Salivation supports oral health, taste sensation, swallowing, and digestion, yet the brain mechanisms that control it remain largely unknown. Here, we identify a salivatory center in the mouse brainstem that integrates sensory and learned anticipatory signals to control salivation. We show that activating choline acetyltransferase-expressing neurons in the inferior salivatory nucleus (IS) is sufficient to trigger saliva secretion. Using fiber photometry, we monitor real-time IS neural responses to mechanosensory and gustatory stimulation and find tight correlation with salivatory output. We further demonstrate that IS neurons receive input from local brainstem circuits, mediating rapid hardwired responses to taste, as well as direct cortical projections. Notably, gustatory cortex input is required for salivatory responses to predictive sensory cues in a Pavlovian conditioning paradigm. Together, our findings define the circuit underlying taste-evoked and anticipatory salivation and provide a foundation for dissecting this autonomic response in health and disease.

Submandibular gland ganglion injection 

Animals were anesthetized with ketamine and xylazine (100 mg/Kg and 10 mg/Kg, intraperitoneal) and given a nonsteroidal anti-inflammatory drug (carprofen 5 mg/Kg, subcutaneous). After surgery, animals received sustained release formulation pain relief (Buprenorphine ER 1 mg/Kg, subcutaneous). Mice were placed on a custom-made platform in the supine position, and the body temperature was maintained between 34 °C and 36 °C. A small incision was made along the midline of the cervical skin, which was gently retracted laterally to expose a submandibular gland duct. A 2 % Fluorogold (Fluorochrome) solution was loaded into a pulled glass capillary needle (length 9 cm, I.D. 0.53 mm, O.D. 1.14 mm) and unilaterally injected into the largest visible SMG ganglion on the salivary duct using a nano-injector system (Drummond, Nanoject III) at 5 nL/sec to a total volume of 100 nL. The incision was closed with a polyglycolic acid suture (Stoelting, Visorb 4-0, 50496) and 7 days later, animals were sacrificed for histology. 

Stereotaxic injection

All stereotaxic surgery procedures were carried out under aseptic environments and all tools for surgery were autoclaved before use. Mice were anesthetized with a mixture of ketamine and xylazine (100 mg/Kg and 10 mg/Kg, intraperitoneal) and received carprofen (5 mg/Kg, subcutaneous). After surgery, animals received sustained release formulation pain relief (Buprenorphine ER 1 mg/Kg, subcutaneous). Mice were then placed onto a stereotaxic frame (Stoelting) with a closed-loop heating system to maintain body temperature between 34 °C and 36 °C. A small incision was made to expose the skull and a craniotomy (< 1 mm) was made at the injection site using a drill (Roboz, RS-6300A). The viral vectors were loaded into pulled glass capillaries and injected using nano-injector system (Drummond) at 5 nL/sec to the total of 90 nL (AAV-Syn-Flex-GCaMP6s), 150 nL (AAV-EF1a-hChR2(H134R)-mCherry) or 300 nL (AAV-CaMKIIa-hM4D(Gi)-mCherry). Injections were performed unilaterally for fiber photometry and optogenetic stimulation experiments, and bilaterally for the gustatory cortex silencing experiments (total 4 spots). The coordinates were: inferior salivary nucleus (bregma −5.88 mm; lateral 1.1 mm; ventral 3.8 mm); two bilateral cortical fields for gustatory cortex silencing (bregma - 0.3 mm; lateral +/- 4.25 mm; ventral 2.6 mm and bregma 1.7 mm; lateral +/- 3.1 mm; ventral 1.75 mm). For fiber photometry and optogenetic stimulation, a commercially prepared optical fiber from Doric Lenses (400 μm O.D., NA = 0.48) was implanted ~100 μm above the GCaMP or ChR2 virus injection site. All implants were secured onto the skull using dental cement (C&B Metabond, S380). A custom-made head-post was placed to facilitate handling during set up for experiments. After surgery, the animals were returned to their home cages and allowed to recover for at least 14 days before any experiments were performed. In this manuscript, we use ‘Flex’ as a unified term to refer to both double-floxed inverted orientation (DIO) and Flip-Excision (FLEx) configurations for Cre-dependent reporters.

Intraoral cannulation

Tastant solutions were delivered via an intraoral cannula. The cannula implantation was based on an original procedure on rats since adapted for mice. A curved needle (Miltex, Regular Surgeon’s Needle ½ circle, Size 17, MS192-17) attached to silicon tubing (Braintree Scientific, Microrenathane Tubing O.D. 0.037", I.D. 0.023") pierced the gum near the upper left 4th molar. This catheter tubing then exited through the lateral surface of the skull. The mouth-end of catheter tubing was trimmed and the loose top end (~1”) was secured onto the skull surface with dental cement. For the 2-3 days following implantation, animals were given daily antibiotic ointments to prevent infection.

Retrograde monosynaptic tracing

Stereotaxic injections were performed with 100 nL mixture (1:1) of AAV8-CAG-FLEX-TVA-mCherry and AAV8-CAG-FLEX-oG-WPRE-SV40-PA, which was injected into the IS coordinates of ChAT-Cre animals. 4 weeks later, 50 nL of EnvA-pseudotyped RabiesΔG-eGFP (N2c) was injected into the same locale. The animals were sacrificed 14 days later for histological analysis.

Photostimulation and saliva collection

For channelrhodopsin-2 (ChR2) stimulation experiments, 473 nm light pulses were delivered using an Optogenetics-LED system (Prizmatix) through a single optogenetics fiber with a rotary joint. Stimulation was controlled through Pulser software (Prizmatix), and laser intensity was maintained between 8.87 mW at the fiber tip. To measure saliva flow during optogenetic photostimulation, we anesthetized animals with a ketamine/xylazine mixture (100 mg/Kg and 10 mg/Kg, intraperitoneal) and placed pre-weighed pH test papers (Micro Essential Lab MES445) under the tongue, and saliva was wicked for 10 minutes without light stimulation. The paper was then returned to the pre-weighed tube for post-collection weighing. pH of mouse saliva is typically between 7.0 and 8.7, but immediately after salivation, it can rise to a range of 9.0 to 10.0. This change in pH can be observed instantly by the color change on the pH paper upon contact with the saliva. After a 30-minute interval, a new pre-weighed pH test paper was placed under the tongue for the light stimulation condition. During the subsequent 10 minutes, 473 nm light was delivered to the inferior salivatory nucleus via the implanted optical fiber. Saliva was again collected on the pH test paper and weighed. Saliva volume for each photostimulation condition was calculated based on the weight difference before and after the collection. To ensure that anesthesia does not affect saliva secretion, we alternated the order of control and illumination sessions. 

Fiber photometry and stimuli presentation

We measured the activity of IS neurons in freely-moving animals using fiber photometry during intraoral stimulation. Bulk GCaMP fluorescence signals were measured as described previously. Briefly, for fiber photometry, a real-time processor from Tucker Davis Technologies (RZ5P) modulated 465 nm and 405 nm LEDs at 200 and 330 Hz, respectively, and were offset by 7 mA. LED currents were adjusted between 200 and 220 mV from the LEDs driver (four ports mini-cube, Doric) for each signal and excited a through fiber optic patch cord (400 µm diameter, Doric Lenses). Taste delivery timing was synchronized with the photometry system via TTL signals input into the RZ5P processor. Emitted fluorescence was collected and focused onto a Doric photoreceiver, then amplified and digitized by the RZ5P for further analysis. Experiments were performed at least 14 days after surgical procedures to give animals ample time for recovery. After at least 7 days of habituation in the fiber photometry arena, stimuli were delivered via implanted intraoral cannula using a pressurized perfusion system (AutoMate Scientific). A typical panel consisted of a 7-minute session of tastants presented in the following order: 30 mM acesulfame potassium (AceK; sweet), 3 mM quinine (bitter), 50 mM citric acid (sour), 30 mM monopotassium glutamate and 1 mM inosine monophosphate (MPG+IMP) mix (umami), and 100 mM NaCl (salty). Each infusion trial consisted of a 20 second delivery of a single taste solution using a pressure-based system (3–5 psi). Between taste deliveries, a 50 second flow of artificial saliva (6 mM NaHCO₃, 6 mM KHCO₃, 4 mM NaCl, 10 mM KCl, 3.6 mM HCl, 0.5 mM MgCl₂ 6H₂O, 0.24 mM K₂HPO₄, 0.24 mM KH₂PO₄, 0.5 mM CaCl₂ 2H₂O) at the same pressure was used to wash out residual taste stimuli. Pressurized perfusion device used for tastant presentation was set at a high flow rate (23 µL/sec) to discount post-ingestive effects. Under these perfusion settings, animals are unable swallow the solutions (data not shown), therefore we expect minimal exposure of tastants at the back of the tongue. For slower infusion conditions in Figure S2D, flow rate was set to 7 µL /sec, at which none of the solutions dripped out of the mouth. Animals underwent at least 5 days of daily training sessions with all five taste solutions prior to the test date when fiber photometry recordings were made. Additional conditions, including air infusion and standard chow presentation, were included to assess mechanosensation-induced salivation.

The collected photometry data were analyzed by custom python code and Fiber Photometry Modular Analysis Tool (pMAT) software. Z-scored ΔF/F signals were calculated in pMAT by normalizing to the median of a 5-second baseline window (–5 to 0 s), using a bin size of 100 ms. The resulting data were exported as a CSV file and imported into MATLAB, where the average trace across animals was plotted. Shaded regions around the trace represent the standard error of the mean (SEM). The area under the curve (AUC) was calculated by integrating fluorescence signals under identified calcium transients.

Chemogenetic silencing of the gustatory cortex

For chemogenetic silencing of the gustatory cortex, animals were trained with five taste (sweet, bitter, sour, umami and salty) deliveries before initiating fiber photometry recording experiments. On the test day, fiber photometry was performed with taste solution trials after DMSO injection (500 µL/Kg, 5 % DMSO). Subsequently, on the same day, fiber photometry was performed on animals after clozapine-n-oxide (CNO) injection (Tocris, 49-361-0, 1 mg/Kg in 5 % DMSO, intraperitoneal) with the same taste regimen. 

Classical (Pavlovian) conditioning

Animal was placed in a behavior arena inside a sound-isolation chamber (Med Associates). Each animal received 3 days of water infusion to acclimate with the behavior arena and with the solution delivery via the intraoral cannula. Each day consisted of 3 test sessions of 5 trials, with a 30-minute rest between sessions. Each trial consisted of 10 cycles of 2 seconds of taste delivery and 58 seconds of rest. 

For classical conditioning, we trained animals for 7 days to associate a taste solution (30 mM AceK for the sweet group, or 3 mM quinine for the bitter group) with the conditioned stimuli. The cues consisted of a 1 second-long tone (0.25 Hz, 1 second later taste solutions were delivered for 2 seconds. This was repeated for 10 cycles with a 1 minute rest between each trial.

Upon completion of training, animals underwent test sessions consisting of a 1-second auditory beep followed by a 59-second rest period, repeated three times. Each animal received two test sessions: one following infusion of vehicle and the other following administration of clozapine-N-oxide (1 mg/kg in 5 % DMSO). The order of drug administration was counterbalanced across animals.

Immunohistochemistry
Animals were sacrificed by CO₂ inhalation and perfused with phosphate-buffered saline (PBS) and then 4 % paraformaldehyde (PFA, Electron Microscopy Sciences, 15714). After perfusion, brains were post-fixed in 4 % PFA overnight at 4 °C and sectioned coronally at 40 µm thickness. For the Fluorogold antibody staining, brain sections were permeabilized and blocked with 10 % normal donkey serum (Millipore, S30-100ML) in PBS with 0.5 % Triton X-100 (Sigma) at room temperature for 2 hr. Brain tissues were stained overnight at 4 °C with anti-Fluorogold primary antibody (Sigma Aldrich, AB153-I, 1: 1000 dilution), and labeled with fluorescence tagged Alexa fluor 488 secondary antibody (Jackson ImmunoResearch, 711-545-152, 1:1000 dilution) at room temperature for 1 hr. Tissues were then washed with PBS and mounted with Fluoromount-G (SouthernBiotech, OB100-01). Images were collected using a Leica DM6B Widefield epifluorescent microscope in the in the Biological Imaging Facility at Northwestern University. Cell numbers were counted manually.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical tests were performed using unpaired t test or paired t test when appropriate. Applied statistical methods are indicated in the figure legends. p < 0.05 was considered to be statistically significant.