Chemosensory modulation of neural circuits for sodium appetite

Sangjun Lee

Vineet Augustine

Yuan Zhao

Haruka Ebisu

Brittany Ho

Dong Kong

Yuki Oka

Division of Biology and Biological Engineering, California Institute of Technology, Pasadena,

California, USA

Department of Neuroscience, Tufts University School of Medicine, Boston, Massachusetts, USA

Abstract

Sodium is the main cation in the extracellular fluid that regulates various physiological functions.

Sodium-depletion in the body elevates the hedonic value of sodium taste, which drives animals

toward sodium consumption

. Conversely, oral sodium detection rapidly promotes satiation of

sodium appetite

, suggesting that chemosensory signals have a central role in sodium appetite

and its satiety. Nevertheless, the neural basis of chemosensory-based appetite regulation remains

poorly understood. Here, we dissect genetically-defined neural circuits in mice that control sodium

intake by integrating sodium taste and internal depletion signals. We show that a subset of

excitatory neurons in the pre-locus coeruleus (pre-LC) that express prodynorphin (PDYN) serve as

a critical neural substrate for sodium intake behavior. Acute stimulation of this population

triggered robust sodium ingestion even from rock salt by transmitting negative valence signals.

Inhibition of the same neurons selectively reduced sodium consumption. We further demonstrate

that peripheral chemosensory signals rapidly suppressed these sodium appetite neurons.

Simultaneous

in vivo

optical recording and gastric infusion revealed that sensory detection of

sodium, but not sodium ingestion

per se

, is required for the acute modulation of pre-LC PDYN

neurons and satiety of sodium appetite. Moreover, retrograde virus tracing showed that sensory

modulation is partly mediated by specific GABAergic neurons in the bed nucleus of the stria

terminalis. This inhibitory neural population is activated upon sodium ingestion, and sends rapid

inhibitory signals to sodium appetite neurons. Together, this study reveals a dynamic circuit

diagram that integrates chemosensory signals and the internal need to maintain sodium balance.

The precise regulation of sodium balance is crucial for any species

–

. Sodium depletion is

detected by the brain through hormonal actions and visceral afferent signals, and elevates the

valence of sodium taste signals

. This valence shift triggers a strong motivational drive

toward sodium consumption. Sodium taste also has a pivotal role for satiety of sodium

appetite because oral sodium intake provides quicker satiety compared to the non-oral

Users may view, print, copy, and download text and data-mine the content in such documents, for the purposes of academic research,

subject always to the full Conditions of use:http://www.nature.com/authors/editorial_policies/license.html#terms

Correspondence and requests for materials should be addressed to Y.O. (yoka@caltech.edu).

Author Contributions

S.L. and Y.O. conceived the research program and designed experiments. S.L. carried out the experiments and analyzed the data with

help from V.A. and Y.O. H.E. and B.H. performed intragastric surgery. Y.Z. performed all slice patch clamp recordings. D.K.

generated and maintained PDYN-GFP animals. S.L. and Y.O. wrote the paper. Y.O. supervised the entire work.

The authors declare no competing financial interests.

HHS Public Access

Author manuscript

Nature

. Author manuscript; available in PMC 2020 April 03.

Published in final edited form as:

Nature

. 2019 April ; 568(7750): 93–97. doi:10.1038/s41586-019-1053-2.

Author Manuscript

sodium administration (e.g., gastric sodium load)

. These studies indicated the

importance of sodium orosensory signals for the regulation of sodium appetite and its

satiety.

Multiple brain sites including the hypothalamus, amygdala, and hindbrain regulate sodium

consumption

. For example, sodium-depletion stimulates neurons in the lamina terminalis

(LT) and hydroxysteroid dehydrogenase (HSD2)-expressing neurons in the nucleus of

solitary tract (NTS) through a combinatorial action of angiotensin II or/and aldosterone, two

major hormones that regulate body fluid balance

. Recent neural manipulation studies

further confirmed the contribution of LT and HSD2 neurons to sodium intake

–

. In these

studies, however, sodium appetite was only observed under water-deprived conditions

with additional signaling

, representing complex regulatory mechanisms of the appetite.

To gain a circuit-level understanding of sodium appetite, we first searched for a neural

population that selectively controls sodium intake. The pre-LC is a hindbrain nucleus that

receives interoceptive information from the NTS, integrates inputs from other nuclei, and

relays them to the forebrain structures

. Because pre-LC neurons are activated under

sodium-depletion

, we hypothesized that this nucleus serves as a central node that controls

the drive for sodium consumption. Consistent with this notion, neurons in the pre-LC

strongly expressed c-Fos, a neuronal activity marker, under sodium-depletion. The activation

was not observed in sated, water-deprived, or sodium-rescued animals after depletion (Fig.

1a and Extended Data Fig. 1a–c). Our histological analysis revealed that about 60% of pre-

LC excitatory neurons are activated under sodium-depleted conditions (Fig. 1b). We further

screened genetic markers and identified that PDYN expression faithfully (>90%) overlaps

with sodium-depletion-activated neurons (pre-LC

PDYN

neurons, Fig. 1c and Extended Data

Fig 1d). Notably, pre-LC

PDYN

neurons co-expressed Foxp2, a genetic marker for sodium

depletion-sensitive neurons in the pre-LC

(Extended Data Fig. 1e). We next investigated

the functional significance of these neurons for sodium appetite by gain- and loss-of-

function approaches. For optogenetic activation of pre-LC

PDYN

neurons, we infected adeno-

associated virus (AAV) encoding Cre-dependent channelrhodopsin (AAV-DIO-ChR2) into

the pre-LC of PDYN-Cre animals. This neural manipulation triggered robust sodium

ingestion from a high concentration of NaCl solution (0.5 M) that is normally aversive under

sated conditions (Fig. 1d, e) or even from rock salt (Fig. 1f, Supplementary Video 1, and 2).

Importantly, the appetite was sodium specific and was observed regardless of sex or time of

the day (Extended Data Fig.2). Sodium consumption required concurrent stimulation of pre-

PDYN

neurons with sodium presentation (Fig. 1g). These data demonstrate that the

ongoing activity of pre-LC

PDYN

neurons is required for driving sodium appetite.

Moreover, loss-of-function studies revealed the functional necessity of pre-LC

PDYN

neurons

for sodium appetite. Photoinhibition of pre-LC

PDYN

neurons specifically suppressed sodium

ingestion under sodium-depleted conditions (Fig. 1h and Extended Data Fig. 3a–d). Similar

results were obtained using chemogenetic inhibition (Extended Data Fig 3e–h). These gain-

of-function and loss-of-function experiments demonstrate that pre-LC

PDYN

neurons have a

critical role in sodium appetite and intake.

Lee et al.

Page 2

Nature

. Author manuscript; available in PMC 2020 April 03.

Author Manuscript

Classical behavioral studies suggest a model that nutrient deficiency evokes negative internal

states, which drives animals toward consumption to alleviate such discomfort

. To

investigate whether sodium appetite neurons encode a specific valence, we used a two-

chamber real-time place preference assay. We found that photostimulation of pre-LC

PDYN

neurons significantly reduced occupancy time in the compartment paired with light (Fig. 2a).

Thus, the activation of pre-LC

PDYN

neurons is an aversive stimulus to animals. We next

tested if animals would perform a task to reduce the aversive state mediated by pre-LC

PDYN

neurons. For this purpose, we used an operant assay where each lever-press pauses

continuous photostimulation of the pre-LC (Fig. 2b, left panel). Indeed, animals exhibited

robust lever-press behavior to stop stimulation (Fig. 2b right panel and Extended Data Fig

4). Thus, pre-LC

PDYN

neurons transmit a negative valence signal upon activation.

Central appetite circuits receive various sensory and behavioral modulations on a real-time

basis

–

. To investigate the regulatory mechanisms of sodium appetite neurons

in vivo

, we

utilized fiber photometry recording from pre-LC

PDYN

neurons in awake animals during

sodium consumption (Fig. 3a and Extended Data Fig 5a). Sodium-depleted animals were

given access to various solutions while recording GCaMP6s fluorescent signals. We found

that the activity of pre-LC

PDYN

neurons was rapidly and persistently suppressed upon

sodium ingestion (Fig. 3a, c and Extended Data Fig 5b). This robust inhibition was not

observed when animals licked water or an empty bottle (Fig. 3b, c, and Extended Data Fig

5c, d). We next examined if the persistent inhibition is selectively driven by chemosensory

detection of sodium. In contrast to NaCl, no suppression was observed by KCl (0.5 M, Fig.

3d and Extended Data Fig 5e), excluding the possibility that the effect is induced by

osmolality changes. Importantly, blocking the sodium taste receptor by amiloride

fully

abolished NaCl-induced suppression of pre-LC

PDYN

neurons (Fig. 3e and Extended Data

Fig 5e). Moreover, a brief contact to NaCl was sufficient to induce robust suppression for

several minutes (Fig. 3f). These results suggest that oral chemosensory signals, likely

mediated by the taste system, mediate acute modulation of sodium appetite neurons.

Given the functional significance of the pre-LC

PDYN

population for sodium intake, we

hypothesized that inhibition of these neurons contributes to satiety of sodium appetite. We

examined this possibility using photometry recording combined with intragastric (IG)

infusion that allows sodium administration without stimulating orosensory systems (Fig. 4a).

Surprisingly, we found that gastric preloading of NaCl in sodium-depleted mice did not

affect subsequent sodium ingestion, whereas oral NaCl contact quickly quenched the

appetite (Fig. 4b left panel). Only after a long period of IG infusion, significant appetite

reduction was observed (41.3 ± 6.0 % after 2 hrs, n = 6 mice). In sharp contrast, IG water

and glucose infusion in thirsty and hungry animals, respectively, suppressed water/food

consumption shortly after the infusion (Fig. 4b middle and right panels). These results

highlight two key aspects of satiety regulation. First, sodium chemosensory inputs are

critical for rapid satiety of sodium appetite. Second, individual appetite circuits appear to

receive temporally-distinct modulations from post-oral signals.

If pre-LC

PDYN

neurons are involved in taste-mediated satiety, we anticipated that they

should be solely suppressed by oral detection of sodium. Consistent with our hypothesis,

while oral NaCl consumption drastically reduced the neural activity of pre-LC

PDYN

neurons

Lee et al.

Page 3

Nature

. Author manuscript; available in PMC 2020 April 03.

Author Manuscript

(Fig. 4c, e), IG infusion of NaCl in the same set of animals had no inhibitory effect (Fig. 4d,

e), suggesting that oral sodium detection facilitates satiety of sodium appetite via persistent

suppression of pre-LC

PDYN

neurons.

Our results indicate a model that the pre-LC integrates the internal sodium need and sensory

information to regulate real-time sodium appetite. We next dissected the neural circuits

carrying these signals. Because HSD2 neurons in the NTS (NTS

HSD2

) project to the pre-LC

, we tested if pre-LC

PDYN

neurons directly receive the interoceptive information from

the NTS. Using ChR2-assisted circuit mapping (Fig. 5a and Extended Data Fig. 6a), we

found that a majority of recorded pre-LC

PDYN

neurons received monosynaptic excitatory

inputs from NTS

HSD2

neurons. We further examined the functional significance of this

connection for pre-LC activity. Optogenetic stimulation of NTS

HSD2

neurons activated pre-

PDYN

neurons in sated animals (Extended Data Fig. 6b). Conversely, ablation of

NTS

HSD2

neurons by caspase greatly attenuated c-Fos expression in pre-LC

PDYN

neurons

after sodium-depletion (Fig. 5b and Extended Data Fig. 6c, d). These data demonstrate that

the excitatory NTS

HSD2

→

pre-LC

PDYN

connections are necessary and sufficient to activate

pre-LC

PDYN

neurons under sodium-depletion.

We further searched for neural circuits that mediate chemosensory-dependent inhibition by

monosynaptic rabies tracing (SAD-ΔG-BFP) from pre-LC

PDYN

neurons. These tracing

experiments identified several brain regions with most prominent inputs from the dorsal area

of the bed nucleus of the stria terminalis (dBNST), and the central amygdala (Fig. 5c and

Extended Data Fig. 7a). Because the BNST was previously shown to contribute to sodium

consumption

, we focused our functional analysis on the dBNST

→

pre-LC circuit. In

dBNST, a majority of RV-positive neurons were the inhibitory population that also expressed

PDYN (dBNST

PDYN

, Fig. 5d and Extended Data Fig. 7b). We confirmed in slice recording

that dBNST

PDYN

neurons send monosynaptic inhibitory inputs to pre-LC

PDYN

neurons (Fig.

5e). If this dBNST

→

pre-LC circuit mediates rapid satiety signals, dBNST neurons should

be activated upon sodium ingestion. To test this idea, we infected canine adenovirus

(CAV2)-Cre in the pre-LC and AAV-flex-GCaMP6s in the dBNST in order to label

dBNST

→

pre-LC neurons (Extended Data Fig. 7c). As anticipated, optical recording from

dBNST

→

pre-LC neurons demonstrated that they responded upon NaCl intake under

sodium-depleted conditions, which were strongly inhibited by amiloride (Fig. 5f). Taken

together, our results show that pre-LC

PDYN

neurons integrate sensory and internal

information through multiple excitatory and inhibitory inputs.

In this study, we dissected the neural circuitry that controls sodium appetite and consequent

ingestive behavior. Extensive studies in the past decade have shown that sodium intake is

regulated by multiple hindbrain and forebrain structures including the NTS, BNST, and

amygdala

. Interestingly, pre-LC

PDYN

neurons are anatomically connected to these

structures (Fig. 5c and Extended Data Fig. 8) suggesting their integral roles for sodium

appetite. At the molecular level, the central opioid signaling has been shown as a key

regulator of sodium intake

. Because pre-LC

PDYN

neurons express PDYN, a precursor of

-opioid receptor agonists, understanding the action of dynorphin at downstream sites may

provide novel molecular insights for appetite regulations.

Lee et al.

Page 4

Nature

. Author manuscript; available in PMC 2020 April 03.

Author Manuscript

Recent

in vivo

analyses of appetite circuits in rodents revealed that ingestive behaviors

rapidly modulate appetite circuits prior to nutrient absorption in the body

. Thirst, hunger,

and sodium appetite neurons are all inhibited upon ingestion of nutrients in need, whereas

the underlying mechanisms appear to be different for individual appetites. Hypothalamic

hunger neurons are acutely suppressed by food anticipation as well as nutrient detection in

the gut

. Conversely, thirst neurons are inhibited by liquid-gulping motion and post-oral

osmolality changes

. This study revealed another aspect of appetite regulation:

chemosensory signals. Our circuit analysis revealed that pre-LC

PDYN

neurons receive

multiple inputs from upstream neural populations, although not directly from gustatory

regions of the NTS. Defining the genetic identity of these populations and their interactions

in the pre-LC would provide further insights for homeostatic and chemosensory regulations

of the innate drive toward sodium consumption.

METHODS

Animals

All procedures followed animal care guidelines from NIH for the care and use of laboratory

animals and California Institute of Technology Institutional Animal Care and Use

Committee (1694–14). Animals at least six weeks old were used for experiments. The

following mice were purchased from the Jackson Laboratory: C57BL/6J, stock number

00064. Slc17a6-Cre, stock number 016963. Ai75D, stock number 025106, Ai3, stock

number 007903. HSD2-Cre mice were provided by A. and G. Fejes-Tóth (Dartmouth

Medical School). PDYN-GFP mice were provided by D. Kong (Tufts University School of

Medicine). PDYN-Cre mice were provided by B. Lowell (Harvard Medical School) and M.

Krashes (NIH). Ai110 line was provided by D. Anderson (Caltech). Mice were housed on a

13 h: 11h light: dark cycle with ad llbitum access to food and water except for specific

depletion experiments (water, food, sodium). Male and female mice were used for

experiments.

Viral constructs

The following AAV viruses were purchased from the UNC Vector Core AAV1-CAG-flex-

RG (3.0 × 10

genome copies per ml), AAV1-EF1a-flex-TVA-mCherry (6.0 × 10

genome

copies per ml), AAV2-EF1a-DIO-eYFP (4.6 × 10

genome copies per ml), AAV1-EF1a-

DIO-ChR2-mCherry (5.1 × 10

genome copies per ml), AAV5-Ef1a-DIO iC++-eYFP (4.5

× 10

genome copies per ml), AAV5-flex-taCasp3-TEVp (4.5 × 10

genome copies per

ml). The following AAV viruses were purchased from the UPenn virus core, AAV1-hSyn1-

flex-GCaMP6s-WPRE-SV40 (2.28 × 10

genome copies per ml), AAV5-EF1a-DIO-ChR2-

eYFP (3.3 × 10

genome copies per ml), AAV1-EF1a-DIO-ChR2-mCherry (2.0 × 10

genome copies per ml). The following AAV viruses were purchased from Addgene, AAV8-

hSyn-DIO-hM4D(Gi)-mCherry (1.9 × 10

genome copies per ml), AAV5-hSyn-DIO-

mCherry (1 × 10

genome copies per ml), AAV8-Ef1a-DIO-iC++-eYFP (8.5 × 10

genome copies per ml) was purchased from the Stanford Virus vector core. SAD- ΔG-BFP

(1.7 × 10

genome copies per ml) was purchased from Salk. CAV-Cre (1.5 × 10

genome

copies per ml) was purchased from Plateforme de Vectorologie de Montpellier (PVM).

Lee et al.

Page 5

Nature

. Author manuscript; available in PMC 2020 April 03.

Author Manuscript

Surgery

Mice were anaesthetized with a mixture of ketamine (1 mg/mL) and xylazine (10 mg/mL) in

isotonic saline, intraperitoneally injected at 10 μl /g body weight. Ketoprofen was

subcutaneously administered at 5 μl /g body weight. The animal was placed in a stereotaxic

apparatus (Narishige Apparatus) with a heating pad. Surgery was performed as previously

described

. The three-dimensional MRI coordinate system was used as a reference for the

injection site coordinates. Viral constructs were injected using a microprocessor-controlled

injection system (Nanoliter 2000, WPI) at 100 nl /min. The coordinates for pre-LC are AP:

−9000, ML: ±1000, DV: −3900 (60–100 nl injection), for dBNST are AP: −3100 ML: 1100

DV: −3600 (100 nl injection), for NTS are AP: −10800 ML: ±150 DV: −5100, −5300 (100–

300 nl injection each).

For optogenetic experiments, implants were made with a 200 μm fiber bundle (FT200EMT,

Thorlabs) glued to a ceramic ferrule (CF230 or CFLC230, Thorlabs). For photometry,

customized implants (400 μm diameter, Doric Lenses) were used. A fiber implant was

placed 200–300 μm (for optogenetic) or 0–50 μm (for photometry) above the virus injection

site. Histology position of fiber implant was confirmed after data collection. Data from

implant disposition was not included. For IG infusion, catheter construction and

implantation closely followed as described previously

. IG catheters were custom made

using silastic tubing (Dow Corning, 508–002), tygon tubing (Instech, BTPE-25) and pinport

(Instech, PNP3F25–50). For photometry recording, IG surgery was performed after animals

recovered from the initial implantation of an optic fiber. After surgery, all animals were

placed in a clean cage placed on a heating pad overnight and then were housed in the animal

facility. Behavioral and histological assays were performed after at least 10 days of recovery.

For ablation experiments, AAV-flex-taCasp3-TEVp or AAV-hSyn-DIO-mCherry (control)

was injected. These animals were sodium-depleted after 2–3 weeks of recovery. At the end

of experiments, all animals were sacrificed for histological examination. For fiber

implantation experiments, we occasionally observed that the position of an implanted fiber

shifted in the hindbrain due to cranial deformation.

Optogenetic and chemogenetic manipulations

For ChR2 photostimulation. 473 nm laser pulses (20ms, 20Hz) were delivered via an optic

cable (MFP-FC-ZF, Doric Lenses) using a pulse generator (Sapphire 9200 from Quantum

composers or SYS-A310 from WPI). The laser intensity was maintained at 5–10 mW at the

tip of the fiber. Unless otherwise noted, photostimulation was delivered for 1 s at 3 s

intervals throughout the behavior session. For iC++ photoinhibition

, 473 nm laser was

continuously turned on throughout the session at 3 mW at the tip of the fiber. For

chemogenetic manipulation

(Extended Data Fig.3e–h), CNO (Sigma, 10 mg/kg) or vehicle

(water) was administered intraperitoneal 20 min before the sodium consumption experiment.

Preference assay

To induce sodium appetite, animals were injected with furosemide (Sigma) at a dose of

50mg/kg body weight. Low sodium diet (TD. 90228, ENVIGO) was provided for 2 days

after the injection of furosemide. For water-restriction experiments, animals were kept in

their home cage without water, and were provided with 1 mL of water daily. For food

Lee et al.

Page 6

Nature

. Author manuscript; available in PMC 2020 April 03.

Author Manuscript

restriction experiments, animals were deprived of food up to 24 hrs with normal water

provided. All assays were performed in a custom gustometer (

Dialog

Instruments) or Biodaq

monitoring system (Research Diets Inc)

. All sodium-depleted animals were trained

in a gustometer before experiments. Animals which licked at least 150 licks during the 30-

min session were used for further behavioral assays. After every sodium-depletion round,

animals were recovered for at least 4 days with the normal diet.

For appetite specificity assay (Extended Data Fig 2c), three different solutions were

presented to animals during the same session, and their preference was measured as a lick

number. For each trial, 20 sec of photostimulation was delivered to the animal with an inter-

trial-interval of 60 secs. We used water, 0.5 M KCl, 0.5 M NaCl, or 0.5 M NaCl + 30 μM

amiloride for preference assay. For sodium consumption assay (Fig. 1e, g, and h), animals

were given ad libitum access to 0.5 M NaCl, water or an empty spout for either 7.5 or 30

min.

For photometry recording, animals were given either 5 or 10 min access to stimuli. To

examine sodium specific responses of pre-LC

PDYN

neurons (Fig. 3d, e), animals were

presented with three solutions during the session. First, animals were given 5 min access to

0.5 M KCl (Fig. 3d) or 0.5 M NaCl + 0.1 mM amiloride (Fig. 3e). Then animals had 5 min

access to water. Finally, animals were given 5 min access to 0.5 M NaCl. The interval

between trials was 5 min.

Rock salt intake behavioral assay

Sodium-depleted animals were acclimatized for 1 hour in an acrylic box (50 cm X 25 cm X

25 cm) with a rock salt (Halite Himalayan Crystal Salt). Then the lick events of rock salt

were monitored for 30 min using a webcam under sated, sodium depleted, or

photostimulated conditions. The start and end of bouts were manually annotated and

quantified.

Real-time place preference

Real-time place preference was performed in a two-chamber acrylic box (50 cm x 25cm x

25cm) as described previously

. Each side of the chamber had distinct visual and textural

cues (different size and shape of holes of plastic bin). A custom MATLAB code was used

for real-time optogenetic stimulation and analyzing the place preference. The initial

preference for each animal was determined during the initial 30-min session without

photostimulation, which was followed by three test sessions with photostimulation. Light

(20Hz, 5s ON 5s OFF) was delivered through an optic fiber in the initially-preferred side.

Negative reinforcement assay

To examine if appetite neurons transmit negative valence

, animals were first

acclimatized in the operant conditioning chamber (MedAssociate). After acclimatization, the

animals were trained to press the lever to avoid negative stimulus (foot shock at 0.15–0.18

mV). Each lever press paused the foot shock for 20 sec. Training was completed when an

animal pushed the lever more than 20 times during the 30-min session. In a test session,

Lee et al.

Page 7

Nature

. Author manuscript; available in PMC 2020 April 03.

Author Manuscript

animals were given continuous 20 Hz photostimulation to the pre-LC, which was paused for

20 sec by each lever press. The number of lever presses during the session was quantified.

Intragastric infusion

In Fig. 4b, 0.5 M NaCl, deionized water, or glucose solution (5 M) was infused via an

intragastric catheter. Solutions were delivered at 0.1 ml/ min for 5 min (water and sodium) or

10 min (food) using an infusion syringe pump (NE-300, New Era Pump Systems Inc). 10

min after gastric infusion, animals were given access to nutrients and their consumption was

quantified for 10 min (for water and 0.5 M NaCl), or 30 min (for normal chow). Either air

infusion (for water), or water infusion (for sodium and food) was used as a control stimulus.

In Fig. 4d, either 0.15 M NaCl, water, or air was infused at a rate of 0.1 ml/ min while

recording the neural activity by photometry. For control, oral ingestion (Fig. 4c), the same

set of animals were given access to 0.15 M NaCl.

Fibre photometry

For all photometry assays, animals were acclimatized for 10–15 min in the chamber before

stimuli were presented. Bulk fluorescence signals were collected using fibre photometry as

previously described

. Briefly, data were extracted and subjected to a low-pass filter at

1.8 Hz. A linear function was used to scale up the 405-nm channel signal to the 490-nm

channel signal to obtain the fitted 405-nm signal. The resultant ΔF/F was calculated as (raw

490 nm signal – fitted 405 nm signal)/ (fitted 405 nm signal). ΔF/F was then time-binned by

a factor of 2.5 times the sampling frequency and down-sampled to 10 Hz. For all bouts, the

mean fluorescence for 5 min before the first lick was calculated and subtracted from the

entire session. The licks from the lickometer were simultaneously recorded. The area under

the curve (AUC) was quantified by integrating the baseline-subtracted fluorescence signals

for 30 sec after the start of the bout. For Extended Data Fig 5d, the data were quantified as

ΔF/F change between 1 sec prior to, and at the first lick (0 sec). For IG infusion experiments

(Fig. 4e), AUC was quantified during the 5-min of infusion.

Retrograde Viral tracing

For monosynaptic rabies tracing

of pre-LC

PDYN

, 100 nl of a mixture of AAV1-CAG-flex-

RG and AAV1-EF1a-flex-TVA-mCherry (4:1 ratio) was injected to the pre-LC. Two weeks

later, 200 nl of SAD-ΔG-BFP was injected into the pre-LC. The mice were euthanized a

week later.

To label the dBNST

→

pre-LC circuit, 100nl of CAV-Cre was injected into the pre-LC

followed by the injection of AAV5-DIO-mCherry or AAV1-flex-GCaMP6S into the dBNST.

These animals were used for experiments at least two weeks after the injection.

Histology

Mice were anaesthetized and were perfused with PBS followed by 4% PFA in PBS (pH 7.4).

The brain was dissected and fixed in 4% PFA at 4 °C for overnight. Fixed samples were

sectioned into 100 μm coronal sections using a vibratome (Leica, VT-1000 s). For

immunohistochemistry (IHC), brain sections were incubated in a blocking buffer (10%

Donkey serum, 0.2% Triton-X) for 1–2 hrs. Then sections were incubated with primary

Lee et al.

Page 8

Nature

. Author manuscript; available in PMC 2020 April 03.

Author Manuscript