Pet-1 is required across different stages of life to regulate serotonergic function

Abstract

Transcriptional cascades are required for the specification of serotonin (5-HT) neurons and behaviors modulated by 5-HT. Several cascade factors are expressed throughout the lifespan, which suggests that their control of behavior might not be temporally restricted to programming normal numbers of 5-HT neurons. We used new mouse conditional targeting approaches to investigate the ongoing requirements for Pet-1 (also called Fev), a cascade factor that is required for the initiation of 5-HT synthesis, but whose expression persists into adulthood. We found that Pet-1 was required after the generation of 5-HT neurons for multiple steps in 5-HT neuron maturation, including axonal innervation of the somatosensory cortex, expression of appropriate firing properties, and the expression of the Htr1a and Htr1b autoreceptors. Pet-1 was still required in adult 5-HT neurons to preserve normal anxiety-related behaviors through direct autoregulated control of serotonergic gene expression. These findings indicate that Pet-1 is required across the lifespan of the mouse and that behavioral pathogenesis can result from both developmental and adult-onset alterations in serotonergic transcription.

Full text

1190 VOLUME 13 | NUMBER 10 | OCTOBER 2010 nature neurOSCIenCe
a r t I C l e S
The 5-HT transmitter system in the brain is an essential homeo-
static modulator of neural circuits that shape emotional behaviors in
response to stressors in the environment1. A widely discussed theory
supported by a rich literature emphasizes the importance of 5-HT
function for the maturation of neural circuits and the development of
normal adult emotional behaviors2. Altered serotonergic signaling and
gene expression during embryonic development disrupts cortical den-
dritic arborization3 and the patterning of forebrain afferents4,5. Other
studies show that early postnatal perturbation of the serotonergic
system can cause emotional disorders in adult animals6–8. These find-
ings, together with correlative studies of serotonergic indices and gene
variants in monkeys and humans, support the idea that alterations in
serotonergic function are involved in establishing vulnerability for
several mood and neurological disorders1,9.
The likelihood that altered serotonergic function during develop-
ment contributes to behavioral pathogenesis has stimulated interest
in the genetic mechanisms that direct the formation of the 5-HT
system10. A cascade (Supplementary Fig. 1) of transcriptional regu-
lators has been identified that progressively restricts multi-potent
neuronal progenitors to a 5-HT neuron fate in the embryonic ventral
hindbrain11. Genetic targeting of factors in the cascade causes alter-
ations in 5-HT-modulated emotional responses in adults12,13, and
this finding provides a link between the transcriptional regulation of
5-HT neuron birth and adult behavior. Nevertheless, the mechanisms
through which transcription factors in the cascade regulate behavior
are poorly understood and might not be simply the result of program-
ming normal 5-HT neuron numbers and 5-HT levels. For example,
although all of the factors known to compose the cascade are nece-
ssary for the initiation of 5-HT synthesis at the cell fate specification
stage, transcriptional control of subsequent steps in the maturation
of the 5-HT system might also be crucial for programming normal
5-HT-modulated behaviors. However, whether or not factors in the
cascade are responsible for additional transcriptional events in the
maturation of the system has not been investigated. Furthermore, it
is not known whether the critical period for transcription directed by
these developmental determinants extends into adulthood to regulate
the maintenance of 5-HT signaling and preserve behavioral integrity.
The concept of a transcriptional maintenance mechanism could be
crucial for understanding the regulation of behavioral and psychiatric
pathogenesis as drug, toxin and dietary perturbation studies in adults
including humans demonstrate the importance of ongoing presynaptic
serotonergic function in emotional and behavioral processing9.
Expression of the rodent Pet-1 ETS domain transcription factor
(human orthologue, FEV) is restricted in the CNS to 5-HT neurons
and is induced in postmitotic precursors just before the initiation of
5-HT synthesis in the ventral hindbrain14. Pet-1 has a crucial role in
the cascade through its coordinate induction of the enzymatic path-
way responsible for 5-HT synthesis in immature postmitotic precur-
sors12. Interestingly, Pet-1 expression seems to continue undiminished
in all adult 5-HT neurons14. This persistent expression suggests that
Pet-1 might be required for events in 5-HT neuron maturation that
occur subsequent to their specification and possibly in adulthood
for transcriptional maintenance of the 5-HT system. Here, we used
new 5-HT neuron-specific and temporally restricted conditional
targeting approaches to investigate requirements for continued
Pet-1–dependent transcription in the 5-HT system.
RESULTS
Conditional deletion of Pet-1 after 5-HT neuron generation
To investigate the function of Pet-1 after its initial role in 5-HT neuron
generation, we inserted two loxP sites into introns on each side of
exon 3, which encodes most of the Pet-1 protein coding sequences
1Case Western Reserve University, School of Medicine, Department of Neurosciences, Cleveland, Ohio, USA. 2Present address: Department of General Zoology and
Neurobiology, Ruhr University, Bochum, Germany. Correspondence should be addressed to E.S.D. (esd@case.edu).
Received 10 March; accepted 23 July; published online 5 September 2010; doi:10.1038/nn.2623
Pet-1 is required across different stages of life to
regulate serotonergic function
Chen Liu1, Takashi Maejima1,2, Steven C Wyler1, Gemma Casadesus1, Stefan Herlitze1,2 & Evan S Deneris1
Transcriptional cascades are required for the specification of serotonin (5-HT) neurons and behaviors modulated by 5-HT. Several
cascade factors are expressed throughout the lifespan, which suggests that their control of behavior might not be temporally restricted
to programming normal numbers of 5-HT neurons. We used new mouse conditional targeting approaches to investigate the ongoing
requirements for Pet-1 (also called Fev), a cascade factor that is required for the initiation of 5-HT synthesis, but whose expression
persists into adulthood. We found that Pet-1 was required after the generation of 5-HT neurons for multiple steps in 5-HT neuron
maturation, including axonal innervation of the somatosensory cortex, expression of appropriate firing properties, and the expression
of the Htr1a and Htr1b autoreceptors. Pet-1 was still required in adult 5-HT neurons to preserve normal anxiety-related behaviors through
direct autoregulated control of serotonergic gene expression. These findings indicate that Pet-1 is required across the lifespan of the
mouse and that behavioral pathogenesis can result from both developmental and adult-onset alterations in serotonergic transcription.
© 2010 Nature America, Inc. All rights reserved.

nature neurOSCIenCe VOLUME 13 | NUMBER 10 | OCTOBER 2010 1191
a r t I C l e S
including the ETS DNA-binding domain (Fig. 1). In situ hybridi-
zation (Fig. 1b,c) and quantitative reverse transcriptase PCR
(RT-qPCR; Supplementary Fig. 2) indicated that Pet-1 expression was
indistinguishable in mice carrying either one copy of the wild-type
or loxP-flanked Pet-1 allele together with a constitutive null allele.
Excision of exon 3 generated a deleted Pet-1 allele (Δ) in which all
protein coding sequences except those encoding 42 amino acids at
the N terminus were eliminated (Fig. 1a). In situ hybridization
showed that Pet-1 was not expressed in mice heterozygous for the
deleted and null Pet-1 alleles (Fig. 1d). Furthermore, Pet-1loxP/−
mice had normal numbers of tryptophan hydroxylase (Tph)-positive
neurons but Pet-1Δ/− mice did not (Supplementary Fig. 2). Thus, the
unrecombined loxP-flanked allele is functionally equivalent to the
wild-type allele and the conditionally deleted Pet-1 allele is functionally
equivalent to the constitutive null Pet-1 allele.
We crossed loxP-flanked Pet-1 mice with ePet::Cre transgenic
mice15, which express Cre recombinase only in postmitotic 5-HT
neurons, to generate Pet-1 early conditional knockout mice (Pet-1loxP/−
ePet::Cre, designated Pet-1eCKO). Nearly all 5-HT neurons derived from
progenitors in rhombomeres 1 and 2 are born by embryonic days (E)10
and 11, respectively16,17. However, reduced serotonergic gene expression
is not seen in mice carrying this ePet::Cre transgene until about E12.5
(ref. 18). Thus, Pet-1 expression should be maintained in Pet-1eCKO
mice for about 2 d after fulfilling its early role in 5-HT neuron genera-
tion (Fig. 1e). On the basis of this reasoning, we predicted that normal
numbers of 5-HT neurons would be generated in Pet-1eCKO mice.
Consistent with our expectation, we found that expression of Pet-1
and other 5-HT neuron markers including tryptophan hydroxy-
lase 2 (Tph2), serotonin transporter (Slc6a4, also called Sert) and
another transcription factor, Gata3, was indistinguishable between
Pet-1eCKO mice and wild-type controls in the anterior hindbrain at
E11.5 (Fig. 1f,h and Supplementary Fig. 3). Furthermore, immuno-
histochemistry using antibodies to 5-HT showed that Pet-1eCKO mice
had the same number of 5-HT+ neurons as did controls (Fig. 1j,k). By
contrast, Pet-1−/− mice had very few 5-HT+ cells at this stage (Fig. 1l).
We did not observe a decrease in Pet-1 transcripts in Pet-1eCKO mice
until E12.5 (Fig. 1g). Concomitant with the conditional deletion of
Pet-1 at E12.5, the expression of Tph2, Slc6a4 and 5-HT in Pet-1eCKO
mice was diminished (Supplementary Fig. 3), whereas the expression
of Gata3 was not altered (Fig. 1i). Pet-1 is not required for the survival
of 5-HT neurons19, and all Pet-1-deficient cells were present in the
brains of Pet-1eCKO mice through adulthood (Supplementary Fig. 4).
These findings indicate that Pet-1eCKO mice can be used to investigate
Pet-1 function in 5-HT system maturation after it has fulfilled its
initial role in 5-HT neuron generation.
Continued Pet-1 function controls serotonergic innervation
Immediately after the birth of 5-HT neurons, maturation of the
5-HT system depends on proper cell body migration, axon path-
finding and innervation in terminal fields20. To investigate the role
of Pet-1 in these maturation events, we used Cre-mediated activa-
tion of the R26RYfp (ref. 21) reporter allele to permanently mark
Pet-1-deficient 5-HT neurons in Pet-1eCKO mice (Fig. 2). Pet-1
deletion was spared in a small subset (~15%) of 5-HT neurons in
Pet-1eCKO mice (Supplementary Fig. 4) and therefore Pet-1-deficient
Tph− cells were situated side by side with untargeted Tph+ 5-HT
neurons (Fig. 2a). Examination of R26R-Yfp+ cells in Pet-1eCKO mice
showed that these Pet-1-deficient 5-HT neurons extended axons
from their cell bodies, similar to intermingled wild-type Tph+ cells,
and therefore Pet-1 was not essential for proximal axonal outgrowth
(Fig. 2a). Immunostaining for yellow fluorescent protein (YFP) in
Pet-1eCKO mice revealed that axon bundles from Pet-1-deficient
5-HT neurons crossed the midbrain-hindbrain boundary and
entered the midbrain at E14.5 (Fig. 2b).
To determine whether these axons could properly reach their
forebrain targets, we performed a retrograde tracing experiment
by injecting tracer into the somatosensory cortex, which receives
serotonergic afferents mainly from 5-HT neurons in the dorsal raphe
a
c d
e
f g
h
j k l
i
Pet-1 transcript
loxP allele
Deleted allele
1 2
1 2
P3
P4 P2 P6
P6
P3
loxP
loxP
b+/–
Pet-1 Pet-1 Pet-1
loxP/– ∆/–
loxP
P1
1 2 3
3
Wild-type allele
5′UTR 3′UTR
Pet-1–/–
Pet-1
Gata3 Gata3 Gata3 Gata3
Pet-1 Pet-1 Pet-1
Specification Maturation
E11.5 E12.5
E11.5
E11.5
Control
Control
5-HT 5-HT 5-HT
Control
E12.5
P21 P50
Adult maintenance
Pet-1eCKO
Pet-1WT
Pet-1eCKO
Pet-1eCKO Pet-1 null
Pet-1eCKO
ETS
Figure 1 Conditional deletion of Pet-1 after specification of 5-HT neuronal fate.
(a) Targeting strategy. From the top, schematic of the Pet-1 mRNA; wild-type Pet-1 allele (+);
loxP-flanked Pet-1 allele (loxP); and the conditionally deleted Pet-1 allele (Δ). (b–d) In situ
hybridization to detect Pet-1 transcripts in the DRN of mice heterozygous for the Pet-1 null
allele and the wild-type (b), loxP-flanked (c) or conditionally deleted (d) Pet-1 alleles. (e) Time frame of Pet-1 expression in Pet-1−/−, Pet-1eCKO (Pet-1loxP/−
ePet::Cre) and wild-type mice. (f–i) In situ hybridization to detect Pet-1 and Gata3 mRNAs in control (Pet-1loxP/+ ePet::Cre) and Pet-1eCKO mice at either
E11.5 (f,h) or E12.5 (g,i). (j–l) 5-HT immunostaining in control, Pet-1eCKO and Pet-1−/− mice at E11.5. Scale bars represent 100 μm (i,l) and 200 μm (d).
© 2010 Nature America, Inc. All rights reserved.

1192 VOLUME 13 | NUMBER 10 | OCTOBER 2010 nature neurOSCIenCe
a r t I C l e S
nucleus22 (DRN; Fig. 2c). In 3-week-old wild-type and Pet-1eCKO mice,
such injection resulted in the retrograde labeling of both 5-HT and
non-5-HT neurons in the DRN (Fig. 2d–f). The tracer intensity
in Pet-1-deficient 5-HT neurons was comparable to that in other
retrogradely labeled cells, suggesting that Pet-1-deficient cells
could retrogradely transport tracer (Fig. 2f). We found a significant
decrease in the total number of labeled cells across the entire DRN
in Pet-1eCKO mice (Fig. 2g,h). In wild-type mice, most retrogradely
labeled cells were 5-HT neurons (Fig. 2i,k). In contrast, far fewer
Pet-1-deficient YFP+ 5-HT neurons were labeled by similar tracer
injections (Fig. 2j,l), indicating that the serotonergic innervation
of the somatosensory cortex from DRN 5-HT neurons was signifi-
cantly disrupted in the brains of Pet-1eCKO mice.
Continued Pet-1 function controls autoreceptor pathways
Another key event in the maturation of 5-HT neuron function in the
brain is the acquisition of 5-HT neuron-specific firing properties23.
To study whether continued Pet-1 function is required for normal
firing of 5-HT neurons, we analyzed Pet-1-deficient R26R-Yfp+ cells
in postnatal brain slices from Pet-1eC KO mice by using whole-cell
recordings under current-clamp conditions. Compared with the aged-
matched controls, many Pet-1-deficient cells showed increased spon-
taneous firing of action potentials (Fig. 3a–c).
This increased excitability could result from
alterations in Htr1a autoreceptor signaling,
which normally inhibits the firing of 5-HT
neurons through negative feedback inhibi-
tion24. As previously described25, activation
of the Htr1a receptor with its specific agonist
Figure 2 Continued Pet-1 function is required
for maturation of serotonergic axonal innervation
patterns. (a) Co-immunostaining for YFP and
Tph in adult Pet-1eCKO mice. White arrows,
untargeted 5-HT neurons; arrowheads,
proximal axons extending from cell bodies
of Pet-1–deficient 5-HT neurons. (b) YFP
immunostaining of Pet-1-deficient 5-HT
neuron axon bundles at E14.5 in Pet-1eCKO
mice. Arrowheads mark axons that have crossed
the midbrain-hindbrain boundary (dashed lines)
and entered the midbrain. (c) Schematic of the
retrograde tracing experiment. (d) Tracer-labeled
cells in the DRN. (e) YFP immunolabeling of
Pet-1–deficient 5-HT neurons in the DRN.
(f) Merge of d and e. (g–l) Significantly fewer
retrogradely labeled cells were found in the
DRN of the Pet-1eCKO brain (g,h; 49.9 ± 3.1%,
mean ± s.e.m., relative to control, n = 6 for
each genotype). Overlay of tracer signal with YFP
immunostaining (i,j) showed that 83.0 ± 1.8%
(mean ± s.e.m.) of the retrogradely labeled DRN cells in control mice were Yfp+ 5-HT neurons (k), whereas only 19.3 ± 1.7% YFP+ Pet-1-deficient 5-HT
neurons (l) were labeled by the same tracer injection in Pet-1eCKO mice. P < 0.001, two tailed t test. Scale bars represent 20 μm (a,f) and 200 μm (l).
ab
cd
ef
gh
ij
kl
Pet-1eCKO Pet-1eCKO Pet-1eCKO
Tph/R26R-Yfp R26R-Yfp
Yfp Yfp
E14.5
Control
Tracer
Tracer
Tracer
Merge Merge
a
d
g h i j
e f
+/+
40
20
mV
pA
mV
pA
mV
–130 –110 –90 –50 –130 –110 –70 –50
–200
200
+/+
eCKO
150
Change of current (pA)
100
50
0
1 10 1 10
–400
8-OH-DPAT
8-OH-DPAT
8-OH-DPAT (µM)
Control
Control
Control
–600
–200
–400
–600
0
0 1 2
S
3 4 5
–20
–40
–60
b c
40
20
mV
0
012
S
3 4 5
–20
–40
–60
12
10
Firing frequency (Hz)
8
+/+
*
***
0
2
4
6
Pet-1eCKO
Pet-1eCKO
Pet-1eCKO
Htr1a Htr1a Htr1b Htr1b
Control Pet-1eCKO
*
Figure 3 Continued Pet-1 function is required
for 5-HT neuron firing properties and inhibitory
autoreceptor function. (a,b) Whole-cell current-
clamp recordings measuring spontaneous firing
of YFP+ neurons with indicated genotypes.
(c) Quantification of firing frequencies in a and
b (+/+, n = 12; Pet-1eCKO, n = 19; *P < 0.05,
two tailed t-test). (d,e) Whole-cell voltage-clamp
recordings measuring current changes in YFP+
neurons induced with 1 or 10 μM of the Htr1a
receptor agonist 8-OH-DPAT. A ramp voltage was
applied at 200 mV s−1. The intersection voltage,
−87 mV, of the control and 8-OH-DPAT traces
in d was close to the estimated K+ equilibrium
potential (−99 mV), considering that recordings
were not corrected for the liquid junction
potential of around 10 mV. (f) Quantification
of current changes at −110 mV in d and e
(*P < 0.05, ***P < 0.001; two-tailed t-test).
(g–j) In situ hybridization of Htr1a (g,h) and
Htr1b (i,j) in control and Pet-1eCKO mice. Scale
bar, 200 μm. Error bars show mean ± s.e.m.
© 2010 Nature America, Inc. All rights reserved.

nature neurOSCIenCe VOLUME 13 | NUMBER 10 | OCTOBER 2010 1193
a r t I C l e S
8-OH-DPAT elicited strong inwardly rectifying potassium currents in
control YFP+ 5-HT neurons under voltage clamp (Fig. 3d). By con-
trast, neither low (1 μM) nor high (10 μM) concentrations of 8-OH-
DPAT elicited a change in baseline currents in Pet-1-deficient YFP+
5-HT neurons (Fig. 3e,f). To investigate the mechanism that accounts
for the loss of 8-OH-DPAT responses, we used in situ hybridization
to examine the expression of the Htr1a receptor and found greatly
decreased levels of Htr1a mRNA in the Pet-1eCKO DRN (Fig. 3g,h).
A second prominent serotonergic autoreceptor that regulates 5-HT
release in serotonergic presynaptic terminals is the Htr1b receptor26.
In situ hybridization in Pet-1eCKO mice revealed that Pet-1 was also
required for expression of the Htr1b gene (Fig. 3i,j). The residual
expression of Htr1a and Htr1b mRNAs is probably from the remain-
ing untargeted 5-HT neurons in Pet-1eCKO mice as the expression of
the two autoreceptors in the DRN was almost completely abolished
in Pet-1−/− mice (Supplementary Fig. 5). Our in situ hybridization
studies (data not shown) indicated that the expression of Htr1a and
Htr1b in nearly all 5-HT neurons begins after E14, which is consistent
with their onset in the forebrain at E14.5 (ref. 27). Thus, our find-
ings show that ongoing Pet-1 expression is required, well after it has
completed its role in the initiation of 5-HT synthesis in immature
precursors, for maturation of essential serotonergic autoreceptor
characteristics that control firing patterns and transmitter release.
Gata3 is not required for Htr1a autoreceptor responses
We investigated whether the establishment of normal 5-HT neuron
firing properties requires parallel ongoing activity of other seroto-
nergic developmental control genes, or whether Pet-1 has a unique
role in this process. Germ line targeting of the zinc finger transcrip-
tion factors Gata2 and Gata3 has shown that both are required for 5-
HT neuron differentiation28. We found that Gata2 protein expression
began to decline in differentiated 5-HT neurons at E12.5 and was not
detectable at E14.5. By contrast, Gata3 expression persisted in all 5-
HT neurons through adulthood (Supplementary Fig. 6).
To compare the role of Gata3 in differentiated 5-HT neurons with
that of Pet-1, we crossed Gata3loxP/loxP mice29 with ePet::Cre mice to
generate Gata3 conditional knockout mice (Gata3loxP/loxP ePet::Cre,
designated as Gata3eCKO). Similar to Pet-1, Gata3 was not required
for cell survival as normal numbers of Gata3-deficient 5-HT neurons
survived in the adult brain (Fig. 4). However, there was a substantial
reduction in the number of Tph immunoreactive cells (Fig. 4b,d),
levels of 5-HT (Fig. 4f,g) and expression of several other 5-HT genes
(Fig. 4e) in the DRN of Gata3eCKO mice. Persistent expression of
Gata3 and Pet-1 seems to be maintained by independent regulatory
pathways, as neither of them was required for the other’s expression
(Figs. 1i and 4c). These findings suggest that Gata3 and Pet-1 function
in parallel pathways to coordinate the expression of normal levels of
serotonergic gene expression and 5-HT in the brain.
Although Gata3 and Pet-1 share several common transcrip-
tional targets, Gata3-deficient 5-HT neurons showed normal Htr1a
expression (Fig. 4e), indicating that Pet-1 and Gata3 regulate distinct
sets genes in 5-HT neurons. These findings also suggest that Gata3
might not be required for serotonergic firing characteristics. Indeed,
whole-cell recordings of slices from Gata3eCKO mice showed firing
properties and Htr1a agonist responses typical of wild-type 5-HT
neurons (Fig. 4h,i).
Targeting of Pet-1 in the adult ascending 5-HT system
Having shown that ongoing Pet-1 function is needed for multiple
steps in the maturation of the 5-HT system, we sought to determine
whether a Pet-1–dependent transcriptional program still operates in
adulthood to support serotonergic function and 5-HT-modulated
d
e f g
h i
40
10
Protein (ng mg–1)
8
6
4
2
0
20
mV
0
012
s
345
–20
–40
–60
Aadc
Sert
Tph2
Vmat2
Htr1a
pA
mV
120
100
80
60
40
20
0
Relative expression
9,000
6,000
Number of Tph+ cells in individual B nuclei
3,000
0B9 B8 B7 B6 B5 B4 B3 B2/B1
Control
Gata3eCKO
*** **
–130
–110
–90
–50
8-OH-DPAT
Control –200
–400
–600
*
** **
**
Forebrain
Spinal cord
5-HIAA
5-HT
a
b
c
Pet-1 Pet-1
Control Gata3eCKO
Tph
R26R-Yfp R26R-Yfp
Tph
25
20
15
10
5
0
*** **
Forebrain
Spinal cord
Control Gata3eCKO
Figure 4 Continued Gata3 expression is needed to maintain 5-HT gene
expression but not autoreceptor function. (a) YFP immunostaining in adult
DRN. (b) Tph immunostaining. (c) In situ hybridization of Pet-1 mRNA.
(d) Counts of Tph+ cell bodies in control versus Gata3eCKO mice in
individual adult B nuclei (n = 3 for each genotype). (e) RT-qPCR of Ddc,
Slc6a4, Tph2, Slc18a2 and Htr1a mRNAs in control versus Gata3eCKO
mice (control n = 7, normalized to 100%; Gata3eCKO n = 11; *P < 0.05,
**P < 0.01, two-tailed t-test). (f,g) HPLC analysis of 5-HT (f) and 5-HIAA
(g) levels in forebrain and spinal cord of control (n = 7) and Gata3eCKO
(n = 5) mice (**P < 0.01, ***P < 0.001, two-tailed t-test). (h) Whole-cell
current-clamp recordings of spontaneous firing in R26R-YFP+ Gata3-
deficient cells. (i) Whole-cell voltage-clamp recordings of current changes
in response to 8-OH-DPAT in R26R-YFP+ Gata3-deficient cells. Scale
bars, 200 μm. Error bars show s.e.m. except for s.d. in d.
© 2010 Nature America, Inc. All rights reserved.

1194 VOLUME 13 | NUMBER 10 | OCTOBER 2010 nature neurOSCIenCe
a r t I C l e S
behaviors (Fig. 5a). Thus, we used Pet-1 regu-
latory elements to generate a transgene that
directed tamoxifen-inducible CreERT2 (Cre
recombinase–estrogen receptor ligand bind-
ing domain, ref. 30) expression specifically in
brain 5-HT neurons. To identify founder lines
expressing inducible Cre activity, we gave
pregnant females a single dose of tamoxifen
(150 μg per g, intraperitoneal) at gestational
stage E11.5 and scored embryos for Cre-
activated β-galactosidase expression from the
R26R
β
gal allele at E16.5 (Supplementar y
Fig. 7). Eight-week-old Cre reporter mice
(R26R
β
gal+ePet::CreERT2) were then given a
single daily dose of tamoxifen or vehicle for
five consecutive days. Most 5-HT neurons in
the DRN and median raphe nucleus (MRN)
contained β-galactosidase 5 or 30 days after
the last tamoxifen injection. Recombination
was strictly dependent on tamoxifen treat-
ment as we found no Cre activity in the
absence of tamoxifen. We did not detect Cre
activity in other regions of the CNS after
tamoxifen injections (Supplementary Fig. 7).
In one of the transgenic lines, designated
ePet::CreERT2ascend, we found differential
targeting efficacies after treatment of adult
mice with tamoxifen between 5-HT neuron
raphe nuclei that give rise to ascending
and descending 5-HT systems (Fig. 5b,c).
Double-labeling to detect β-galactosidase
and Tph revealed tamoxifen-activated Cre
activity in up to 80% of 5-HT neurons in the DRN (B6, B7), MRN
(B5, B8) and B9 nucleus. By contrast, far fewer 5-HT neurons
contained β-galactosidase in the medullary nuclei (B1–B3) after
tamoxifen treatment (Fig. 5d,e).
To determine the efficacy of the ePet::CreERT2ascend line for excision
of Pet-1 in Pet-1loxP/− mice, we crossed ePet::CreERT2ascend, Pet-1−/−
and Pet-1loxP/loxP mice to generate Pet-1loxP/− ePet::CreERT2ascend mice,
designated as Pet-1aCKO. Six-to-eight-week-old Pet-1aCKO mice were
given single daily tamoxifen treatments for 5 d and killed 5 or 30 d
after tamoxifen treatments for evaluation of Pet-1 expression. In situ
hybridization showed that tamoxifen treatment abolished the majority
of Pet-1 expression in the adult DRN (B6 and B7 nuclei), MRN
(B5 and B8 nuclei) and the B9 group of 5-HT neurons (Fig. 5f,g). By
contrast, Pet-1 mRNA was not decreased in the B1–B3 groups of 5-HT
neurons in the ventral medulla (Fig. 5h,i). The reduction of Pet-1
in Pet-1aCKO mice treated with tamoxifen was further quantified by
RT-qPCR, which revealed a >70% loss of Pet-1 mRNA in pontine tissue
containing the B5–B9 serotonergic nuclei, but no significant change
in tissue containing the medullary B1–B3 nuclei (Fig. 5j). The loss
of Pet-1 mRNA in the B5–B9 nuclei showed that ePet::CreERT2ascend
could be used for highly reproducible and stage-specific disruption
of Pet-1 expression in adult ascending 5-HT neurons.
Adult Pet-1 is required for normal anxiety-like behaviors
Germ line targeting of Pet-1 results in increased anxiety-like behaviors in
the adult. However, it remains unclear whether Pet-1–dependent transcrip-
tion is needed only during development or also in adulthood to modulate
normal anxiety responses. To address this question, we treated 6–8-week-
old Pet-1aCKO mice with tamoxifen to delete Pet-1 in the ascending 5-HT
system. We then investigated the effect of adult Pet-1 deletion on anxiety-
related behaviors 4 weeks after the last tamoxifen treatment.
It was recently shown, using an administration protocol similar to
that used here, that tamoxifen does not alter anxiety-related behaviors
in mice31. We verified this finding on a separate cohort of wild-type
mice treated with either vehicle or tamoxifen (Supplementary Fig. 8).
We then tested tamoxifen-treated control and Pet-1aCKO mice on the
elevated plus maze test and found that tamoxifen-treated Pet-1aCKO
mice spent significantly less time in and initiated fewer entries into
the open unprotected arms of the maze than did tamoxifen-treated
control mice (Fig. 6a,b). In addition, these mice spent less time in the
hub area but significantly more time in the closed arm (Fig. 6c,d). We
found no differences between genotypes in overall explorative acti-
vities determined as the number of total open/closed arm entrances
(Fig. 6e). Their increased avoidance of the aversive properties of
height and openness suggests that tamoxifen-treated Pet-1aCKO mice
show augmented anxiety-like behavior. To further study this behavior,
we tested the same mice in the light↔dark exploration paradigm,
which presents the mice with a similar conflict between the desire to
explore a novel environment and the aversive features of a brightly
illuminated open field. As compared to tamoxifen-treated littermate
controls, tamoxifen-treated Pet-1aCKO mice spent significantly more
time in the dark chamber (Fig. 6f,g) with a trend towards reduced
latency to enter the dark area from the beginning of the test (Fig. 6h).
The increased time in the dark area and avoidance of the bright open
areas supported the idea that tamoxifen-treated Pet-1aCKO mice
show increased anxiety-like behavior. Finally, tamoxifen-treated
Pet-1aC KO mice spent significantly less time than controls in the
center of an open field (Fig. 6i), whereas overall locomotor activity
Figure 5 Stage-specific disruption of Pet-1 in the adult ascending 5-HT system. (a) Adult stage-
specific deletion of Pet-1 in Pet-1aCKO mice. (b,c) Co-immunostaining for β-galactosidase and Tph
in adult DRN (b) and medullary raphe (c). (d,e) Percentage of TPH+ cells expressing CreER-activated
β-galactosidase in individual adult B nuclei (d) and in 5-HT neurons of ascending versus descending
pathways (e; 68.5 ± 9.5% in the pons (B4–B9) and 12.1 ± 6.8% in medullary nuclei (B1–B3); n = 7,
mean ± s.d., ***P < 0.001, two-tailed t-test). (f–i) In situ hybridization of Pet-1 mRNA in coronal
sections from adult Pet-1aCKO mice treated with tamoxifen (TM) or vehicle (Veh). (j) RT-qPCR of Pet-1
mRNA in tamoxifen-treated adult Pet-1aCKO mice (pons, n = 30, 28.0 ± 2.5% relative to control,
n = 35; medulla, n = 12, 105.9 ± 5.8% relative to control, n = 15). Each dot represents a sample
from the indicated group; data shown are mean ± s.e.m., two-tailed t-test. Scale bar, 200 μm.
100
R26R-βgal Tph Merge
MergeTph
R26R-βgal
100
50
0
150
Specication
E11.5
Pet-1aCKO
Pet-1WT
Pet-1 Pet-1
Pet-1 Pet-1
ePet::CreERT2ascend; Pet-1loxP/loxP
E12.5
a
f
h
j
i
g
b
c
d e
P21 +TM
+TM
+Veh
+Veh
P50
Maturation Adult maintenance
80
βgal+/Tph+%
βgal+/Tph+%
Percentage of Pet-1
B9 B8 B7 B3 B2/1 Pons
***
Medulla Pons
n = 30, P < 0.0001
Medulla
n = 12, P = 0.516
B6/4B5
60
40
20
0
100
80
60
40
20
0
ePet::CreERT2ascend; Pet-1loxP/–
© 2010 Nature America, Inc. All rights reserved.

nature neurOSCIenCe VOLUME 13 | NUMBER 10 | OCTOBER 2010 1195
a r t I C l e S
was not different (Fig. 6j). A second, independent cohort mice
showed similar significant increases in all three tests of anxiety-like
behaviors in tamoxifen-treated Pet-1aCKO mice (Supplementary
Fig. 9). Furthermore, the increased anxiety-like behavior seen in
all three tests depended on reduced Pet-1 levels, as we found no
differences in a separate cohort of Pet-1aCKO mice treated with vehicle
(Supplementary Fig. 10). Overall growth measured by body weight
was similar in control and Pet-1aCKO mice after tamoxifen treatments
(data not shown).
Adult Pet-1 is required for serotonergic gene expression
The altered anxiety-like behaviors in tamoxifen-treated Pet-1aCKO
mice show that Pet-1 is required in adulthood to maintain serotonergic
function. To explore the mechanisms that underlie the alterations in
serotonergic function, we first measured levels of brain 5-HT and its
metabolite 5-HIAA in Pet-1aCKO and control mice killed 5 days after
tamoxifen treatments. High-performance liquid chromatography
(HPLC) analysis showed that both 5-HT and 5-HIAA were significantly
decreased in the forebrain of tamoxifen-treated Pet-1aCKO mice
(Fig. 7a,b), but, as predicted, 5-HT levels were not altered in the spinal
cord (data not shown), which is innervated by the descending 5-HT
system. Western blotting using a monoclonal antibody to both Tph1
and Tph2 indicated that levels of Tph were reduced by about 50%
in tamoxifen-treated Pet-1aCKO mice relative to controls (Fig. 7c,d).
Consistent with these findings, we found a comparable decrease in Tph2
mRNA levels in tamoxifen-treated Pet-1aCKO mice as early as 5 d after
the last tamoxifen treatment (Fig. 7e,f). Furthermore, Tph2 mRNA was
still decreased 30 d after the last treatment and we did not find a com-
pensatory increase in Tph1 expression with the loss of Tph2 (Fig. 7e).
Together, these findings indicate that Pet-1 is required in adult 5-HT
neurons to regulate 5-HT synthesis by maintaining Tph2 expression.
To determine whether disruption of the Htr1a autoreceptor path-
way might have contributed to the abnormal anxiety-like behavior,
we performed whole-cell recordings in slices from tamoxifen-treated
Pet-1aCKO mice but found that Pet-1 was no longer required in the
adult brain for spontaneous firing, autoreceptor agonist responses
a c d eb f g h i j
40
Open time %
30
10
Control
aCKO
Control
aCKO
Control
aCKO
Control
aCKO
Control
aCKO
Control
aCKO
Control
aCKO
Control
aCKO
Control
aCKO
Control
aCKO
20
0
**
80
Close time %
60
20
40
0
*** 40
Hub time %
30
10
20
0
*
20
Total
entrances
15
5
10
0
200
Duration
in dark
150
50
100
Time (s)
0
*
1.5
Dark/light
ratio
0.5
1.0
Ratio
0
**
50
Latency
20
10
40
30
Time (s)
0
60
Open
entrances %
40
20
0
***
40
Time in
center
Open field
Dark ↔ light exploration
Elevated plus maze
10
20
30
Time (s)
0
*
6,000
4,000
2,000
Distance
moved
Cm
0
Figure 6 Disruption of Pet-1-dependent transcription in the adult ascending 5-HT system causes elevated anxiety-like behavior. Six-to-eight-week-old
male Pet-1aCKO mice (n = 12) and their littermate controls (Pet-1loxP/−, n = 12) were treated with tamoxifen for 5 consecutive days and then acclimated
for another 4 weeks before behavioral testing. (a–e) Elevated plus maze. *P < 0.05, **P < 0.01, ***P < 0.001; two-tailed t-test. (f–h) Dark ↔ light
exploration. *P < 0.05, **P < 0.01; two-tailed t-test. (i,j) Open field. *P < 0.05; two-tailed t-test. Error bars represent s.e.m.
Figure 7 5-HT synthesis and Slc6a4 expression
are maintained in the adult ascending 5-HT
system through positively autoregulated direct
Pet-1 transactivation. (a,b) HPLC analysis of 5-HT
and 5-HIAA levels in the forebrain of tamoxifen-
treated control (n = 7) and Pet-1aCKO (n = 7)
mice (***P < 0.001, two-tailed t-test).
(c,d) Western blotting analysis of Tph protein
in DRN of tamoxifen-treated control (n = 8) and
Pet-1aCKO mice (n = 7; 50.3 ± 4.0% relative to
the control; **P < 0.01, two-tailed t-test).
(e) RT-qPCR analysis of serotonergic gene
expression in Pet-1aCKO or control mice either 5 days
(control, n = 11; Pet-1aCKO, n = 11) or 30 days
(control n = 14; Pet-1aCKO, n = 16) after treatment
with tamoxifen or vehicle (*P < 0.05, **P < 0.01,
***P < 0.001). (f–i) In situ hybridization to detect
Tph2, Slc6a4, Maob and Lmx1b mRNAs in
coronal sections from Pet-1aCKO mice treated with
tamoxifen or vehicle. (j) Whole-cell current-clamp
recordings of spontaneous firing in YFP+ Pet-1-
deficient cells in tamoxifen-treated Pet-1aCKO mice.
(k) Whole-cell voltage-clamp recordings of current
changes in response to 8-OH-DPAT in YFP+ Pet-1-
deficient cells. (l) RT-qPCR analysis of chromatin
immunoprecipitations. Values represent fold
enrichment in binding to the indicated regions
as compared to negative control region (Untr17).
Untr, untranscribed genomic region; #, P < 0.0001
for Tph2, Slc6a4, Pet-1 versus Untr17 or Slc6a4 intron, one-way ANOVA with Bonferroni’s Multiple Comparison Test. (m,n) In situ hybridization to detect
CreERT2 mRNA in adult Pet-1aCKO mice treated with tamoxifen or vehicle. Scale bars, 200 μm. Error bars represent s.e.m except for s.d. in (l).
e f g h i
*
*
**
*
**
*
*
**
**
*
*
150
Tamoxifen
Tph2
Sert Sert
Tph2 Maob
Lmx1b
Maob
Lmx1b
Vehicle Tamoxifen Vehicle
Pet-1
Tph2
Sert
Vglut3
Gch1
Aadc
Vmat2
Maob
Htr1a
Tph1
Cpne7
Gata3
Lmx1b
100
Relative expression % Protein (ng mg–1)
50
0
j k l m n
TamoxifenTamoxifen
CreER CreER
Vehicle
40
20
0
mV
–20
–40
–130
–110
–90
–60
–50
–100
–200
–300
–400
0
Untr17
Sert
Tph2
Pet-1
Sert-intron
6##
#
12
0123
s
4 5
Control
mV
Fold enrichment
pA
8-OH-DPAT
5 d after TM
30 d after TM
a b c d
Forebrain 5-HT Forebrain 5-HIAA Control Tph protein
Relative fold change
1.5
**
0.5
0
1.0
Pet-1aCKO
Tph
β-action
***
***
aCKOControl aCKOControl aCKOControl
14 5.0
4.5
3.5
0
4.0
12
10
8
0
© 2010 Nature America, Inc. All rights reserved.

1196 VOLUME 13 | NUMBER 10 | OCTOBER 2010 nature neurOSCIenCe
a r t I C l e S
(Fig. 7j,k), or Htr1a gene expression (Fig. 7e). To explore other
potential deficits beyond the loss of 5-HT, we analyzed the set of
additional genes that are known to depend on Pet-1 in the embryonic
hindbrain as well as other genes that are important for 5-HT synthesis
and metabolism (Fig. 7e,f–i). The expression of Slc6a4 and vesicular
glutamate transporter 3 (Slc17a8, also called Vglut3) were also sig-
nificantly reduced (Fig. 7e,h), but for Slc17a8, decreased expression
was not observed until 30 d after tamoxifen treatments (Fig. 7e).
In contrast to the marked decreases in Tph2, Slc6a4 and Slc17a8
expression, Pet-1 was no longer required in adulthood to maintain
the expression of Ddc (also called Aadc and Maob) and Slc18a2 (also
called Vmat2) mRNA (Fig. 7e), even though their embryonic expres-
sion depends on Pet-1 function (Supplementar y Figure 3).
Direct autoregulation of serotonergic gene expression
The studies presented so far do not distinguish direct from indirect
transcriptional control of target genes by Pet-1. We previously iden-
tified a consensus Pet-1-binding sequence, GGAAR(T), upstream
of Slc6a4 and showed that Pet-1 protein interacted with this site
in vitro14. Further analyses identified conserved putative Pet-1 ETS
binding sites in highly conserved upstream regulatory regions of Tph2
and Slc6a4 genes (Supplementar y Fig. 11). To probe the mechanism
through which Pet-1 regulates Tph2 and Slc6a4 in 5-HT neurons, we
investigated the possibility that Pet-1 directly regulates their transcrip-
tion by interacting with conserved upstream regulatory elements.
Because we have been unable to prepare a suitable Pet-1 antibody
for chromatin immunoprecipitation (ChIP), we generated a new
transgenic mouse line that expressed a myc-epitope-tagged Pet-1
protein in the brains of Pet-1−/− mice with Pet-1 promoter/enhancer
sequences15. Expression of the ePet::mycPet-1 transgene recapitulated
endogenous Pet-1 expression in both developing and adult hindbrain,
resulting in the rescue of normal numbers of 5-HT neurons in
Pet-1−/− mice (Supplementary Fig. 12). We used this rescue
line for ChIP to determine whether Pet-1 directly interacted with
Tph2 and Slc6a4 promoter sequences in vivo. Chromatin was
harvested from E12.5 mouse hindbrain before extensive 5-HT neuron
dispersion began to scatter these cells, although the dissected tissue
was still largely composed of non-serotonergic cells. Sheared chro-
matin was immunoprecipitated with an anti-myc antibody and ana-
lyzed by RT-qPCR for anti-myc enrichment of genomic fragments
that included predicted Pet-1–binding sites upstream of Tph2 and
Slc6a4 as well as in the intron of Slc6a4 (+11390). Compared to the
control, we found no enrichment near the Slc6a4 intron sequence. By
contrast, upstream Tph2 and Slc6a4 sequences showed several-fold
enrichment compared to both the negative control region untr17 and
the Slc6a4 intron sequence in two independent immunoprecipitation
assays (Fig. 7l and data not shown).
Finally, we used the ePet::CreERT2ascend transgene as a reporter for
Pet-1–dependent regulation of its own enhancer in Pet-1aCKO mice.
Expression of CreERT2 in the adult DRN was significantly reduced
in Pet-1aCKO mice treated with tamoxifen, but not in Pet-1aCKO mice
treated with vehicle, showing that adult expression of Pet-1 depends on
positive autoregulation (Fig. 7m,n). Inspection of the upstream Pet-1
promoter/enhancer sequences32 revealed conserved Pet-1 consensus
binding sites at −465 and −621 relative to the predicted transcription
start site (Supplementary Fig. 11). ChIP for genomic fragments with
these binding sites revealed a tenfold enrichment relative to control
immunoprecipitations (Fig. 7l and data not shown). These find-
ings suggest that transcriptional regulation of 5-HT synthesis and
serotonergic gene expression in adulthood depends on direct positive
autoregulatory maintenance of Pet-1 expression.
DISCUSSION
In this study, we tested the idea that Pet-1, a key component of an
embryonic transcriptional cascade that generates 5-HT neurons in
the ventral hindbrain, continues to regulate subsequent milestones
in 5-HT system maturation and 5-HT function in adulthood. We
have shown that Pet-1 function is not restricted to the induction of
serotonergic characteristics in embryonic 5-HT neuron precursors.
Instead, ongoing Pet-1–directed transcription is required across the
lifespan for multiple regulatory events that shape and maintain the
serotonergic neurotransmitter system. Our findings also support the
idea that the etiology of behavioral pathogenesis is not limited to dys-
function of the serotonergic system during development but may also
result from adult-onset alterations in serotonergic transcription.
These and earlier findings12 define three general but distinct stages
of Pet-1 function. The initial stage occurs during serotonergic neuro-
genesis, when Pet-1 regulates a late phase of 5-HT neuron genera-
tion by coordinating the induction of key serotonergic genes that are
required for transmitter synthesis, reuptake and vesicular transport
in immature postmitotic precursors12. Here, we uncovered a second
stage of Pet-1 function using a conditional targeting approach that
did not interfere with Pet-1 expression until about 2 d after the
completion of serotonergic neurogenesis. This transcriptional stage
coincides with the prolonged period of 5-HT neuron maturation,
during which these cells must negotiate complex axonal growth and
pathfinding decisions and acquire their characteristic firing pro-
perties. We identified multiple requirements for Pet-1 at this second
stage, which showed that Pet-1 is essential for proper 5-HT system
maturation. For example, retrograde tracing of R26R-YFP-marked
Pet-1–deficient 5-HT neurons revealed a substantial deficit in the
number of serotonergic projections to the somatosensory cortex.
However, initial serotonergic axon-like outgrowth did not appear to be
compromised in Pet-1eCKO mice, which suggests that Pet-1–dependent
transcription regulates subsequent pathfinding decisions that help
to build the ascending serotonergic system. The innervation defects
in Pet-1eCKO mice were probably not contributed to by the reduction
in brain 5-HT. Although pharmacological disruption of embryonic
5-HT signaling alters neuronal organization in the presubicular
cortex33 and 5-HT regulates thalamocortical axon pathfinding by
modulating axonal responsiveness to guidance cues5, recent studies
of Tph2-targeted mice, which lack 5-HT synthesis in the brain, did
not find defects in serotonergic innervation patterns34.
We also identified a special role for continued Pet-1–directed trans-
cription, not shared by Gata3, in regulating the maturation of 5-HT
neuron firing frequency through the control of 5-HT autoreceptor–
mediated inhibitory responses. Similar to the innervation defects,
the defects in firing frequencies and autoreceptor-mediated inhibi-
tory responses were probably not caused simply by reduced 5-HT, as
5-HT is also reduced in the brains of Gata3eCKO mice. Instead, our
findings showed that Pet-1 transcriptionally controls spontaneous
firing frequency and inhibitory responses by regulating expression
of the Htr1a autoreceptor gene. In addition, Pet-1 was required for
expression of the presynaptic Htr1b autoreceptor. Because Htr1a and
Htr1b are not normally expressed until several days after 5-HT neuron
generation, our findings support the hypothesis that persistent Pet-1-
directed transcription is essential for maturation steps during which
5-HT neurons acquire key functional characteristics.
We identified a third stage of Pet-1 function using a tamoxifen-
inducible targeting approach that resulted in a severe and selective
reduction in Pet-1 in the adult ascending 5-HT system. Significantly,
the targeting of Pet-1 in adult 5-HT neurons revealed that this late
stage of Pet-1 function is required in the adult ascending 5-HT system
© 2010 Nature America, Inc. All rights reserved.

nature neurOSCIenCe VOLUME 13 | NUMBER 10 | OCTOBER 2010 1197
a r t I C l e S
to maintain emotional behaviors. Our finding that tamoxifen-treated
Pet-1aCKO mice showed altered emotional behavior was supported by
three tests of rodent anxiety-related behavior performed on two inde-
pendent cohorts of mice that were tested several months apart. The
simplest interpretation of our findings is that the accompanying reduc-
tion in 5-HT levels in the ascending serotonergic system accounted
for the elevated anxiety in tamoxifen-treated Pet-1a CKO mice.
However, the literature regarding the effect of neurotoxin-mediated
depletion of 5-HT on anxiety in adult rats is conflicting, with both
anxiogenic and anxiolytic effects being reported depending on
experimental design35–37. Furthermore, as Pet-1 is likely to control a
network of downstream transcriptional targets, the increased anxiety-
like behavior in tamoxifen-treated Pet-1aCKO mice may have resulted
from multiple alterations in adult serotonergic function. We found a
reduction in Slc6a4 and Slc17a8 gene expression in these mice, which
suggests that there were complex changes in the 5-HT neuron genetic
network. A recent study showed that at least part of the ascending
5-HT system engages in dual serotonergic/glutamatergic fast synaptic
transmission38. Reduced expression of Slc17a8 in tamoxifen-treated
Pet-1aCKO mice was not observed until 30 d after the tamoxifen
treatments, indicating that adult loss of Pet-1 expression could have
elicited gradual changes in glutamatergic transmission that contributed
to the behavioral phenotype of tamoxifen-treated Pet-1aCKO mice.
The normal expression of the Htr1a autoreceptor and Htr1a-
mediated inhibitory responses in tamoxifen-treated Pet-1aCKO mice
indicates that the increased anxiety-like behaviors that followed adult
deletion of Pet-1 were not due to deficiencies in Htr1a signaling. This
result is consistent with the findings that although germline targeting
of Htr1a leads to increased anxiety-related behaviors39,40, reduced
Htr1a signaling in adulthood does not41. These findings suggest
that anxiety-like behavior in tamoxifen-treated Pet-1aCKO mice may
be caused by a different process than that responsible for increased
anxiety in Pet-1−/− mice. Nevertheless, our findings provide the first
direct evidence in support of the concept that adult 5-HT–modulated
behaviors are not hardwired during development but are transcrip-
tionally regulated in the adult brain. Moreover, they highlight the
potential importance of perturbations in serotonergic transcription
at any stage of life in emotional pathogenesis.
Several characteristics of the expression and function of Pet-1
suggest that it is a terminal selector gene analogous to the
Caenorhabditis elegans ETS terminal selector gene, ast-1, which
coordinates the induction and maintenance of dopamine synthesis
and transport in postmitotic neurons through a common conserved
terminal selector motif42. Consistent with the fundamental proper-
ties of a terminal selector gene, Pet-1 is expressed throughout the life
of postmitotic 5-HT neurons and is required not only to determine
serotonergic identity but also to maintain it. However, like a terminal
selector gene, it is not required for generic neuronal identity. Further
key features of Pet-1 that fit with its classification as a terminal selector
gene are that it directly regulates and maintains expression of
terminal differentiation genes that define serotonergic-type identity
and positively autoregulates its own expression, all through conserved
ETS binding motifs43. It remains to be determined whether Pet-1
induces other transcription factors that then function cooperatively
in a feedforward loop to control serotonergic identity.
Our findings raise the question of why Pet-1 is still needed in adult
5-HT neurons for regulation of a subset of its known embryonic
targets but not for others such as Ddc, Slc18a2 and Htr1a. Expression
of Tph2 and Slc6a4 in the adult is restricted to Pet-1–expressing 5-HT
neurons and is rate-limiting for the essential serotonergic functions
of 5-HT synthesis and reuptake. The expression of Tph2 and Slc6a4
in the adult DRN are regulated by external stimuli such as selective
serotonin reuptake inhibitors and different stress paradigms44,45.
In addition, the expression of Tph2 and Slc6a4 in vivo are sensitive
to the levels of Pet-1 expression46. We hypothesize (Supplementary
Fig. 1) that environmentally induced alterations in the expression of
Tph2 and Slc6a4 might be mediated through direct transcriptional
activation by Pet-1, which itself is subject to extrinsic regulation47,
thereby providing an efficient homeostatic transcriptional mechanism
that acts throughout life to alter serotonergic function in response to
environmental challenges.
METHODS
Methods and any associated references are available in the online version
of the paper at http://www.nature.com/natureneuroscience/.
Note: Supplementary information is available on the Nature Neuroscience website.
ACKNOWLEDGMENTS
We thank L. Landmesser for suggestions on retrograde tracing; S. Dymecki and
P. Chambon for CreERT2 plasmids; Q. Ma for the mycPet-1 vector; J. Zhu for the
Gata3loxP/loxP mice; K. Lobur for assistance with genotyping of mice; J. Reeves
for behavioral testing in the Case Rodent Behavior Core; and L. Landmesser,
S. Maricich and J. Silver for comments on the manuscript. This work was supported
by grants MH062723 and MH078028 to E.S.D. (US National Institutes of Health).
AUTHOR CONTRIBUTIONS
E.S.D. conceived the project. C.L. made the transgenic and targeting constructs,
characterized all new mouse lines and generated all histological, RT-PCR and
retrograde tracing data and images. C.L. and S.C.W. performed western blot
analyses. C.L. and G.C. performed behavioral analyses. T.M. and S.H. generated the
electrophysiology data. C.L., S.H., T.M., G.C., S.C.W. and E.S.D. analyzed the data.
E.S.D. and C.L. designed the experiments and wrote the manuscript.
COMPETING FINANCIAL INTERESTS
The authors declare no competing financial interests.
Published online at http://www.nature.com/natureneuroscience/.
Reprints and permissions information is available online at http://www.nature.com/
reprintsandpermissions/.
1. Holmes, A. Genetic variation in cortico-amygdala serotonin function and risk for
stress-related disease. Neurosci. Biobehav. Rev. 32, 1293–1314 (2008).
2. Ansorge, M.S., Hen, R. & Gingrich, J.A. Neurodevelopmental origins of depressive
disorders. Curr. Opin. Pharmacol. 7, 8–17 (2007).
3. Vitalis, T., Cases, O., Passemard, S., Callebert, J. & Parnavelas, J.G. Embryonic
depletion of serotonin affects cortical development. Eur. J. Neurosci. 26, 331–344
(2007).
4. Salichon, N. et al. Excessive activation of serotonin (5-HT) 1B receptors disrupts
the formation of sensory maps in monoamine oxidase a and 5-ht transporter knock-
out mice. J. Neurosci. 21, 884–896 (2001).
5. Bonnin, A., Torii, M., Wang, L., Rakic, P. & Levitt, P. Serotonin modulates the
response of embryonic thalamocortical axons to netrin-1. Nat. Neurosci. 10,
588–597 (2007).
6. Ansorge, M.S., Zhou, M., Lira, A., Hen, R. & Gingrich, J.A. Early-life blockade of
the 5-HT transporter alters emotional behavior in adult mice. Science 306,
879–881 (2004).
7. Maciag, D. et al. Neonatal antidepressant exposure has lasting effects on behavior
and serotonin circuitry. Neuropsychopharmacology 31, 47–57 (2006).
8. Gross, C. et al. Serotonin1A receptor acts during development to establish normal
anxiety-like behaviour in the adult. Nature 416, 396–400 (2002).
9. Jans, L.A., Riedel, W.J., Markus, C.R. & Blokland, A. Serotonergic vulnerability and
depression: assumptions, experimental evidence and implications. Mol. Psychiatry
12, 522–543 (2007).
10. Scott, M.M. & Deneris, E.S. Making and breaking serotonin neurons and autism.
Int. J. Dev. Neurosci. 23, 277–285 (2005).
11. Cordes, S.P. Molecular genetics of the early development of hindbrain serotonergic
neurons. Clin. Genet. 68, 487–494 (2005).
12. Hendricks, T.J. et al. Pet-1 ETS gene plays a critical role in 5-HT neuron development
and is required for normal anxiety-like and aggressive behavior. Neuron 37,
233–247 (2003).
13. Dai, J.X. et al. Enhanced contextual fear memory in central serotonin-deficient
mice. Proc. Natl. Acad. Sci. USA 105, 11981–11986 (2008).
14. Hendricks, T., Francis, N., Fyodorov, D. & Deneris, E. The ETS domain factor Pet-1
is an early and precise marker of central 5-HT neurons and interacts with a conserved
element in serotonergic genes. J. Neurosci. 19, 10348–10356 (1999).
© 2010 Nature America, Inc. All rights reserved.

1198 VOLUME 13 | NUMBER 10 | OCTOBER 2010 nature neurOSCIenCe
a r t I C l e S
15. Scott, M.M. et al. A genetic approach to access serotonin neurons for in vivo and
in vitro studies. Proc. Natl. Acad. Sci. USA 102, 16472–16477 (2005).
16. Jacob, J. et al. Transcriptional repression coordinates the temporal switch from
motor to serotonergic neurogenesis. Nat. Neurosci. 10, 1433–1439 (2007).
17. Pattyn, A. et al. Coordinated temporal and spatial control of motor neuron and
serotonergic neuron generation from a common pool of CNS progenitors. Genes
Dev. 17, 729–737 (2003).
18. Zhao, Z.Q. et al. Lmx1b is required for maintenance of central serotonergic neurons
and mice lacking central serotonergic system exhibit normal locomotor activity.
J. Neurosci. 26, 12781–12788 (2006).
19. Krueger, K.C. & Deneris, E.S. Serotonergic transcription of human FEV reveals direct
GATA factor interactions and fate of Pet-1-deficient serotonin neuron precursors.
J. Neurosci. 28, 12748–12758 (2008).
20. Lidov, H.G. & Molliver, M.E. An immunohistochemical study of serotonin neuron
development in the rat: ascending pathways and terminal fields. Brain Res. Bull.
8, 389–430 (1982).
21. Srinivas, S. et al. Cre reporter strains produced by targeted insertion of EYFP and
ECFP into the ROSA26 locus. BMC Dev. Biol. 1, 4 (2001).
22. Vertes, R.P., Fortin, W.J. & Crane, A.M. Projections of the median raphe nucleus
in the rat. J. Comp. Neurol. 407, 555–582 (1999).
23. Le Francois, B., Czesak, M., Steubl, D. & Albert, P.R. Transcriptional regulation at
a HTR1A polymorphism associated with mental illness. Neuropharmacology 55,
977–985 (2008).
24. Fornal, C.A. et al. Single-unit responses of serotonergic dorsal raphe neurons to
5–HT1A agonist and antagonist drug administration in behaving cats. J. Pharmacol.
Exp. Ther. 270, 1345–1358 (1994).
25. Bayliss, D.A., Li, Y.W. & Talley, E.M. Effects of serotonin on caudal raphe neurons:
activation of an inwardly rectifying potassium conductance. J. Neurophysiol. 77,
1349–1361 (1997).
26. Sari, Y. Serotonin1B receptors: from protein to physiological function and behavior.
Neurosci. Biobehav. Rev. 28, 565–582 (2004).
27. Bonnin, A., Peng, W., Hewlett, W. & Levitt, P. Expression mapping of 5-HT1
serotonin receptor subtypes during fetal and early postnatal mouse forebrain
development. Neuroscience 141, 781–794 (2006).
28. Craven, S.E. et al. Gata2 specifies serotonergic neurons downstream of sonic
hedgehog. Development 131, 1165–1173 (2004).
29. Zhu, J. et al. Conditional deletion of Gata3 shows its essential function in T(H)1-
T(H)2 responses. Nat. Immunol. 5, 1157–1165 (2004).
30. Feil, R., Wagner, J., Metzger, D. & Chambon, P. Regulation of Cre recombinase
activity by mutated estrogen receptor ligand-binding domains. Biochem. Biophys.
Res. Commun. 237, 752–757 (1997).
31. Vogt, M.A. et al. Suitability of tamoxifen-induced mutagenesis for behavioral
phenotyping. Exp. Neurol. 211, 25–33 (2008).
32. Scott, M.M., Krueger, K.C. & Deneris, E.S. A differentially autoregulated Pet-1
enhancer region is a critical target of the transcriptional cascade that governs
serotonin neuron development. J. Neurosci. 25, 2628–2636 (2005).
33. Janusonis, S., Gluncic, V. & Rakic, P. Early serotonergic projections to Cajal-Retzius
cells: relevance for cortical development. J. Neurosci. 24, 1652–1659 (2004).
34. Gutknecht, L. et al. Deficiency of brain 5-HT synthesis but serotonergic neuron
formation in Tph2 knockout mice. J. Neural Transm. 115, 1127–1132 (2008).
35. Ludwig, V. & Schwarting, R.K. Neurochemical and behavioral consequences of striatal
injection of 5,7-dihydroxytryptamine. J. Neurosci. Methods 162, 108–118 (2007).
36. Pum, M.E., Huston, J.P. & Muller, C.P. The role of cortical serotonin in anxiety and
locomotor activity in Wistar rats. Behav. Neurosci. 123, 449–454 (2009).
37. Sommer, W. et al. Local 5,7-dihydroxytryptamine lesions of rat amygdala: release
of punished drinking, unaffected plus-maze behavior and ethanol consumption.
Neuropsychopharmacology 24, 430–440 (2001).
38. Varga, V. et al. Fast synaptic subcortical control of hippocampal circuits. Science
326, 449–453 (2009).
39. Heisler, L.K. et al. Elevated anxiety and antidepressant-like responses in serotonin
5–HT1A receptor mutant mice. Proc. Natl. Acad. Sci. USA 95, 15049–15054
(1998).
40. Ramboz, S. et al. Serotonin receptor 1A knockout: an animal model of anxiety-
related disorder. Proc. Natl. Acad. Sci. USA 95, 14476–14481 (1998).
41. Richardson-Jones, J.W. et al. 5-HT1A autoreceptor levels determine vulnerability to
stress and response to antidepressants. Neuron 65, 40–52 (2010).
42. Flames, N. & Hobert, O. Gene regulatory logic of dopamine neuron differentiation.
Nature 458, 885–889 (2009).
43. Hobert, O. Regulatory logic of neuronal diversity: terminal selector genes and
selector motifs. Proc. Natl. Acad. Sci. USA 105, 20067–20071 (2008).
44. Gardner, K.L., Hale, M.W., Lightman, S.L., Plotsky, P.M. & Lowry, C.A. Adverse
early life experience and social stress during adulthood interact to increase serotonin
transporter mRNA expression. Brain Res. 1305, 47–63 (2009).
45. Shishkina, G.T., Kalinina, T.S. & Dygalo, N.N. Up-regulation of tryptophan
hydroxylase-2 mRNA in the rat brain by chronic fluoxetine treatment correlates with
its antidepressant effect. Neuroscience 150, 404–412 (2007).
46. Lerch-Haner, J.K., Frierson, D., Crawford, L.K., Beck, S.G. & Deneris, E.S.
Serotonergic transcriptional programming determines maternal behavior and
offspring survival. Nat. Neurosci. 11, 1001–1003 (2008).
47. Rivera, H.M., Oberbeck, D.R., Kwon, B., Houpt, T.A. & Eckel, L.A. Estradiol
increases Pet-1 and serotonin transporter mRNA in the midbrain raphe nuclei of
ovariectomized rats. Brain Res. 1259, 51–58 (2009).
© 2010 Nature America, Inc. All rights reserved.

nature neurOSCIenCe
doi:10.1038/nn.2623
ONLINE METHODS
Mice. Animal procedures used in this study were approved by the CWRU School
of Medicine Institutional Animal Care in compliance with the National Institutes
of Health guide for the care and use of laboratory animals.
loxP-flanked Pet-1 mice. An 11-kb genomic fragment that included Pet-1 was
subcloned into a targeting construct designed to insert loxP sites around exon 3.
Several rounds of electroporation and G418 selection were performed on geneti-
cally modified R1 ES cells containing a protamine Cre transgene48. A total of 176
colonies were isolated and screened by Southern blot analysis using an NcoI diges-
tion and a 5′ external probe. Nine positive clones were identified and rescreened
using a KpnI digestion and 3′ external probe. The 5′ probe hybridized to an
8.4-kb fragment in wild-type DNA and an 11.2-kb fragment in targeted DNA.
The 3′ probe hybridized to an 8.0-kb fragment in wild-type DNA and a 10.6-kb
fragment in targeted DNA. Two clones, i5h and i7c, were chosen for blastocyst
injection. All resulting chimeras showed germline transmission and were bred
to mice of mixed 129Sv and C57BL/6 backgrounds. The F1 pups from male
chimeras were screened for mice carrying either a loxP-flanked Pet-1 allele or a
conditionally deleted Pet-1 allele using PCR genotyping with following primers.
p1: 5′-ACTCTGGCTTCCCTTTCTCC-3′; p2: 5′-ACTTGGAGGCCTTTTGCT
CT-3′; p3: 5′-TAGGAGGGTCTGGTGTCTGG-3′; p4: 5′-GCGTCCTTGTGTGTA
GCAGA-3′; p6: 5′-ATGCAAGAAGTTTCGGATGG-3′ (Supplementary Fig. 2).
ePet::CreERT2. DNA sequences encoding a fusion protein of Cre recombi-
nase with a mutated estrogen receptor (CreERT2, a gift from P. Chambon via
S. Dymecki) were first subcloned into the pSG5 vector (Stratagene) between
the
β
-globin intron and the simian virus 40 polyadenylation sequences. The
β
-globin-intron/CreERT2/poly(A) cassette was then released from the pSG5
vector and subcloned downstream of the
β
-globin minimal promoter in a
modified BGZA vector in which LacZ was removed. The
β
-globin promoter/
β
-globin-intron/CreERT2/poly(A) region was subcloned downstream of the
40-kb ePet genomic fragment in the modified pBACe3.6 vector15. The transgene
was released from vector with an AscI digestion and purified for pronuclear
injections into hybrid c57B6/129 zygotes. Founders were identified by PCR with
5′-AAAATTTGCCTGCATTACCG-3′ and 5′-ATTCTCCCACCGTCAGTA
CG-3′ primers.
ePet::mycPet-1. DNA sequences encoding a myc-tagged Pet-1 protein (gift from
Q. Ma) as well as a simian virus 40 polyadenylation region were first subcloned
downstream of the
β
-globin minimal promoter in a modified BGZA vector.
The
β
-globin/Pet-1/poly(A) cassette was then released and subcloned down-
stream of the ePet enhancer sequence in pBACe3.6. The transgene was released
from vector with an AscI digestion and purified for pronuclear injections into
Pet-1−/− fertilized eggs in a mixed C57BL/6 and 129 background. Founders
were identified by PCR with 5′-GGGCCTATCCAAACTCAACTT-3′ and
5′-GGGAGGTGTGGGAGGTTTT-3′primers.
Histology. Fluorescent and diaminobenzidine (DAB) immunohistochemistry
were performed as described46. The following primary antibodies were used:
rabbit anti-5-HT (1:10,000, ImmunoStar), mouse anti-TPH (1:200, Sigma),
rabbit anti-GFP (1:1,000, Invitrogen), rabbit anti-Slc18a2 (1:200, Millipore), goat
anti-CHAT (1:200, Millipore), chicken anti-TH (1:100, Aves), mouse anti-NeuN
(1:500, Millipore), mouse anti-GFAP (1:200, Imgenex), rabbit anti-β-galactosi-
dase (1:5,000, MP Biomedicals), rabbit anti-Cre (1:500, Covance), mouse anti-
myc (9E10, Sigma). Secondary antibodies including FITC, TexRed and Cyanine3
(1:200) were from Jackson ImmunoResearch. Fluorescent and bright field images
were collected using a SPOT RT color digital camera (Diagnostic Instruments)
attached to an Olympus Optical BX51 microscope. Confocal images were taken
on a Zeiss LSM 510 confocal laser microscope.
Retrograde tracing. Six wild-type (ePet::Yfp+) and six Pet-1eCKO mice (~P22)
were deeply anesthetized by 1.5% isoflurane in the air flow. Mice were placed
into a stereotaxic frame and a small opening was made in the skull directly
over the injection site (−0.5 mm, 3 mm, 0.5 mm from bregma). Coordinates
for stereotaxic injections were obtained from the Paxinos mouse brain
atlas. About 1 μl of Texas Red–conjugated dextran (5% diluted in 0.5 × PBS,
3,000 Mw, Invitrogen) was pressure injected into the somatosensory barrel
cortex using a Hamilton syringe. After injection, animals were allowed
to survive for another 3 d before being killed for histology.
In situ hybridization. Gene-specific DNA oligonucleotide primers (Supplementary
Table 1) were designed to amplify ~600-bp fragments using cDNA synthesized
from adult DRN mRNA. Forward and reverse primers contained bacteriophage T7
or T3 promoter sequences at their 5′ ends so that PCR products could be directly
used as templates to synthesize digoxigenin (Roche)-labeled sense and antisense
riboprobes. In situ hybridization was performed as described14.
Western blot analysis. Mice were killed 30 d after the last tamoxifen treatment.
DRN tissue was dissected and homogenized in RIPA buffer containing 1× pro-
teinase inhibitor (Sigma). Proteins were quantified using the BCA Protein Assay
Kit (Pierce). Seven micrograms of each protein extract was separated by 10%
SDS-PAGE (BioRad) and then transferred to a 0.45-μm nitrocellulose mem-
brane (BioRad). The antibodies used were a monoclonal anti-TPH antibody
(1:2,000, Sigma), an HRP-conjugated anti-mouse secondary antibody (1:2,000,
Cell Signaling Technology), and an anti-β-actin antibody (1:3,000, Millipore).
The film was developed and then scanned on a HP Scanjet 8200. The mean band
density was measured using ImageJ (http://rsb.info.nih.gov/ij).
Tamoxifen preparation and treatment. Tamoxifen (Sigma) was dissolved in
corn oil at 20 mg ml−1 according to the Joyner laboratory’s protocol (http://www.
mskcc.org/mskcc/html/77387.cfm). For tamoxifen treatment in the embryo, one
single dose of tamoxifen (150 μg per g body weight) was given to the mother
by intraperitoneal injection or oral gavage at E11.5. For treatment in adults,
5 single daily doses of TM (150 μg per g body weight) were given to adult mice
by intraperitoneal injection.
Electrophysiology. Coronal slices including the the DRN (250 μm thick) were cut
from brainstem of untreated Pet-1eCKO, tamoxifen-treated Pet-1aCKO) and control
mice aged 3–5 weeks. Mice were anesthetized with isoflurane and decapitated.
The brainstem was cooled and sliced in ice cold solution containing (in mM) 87,
NaCl; 75, sucrose; 2.5, KCl; 0.5, CaCl2; 7, MgCl2; 1.25, NaH2PO4; 25, NaHCO3;
and 20, glucose bubbled with 95% O2 and 5% CO2 using a vibratome (VT1000S,
Leica). Slices were stored for at least 1 h at room temperature in recording arti-
ficial cerebrospinal fluid containing (in mM) 124, NaCl; 3, KCl; 2.5, CaCl2; 1.2,
MgSO4; 1.23, NaH2PO4; 26, NaHCO3, and 10, glucose bubbled with 95% O2
and 5% CO2. YFP+ cells were visually identified under an upright microscope
(DMLFSA, Leica) equipped with a monochromator system (Polychrome IV, TILL
Photonics). Whole-cell recordings were made from the cells in the dorsomedial
subregion of the B7 DRN. During recordings, slices were continuously perfused
with the external solution containing 10 μM 6-cyano-7-nitroquinoxaline-2,3-
dione disodium (CNQX), 20 μM -(−)-2-amino-5-phosphonopentanoic acid
(D-AP5) and 20 μM picrotoxin at room temperature. Patch pipettes (2–4 MΩ)
were filled with an internal solution with the following composition (in mM)
140, K-methylsulfate; 4, NaCl; 10, HEPES; 0.2, EGTA; 4, Mg-ATP; 0.3, Na-GTP
and 10, Tris-phosphocreatine (pH 7.3, adjusted with KOH). Membrane currents
or voltages were recorded with an EPC10/2 amplifier (HEKA). The signals were
filtered at 3 kHz and digitized at 50 kHz. PatchMaster software (HEKA) was
used for control of voltage and data acquisition. Off-line analysis was performed
with Igor Pro software (Wavemetrics). The Htr1a agonist (±)-8-Hydroxy-2-dipro-
pylaminotetralin hydrobromide (8-OH-DPAT) was purchased from Tocris and
bath-applied to slices.
Quantitative real-time PCR. Mice were anesthetized with Avertin (0.5 g tri-
bromoethanol per 39.5 ml H20, 0.02 ml per g body weight) and killed by rapid
decapitation. Brains were dissected and placed in RNase-free tissue culture plates.
A sterile razor blade was used to cut a transverse section at bregma area −2.92 mm
and then again at bregma area −5.46 mm to isolate the area containing the DRN
and MRN. The tissue was placed immediately in Trizol (Invitrogen) and RNA
was extracted according to the manufacturer’s manual. Genomic DNA was
removed by DNase I treatment (Roche) and 1 μg RNA was used for first-strand
cDNA synthesis (Invitrogen). For real-time RT-qPCR, a SYBR green detection
system (Molecular Probes), fluorescein calibration dye (Bio-Rad), Platinum Taq
(Invitrogen), specific primers (Supplementary Table 2) and 2 μl of undiluted
cDNA were used in 20-μl PCR reactions. Each reaction was performed in
© 2010 Nature America, Inc. All rights reserved.

nature neurOSCIenCe doi:10.1038/nn.2623
triplicate. All real time RT-PCR reactions were performed in 40 cycles on the
iCycler (Bio-Rad). Relative gene expression and statistics analysis were determined
using the Relative Expression Software Tool (http://www.gene-quantification.de/
rest-paper.html).
HPLC analysis. Tissues were collected as described46. HPLC analysis was per-
formed by the Neurochemistry Core Lab at Vanderbilt University, Center for
Molecular Neuroscience.
Sequence analysis. Three kilobases upstream of the predicted human and mouse
Tph2, Slc6a4 and Pet-1 transcription start sites were compared using ECR browser
tool (http://ecrbrowser.dcode.org/) as described19. The minimum criterion for
significant sequence conservation was 70% identity over 100 bp. Gene annota-
tion information was derived from NCBI (Pet-1, GeneID 260298; Tph2, GeneID
216343; Slc6a4, GeneID 15567). Predicted conserved Pet-1 consensus binding
sites (GGAAR(T)) were identified using rVista 2.0 (http://rvista.dcode.org/).
ChIP assays. Hindbrain tissue from the mesencephalic flexure to the cervical
flexure was removed from 56 E12.5 ePet::mycPet-1 transgenic embryos and
quickly frozen on dry ice. MycPet-1 occupancy of genomic regions was tested by
GenPathway, Inc. using goat anti-Myc antibody (Abcam ab9132) and quantitative
PCR (qPCR) according to their protocols. Binding was tested in triplicate for the
negative control region (untranscribed genomic region Untr17) and regions in
or near predicted Pet-1–binding sites. Data are expressed as fold enrichment for
each sample relative to binding at Untr17. Differences in binding among regions
were calculated using one-way ANOVA with Bonferroni’s Multiple Comparison
Test (Prism 5.0, GraphPad Software). Replication of the entire assay gave similar
results. Sequences of primers used for qPCR and their positions relative to the pre-
dicted conserved Pet-1–binding sites for each of the test genes are shown below.
Primers sequences are underlined. Pet-1-binding sites are shown in bold.
Tph2: TTTCCTGTGGCTTTCTAAAGTTGGAAAAGTACAAATATAATC
TTGTCTATGCCTGTCAAATTGCTGGGTCTGATCAGGTCATAGATGGA
GAGCAATAAAATTGTATCAGAAGAGTATCAAAGGAATGATGGGCCTA
TGGGCATTTCATTTCC
Slc6a4: CCCCTTCTTTCCGCTCTATCTTGATTAGCTAGGTCAGCCTCAG
GTGGTTGCTGGGGAGATTCCAGGCCTACTGTGGTGGACATCCGAAAC
AAGAGATTCCCTGAGAGGGAGGGGTGTGGTAGCCATTTCCTGGGCCT
AAGAAGAAGCCCACAAGGAAGGGAGAGCTTCCTCTTCTGTCACGGTG
TAAACAGAACACAGGCAGACAGACAGATGGCACCGAGAGCTTCC
Slc6a4 intron: CATCCTCAGTCCAGAAGAGAGAGCCCAGCTCCTTCCC
TGTGCCCCGTCCCGGCAGTGAAATGAAGGTACAGCCT
Pet-1: GGAAACCAGGAAATCGAGGAGGGGATGGGTCTCTAGGGACC
TAAAGAGAGTAGGAAAAAAGGAGGGAGAAGGCACGGGGGTGGGCAA
AGATAAAGGGAGCCACGGCAGCGCGGTAGCGCGGCTGGGAGCGCAG
CGACAGGCGAGAGGGAGGGAAGCGGAAAT
Behavioral tests. All tests were carried out in the Case Western Reserve
University Rodent Behavior Core. Six-to-eight-week-old Pet-1aCKO (ePet::CreERT2
Pet-1loxP/−) and littermate controls (Pet-1loxP/−) were treated with tamoxifen for
5 consecutive days. After the last tamoxifen injection, mice were rested for 4 weeks
before testing with access to food and water ad libitum. All tests were performed
during the light cycle between 10:30 a.m. and 6:00 p.m. Equipment was cleaned
thoroughly with 70% ethanol between each test to remove odor cues. The elevated
plus maze test was conducted first because of its sensitivity to prior experience.
Individual tests were performed at least 48 h apart. The tester was blinded to group
identification. Cohort 1 was tested in autumn and Cohort 2 in spring.
The elevated plus maze, equipped with infrared grid and video tracking system
(Med Associates Inc.), was ~1 m high and consisted of two open and two closed
arms forming a cross. Mice were placed in the center of the maze facing the
open arm and their activity was recorded for 5 min. The total time spent in the
open arms, closed arms and hub and the number of entries into each arm were
measured. We did not observe differences in frequency of defecation, urination
and head dips between control and tamoxifen-treated Pet-1aCKO mice.
The light/dark box consists of two square dark gray chambers. The lit open
chamber (20 × 20 cm) was illuminated with a 100-W light 40 cm above the
chamber floor and the dark chamber (15 × 15 cm) was entirely enclosed with a
solid black plastic top. Mice were placed in the open chamber, facing away from
the dark side, and their exploration pattern was tracked for 5 min. Latency to
cross over into the dark chamber and total duration in light were scored. We did
not detect differences in the number of re-entries into the illuminated chamber
between control and tamoxifen-treated Pet-1aCKO mice.
The open field consisted of a 40 cm × 40 cm box in a dimly lit room. Using
EthoVision XT 5.0 (Noldus), the area was digitally subdivided into a 20 cm ×
20 cm center area and a peripheral area. The peripheral area was also divided into
a middle (inner 10 cm) and an outer area (outer 10 cm) to determine thigmotaxic
behavior. Animals were placed in the open field and allowed to explore the enclo-
sure freely for 15 min. During this period locomotor parameters such as total
distance moved, velocity, angular velocity and heading degrees were measured to
determine basic locomotor activity and presence of stereotypies. Frequency and
duration in the center, periphery and outer quadrants were collected to determine
anxiety-like behavior. In addition, data were nested into 5-min bins and distance
moved during each of these 3 periods was recorded to evaluate habituation dif-
ferences across groups.
Statistics. All statistical measures on normally distributed data were done using
either a two-tailed t-test between the control and mutant mice or one-way
ANOVA with Bonferroni’s multiple comparison test to compare means between
all combinations of groups. Statistical analysis in the RT-qPCR experiment was
carried out by using the pair-wise fixed reallocation randomization test (http://
www.gene-quantification.de/rest-paper.html).
48. O’Gorman, S., Dagenais, N.A., Qian, M. & Marchuk, Y. Protamine-Cre recombinase
transgenes efficiently recombine target sequences in the male germ line of mice,
but not in embryonic stem cells. Proc. Natl. Acad. Sci. USA 94, 14602–14607
(1997).
© 2010 Nature America, Inc. All rights reserved.