Detecting CO2-Sensitive Hemichannels in Neurons in Acute Brain Slices

Summary This protocol provides two independent methods to functionally detect the neuronal expression of CO2-sensitive hemichannels. These hemichannels (consisting of connexins 26 or 30) are directly gated by CO2, independent of pH changes and until recently were thought to be only expressed by glia. This protocol outlines a method to change the concentration of CO2 without changing pH, using isohydric solutions and then utilizing this to detect opening and closing of functional hemichannels using whole-cell patch clamp recording and dye loading. For complete details on the use and execution of this protocol, please refer to Hill et al. (2020).


HIGHLIGHTS
Protocols for detecting CO 2sensitive hemichannels in neurons of acute brain slices Use of electrophysiology to look for changes in neuronal firing and input resistance Use of dye loading to confirm neuronal CO 2 -sensitive hemichannel expression Use of immunohistochemistry to confirm connexin expression in neurons of interest SUMMARY This protocol provides two independent methods to functionally detect the neuronal expression of CO 2 -sensitive hemichannels. These hemichannels (consisting of connexins 26 or 30) are directly gated by CO 2 , independent of pH changes and until recently were thought to be only expressed by glia. This protocol outlines a method to change the concentration of CO 2 without changing pH, using isohydric solutions and then utilizing this to detect opening and closing of functional hemichannels using whole-cell patch clamp recording and dye loading. For complete details on the use and execution of this protocol, please refer to Hill et al. (2020).

BEFORE YOU BEGIN
Note: We use the following abbreviations in this protocol:

Electrophysiology Setup
Timing: 1.0-1.5 h 1. The researcher will need access to a whole-cell patch clamp electrophysiology recording rig to be able to record from the neurons of interest in acute brain slices. For the dye-loading experiments, there is an additional requirement for fluorescent illumination which could be part of a patch clamp or calcium imaging rig. The fluorescence needs to be~488 nm with a total magnification of at least x400 (x40 objective and x10 camera magnification) with the facility to record images. 2. Acute brain slices containing the relevant brain area (recommendation of 300-350 mm thickness) first have to be prepared and can then be stored at 34 C in 35 mmHg CO 2 buffer (see recipes below). Slices can be stored for up to 6-8 h post-slicing.
Note: Standard slice preparation protocols will generate brain slices that can be stored and used between 1 to 8 h post slicing. This time will differ depending on the age of the animal and the region of interest. Always allow the slices a minimum of 1 hour for recovery after slicing before experiments.
Note: This protocol can be applied to both mouse and rat brain slices. It can be used with any region of interest and any age of animal as required by your study. We have tested dopaminergic neurons in the substantia nigra pars compacta and GABAergic neurons in the ventral tegmental area using two different ages of mice (P7-10 and P17-21) and showed CO 2 sensitivity. As a control, we also demonstrated that there was no CO 2 sensitivity in the CA1 region of the hippocampus using both dye loading and electrophysiology. See Hill et al (2020) for more details.
3. The researcher will need a 95% O 2 and 5% CO 2 supply, an 100% CO 2 supply, and an 100% O 2 supply. The researcher will need mixer valves so the gases can be mixed to produce the correct levels of CO 2 for the experiments (see below for details) and a pH meter and electrode.
CRITICAL: It is vital to correctly use the mixer valves to ensure a constant pH across the two solutions (20 or 55 mmHg CO 2 , relative to 35 mm Hg) as this protocol requires isohydric (constant pH) conditions in order to isolate pH-independent CO 2 -mediated effects. Always monitor the pH of the solutions throughout the day, to ensure that the pH does not drift over time. Although the mixer valves take more time and monitoring than prebought gases, they allow for much greater experimental flexibility so that bespoke mixtures can be created as required.
Note: The midpoint for connexin 26 hemichannel opening is around 35 mmHg CO 2 . Therefore, if cells are sensitive, then fluctuations above this (to 50 mmHg CO 2 ) and below this (to 20 mmHg CO 2 ) would be expected to alter the open probability of the channels and thus the cells would display a conductance phenotype.
Note: For the electrophysiology, we carry out two different experiments: firstly, we change from 35 to 55 mmHg CO 2 buffer to increase the opening of the hemichannels and in separate experiments we change from 35 to 20 mmHg CO 2 buffer to increase the closing of the hemichannels.

Isohydric Solution Preparation
Timing: 30 min Note: The baseline solution (buffer) is always 35 mmHg CO 2 . Depending on experiment, either 55 mmHg or 20 mmHg CO 2 buffer will also need to be prepared. These buffers differ in their concentrations of NaCl and NaHCO 3 (see Materials and Equipment for further details).
Note: Before starting, ensure that you have the mixers set up correctly (see Figure 1). There are a number of suppliers for mixers, we use PLATON Variable-area flow meters. If you would rather buy bespoke pre-made CO 2 /O 2 mixtures to avoid having to mix your own, you can do this at an extra cost.
Note: For 20 mmHg CO 2 buffer you will need 2 3 0-10 mL/MIN AIR 3 100 (6831), one for the 100% O 2 line and one for the 95% O 2 /5% CO 2 line to mix the two gases in a roughly 1:1 proportion.  (2010)) 5. Generating correct gas mixtures and correcting for pH differences a. The 35 mmHg CO 2 buffer is bubbled with standard 95% O 2 /5% CO 2. The pH for this solution does not need to be altered but should be checked each day to ensure that it is~7.4. Test this first before beginning the mixer preparation for (b) and (c). If the solution is not~pH 7.4, then remake to ensure no errors have been made. b. The 55 mmHg CO 2 buffer is saturated with 9% CO 2 (with the balance being O 2 ) with pH matched to control (35 mmHg CO 2 buffer). This is achieved by mixing the 95% O 2 /5% CO 2 with the 100% CO 2 using the mixer valves. The 35 mmHg CO 2 buffer is bubbled with 95% O 2 /5% CO 2 . The pH of this buffer should not need to be altered and should be~7.4. The 55 mmHg CO 2 buffer is saturated with 9% CO 2 (with the balance being O 2 ) with pH maintained to match control (35 mmHg CO 2 ). This is achieved by mixing the 95% O 2 /5% CO 2 line with a 100% CO 2 line (adjustments are made using the mixer dials). The 20 mmHg CO 2 buffer is saturated with 2% CO 2 (with the balance being O 2 ), with pH maintained to match control (35 mmHg CO 2 ). This is achieved by mixing the 95% O 2 /5% CO 2 line with a 100% O 2 line (adjustments are made using the mixer dials). Created with BioRender.com i. Place the pH electrode into the 55 mmHg CO 2 buffer solution. Make small adjustments to the levels of the 95% O 2 5% CO 2 and 100% CO 2 by turning the black dials at the bottom of the mixer. Allow 5-10 min for equilibration and then read the pH from the meter. ii. Continue this process until the pH matches that of the 35 mmHg CO 2 buffer solution.
iii. Periodically check throughout the day that the pH has not altered. c. The 20 mmHg CO 2 buffer is saturated with 2% CO 2 (with the balance being O 2 ), with pH matched to control (35 mmHg CO 2 buffer). This is achieved by mixing the 95% O 2 5% CO 2 with 100% O 2 using one of the mixer valves. i. Place the pH electrode into the 20 mmHg buffer solution. Make small adjustments to the levels of the 95% O 2 5% CO 2 and 100% O 2 by turning the black dials at the bottom of the mixer. Allow 5-10 min for equilibration and then read the pH from the meter. ii. Continue this process until the pH matches that of the 35 mmHg CO 2 buffer solution.
iii. Periodically check throughout the day that the pH has not altered.
Note: Setting up and calibrating the mixers will take some time on the first day of experiments, but if the mixer dials are left at the end of the experiments, then only minor adjustments will be needed for subsequent experiments.

Preparation of Solutions for Dye Loading Experiments
Timing: 10 min 6. Generate isohydric solutions as outlined above (35 mmHg CO 2 and 55 mmHg CO 2 Note: A 103 stock can be made including NaCl, NaHCO 3 , NaH 2 PO 4 and KCl and stored at 4 C until the day of use. Glucose and CaCl 2 must be added on the morning of each experiment. Note: The buffers were originally confirmed to be either 35, 20, or 55 mmHg CO 2 using a blood gas analyzer. Each of the solutions was tested in triplicate. They have since been used routinely in the lab. If you wanted to use a new level of CO 2 , then you would need to alter the buffer recipe accordingly and measure the level of CO 2 (mmHg) with an appropriate method (such as gas analyzer or CO 2 electrode). We found that the composition of the buffer required to give the desired pCO 2 cannot be predicted using the Henderson Hasselbalch equation, which we think is due to the presence of other ions in the buffers.
CRITICAL: When making the solutions, add all reagents except CaCl 2 and 50 mL of dH 2 0. Bubble this solution for 10 min with 95% O 2 /5% CO 2 , then add CaCl 2 , and top up with the remaining dH 2 0 to avoid any precipitation.

STEP-BY-STEP METHOD DETAILS Electrophysiological Detection of Hemichannels in Neurons in Acute Brain Slices
Timing: 5-6 h This protocol provides a method for testing using electrophysiology whether changes in CO 2 levels can modulate neuronal excitability.
1. Making recordings from identified neurons in acute brain slices a. Perfuse the recording bath (2-3 mL/min) with the 35 mmHg CO 2 buffer (bubbled with 95% O 2 /5% CO 2 ) and allow to reach recording temperature (~30 C). b. Place brain slice in the bath (hold in place using a slice hold down for example, Multichannel systems #64-1418) and allow to equilibrate for 10-15 min c. Make patch clamp recording in current clamp mode from neuron(s) of interest. See Figure 2 for details of the process or refer to the axon guide (Molecular Devices, 2020) for a more thorough overview of the patch clamp technique.
Note: Any standard patch clamp electrophysiology rig can be used for this protocol. There are many possible options for rig set up, but for clarity our slices were visualized using IR-DIC optics with an Olympus BX151W microscope (Scientifica, Bedford UK) and a CCD camera (Hitachi). Voltage recordings were made using an Axon Multiclamp 700B amplifier (Molecular ll OPEN ACCESS Devices, USA) and digitized at 20 KHz (Digidata 1440A; Molecular devices, USA). Data acquisition and analysis were performed using pClamp 10 (Molecular Devices, USA).
Note: Always ensure intracellular solution is filtered directly before use. i. Carry out the standard IV protocol ( Figure 3B) to check the health of the neuron, seal stability and identity (resting membrane potential, firing pattern) and allow cell to equilibrate for 5 min. Can also inject naturalistic current to measure firing rate ( Figure 3C).
CRITICAL: Monitor the bridge balance/ series resistance across recordings to ensure stability. Any recordings where this diverges by more than 20% over the period of recording should be disregarded.
ii. Begin running current steps (50 pA hyperpolarizing step, 1s interval; Figure 3A) so you can monitor the input resistance of recorded neuron. iii. Generate a baseline of current steps over 10-15 min.
2. Changing the level of CO 2 to open or close hemichannels Firstly, identify the neuron of interest using the brightfield image. Then using a glass pipette (5-7 MU resistance) approach the cell with positive pressure on the pipette. We monitor and control the pressure with our mouths, but it is also possible to use a syringe to apply pressure (but there is less control). Upon making contact with the target cell you should see a dimple form on the cell at the end of the pipette (as a result of the positive pressure). To monitor the resistance of the pipette and sealing, a square step voltage of 5-10 mV (at 1 Hz, as shown in A) is applied. The amplitude of the current response gives a measure of the resistance B) Once the dimple is observed, release the pressure and this will allow a high resistance seal to form between the end of the pipette and the cell membrane (some negative pressure may also need to be applied). Since I = V/R (Ohms law), as the resistance of the pipette goes up (as the pipette seals to the membrane), the current response will get smaller until the trace appears flat with only the pipette capacitance transients remaining (B). The seal resistance needs to be at least 1 GU but hopefully higher for a good recording (if the seal is poor then a new recording has to be made from a new cell, using a new pipette). The pipette transients can be zeroed off at this point. Increasing negative pressure is slowly applied to the pipette in order to rupture the cell membrane under the pipette and this allows break through into whole-cell. At this point the capacitance transients for the cell are detected (C). At this stage we would then move from voltage clamp into current clamp mode (the resting potential for a good recording will be~À60 to À70 mV depending on cell type) (D). The contents of the patch pipette will start to dialyze into the cell and it is good to wait for 5-10 min for equilibration before starting recording. See Figure 3 for suggested recording protocols. Created in part with BioRender.com ll OPEN ACCESS STAR Protocols 1, 100139, December 18, 2020 a. Change perfusion of the slice from 35 mmHg CO 2 buffer to either 55 mmHg CO 2 buffer (to open hemichannels) or to 20 mmHg CO 2 buffer (to close hemichannels) i. Continue running current steps for at least 10-15 min ii. Once stable, carry out standard IV and inject naturistic current injections to measure firing rate. b. Optional: Change perfusion of slice back to 35 mmHg CO 2 buffer to attempt reversal of the effect. Note that this can be quite challenging. i. Run current steps for a further 10-15 min Note: Ensure that you save all SIV, firing rate traces, and sweep files of the current steps for analysis. A measure of the size of the voltage response can be taken from each step. Steps can be averaged easily in time-bins if required in Clampfit. SIV measurements can be used to measure resting membrane potential and input resistance and the voltage response to the naturalistic current injection can give a measure of the cells firing.
Note: In our study we chose to record in current clamp (recording the cell's membrane potential) as this allowed us to run the SIV protocol to validate cell type at the start of recording and also inject naturalistic current to look at firing. This protocol could also be used to record the current directly (in voltage clamp), although additional stimulation protocols will need to be used.

Dye Loading of Neurons through Hemichannel Opening
Timing: 5-6 h

. Suggested Stimulation Protocols
In our experiments we run three different current input protocols in order to fully assess the effects of the CO 2 sensitivity on neural function. (A) The first is a repeated hyperpolarizing step of À50 pA, for 100 ms, delivered every second. This is given continuously and can therefore provide a measure of the voltage response and input resistance over time.
(B) The second is a standard stepwise current voltage (IV) protocol, often referred to as a SIV protocol. The current steps to use and the voltage response will depend on the cell type that you are recording from. Here is an example from a CA1 hippocampal pyramidal neuron, where current steps from À300 pA increasing by 50 pA have been introduced. The protocol should be run until a regular firing pattern is induced. This can be used to assess cell identity and health and to measure resting membrane potential and input resistance (from the current steps around the membrane potential).
(C) The third current injection protocol (as used to produce a dynamic IV) injects a naturalistic current into the cell (top panel) and the voltage recorded (bottom panel) can be used to extract to determine the firing rate under a naturalistic state (for more detail, see Badel et al, 2008, Harrison et al, 2015used

OPEN ACCESS
This protocol outlines the fluorescent imaging of neurons that express CO 2 sensitive hemichannels by trapping a cell impermeant dye (which can move through open hemichannels). The protocol is simplified in Figure 4.
3. Preparing a slice for the dye loading of neurons a. Perfuse recording bath with the 35 mmHg CO 2 (control) buffer (bubbled with 95% O 2 /5% CO 2 ) and allow to reach recording temperature. b. Place brain slice in the bath and allow to equilibrate for at least 15 min. Researcher could make initial recordings from neurons or use landmarks to ensure that they are imaging the correct region of the slice.
Note: If performing dye-loading experiments following electrophysiology, you could potentially include in your pipette solution a different membrane impermeable dye so that you could identify with dye loading the neurons you recorded from and found to be sensitive to CO 2 changes.
4. Opening hemichannels a. The control buffer is exchanged for 55 mmHg CO 2 aCSF buffer (hypercapnic) containing 5(6)-carboxy-fluorescein (CBF, 100 mM) for 20 min to allow the CO 2 -sensitive hemichannels to open. 5. Closing hemichannels and trapping dye a. Solution is exchanged for 35 mmHg CO 2 buffer containing CBF (100 mM) for 5 min to allow the hemichannels to close, trapping the dye inside. b. Solution is exchanged to 35 mmHg CO 2 aCSF without the dye for at least for 3 h to reduce the background staining before imaging. 6. Imaging of filled cells a. Images are taken using a CCD camera with 488 nm fluorescence (CoolLED). As CBF rapidly bleaches, images need to be quickly acquired from regions of interest. CBF cannot be fixed using PFA (as it lacks the required groups for cross-linking).
CRITICAL: Do all the dye loading without the microscope light on to avoid bleaching. Prolonged washing of the dye is required for brain slices compared to use of cultured cells or Whole-cell patch clamp recordings can be used to confirm the identity of the target region of cells. Slices were then allowed to equilibrate for 20 min. The control aCSF was then exchanged for 55 mmHg CO 2 aCSF (hypercapnic) containing 5(6)-carboxy-fluorescein (CBF, 100 mM) for 20 min to allow the CO 2 sensitive-hemichannels to open. The solution was then exchanged for 35 mmHg CO 2 aCSF containing CBF (100 mM) for 5 min to allow the hemichannels to close. Finally, the slice was washed with 35 mmHg CO 2 aCSF for 3 h to reduce the background staining before imaging. The dye loading method is based on that described in Huckstepp et al (2010). Created with BioRender.com cell lines due to the thickness of the tissue. The dye rapidly bleaches so it is important to take the images rapidly.
Note: A suitable control is to not change the buffer from 35 mmHg CO 2 and show neurons do not dye load. It is also possible to block dye loading with carbenoxolone (100 mM; an indiscriminate hemichannel blocker) to further show that the effects are hemichannel mediated.
Note: We use CBF as we know it passes through connexin 26 hemichannels. It is not fixable with PFA as it lacks the required groups for cross-linking. There may be other dyes that can pass through the hemichannels and can be fixed. A trial and error approach would have to be used to discover such dyes.

Immunohistochemistry to Confirm Connexin Expression Profile
This protocol outlines how to confirm the connexin expression profile of your region of interest, as further evidence of the observed effects being mediated by CO 2 -sensitive hemichannels (in this case Cx26, see Figure 5).
Note: For the best results with immunohistochemistry, mice/ rats should be cardiac perfused 4% PFA and then post-fixed overnight at 4 C (18-24 h).
7. The slices are then washed 5 times in PBS. 8. They are then blocked for 1 h (1% BSA, 0.4% Triton X-100 in PBS, 400 mL per slice). 9. The slices are washed 5 times in PBS. 10. The primary antibodies: a marker for your cell of interest (at required concentration; Green 488) and an antibody against connexin 26 (1:200, Mouse; Red 594) are added to the slices (400 mL per slice) for 1 h at room temperature (20 C-22 C) on a shaker and then kept at 4 C-8 C overnight (12-16 h) static in the fridge. 11. Slices are washed five times for 5 min with PBS. 12. The corresponding secondary antibodies are then added (400 mL per slice) for 4 h at room temperature (20 C-22 C).
Note: Cover the plate in foil after addition of the secondary antibodies to avoid contact with light. Maintain all further steps in as little light as possible.
13. The slices are washed 5 times with PBS 14. Slices are mounted on glass slides with Vectashield (Vector laboratories, Peterborough UK) and stored in foil at 4 C until imaging.
Note: The earlier the imaging can be done following mounting the better. All our imaging was completed within 2 weeks of the initial immunohistochemistry to ensure the best staining.
Note: Any standardized protocol for imaging can be used. The immunohistochemistry protocol will need to be optimized for the antibodies that you are using and the cells of interest. We image with a Leica 880 confocal microscope and perform image acquisition and processing in the Zen black software suite.

EXPECTED OUTCOMES
This protocol describes a way in which you can look for a CO 2 -sensitivity phenotype among a specific group of neurons of interest across different areas of the brain. Electrophysiology provides the first evidence of a neuronal response to changing the level of CO 2 and as the solutions are kept at a constant pH (isohydric), the changes are completely independent of pH. Connexin 26 and 30 are known to be directly CO 2 sensitive and open in response to raised levels of CO 2 (Huckstepp et al, 2010; ll OPEN ACCESS Meigh et al, 2013). If they are expressed and form functional hemichannels in the cells of interest, then an increase of the CO 2 level would be expected to increase the opening probability of hemichannels and thus cause an increase in conductance (decrease in input resistance), which can be measured by whole-cell patch clamp. The firing rate of the neuron is also a good measure of change, as a drop in input resistance should lead to a decrease in excitability. Figure 6 outlines what the researcher would observe in two cases: firstly, where CO 2 sensitive hemichannels are present (Figure 6; top) and secondly where CO 2 sensitive hemichannels are not present (Figure 6; bottom). In the case of presence, there is a steady reduction in the size of the response to a 50 pA (100 ms) hyperpolarizing current step over time. The full time course can be seen on the left and examples before and after raised CO 2 of the voltage response to the step and SIV injection in the middle. This tells you that the neurons are responding to CO 2 . In the case of no CO 2 sensitive hemichannels present, there is no significant change to conductance or excitability over time.
Note: Always perform the relevant controls, suggestions of appropriate controls are detailed next. First, repeat the experiments but with no change in the level of CO 2 (just by switching solutions) to ensure it is not an artifact of this. Secondly, repeat the experiments with carbenoxolone (100 mM; an indiscriminate hemichannel blocker), if the effect is real and through hemichannels then it will be abolished in this case. Finally, as a control try a different unrelated region of the brain (that has not previously been shown to be CO 2 -insensitive). In our study we used the CA1 region of the hippocampus and showed no electrophysiological phenotype or dye loading.
Dye-loading experiments can then take you a step further to confirming that the changes are mediated by connexin 26. Here, you will use (6)-Carboxy-fluorescein (CBF), as it is known to dye load through connexin 26 in raised CO 2 . If the cells in your field of view have functional hemichannels then they will load with CBF ( Figure 6; top right). If they do not, then no dye loading will be observed ( Figure 6; bottom left).
Finally, immunohistochemistry can be performed to look for co-localization between the expression of CO 2 sensitive connexins (26 and 30) with your neuron / cell of interest by looking for co-localization.

QUANTIFICATION AND STATISTICAL ANALYSIS
All data processing and analysis can be performed in Clampfit which is available with the pClamp suite of software. All imaging was carried with confocal microscopy (Leica 710 and Zen Black for image acquisition and processing). Appropriate statistical tests were chosen based on sample size, whether there were repeated measures and whether the populations were paired or unpaired (Wilcoxon rank sum and Mann Whitney tests respectively). For tests of more than two variables, Kruskal-Wallis ANOVAs were run with Dunn's post hoc multiple comparisons. All tests were run to find significance at the level p < 0.05 and were performed on raw (non-normalized) data.

LIMITATIONS
This protocol provides a way to test for the presence of functional (pH independent) CO 2 -sensitive hemichannels in different cell populations. The combination of electrophysiology with dye loading and immunohistochemistry can provide you with a strong foundation of understanding of expression and functional consequence of CO 2 -sensitivity within your cells. There are three main limitations to the study, the first is that it does not provide a behavioral correlate to the sensitivity, as this was outside of the scope of this preliminary study. The second being the use of immunohistochemistry to study the co-localization of connexin 26 with the marker for your cell of interest. This is a qualitative rather than quantitative method and perhaps a method like FISH could also be used in order to give a more accurate measure of expression. The third limitation comes from the use of electrophysiology as a means of measurement. When recording for prolonged periods (>30 min), cells can sometimes be susceptible to run down. To address this, we performed our recordings within a Step and SIV responses in 35 (black) and 55 (red) mmHg CO 2 from a representative cell, displaying the decrease in input resistance and excitability upon hemichannel opening. (Right) Successful dye loading will appear as green filled neurons. The bottom row represents results from a group of neurons that do not express Cx26 as functional hemichannels. Here, there is no change to conductance or excitability and no successful dye loading. Confirming that there is no presence of functional CO 2 -sensitive hemichannels in these neurons. Reproduced in part using content previously published in Hill et al. (2020); iScience on a CC-BY 4.0 license. Scale bars, 50 mm. Data are presented as means G SEM.

OPEN ACCESS
relatively short time period of time. The changes in CO 2 level that we use are subtle (physiological) and should not therefore cause any cell swelling or lasting distress. However, it can be difficult to routinely get full recovery (Hill et al, 2020). To ensure this partial recovery was not an effect of the initial CO 2 change, our experiments have been performed in both directions -from 35 mmHg CO 2 to 55 mmHg CO 2 , and a small number starting in 55 mmHg CO 2 and changing back to 35 mmHg CO 2 . If raising the CO 2 was opening for example a KATP channel or resulting in cell swelling and this was responsible for the change in input resistance, then the reverse experiments would not have worked.
Note: Before starting our study, there was already a publication (Vandecasteele et al, 2006) highlighting the expression of connexin 26 in our region of interest using single cell PCR. If this data is not available for your region of your interest then you could consider running single cell PCR, western blot, or preliminary IHC to check for connexin expression profiles before undertaking the study to avoid wasting time.

TROUBLESHOOTING Problem
Over the period of recording, the flow of gases may change and hence lead to small changes in pH. It is vital that the experiments are isohydric so that any observed changes can be attributed to the shift in CO 2 concentration. If expected responses are not observed, the first thing to check is the pH of the solutions.

Potential Solution
Two strategies are available for this problem, the first is to regularly monitor the pH of the solutions throughout the day to ensure they remain the same. It is also possible to purchase pre-mixed gases for these experiments, although this is a more expensive option and limits the flexibility that you have in testing various different concentrations that is granted by mixing your own gases.

RESOURCE AVAILABILITY Lead Contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Emily Hill, E.Hill.2@warwick.ac.uk

Materials Availability
No new mouse lines or reagents were generated for this study

Data and Code Availability
This study did not generate new code or structural datasets