening of memoryformation, the current study reveals a molecularmechanism through suppression of forgetting. Wefind that single-session training leads to formationof anesthesia-resistant memory (ARM) and then acti-vation of the small G protein Cdc42 to cause decayor forgetting of ARM within 24 hr. Repetition sup-presses the activation of Cdc42-dependent forget-ting, instead of enhancing ARM formation, leadingto prolonged ARM. Consistently, inhibition of Cdc42activity through genetic manipulation mimicked therepetition effect, while repetition-induced ARMimprovementwas abolishedby elevatedCdc42 activ-ity. Thus, only the first session in repetitive trainingcontributes to ARM formation, while the subsequentsessions are devoted not to acquiring informationbut to inhibiting forgetting.
INTRODUCTION
It is a universal principle that repeated learning of the samemate-
rials can significantly strengthen memory representations and
make them more resistant to forgetting (Ebbinghaus, 1885/
1913; Kandel, 2001; McGaugh, 1966; Menzel and Muller, 1996;
Tully et al., 1994). InDrosophila, aversive conditioning yieldsmul-
tiple identifiablememorycomponents, includinganesthesia-sen-
sitive memory (ASM), anesthesia-resistant memory (ARM), and
protein synthesis-dependent long-term memory (LTM) (Davis,
2005; Heisenberg, 2003; Margulies et al., 2005). While LTM can
only be produced through spaced repetitive training (with resting
intervals between training sessions), either one-session or multi-
ple-sessionmassed training producesASMandARM (Tully et al.,
1994). ARM can be prolonged from less than 24 hr to 4 days
This is an open access article under the CC BY-N
through repetition of ten-sessionmassed training (without resting
interval between training sessions) (Tully et al., 1994). How
massed repetition prolongs ARM remains to be determined.
In psychology, a number of theories have been proposed
to explain the repetition effects on memory, such as the
strength hypothesis and multiple-trace hypothesis (Hintzman,
1988, 2010). In animal models, studies of repetition effects on
memory have focused on mechanisms underlying so-called
memory consolidation (Dudai, 2004; Kandel, 2001; McGaugh,
2000). In other words, studies of repetition effects in psychology
and animalmodeling so far have emphasized changes inmemory
formation. The current study finds that active forgetting, which is
referred to as training-induced or behaviorally evoked memory
decay, makes a major contribution to repetition effects on mem-
ory.We find that repeated training improvesARMby suppressing
Cdc42 activation, and thereby its mediated forgetting, instead
of strengthening ARM formation. In other words, only the first
session of training contributes to ARM formation; the subsequent
trials are devoted not to acquiring information but to inhibiting
forgetting of the ARM acquired through the first trial.
RESULTS
Repetition-Induced ARM Improvement Is Due to SlowerDecay but Not Enhanced FormationIn olfactory aversive conditioning of Drosophila, one-session
training produces two memory components: ASM that lasts
for about 5 hr and ARM that lasts for about 24 hr (Margulies
et al., 2005). In contrast to one-session training, repetition (ten-
session massed training) produces ARM that lasts for 4 days.
To determine the contribution of each subsequent training ses-
sion to ARM formation, we established ARM formation curves
of w1118 control flies (Canton-S flies carrying the w1118 muta-
tion) under different training paradigms: single session (13)
and two (23), four (43), and ten (103) massed sessions.
Because ARM can be readily isolated from ASM through cold-
shock treatment and its formation is largely confined to the first
hour after training (Quinn and Dudai, 1976; Tully et al., 1994),
cold-shock treatments were applied at different time points
within this first hour to determine the amount of ARM formed
at the time of cold shock. Time points for delivering cold shock
Cell Reports 16, 817–825, July 19, 2016 ª 2016 The Author(s). 817C-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Figure 1. Major Effect of Repeated Training
Sessions on ARM Decay but Not Formation
(A) No effect of repetition on ARM formation in
w1118 control flies. Left: schematic of training
paradigms and cold-shock information. Right: no
significant difference was found in 2 hr ARM that
was isolated by cold shock at different time points
(5.5 min, 12.5 min, 33.5 min, or 1 hr) relative to
13 training. n = 8–12.
(B) ARM retention curves of w1118 control flies
after different training paradigms. Compared with
13 training, flies after 23 massed training dis-
played no significant difference in ARM retention at
all time points, while flies after multiple massed
training (43 and 103) exhibited slower ARMdecay
(for the retention curve after 43 massed training,
p = 0.8675, p = 0.0321, p = 0.0003, p < 0.0001, and
p = 0.0115 for 3, 6, 12, 24, and 48 hr, respectively;
for the retention curve after 103 massed training,
p = 0.6590, p = 0.0030, p < 0.0001, p < 0.0001, and
p < 0.0001 for 3, 6, 12, 24, and 48, respectively).
n = 8–12.
(C) No significant difference in 3 hr ARM after 13
and 103massed training was found using either a
standard (60 V, left) or a weak (20 V, right) training
intensity. n = 12.
Results with error bars are means ± SEM. NS,
non-significance (p > 0.05). See also Figure S1 and
Table S1.
were chosen immediately after completion of 13, 23, 43,
and 103 training, which corresponded to 2, 5.5, 12.5, and
33.5 min, plus an additional time point at 60 min (Figure 1A,
left). For all trials, ARM isolated by cold shock at these desig-
nated time points was assayed at hr 2 after training (Figure 1A,
right). The data show that all training paradigms produced the
same time course of ARM formation, irrespective of the number
of training sessions involved. This suggests that the time course
of ARM formation is solely determined by the first session of
training and the subsequent training sessions, no matter how
many are involved, do not contribute to the formation of ARM.
To exclude the potential ceiling effect that could mask the
amount of ARM produced by multiple-session massed training,
we compared 2 hr ARM produced by different training strengths
(weak 20 V and normal 60 V for electric shock). There was no dif-
ference observed in ARM formation between the two training
strengths (Figure S1A).
We then investigated the detailed contribution of each addi-
tional training session toward ARM decay. As the number of
training sessions increased, ARM decay became significantly
slower at time points more than 3 hr after training, that is, at 6,
12, 24, and 48 hr (Figure 1B). ARM formed 3 hr after training
was similar for all paradigms, even using the weaker training
strength (Figure 1C). Taken together, the data presented
suggest that the repetition-induced ARM improvement re-
sulted from inhibition of forgetting, not from enhanced memory
formation.
ARM Decay Is Regulated by Cdc42 ActivityBidirectionallyWe next aimed to determine the molecular mechanisms of ARM
forgetting. Studies have increasingly focused on the biological
818 Cell Reports 16, 817–825, July 19, 2016
basis of active forgetting using different animal models (Akers
et al., 2014; Berry et al., 2012, 2015; Hadziselimovic et al.,
2014; Inoue et al., 2013; Shuai et al., 2010, 2015). In Drosophila,
because the small G protein Rac1 is reported to regulate active
forgetting of ASM and does not affect ARM (Shuai et al., 2010),
we assayed the effects of other Rho family members, Cdc42
and RhoA (Figure S1B). We found that Cdc42 specifically regu-
lates ARM forgetting without affecting ARM formation, as well
as without affecting ASM formation and forgetting.
Forgetting is generally characterized as two independent
forms: time-based passive memory decay and interference-
induced memory loss (Jonides et al., 2008; Wixted, 2004). To
determine its role in forgetting, Cdc42 activity was manipulated
in neurons through the acutely induced expression of two
mutants of Cdc42 (Luo et al., 1994; Osterwalder et al., 2001):
dominant-negative Cdc42(N17) for inhibiting endogenous activ-
ity by competing for upstream activators and constitutively
active Cdc42(V12) for elevating the activity through persistent
activation as a consequence of its abolished intrinsic guanosine
triphosphatase (GTPase) activity. The effectiveness of the ge-
netic manipulation of Cdc42 activity was confirmed by western
blotting (Figures 2A and 2C).
Wecompared retention curves at different timepoints after both
one-session training and ten-session massed training (Figures 2B
and 2D). Inhibition of Cdc42 activity in Cdc42(N17)-expressing
flies (UAS-Cdc42(N17)/+; elav-GS/+, RU486+) led to significantly
slower ARM decay in both cases compared with uninduced con-
trol flies (UAS-Cdc42(N17)/+; elav-GS/+, RU486�) (Figure 2B).
Conversely, elevated Cdc42 activity in Cdc42(V12)-expressing
flies (elav-GS/UAS-Cdc42(V12), RU486+) acceleratedARMdecay
(Figure 2D). Thus, Cdc42 activity regulates ARM decay or
forgetting.
Figure 2. Cdc42-Mediated Bidirectional Regulation of ARM Decay
To test ARM, flies were subjected to 2 min cold shock at 1 hr before testing. RU486 drug feeding was used to induce overexpression of specific transgenes.
(A) Representative western blot (left) and statistical graph (right). Compared to uninduced controls (RU486�), Cdc42(N17)-expressing flies (RU486+) showed
significant inhibition of Cdc42 activity. n = 6. *p < 0.05.
(B) ARM retention curves after 13 training (left) and 103massed training (right). Cdc42(N17)-expressing flies (RU486+) displayed apparently slower decay after
3 hr in contrast to uninduced controls (RU486�). For 13 training, p > 0.05 for 2 and 3 hr and p = 0.0006, p = 0.0049, and p = 0.0234 for 6, 9, and 12 hr; for 103
massed training, p > 0.05 for 2 and 3 hr and p = 0.0016, p = 0.0007, and p < 0.0001 for 12, 24, and 36 hr. n = 8.
(C) Increased Cdc42 activity in Cdc42(V12)-expressing flies (RU486+) compared with control group (RU486�). n = 6. *p < 0.05.
(D) Accelerated decay of ARM in Cdc42(V12)-expressing flies (RU486+) relative to controls (RU486�) after both 13 training (left) and 103massed training (right).
For 13 training, p > 0.05 for 2 and 3 hr and p = 0.0001, p = 0.0024, and p < 0.0001 for 6, 9, and 12 hr; for 103massed training, p > 0.05 for 2 and 3 hr and p = 0.0045,
p = 0.0014, and p = 0.0084 for 12, 24, and 36 hr. n = 8.
Results with error bars are means ± SEM. See also Figures S2–S4.
However, alteration of Cdc42 activity did not affect ARM for-
mation. ARM remained similar 2 and 3 hr after training (even
with weaker training strength) (Figure S2) in induced and unin-
duced transgenic flies for either inhibited or elevated Cdc42 ac-
tivity (see performance index [PI] at 2 and 3 hr in Figures 2B and
2D). From the decay curves of total memory (ASM+ARM) (Fig-
ures S3A and S3B), it appears that both ASM formation (3 min
after training) and its decay (from 3 min to 3 hr after training)
are independent of Cdc42 activity, because total memory is
almost identical during the time window critical for assaying
ASM for the controls and the transgenic flies with altered
Cdc42 activity. We then checked ASM retention by subtracting
ARM from total memory and found ASM decay was not affected
by Cdc42 activity (Figures S3C and S3D). Thus, Cdc42 activity
specifically regulates ARM decay or time-based forgetting.
To verify the physiological significance of the observed phe-
notypes, two independent RNAi fly strains, HMS01502 and
HMS02553, were used to knock down endogenous Cdc42.
Acute knockdown of Cdc42 led to normal formation of 3 hr
ARM but significantly slower decay 24 hr after ten-session
massed training (Figure 3). This effect was further confirmed in
one-session training (Figure S3E). Thus, endogenous Cdc42 ac-
tivity regulates ARM decay without affecting its formation.
Three additional experiments were included to solidify the
conclusion. First, we excluded any direct effects from drug
feeding (RU486+, for inducing targeted gene expression) per
se, because feeding of RU486 in parental control flies did not
yield relevant phenotypes (Figures S4A and S4B). Second, we
ruled out influence of the cold-shock regimen, because we ob-
tained similar Cdc42 phenotypes with cold-shock treatment at
different time points (Figures S4C and S4D). Third, we verified
that improved ARM was not a result of formation of protein syn-
thesis-dependent LTM (Figure S4E).
Because ARM formation is mapped to a brain region of the
mushroom body (MB) (Isabel et al., 2004; Lee et al., 2011), an
MB-specific driver, MB-GeneSwitch (MB-GS) Gal4 (Mao et al.,
2004), was used to target expression of Cdc42 mutants. Acutely
induced expression of Cdc42(N17) within MB neurons in adult
flies (UAS-Cdc42(N17)/+; MB-GS/+, RU486+) had no effect
on total memory and isolated ARM at 3 hr but slowed the
decay of total memory and ARM at later time points (12 and
24 hr) after ten-session massed training, relative to uninduced
Cell Reports 16, 817–825, July 19, 2016 819
Figure 3. Prolonged ARM Caused by Acute Knockdown of Endog-
enous Cdc42
Slower decay of ARM by knocking down Cdc42 expression after 103massed
training. Two RNAi stocks were used: UAS-Cdc42-RNAiHMS01502 (left)
and UAS-Cdc42-RNAiHMS02553 (right). The RU486� group served as control.
n = 8–12. Results with error bars are means ± SEM. *p < 0.05. NS, non-sig-
nificance (p > 0.05).
Figure 4. Cdc42-Mediated Bidirectional Regulation of ARMDecay in
the MB
(A) Prolonged ARM caused by inhibiting Cdc42 activity within MB neurons.
Relative to uninduced controls (RU486�), Cdc42(N17)-expressing flies
(RU486+) displayed slower decay of both total memory (left) and ARM (right)
after 3 hr.
(B) Accelerated decay of ARMby increasing Cdc42 activity withinMB neurons.
Compared to uninduced controls (RU486�), Cdc42(V12)-expressing flies
(RU486+) sped up ARM decay after 3 hr.
n = 12. Results with error bars are means ± SEM. *p < 0.05. NS, non-signifi-
cance (p > 0.05).
control flies (UAS-Cdc42(N17)/+; MB-GS/+, RU486�) (Fig-
ure 4A). Conversely, induced expression of Cdc42(V12) in the
MB (MB-GS/UAS-Cdc42(V12), RU486+) did not affect total
memory and ARM at 3 hr but sped up its decay at 12 and
24 hr (Figure 4B). Therefore, the effects of Cdc42 on ARM decay
are confined to the MB.
Interference-Based Forgetting Is Regulated by Cdc42Activity BidirectionallyUp to now, our study focused on time-based ARM decay. To
determine how Cdc42 activity affects interference-based forget-
ting, the following interference paradigm was used to assay the
effects of retroactive interference (Melton and von Lackum,
1941): ten-session massed aversive conditioning (interference)
with a pair of novel odors (ethyl acetate [EA] versus isoamyl
acetate [IA]) was performed immediately after the ordinary
ten-session massed training (3-octanol [OCT] versus 4-methyl-
cyclohexanol [MCH]) (Figure 5A). The consequences of such
interference were evaluated by assaying ARM retention (OCT
versus MCH) at 2 and 3.6 hr. A recent study found a generaliza-
tion phenomenon in appetitive LTM (Ichinose et al., 2015). It rai-
ses a possibility that 103 EA versus IA training may affect the
choice between OCT andMCH. As shown in Figure S5, this pos-
sibility was ruled out from our case. In uninduced control flies
(UAS-Cdc42(N17)/+; elav-GS/+ and elav-GS/UAS-Cdc42(V12),
RU486�), interference did not perturb the 2 hr retention of
ARM but caused significantly lower ARM at 3.6 hr, showing
that interference does not affect ARM formation but accelerates
ARM decay or causes interference-based forgetting (Figures 5B
and 5C). This interference-based ARM forgetting was also
observed in w1118 control flies (data not shown).
In transgenic flies with altered Cdc42 activities, 2 hr ARM (for-
mation) was also not affected by interference training (Figures 5B
and 5C, left). However, the interference-induced forgetting
at 3.6 hr was suppressed by inhibition of Cdc42 activity in
Cdc42(N17)-expressing flies (Figure 5B, right) and enhanced
820 Cell Reports 16, 817–825, July 19, 2016
by elevation of Cdc42 activity in Cdc42(V12)-expressing flies
(Figure 5C, right).
Time-Based and Interference-Based Forgetting of ARMCorrelate with Cdc42 ActivityTaken together, we found that genetic manipulation of Cdc42
activity specifically regulates ARM forgetting without affecting
ARM formation, as well as ASM formation and forgetting. To
establish the behavioral relevance of Cdc42 activity, we per-
formed western blotting to assess whether Cdc42 is activated
to mediate passive decay and interference of ARM. Fly heads
were collected, for single-session conditioning, 3 hr after the
training at which ARM decay begins or immediately after the
interference training (ten-session OCT-MCH massed training
followed by ten-session EA-IA massed interference training) at
which forgetting is evident (Figure 5B). In the single session,
naive flies serve as the control, while in interference assaying,
two repetitive ten-session massed trainings (no difference of
3.6 hr ARM compared with ten-session massed training) (Fig-
ure S6A) are used as the control. In either case, guanosine
triphosphate (GTP)-bound Cdc42 (activated) was significantly
Figure 5. Cdc42-Mediated Bidirectional
Regulation of Interference-Based Forgetting
of ARM
(A) Schematic of the training protocol used for
interference-based forgetting of ARM.
(B and C) Left: no effect of interference training
(EA versus IA) on 2 hr ARM retention of the
prior memory (OCT versus MCH) in both induced
(RU486+) and uninduced (RU486�) flies. Right:
accelerated ARM decay at 3.6 hr in uninduced
controls (RU486�) was suppressed in Cdc42(N17)-
expressing flies (RU486+) (B) and sped up
in Cdc42(V12)-expressing flies (RU486+) (C).
n = 8–12. Results with error bars aremeans ± SEM.
*p < 0.05. NS, non-significance (p > 0.05).
See also Figure S5.
increased (Figure 6). In other words, Cdc42 activity is elevated
3 hr after single-session conditioning (no significant change at
2 hr) (Figure S6B). This activation is supposedly responsible for
triggering ARM forgetting, because inhibition of this activity pro-
longs ARM (Figure 2). Cdc42 activity is also evoked by interfer-
ence training, which presumably leads to interference-based
ARM forgetting as blockade of this activity suppresses forgetting
(Figure 5B).
The Repetition Effect in Improving ARM Is Regulated byCdc42 ActivityThe finding of Cdc42-dependent forgetting for ARM raises the
possibility that repetitive massed training-induced prolonged
ARM is the result of inhibition of the activation of Cdc42-depen-
dent forgetting. In other words, the decay of single-session-
induced ARM is caused by training-evoked Cdc42 activation,
while ten-session massed repetition suppressed activation,
thereby prolonging ARM by blocking forgetting. Two lines of
evidence presented below support this idea.
First, we performed western blotting using extracts from whole
heads. Because the extracts were from whole heads rather than
just from cells in the MB, where memory-specific changes are
C
thought to reside, the changes we
observed in Cdc42 activity cannot be
attributed specifically to cells involved
in memory formation and must reflect
more widespread changes than those
related to memory dynamics. Never-
theless, as shown earlier, one-session
training led toan increase inCdc42activity,
relative to thenaivecontrol (Figure7A). This
Cdc42 activation was suppressed in
ten-session massed training (Figure 7A).
Second, to verify the behavioral con-
sequences of such suppression of
Cdc42 activation, Cdc42 activity was
genetically manipulated. Inhibition of
Cdc42 in Cdc42(N17)-expressing flies
(UAS-Cdc42(N17)/+; elav-GS/+, RU486+)
yielded a single training-induced ARM at
24 and 48 hr similar to that produced by
ten sessions of repetitive massed training in uninduced control
flies (UAS-Cdc42(N17)/+; elav-GS/+, RU486�) (Figure 7B).
Conversely, ten-session massed training-induced 24 hr ARM in
flies with elevated Cdc42 activity (elav-GS/UAS-Cdc42(V12),
RU486+) was similar to that induced by single-session training in
uninduced control flies (elav-GS/UAS-Cdc42(V12), RU486�) (Fig-
ure 7C). Thus, Cdc42-dependent forgetting regulates the repeti-
tion effect in ARM improvement.
DISCUSSION
The current work identifies activation of Cdc42 as a signal trans-
duction mechanism mediating the time-based and interference-
based forgetting of ARM. This activation exerts no effects on
ARM formation or on either formation or forgetting of ASM.
ARM could be prolonged dramatically through repetition of
training, and this improvement is achieved through inhibition of
Cdc42-dependent forgetting, rather than through enhancement
of memory formation.
Memory can be divided into labile and consolidated forms in
both invertebrates and vertebrates (DeZazzo and Tully, 1995;
Kandel, 2001; McGaugh, 2000). In Drosophila, a single trial of
ell Reports 16, 817–825, July 19, 2016 821
Figure 6. Cdc42 Activity Evoked by Time-
Based and Interference-Based Forgetting
of ARM
Representative western blots (left) and statistical
graph (right).
(A) In contrast to naive flies, Cdc42 activity was
increased 3 hr after 13 training.
(B) Interference training (interference+) increased
Cdc42 activity immediately after training relative to
control group without interference (interference�).
n = 6. Results with error bars are means ± SEM.
*p < 0.05. See also Figure S6.
aversive conditioning training results in at least two memory
components, including a labile ASM that lasts several hours
and a consolidated ARM that extends close to 24 hr (Quinn
and Dudai, 1976; Tully et al., 1994). ASM formation requires
the cyclic AMP pathway, which is well established in thememory
formation of other organisms (Davis, 2005; Elgersma and Silva,
1999; Kandel, 2001; Margulies et al., 2005). A previous study
demonstrates a Rac1-dependent signaling pathway for forget-
ting of labile ASM independent from formation (Shuai et al.,
2010). To date, the radish gene is a well-known and distinct mo-
lecular link to the poorly understood ARM formation mechanism
(Folkers et al., 1993, 2006). In the current study, we show that a
Cdc42-dependent pathway is responsible for ARM forgetting
but has no apparent influence on formation. Combined with
the previous studies inDrosophila (Berry et al., 2012, 2015; Shuai
et al., 2010, 2011, 2015), it appears that not only for ASMbut also
for ARM, there are molecular mechanisms devoted specifically
to their forgetting, while their formation is mediated through
distinct signaling pathways. Different forgetting mechanisms
have been reported in other organisms (Akers et al., 2014; Had-
ziselimovic et al., 2014; Inoue et al., 2013). It will be of interest to
study whether there is a distinct forgetting mechanism for LTM.
Besides the forgetting mechanism in Drosophila, Cdc42 and
Rac1 play distinct roles in neuronal plasticity and memory in
other animal models in a pattern consistent with that in flies.
A study of Aplysia shows that Cdc42, but not Rac1, is required
in long-term facilitation (Udo et al., 2005). Studies of mice reveal
that loss of Rac1 specifically impairs working memory (Ha-
ditsch et al., 2009), while loss of Cdc42 specifically impairs
remote memory (Kim et al., 2014). Together with studies from
Drosophila, these findings suggest that Cdc42 and Rac1 might
be generally designed for memory processing of labile and
consolidated forms, respectively.
Cdc42 and Rac1 share 70% amino acid identity and belong to
the same family of Rho GTPases (Luo et al., 1994). How can they
play distinct roles in forgetting? This idea can be taken further by
considering their well-established roles in regulating actin-based
protrusions of both non-neuronal cells and neurons, because
Cdc42 and Rac1 regulate the formation of filopodia and lamelli-
podia, respectively (Garvalov et al., 2007; Hall, 1998; Heasman
and Ridley, 2008; Irie and Yamaguchi, 2002; Luo, 2000). These
different roles of Cdc42 and Rac1 in actin-based protrusions of
822 Cell Reports 16, 817–825, July 19, 2016
neurons may help to explain their different roles in memory
forgetting.
Our study of ARM reveals a major contribution from inhibition
of forgetting in repetitive training-inducedmemory improvement.
We found that ARM formation curves are almost identical among
one-session training and two-, four-, and ten-session massed
trainings in w1118 control flies (Figures 1A and S1A). ARM reten-
tion at 3 hr is also indistinguishable between one-session training
and multiple-session massed training, even with weak training
intensity (Figures 1B and 1C). As the time passes, w1118 control
flies exhibit significantly slower decay when the number of
repeated training sessions increases. This suggests that subse-
quent repeated training sessions mainly contribute to inhibition
of forgetting of ARM but not formation. This conclusion is sup-
ported consistently by multiple lines of evidence. First, ARM for-
mation is generally believed to be largely completed within the
first hour after training (Quinn and Dudai, 1976; Tully et al.,
1994). This is consistent with our data in Figure 1. Within this
time window, ARM is gradually formed and repetition exerts no
effect on ARM formation. Second, to rule out a potential ceiling
effect, we use weak training intensity (20 V) to confirm that
no enhancement of ARM formation is induced by repetition (Fig-
ure S1). Third, one-session training leads to activation of Cdc42
3 hr later, presumably for erasing ARM, whereas ten-session
massed training inhibits activated Cdc42, supposedly leading
to extension of ARM (Figure 7A). Fourth, acutely induced inhibi-
tion of Cdc42 activity makes one-session training-induced ARM
decay occur at a rate similar to that induced by ten-session
training, while elevated Cdc42 activity accelerates ten-session-
induced ARM decay like that induced by one-session training
(Figures 7B and 7C). In addition, a previous study revealed that
if ARM formation is enhanced, it could be observed in the first
3 hr after training (Horiuchi et al., 2008). Even on basis of such
strong supports, we may not rule out a possibility of a masked
formation effect, as implicated in one study (Aso et al., 2012).
Together, our findings provide an insight into the molecular
and cognitive mechanisms underlying repetition-induced mem-
ory improvement: the first training mainly encodes memory for-
mation, while the subsequent trainingsmainly encode forgetting.
In other words, ARM formation is solely a result of the first
training session, and the rest of the training sessions make
no contribution toward acquiring additional memory but are
Figure 7. Cdc42-Activity-Mediated Repeti-
tion Effect of ARM
(A) Representative western blots (left) and statisti-
cal graph (right). Compared to naive flies, Cdc42
activity significantly increased 3 hr after 13 training
but not 103 massed training. n = 6.
(B andC) Uninduced control flies (RU486�) showed
apparently slower decay of ARM after 103massed
training compared with 13 training. This repetition-
induced ARM improvement was mimicked in
Cdc42(N17)-expressing flies (RU486+) after 13
training (B) andabolished inCdc42(N17)-expressing
flies (RU486+) after 103massed training (C). n = 8.
Results with error bars are means ± SEM. *p <
0.05. NS, non-significance (p > 0.05).
devoted to inhibiting forgetting of memory acquired through the
first training.
EXPERIMENTAL PROCEDURES
Fly Stocks
Flies were cultured on a 12 hr light-dark cycle schedule at 25�C and 60% rela-
tive humidity with standard medium. elav-GS and MB-GS were gifts from
Dr. Ronald L. Davis. The stocks acquired from the Bloomington Stock Center
were as follows: UAS-Cdc42(V12), UAS-Cdc42(N17), and UAS-RhoA(N19).
Two RNAi stocks were acquired from Tsinghua Fly Center: UAS-Cdc42-
RNAiHMS01502 and UAS-Cdc42-RNAiHMS02553. The flies (Canton-S flies car-
rying the w1118 mutation) served as controls for some experiments.
Behavioral Assays
Pavlovian Olfactory Aversive Conditioning
Two- to five-day-old flies were raised for behavior experiments under a clas-
sical Pavlovian olfactory conditioning procedure (Tully et al., 1994; Tully and
Quinn, 1985). Briefly, flies were first transferred to a behavioral room at 25�Cand 60% relative humidity to adapt to the environment for 30 min. During
training, around 100 flies were successively exposed to two aversive odors
(used as conditional stimulus [CS]), OCT (1.53 10�3 in dilution; Sigma-Aldrich)
and MCH (1.0 3 10�3 in dilution; Fluka), for 60 s, with 45 s fresh air after each
odor. Flies were exposed to the first odor paired with 12 pulses of electric foot
shock (used as unconditioned stimulus [US]) at 60 V (CS+) followed by a sec-
ond odor without shock (CS�). This process generated a standard one-cycle
training session (13 training). Ten sequential cycles without inter-trial intervals
constituted massed training (103 training). In some experiments, to weaken
the training intensity, the number of electric shock pulses was modulated
from 12 to 2 or the voltage was altered from 60 to 20 V. To measure memory,
trained flies were transferred into a T maze, where they were allowed 2 min to
choose between two odors CS+ and CS�. Memory retention was quantified
by a PI calculated from the fraction of flies in the two T-maze arms. A PI of
100 indicated all flies made the right choice to avoid the odor paired with
shock, while a PI of 0 manifested no memory retention, with a 50:50 distribu-
C
tion between the arms. To balance naive odor bias,
two reciprocal groups were trained and tested
simultaneously; one group was trained to asso-
ciate OCT with shock, and the other was trained
to associate MCH with shock. The complete PI
was defined as the average of the two groups.
For 3 min memory, flies were tested immediately
after training. For longer memory retention, flies
were placed in a vial (with the same content they
had been kept in before training) until the test.
Massed Interference Conditioning
Massed retroactive interference was carried out
immediately after the initial massed training
(OCT/MCH) by transferring flies to new conditioning with a novel pair of odors,
EA (2.0 3 10�3 in dilution; Alfa Aesar) and IA (2.0 3 10�3 in dilution; Avocado
Research Chemicals), as CS+/CS�. In experiments of massed interference
conditioning, ten-session massed training (OCT/MCH) followed immediately
by a novel ten-session massed training (EA/IA) served as interference+,
whereas only one set of ten-session massed training (OCT/MCH) without
following interference served as interference�. Specifically for western blot,
two sets of ten-sessionmassed trainings (OCT/MCH) served as interference�.
Cold-Shock Regimen
The procedures were described previously (Tully et al., 1994). For cold-shock
experiments to measure ARM, trained flies were transferred to a prechilled
glass vial in an ice-water mixture (�0�C) and remained there for 2 min. After
the treatment, the anesthetized flies were permitted to recover in vials (with
the same content as before training) for 1 hr before the test. In all experiments,
ARM was tested at different time points (1, 2, 3, 6, 9, 12, 24, 36, or 48 hr) after
training, and the cold shock was given to flies 1 hr before the memory test,
except for the experiments shown in Figures 1A, S1A, S4C, and S4D.
Drug Feeding Treatment
RU486 drug feeding was described previously (Huang et al., 2012). Flies in the
RU486� group were fed control solution (5% glucose and 3% ethanol). In
contrast, flies in the RU486+ group were fed 500 mM RU486 (Mifepristone,
J&K Scientific) dissolved in control solution. In all experiments involving elav-
GS or MB-GS flies were fed RU486� or RU486+ solution for 2 days before
and after training until memory retention was tested. For cycloheximide
(CXM) feeding, flies were fed with (CXM+) or without (CXM�) 35 mM (Sigma)
dissolved in control solution as described previously (Huang et al., 2012).
Cdc42 Activity Assay
Briefly, for each sample, about 400 fly heads were collected and homogenized
in Mg2+ lysis buffer (EMD Millipore). After centrifugation at 13, 000 rpm for
15 min at 4�C, the supernatant was divided into two parts. A small fraction
was used for total Cdc42 detection by western blot. A large fraction was incu-
bated with PAK-PBD beads (Cytoskeleton), which can bind the GTP-bound
ell Reports 16, 817–825, July 19, 2016 823
form of Cdc42 for 1 hr at 4�C. Then, GTP-bound Cdc42 pulled down by beads
was examined by western blot. The antibody of Cdc42 (1:200 dilution; Santa
Cruz) was used as the primary antibody in western blot for both total and
GTP-bound Cdc42 detection. The second antibody was purchased from
Cell Signaling Technology. Data analysis was performed by NIH ImageJ
software.
Statistics
Statistical analysis was performed using Prism (GraphPad). All data in
graphs and tables are normally distributed and shown as means ± SEM.
Bonferroni post hoc tests were performed when one-way or two-way
ANOVA revealed significant differences among multiple groups. Unpaired,
two-tailed t tests were used to identify significant differences in experiments
comparing two conditions. Any p values <0.05 were considered statistically
significant and marked with an asterisk, and NS indicates non-significance
(p > 0.05).
SUPPLEMENTAL INFORMATION
Supplemental Information includes six figures and one table and can be found
with this article online at http://dx.doi.org/10.1016/j.celrep.2016.06.041.
AUTHOR CONTRIBUTIONS
X.Z., Q.L., and Y.Z. conceived and designed this project. X.Z. and L.W. per-
formed all behavior experiments. Q.L. performed western blotting and
analyzed the data. X.Z., Q.L., Z.-J.L., and Y.Z. wrote the manuscript.
ACKNOWLEDGMENTS
We thank Dr. Ronald L. Davis, Bloomington Stock Center and Tsinghua Fly
Center, for fly stocks and Wenjuan Wu for sharing the experimental technique
of Cdc42 activity assay. This work was supported by grants from the National
Basic Research Project (973 program) of the Ministry of Science and Technol-
ogy of China (2013cb835100 to Y.Z.), the National Science Foundation
of China (91332207 to Y.Z.), the Tsinghua-Peking Joint Center for Life Sci-
ences, and the Tsinghua University Initiative Scientific Research Program
(20111080956 to Y.Z.).
Received: March 3, 2016
Revised: April 29, 2016
Accepted: June 7, 2016
Published: July 7, 2016
REFERENCES
Akers, K.G., Martinez-Canabal, A., Restivo, L., Yiu, A.P., De Cristofaro, A.,
Hsiang, H.L., Wheeler, A.L., Guskjolen, A., Niibori, Y., Shoji, H., et al. (2014).
Hippocampal neurogenesis regulates forgetting during adulthood and infancy.
Science 344, 598–602.
Aso, Y., Herb, A., Ogueta, M., Siwanowicz, I., Templier, T., Friedrich, A.B., Ito,
K., Scholz, H., and Tanimoto, H. (2012). Three dopamine pathways induce
aversive odor memories with different stability. PLoS Genet. 8, e1002768.
Berry, J.A., Cervantes-Sandoval, I., Nicholas, E.P., and Davis, R.L. (2012).
Dopamine is required for learning and forgetting in Drosophila. Neuron 74,
530–542.
Berry, J.A., Cervantes-Sandoval, I., Chakraborty, M., and Davis, R.L. (2015).
Sleep facilitates memory by blocking dopamine neuron-mediated forgetting.
Cell 161, 1656–1667.
Davis, R.L. (2005). Olfactory memory formation in Drosophila: from molecular
to systems neuroscience. Annu. Rev. Neurosci. 28, 275–302.
DeZazzo, J., and Tully, T. (1995). Dissection of memory formation: from behav-
ioral pharmacology to molecular genetics. Trends Neurosci. 18, 212–218.
Dudai, Y. (2004). The neurobiology of consolidations, or, how stable is the
engram? Annu. Rev. Psychol. 55, 51–86.
824 Cell Reports 16, 817–825, July 19, 2016
Ebbinghaus, H.E. (1885/1913). Memory: A Contribution to Experimental Psy-
chology (Dover).
Elgersma, Y., and Silva, A.J. (1999). Molecular mechanisms of synaptic plas-
ticity and memory. Curr. Opin. Neurobiol. 9, 209–213.
Folkers, E., Drain, P., and Quinn, W.G. (1993). Radish, a Drosophila mutant
deficient in consolidated memory. Proc. Natl. Acad. Sci. USA 90, 8123–8127.
Folkers, E., Waddell, S., and Quinn, W.G. (2006). The Drosophila radish gene
encodes a protein required for anesthesia-resistant memory. Proc. Natl.
Acad. Sci. USA 103, 17496–17500.
Garvalov, B.K., Flynn, K.C., Neukirchen, D., Meyn, L., Teusch, N., Wu, X.,
Brakebusch, C., Bamburg, J.R., and Bradke, F. (2007). Cdc42 regulates cofilin
during the establishment of neuronal polarity. J. Neurosci. 27, 13117–13129.
Haditsch, U., Leone, D.P., Farinelli, M., Chrostek-Grashoff, A., Brakebusch, C.,
Mansuy, I.M., McConnell, S.K., and Palmer, T.D. (2009). A central role for the
small GTPase Rac1 in hippocampal plasticity and spatial learning and mem-
ory. Mol. Cell. Neurosci. 41, 409–419.
Hadziselimovic, N., Vukojevic, V., Peter, F., Milnik, A., Fastenrath, M., Fenyves,
B.G., Hieber, P., Demougin, P., Vogler, C., de Quervain, D.J., et al. (2014).
Forgetting is regulated via Musashi-mediated translational control of the
Arp2/3 complex. Cell 156, 1153–1166.
Hall, A. (1998). Rho GTPases and the actin cytoskeleton. Science 279,
509–514.
Heasman, S.J., and Ridley, A.J. (2008). Mammalian Rho GTPases: new
insights into their functions from in vivo studies. Nat. Rev. Mol. Cell Biol. 9,
690–701.
Heisenberg, M. (2003). Mushroom body memoir: from maps to models. Nat.
Rev. Neurosci. 4, 266–275.
Hintzman, D.L. (1988). Judgments of frequency and recognition memory in a
multiple-trace memory model. Psychol. Rev. 95, 528–551.
Hintzman, D.L. (2010). How does repetition affect memory? Evidence from
judgments of recency. Mem. Cognit. 38, 102–115.
Horiuchi, J., Yamazaki, D., Naganos, S., Aigaki, T., and Saitoe, M. (2008). Pro-
tein kinase A inhibits a consolidated form of memory in Drosophila. Proc. Natl.
Acad. Sci. USA 105, 20976–20981.
Huang, C., Zheng, X., Zhao, H., Li, M., Wang, P., Xie, Z., Wang, L., and Zhong,
Y. (2012). A permissive role of mushroom body a/b core neurons in long-term
memory consolidation in Drosophila. Curr. Biol. 22, 1981–1989.
Ichinose, T., Aso, Y., Yamagata, N., Abe, A., Rubin, G.M., and Tanimoto, H.
(2015). Reward signal in a recurrent circuit drives appetitive long-termmemory
formation. eLife 4, e10719.
Inoue, A., Sawatari, E., Hisamoto, N., Kitazono, T., Teramoto, T., Fujiwara, M.,
Matsumoto, K., and Ishihara, T. (2013). Forgetting in C. elegans is accelerated
by neuronal communication via the TIR-1/JNK-1 pathway. Cell Rep. 3,
808–819.
Irie, F., and Yamaguchi, Y. (2002). EphB receptors regulate dendritic spine
development via intersectin, Cdc42 and N-WASP. Nat. Neurosci. 5, 1117–
1118.
Isabel, G., Pascual, A., and Preat, T. (2004). Exclusive consolidated memory
phases in Drosophila. Science 304, 1024–1027.
Jonides, J., Lewis, R.L., Nee, D.E., Lustig, C.A., Berman, M.G., and Moore,
K.S. (2008). The mind and brain of short-term memory. Annu. Rev. Psychol.
59, 193–224.
Kandel, E.R. (2001). The molecular biology of memory storage: a dialogue
between genes and synapses. Science 294, 1030–1038.
Kim, I.H., Wang, H., Soderling, S.H., and Yasuda, R. (2014). Loss of Cdc42
leads to defects in synaptic plasticity and remote memory recall. eLife 3, 1–16.
Lee, P.T., Lin, H.W., Chang, Y.H., Fu, T.F., Dubnau, J., Hirsh, J., Lee, T., and
Chiang, A.S. (2011). Serotonin-mushroom body circuit modulating the forma-
tion of anesthesia-resistant memory in Drosophila. Proc. Natl. Acad. Sci. USA
108, 13794–13799.
Luo, L. (2000). Rho GTPases in neuronal morphogenesis. Nat. Rev. Neurosci.
1, 173–180.
Luo, L., Liao, Y.J., Jan, L.Y., and Jan, Y.N. (1994). Distinct morphogenetic
functions of similar small GTPases: Drosophila Drac1 is involved in axonal
outgrowth and myoblast fusion. Genes Dev. 8, 1787–1802.
Mao, Z., Roman, G., Zong, L., and Davis, R.L. (2004). Pharmacogenetic rescue
in time and space of the rutabaga memory impairment by using Gene-Switch.
Proc. Natl. Acad. Sci. USA 101, 198–203.
Margulies, C., Tully, T., and Dubnau, J. (2005). Deconstructing memory in
Drosophila. Curr. Biol. 15, R700–R713.
McGaugh, J.L. (1966). Time-dependent processes in memory storage. Sci-
ence 153, 1351–1358.
McGaugh, J.L. (2000). Memory—a century of consolidation. Science 287,
248–251.
Melton, A.W., and von Lackum, W.J. (1941). Retroactive and proactive inhibi-
tion in retention. Evidence for a two-factor theory of retroactive inhibition. Am.
J. Psychol. 54, 157–173.
Menzel, R., and Muller, U. (1996). Learning and memory in honeybees: from
behavior to neural substrates. Annu. Rev. Neurosci. 19, 379–404.
Osterwalder, T., Yoon, K.S., White, B.H., and Keshishian, H. (2001). A condi-
tional tissue-specific transgene expression system using inducible GAL4.
Proc. Natl. Acad. Sci. USA 98, 12596–12601.
Quinn, W.G., and Dudai, Y. (1976). Memory phases in Drosophila. Nature 262,
576–577.
Shuai, Y., Lu, B., Hu, Y., Wang, L., Sun, K., and Zhong, Y. (2010). Forgetting is
regulated through Rac activity in Drosophila. Cell 140, 579–589.
Shuai, Y., Hu, Y., Qin, H., Campbell, R.A., and Zhong, Y. (2011). Distinct
molecular underpinnings of Drosophila olfactory trace conditioning. Proc.
Natl. Acad. Sci. USA 108, 20201–20206.
Shuai, Y., Hirokawa, A., Ai, Y., Zhang,M., Li, W., and Zhong, Y. (2015). Dissect-
ing neural pathways for forgetting in Drosophila olfactory aversive memory.
Proc. Natl. Acad. Sci. USA 112, E6663–E6672.
Tully, T., and Quinn, W.G. (1985). Classical conditioning and retention
in normal and mutant Drosophila melanogaster. J. Comp. Physiol. A 157,
263–277.
Tully, T., Preat, T., Boynton, S.C., and Del Vecchio, M. (1994). Genetic dissec-
tion of consolidated memory in Drosophila. Cell 79, 35–47.
Udo, H., Jin, I., Kim, J.H., Li, H.L., Youn, T., Hawkins, R.D., Kandel, E.R., and
Bailey, C.H. (2005). Serotonin-induced regulation of the actin network for
learning-related synaptic growth requires Cdc42, N-WASP, and PAK inAplysia
sensory neurons. Neuron 45, 887–901.
Wixted, J.T. (2004). The psychology and neuroscience of forgetting. Annu.
Rev. Psychol. 55, 235–269.
Cell Reports 16, 817–825, July 19, 2016 825
Cell Reports, Volume 16
Supplemental Information
Cdc42-Dependent Forgetting Regulates Repetition
Effect in Prolonging Memory Retention
Xuchen Zhang, Qian Li, Lianzhang Wang, Zhong-Jian Liu, and Yi Zhong
Figure S1: Related to Figure 1
(A) Schematic of training paradigms and cold shock information (left). w1118 control flies were subjected to four training
paradigms (1×, 2×, 4× and 10×massed training) under weak training intensity (20V). Starting time point (0 min) to
measure memory retention was set as the time point immediate after the first training of all training paradigms. No
significant difference was found in 2-hr ARM isolated by cold shock at different time points (33.5 min or 1 hr) (right).
n=8-12. Results with error bars are means ± SEM. n.s. indicates non-significance (P > 0.05).
(B) Prolonged memory retention caused by inhibition of Cdc42 (Cdc42(N17) flies, RU486+) but not inhibition of RhoA
(RhoA(N19) flies, RU486+). RU486- served as controls. n=8-12. Results with error bars are means ± SEM. Asterisk, P <
0.05. n.s. indicates non-significance (P > 0.05).
Figure S2: No effect on ARM formation by acute genetic manipulation of Cdc42. Related to Figure 2.
(A, B) Total memory and ARM retention at 3 hr after weak massed training using reduced pulse number (left) or voltage
(right). Compared with uninduced controls (RU486-), no performance difference was found in either Cdc42(N17)-
expressing flies (A) or Cdc42(V12)-expressing flies (B). n=8. Results with error bars are means ± SEM. n.s. indicates non-
significance (P > 0.05).
Figure S3: Related to Figure 2
(A) Total memory retention curves after 1×training (left) and 10×massed training (right). In contrast to uninduced controls
(RU486-), Cdc42 (N17)-expressing flies (RU486+) displayed similar ARM retention in the first 3 hrs after training but
slower decay in the later time periods. For 1×training, P > 0.05 for 3 min, 1 hr, 2 hr and 3 hr; P = 0.0184, 0.0149 and 0.0007
for 6 hr, 9 hr and 12 hr, respectively. For 10×massed training, P > 0.05 for 2 hr and 3 hr; P = 0.0016, 0.0205 and 0.0331
for 12 hr, 24 hr and 36 hr, respectively. n=8. Error bars are SEM.
(B) Total memory retention curves after 1×training (left) and 10×massed training (right). Compared to uninduced controls
(RU486-), Cdc42(V12)-expressing flies (RU486+) displayed similar ARM retention in the first 3 hrs after training but
accelerated decay in the later time periods. For 1×training, P > 0.05 for 3 min, 1 hr, 2 hr and 3 hr; P = 0.0001, 0.0070 and
0.0207 for 6 hr, 9 hr and 12 hr, respectively. For 10×massed training, P > 0.05 for 2 hr and 3 hr; P = 0.0060, 0.0022 and
0.0179 for 12 hr, 24hr and 36 hr, respectively. n=8. Error bars are SEM.
(C and D) No effect of Cdc42 on ASM retention. Inhibition (C) or elevation (D) of Cdc42 activity did not affect ASM
retention at 2 hr, 3 hr and 6 hr after one-session training. n=8-10. Results with error bars are means ± SEM. n.s. indicates
non-significance (P > 0.05).
(E) Prolonged ARM caused by knockdown of Cdc42. Acutely knocking down endogenous Cdc42 (RU486+) showed a
slower decay of ARM after one-session training, in contrast to RU486- controls. n=8-10. Results with error bars are means
± SEM. Asterisk, P < 0.05. n.s. indicates non-significance (P > 0.05).
Figure S4: Related to Figure 2
(A and B) Cdc42-mediated regulation of ARM decay compared with parental controls. (A) Cdc42(N17)-expressing flies
(dark bar) displayed significantly higher ARM retention at 24 hr but not at 3 hr after 10×massed training in contrast to
parental controls (elav-GS/+ and UAS-Cdc42(N17)/+). (B) Cdc42(V12)-expressing flies (dark bar) exhibited remarkably
lower ARM retention at 24 hr but not at 3 hr than parental controls (elav-GS/+ and UAS-Cdc42(V12)/+). n=8. Results
with error bars are means ± SEM. Asterisk, P < 0.05. n.s. indicates non-significance (P > 0.05).
(C and D) Cdc42 effects on ARM retention using another cold-shock paradigm. Flies were subjected to 2-min cold shock
at 2 hr after 10-session massed training and rested till the test at different time points. (C) Cdc42(N17)-expressing flies
(RU486+) showed slower decay of ARM than controls (RU486-). n=8-12. (D) Cdc42(V12)-expressing flies (RU486+)
displayed accelerated ARM decay compared with controls (RU486-). n=12. Results with error bars are means ± SEM.
Asterisk, P < 0.05. n.s. indicates non-significance (P > 0.05).
(E) Component analysis of the extended memory in Cdc42(N17)-expressing flies. Cdc42(N17)-expressing flies (RU486+)
showed prominently higher 24-hr ARM retention than uninduced control flies (RU486-) after 10×massed training, with
(CXM+) or without (CXM-) CXM feeding. n=8. Results with error bars are means ± SEM. Asterisk, P < 0.05. n.s.
indicates non-significance (P > 0.05).
Figure S5: Related to Figure 5
No effect of 10×EA/IA training on OCT/MCH choice. Flies after 10×EA/IA massed training showed no difference
relative to naive control in choice between OCT and MCH at 3.6 hr after training. n=8. Results with error bars are means
± SEM. indicates non-significance (P > 0.05).
Figure S6: Related to Figure 6
(A) No significant difference of 3.6-hr ARM between 10× and 20× massed training. n=8. Results with error bars are
means ± SEM. indicates non-significance (P > 0.05).
(B) Representative Western blots (left) and statistical graph (right). In contrast to naive flies, Cdc42 activity was not
significantly changed at 2 hr after 1×training. n=5. Results with error bars are means ± SEM. indicates non-significance
(P > 0.05).
Table S1. Task-Relavant odor avoidance. Related to Firgure 1
Flies used in Figure 1 were tested for odor avoidance in contrast to air at 25°C. At 3 hr or 24 hr after shock, there were no
significant difference between 1×shock group and 4×shock group. Data are shown as means ± SEM. Unpaired t-test was
used for statistics, n≥6.
Flies
Odor avoidance (odor VS. air) (25℃)
OCT
（1.5×10-3）
MCH
（1.0×10-3）
w1118 control flies (1×shock, 3 hr) 30.35 ± 2.4 36.45 ± 5.9
w1118 control flies (4×shock, 3 hr) 33.39 ± 2.9 44.93 ± 6.1
w1118 control flies (1×shock, 24 hr) 34.36 ± 6.8 52.04 ± 4.6
w1118 control flies (4×shock, 24 hr) 35.64 ± 4.2 55.43 ± 2.4