All procedures involving animals were approved by and performed in accordance with the rules and recommendations of the Institutional Animal Care and Use Committee of the University at Buffalo.

The generation and genotyping of the KO_b/c mouse on a C57BL/6J background have been previously reported (Salerno et al., 2019). For this study, heterozygous parents were produced by backcrossing heterozygous mice with verified C57BL/6J wild-type mice (Jackson Laboratory, Bar Harbor, ME). Heterozygous progeny (F6 to F17 generation, making them at least 99% genetically identical) were crossed to produce experimental WT and KO_b/c mice for this study. As previously reported, KO_b/c mice exhibit increased mortality [50% at 5 weeks (Salerno et al., 2019)] so we were constrained to working with mice between 4–5 weeks of age in order to maximize the likelihood of survival during study, while still using mice past the age of full structural and functional kidney development (Satlin et al., 2003; Seely, 2017).

WT and KO_b/c mice between 4-5-weeks old were housed in metabolic chambers (Tecniplast, 3600M021) and given standard powered rodent chow (Teklad 8604). After a 1-day acclimation period with plain tap water, experimental groups were subjected to 1–3 days (duration depending on the experiment) of control or MAc-challenged conditions. MAc was induced by adding 0.28 M ammonium chloride (NH₄Cl) + 0.5% sucrose (for palatability) in drinking water (tap water). Control groups were given drinking water containing only 0.5% sucrose for the same 1–3-day duration. NH₄Cl administration is a common method for induction of MAc and enhancement of ammoniagenesis in both humans and animals (Rector et al., 1955; Amlal et al., 2004; Alam et al., 2020). All mice had free access to powdered standard rodent chow for the duration of the experiment. During the acclimation period, body weight, food intake, and fluid intake were monitored. During the experimental period, urine and fecal excretion were also monitored, and urine was collected each day under mineral oil (to prevent evaporation). Urine samples were spun at 10,000 RCF for 5 min to remove any solid particles and frozen at −80°C until urinalysis. Since repeated measures of metabolic and urinary parameters were possible in mice subjected to experiments of 2-, and 3-day durations before sacrifice for plasma electrolyte analysis, this led to a larger sample size for these parameters for day 1, compared to day 2 or day 3. Sample sizes for mice kept under 1, 2, or 3 days of control or MAc-challenged conditions are shown in Table 1, which relates to the data presented in Table 2 and Figure 7.

The breathing activity of unrestrained mice was measured without anesthesia as described previously (Morse et al., 2012). Individual animals were kept in a plexiglass plethysmography chamber connected to an empty reference chamber for buffering changes in atmospheric pressure (PLY4213, Buxco Research Systems, Wilmington, NC). The plethysmograph was connected to a Rodent Bias Flow Supply (BFL0250) to draw expired CO₂ out of the chamber and provide a constant flow of room air at a flow rate of 2.5 L/min. Signals were collected with a barometric pressure sensor attached to a port on the reference chamber. The sensor was connected to a MAX 1500 preamplifier, and signals were visualized and collected using BioSystem XA software. The chamber and software were calibrated for mice at the beginning of each experiment using an injection of 1 mL of air through the base port of the chamber. To limit interference, exclusion parameters were set as follows: Max expiration time–0.3 s, Min Tidal volume–0.02 mL, Min Inspiration time–0.03 s. Recordings were taken once mice settled into a period of quiet breathing, after a minimum 30-min acclimation period. Each recording consisted of 15 min, during which tidal volume, frequency, and minute volume were averaged over 5 s intervals. Periods of activity or sniffing were noted and excluded from analysis. Tidal volume and minute volume were normalized to body-weight. Baseline parameters were gathered by averaging data from 2 consecutive days of mice housed in unchallenged conditions, prior to each mouse being given a 3-day MAc challenge.

To obtain repeated daily blood pCO₂ and pH measurements in mice without anesthesia, mice were restrained using 50 mL conical tube, modified by cutting a breathing hole at the conical end and a hole for tail access through the screw cap. After allowing 5 min for the mice to acclimate, an incision was made in the lateral tail vein using a 18G needle. 6–10 µL of blood was drawn from the incision using a P10 pipette and immediately transferred to a 1.5 mL Eppendorf containing 25 µL of prewarmed (37°C) mineral oil to prevent degassing of the blood sample. pCO₂ was immediately measured using a micro-carbon dioxide electrode (MI-720; Microelectrodes Inc, Bedford NH), that had been calibrated in 1L CO₂/HCO₃ ⁻-containing solutions gassed to 3% (22.8 mmHg), 5% (38 mmHg), or 8% (60.8 mmHg) CO₂ concentrations at pH 7.4 and maintained at 37°C (in mM: 5 HEPES, 10 glucose, 5 KCl, 0.8 MgCl₂, 1.35 CaCl₂, [127, 119, or 106] NaCl, [13, 21, or 34] NaHCO₃; the varying NaCl and NaHCO₃ concentrations represent those used for the 3%, 5%, and 8% CO₂ solutions, respectively). Following measurement of pCO₂, pH was immediately measured using a micro-pH electrode (Orion 9810BN, Thermo Fisher Scientific). Calculation of [HCO₃ ⁻] was derived from the Henderson-Hasselbach equation ( p H = 6.1 + log   H C O 3 − 0.03 ∗ p C O 2 ). Specifically [HCO₃ ⁻] = 0.03 x (pCO₂ in mmHg) x 10^(pH–6.125).

Mice were anesthetized with isoflurane (5%, inhaled) and blood drawn by cardiac puncture was immediately transferred into a blood metabolite test card and analyzed using an Epoc reader according to the manufacturer’s instructions (both Siemens Medical Solutions United States Inc., Malvern, PA).

24-h urine collections obtained from mice housed in metabolic cages for 1–3 days (see ‘Metabolic Studies’ above) were used to assess daily NH₄ ⁺ excretion, daily TA excretion, and daily urine pH. Since mice were sacrificed after each experimental duration for electrolyte analysis, there are larger sample sizes for day 1 than for day 2 or day 3 within the urinalysis data set (Figure 7). Urine NH₄ ⁺ concentrations were assessed using a commercially available kit (AA0100, Sigma-Aldrich, St. Louis, MI) that was modified for use in 96-well plates, with samples diluted 1:100 and ran in triplicate. Daily NH₄ ⁺ excretion rates were calculated using 24-h urine NH₄ ⁺ concentrations and volumes. Titratable acid (TA) content was measured as described previously (Chan, 1972; Lee et al., 2009). Briefly, 25 µL of 0.1 M HCl was added to 25 µL of urine, boiled for 2 min, and let cooled for 1 min. The volume of 0.4 M NaOH needed to bring the sample pH to 7.4 was quantified. Samples of distilled (DI) water were run in parallel. The total number of moles needed to bring samples to pH 7.4, less the moles needed to bring the DI water to pH 7.4, was multiplied by 24-h urine volumes to yield daily TA excretion. Urine pH was measured using a micro-pH electrode (Orion 9810BN, Thermo Fisher Scientific).

The Total-NBCe1 antibody (affinity for all NBCe1 protein variants) was purchased from Elabscience Biotechnology, Inc. (E-AB-14348, Houston TX; raised in rabbit), and has been previously validated (Salerno et al., 2019). The Total-NBCe1 antibody was used at a 1:1,000 concentration for Western blot. The NBCe1-A specific antibody generously gifted by Dr. Michael Romero (Mayo Clinic, Rochester MN; raised in chicken) has been previously validated (Lee et al., 2018), and was used at a 1:1,000 concentration for Western blot. The NBCe1-B/-C antibody was purchased from Santa Cruz Biotechnology, Inc. (sc-515543, Dallas TX; raised in mouse), and was used at a 1:1,000 concentration for Western blot and 1:100 dilution for immunohistochemistry (IHC). The NBCe1-B/C antibody was raised against an epitope that corresponds to the first 41 amino acids of the unique amino-terminal sequence of NBCe1-B/C (MEDEAVLDRGASFLKHVCDEEEVEGHHTIYIGVHVPKSYRR). The Phox2B antibody was purchased from Thermo Fisher (PA5-115754; raised in rabbit), and was used at a 1:500 concentration for IHC. The glutamine synthetase (GS) antibody was purchased from Abcam (ab73593; raised in rabbit), and was used at a 1:20,000 concentration for IHC. The phosphoenolpyruvate carboxykinase (PEPCK) antibody was purchased from Cayman Chemicals (10004943; raised in rabbit), and was used at a 1:10,000 dilution for IHC.

The secondary HRP-conjugated anti-chicken immunoglobulin antibody was purchased from Thermo Fischer Scientific (A16054; raised in goat), and used at a 1:2,000 dilution for Western blot and 1:1,000 dilution for IHC. The secondary HRP-conjugated anti-mouse immunoglobulin antibody was purchased from MP Biomedicals (55563, Solon OH; raised in goat), and used at a 1:2,000 dilution for both Western blot and 1:1,000 dilution for IHC. The secondary HRP-conjugated anti-rabbit immunoglobulin antibody was purchased from MP Biomedicals (55685, Solon OH; raised in goat), and used at a 1:1,000 dilution for IHC.

Oocytes were extracted from female Xenopus laevis frogs (Xenopus Express, Brooksville, FL) as described elsewhere (Musa-Aziz et al., 2010). In brief, ovaries of anesthetized frogs (0.2% tricaine) were surgically removed and washed in Ca²⁺-free “NRS” buffer (in mM: 82 NaCl, 2 KCl, 20 MgCl₂, 5 HEPES, adjusted to pH 7.45 with NaOH). Frogs were then euthanized via exsanguination. Oocytes were extracted from ovaries by digestion in NRS buffer plus 2 mg/mL type-IA collagenase (C2674: Sigma-Aldrich, St. Louis, MO) for 15–35 min, and washed in NRS buffer, followed by ND96 solution, and finally with OR3 medium. OR3 medium contains 14 g/L of Leibovitz’s L-15 medium (10–045-CV, Thermo Fisher Scientific), 100 U/mL penicillin, 100 μg/mL streptomycin, 5 mM HEPES, and adjusted to pH 7.5 using NaOH. The osmolality of OR3, measured using a Vapro vapor pressure osmometer (Wescor, Logan, UT), was adjusted to 195 ± 5 mOsmol/kg osmolality with H₂O.

The construction of the plasmids for expression of WT human NBCe1-A or NBCe1-B in Xenopus oocytes has been described previously (Lee et al., 2011; Myers et al., 2016b). Each construct included a carboxy-terminal enhanced green fluorescent protein (EGFP) tag (NBCe1-A-EGFP.pGH19 and NBCe1-B-EGFP.pGH19) used to confirm expression in oocytes before homogenization. Plasmids were transformed into E. coli and cultured overnight. After isolation, plasmids were linearized using NotI, followed by purification using the MinElute PCR Purification Kit (28004, Qiagen). Linearized DNA was transcribed to capped cRNA using the T7 mMessage mMachine Transcription kit (Invitrogen, Carlsbad, CA). cRNA was further purified with the RNeasy MinElute Cleanup kit (74204, Qiagen) before quantification using a Nanodrop 2000 (Thermo Fisher Scientific). A 25 nL bolus of cRNA (or water for negative control) was injected into oocytes using a Nanoject III programmable nanoliter injector (Drummond Scientific Co., Broomall, PA).

Xenopus oocytes: Oocytes were prepared as previously described (Myers et al., 2016a). In brief, cells were incubated for 3 days in OR3 medium to allow for protein expression and cells were homogenized in TBS buffer containing 1% Triton X-100 and Complete Protease Inhibitor Cocktail. Yolk platelets and other insoluble components were pelleted by low-speed centrifugation: 850x RCF for 5 min. The supernatant was mixed with gel-loading buffer and the equivalent of ¼ an oocyte was loaded into each well for Western blotting.

Mouse kidneys: Excised kidneys were placed into of ice-cold homogenization buffer (HB; in mM: 100 NaCl, 25 HEPES, 250 sucrose, pH 7.4) plus cOmplete Protease Inhibitor Cocktail (Pierce A32963, Thermo Fisher Scientific). For separation of cortical and medullary protein samples, the cortex and medulla of freshly excised kidneys were separated under a dissecting microscope, guided by the contrast in color of the two zones, prior to homogenization. Whole kidney, cortical, or medullary membrane protein homogenates were prepared from dissected tissue using fractional centrifugation techniques. Briefly, tissue was homogenized with a handheld homogenizer (D1000, Benchmark Scientific, Edison, NJ), resuspended in 5 mL of HB, and spun at 1075 RCF for 15 min with the resulting supernatant retained. Membrane fragments were precipitated from the supernatant by ultracenxtrifugation using a Beckman Optima L-70 Ultracentrifuge (Beckman Coulter Inc., Indianapolis, IN) at 204,300x RCF for 45 min. The resulting pellet was resuspended in 300 µL HB. Homogenate protein concentrations were determined using the BCA colorimetric assay (Pierce 23227, Thermo Fisher Scientific), modified for microplate conditions. Immediately before gel loading, samples were pre-treated with a 0.1 M dithiothreitol (DTT) solution and denatured with LDS (Invitrogen NP0007, Thermo Fisher Scientific). DTT treatment is necessary to remove interfering endogenous mouse immunoglobulin for NBCe1-B probed kidney blots since the secondary antibody recognizes mouse immunoglobulin which, when intact, interferes with the measurement of NBCe1-B abundance (Brady; unpublished observations). DTT reduces the disulfide bonds of intact immunoglobulin resulting in lower molecular weight subunits that do not interfere with the NBCe1 band (Ahmad-Zadeh et al., 1971). DTT treatment is also known to dissociate NBCe1 dimers (Kao et al., 2008), resulting in only bands representing NBCe1 monomer abundance. Protein (10 µg/lane, unless specified) was resolved on a 3%–8% Tris-Acetate gel (Invitrogen EA0378BOX, Thermo Fisher Scientific) and transferred onto a PVDF membrane (Invitrogen LC 2005, Thermo Fisher Scientific). For studies of NBCe1-A and NBCe1-B abundance, age- and sex-matched littermates were paired between control and experimental groups, and each sample was run in duplicate on the same gel as its pair. To confirm even protein loading, prior to antibody application, each blot was treated with the reversible Memcode total-protein stain and blots with clearly unequal loading were discarded (Pierce 24585, Thermo Fisher Scientific). Additionally, Memcode stained blots were imaged using the visible light setting of a myECL imager (Thermo Fisher Scientific), and used to normalize differences in protein between lanes (Janes, 2015). After the total protein was imaged and the stain was erased, the PVDF membrane was incubated overnight in Tris-buffered saline (TBS; in mM: 10 HCl, 150 NaCl) containing 0.1% Tween-20 (0.1% TBS-T) and 4% milk powder at 4°C. The next day, the membrane was probed with primary antibody diluted in 2% milk powder prepared in 0.1% TBS-T, followed by an HRP-conjugated secondary antibody diluted similarly. Immunoreactive protein bands were visualized and imaged using ECL2 reagent (Pierce 32106, Thermo Fisher Scientific) and the chemiluminescent signal was imaged using the myECL imager after confirming the absence of pixel saturation. Densitometry was performed using FIJI software on both total-protein stained and antibody probed blots (Schindelin et al., 2012; Janes, 2015). The images of the total-protein stained blots were background subtracted using the FIJI rolling ball subtraction algorithm with a 100-pixel radius. Total-protein was quantified using densitometry of the entire lane profile plotted from rectangles drawn around each lane of the blot. Individual band densitometry results from the antibody-probed blots were normalized to their respective total-protein results and these normalized ratios were subjected to statistical analysis.

Kidneys and brainstems were excised post-euthanasia and immediately placed in 4%-paraformaldehyde for 24-h at 4°C (kidneys cut transversely in half). Tissue was then stored in a 70% ethanol solution at 4°C until embedding. Before paraffin embedding of the brain, a transverse cut was made at the level of the rostral medulla/caudal pons, suggested as the optimal location for obtaining sections containing the RTN (Lavezzi et al., 2019). Tissue was embedded in paraffin blocks using standard embedding procedures. Briefly, tissue was dehydrated through incubations in 80% and 95% ethanol, 45 min each, and followed by 3 changes of 100% ethanol, 1 h each. Tissue was cleared through 2 changes of xylene, 1 h each, and placed in molten paraffin overnight (H-PF, General Data, Cincinnati, OH). Tissue was then embedded, cut side down, in paraffin blocks. Tissue was sectioned at a thickness of 5 μm, mounted on frosted slides, and dried at 37°C overnight.

Standard immunoperoxidase procedures were used for chromogenic immunostaining. Briefly, sections were deparaffinized in xylene, rehydrated in decreasing ethanol concentrations, and rinsed in cool tap water. Sections were incubated in a Tris/EDTA solution (in mM: 10 Tris-base, 1 EDTA; pH 9–9.1) at 95°C for 40 min and let cooled at room temperature for 20 min. Endogenous peroxidase activity was blocked using Peroxidase Suppressor (35000, Thermo Fisher Scientific) for 10 min. Sections were rinsed 2x for 3 min in TBS with 0.05% Tween 20 (0.05% TBS-T). The sections were blocked with Rodent Block M (for NBCe1-B and Phox2B staining; Biocare Medical, Pacheco, CA) or 10% normal goat serum (for PEPCK and GS staining; Thermo Fisher Scientific) for 30 min. For PEPCK and GS staining, sections were additionally permeabilized in 0.5% Triton X-100 prepared in TBS for 15 min. Sections were incubated overnight in a humidified chamber at 4°C with primary antibody dilutions prepared in a TBS solution with 5% BSA (5% TBS-BSA). The next day, sections were washed 3x 5-min with 0.05% TBS-T. Sections were incubated with HRP-conjugated secondary antibody diluted in 5% TBS-BSA for 1 h at room temperature, and again washed 3x 5-min with 0.05% TBS-T. Sections were then exposed to diaminobenzidine (DAB) for 10 min. To counterstain nuclei, sections were rinsed with DI water and stained with hematoxylin (Abcam; hematoxylin ab220365) for 1 min and rinsed with DI water. All sections were dehydrated with graded ethanol solutions and xylene, and mounted for light microscopy. Comparisons of labeling were made only between sections from the same immunohistochemistry experiment, treated with identical reagents. In some replicates, sections were included that were only treated with primary or secondary antibodies as internal controls. Images were taken with a Leica DM 6B upright microscope with identical capture settings between slides from the same experiment. For analysis of NBCe1-B/C expression intensity in brainstem IHC images in Figure 2, images were converted to 8-bit greyscale images and inverted, and the average pixel intensity was measured within a region of interest (ROI) containing the tissue section using FIJI. Similar measurements were made for a background ROI, which was subtracted from the tissue section ROI measurement. For quantification of PEPCK and GS expression in Figures 8, 9, high magnified images were taken across the cortical and OSOM regions (5-images per region). Individual PTs were manually traced (identified by the presence of a prominent apical brush border; in contrast, distal tubules have sparse microvilli, which makes the lumen appear to be wider), and images were converted to 8-bit greyscale images and inverted. A 30–255-bit threshold was applied to all images in order to limit inclusion of background pixels, and the mean pixel intensity of selected tubules was measured. The average pixel intensity across all 5 images within the cortical or OSOM regions was normalized to the average in the WT cortex.

Results are presented as mean ± SEM, with n referring to the number of animals studied. In all analyses the threshold of p < 0.05 was used to determine statistical significance. Normality of data was tested in GraphPad (v9.0) using the D'Agostino-Pearson omnibus test and statistical comparisons between two groups (i.e., WT and KO_b/c mice) with normal distributions were performed using Student’s two-tailed unpaired t-test. Non-normal distributions were only observed in the control NH₄ ⁺ excretion data set (Figure 7A, left) and thus for these comparisons only, the Mann-Whitney test (2-tailed, unpaired) was used. p-values were Bonferroni corrected for multiple comparisons when appropriate. For plethysmography and blood gas experiments (Figures 5, 6), in which repeated measures in individual mice were possible, repeated measures ANOVA (RM-ANOVA) was used to assess the response of WT and KO_b/c groups to MAc, while comparisons between WT and KO_b/c mice at each timepoint was assessed using Student’s unpaired two-tailed t-test. Since in these experiments the percent change from baseline was also calculated, animals with outliers (defined as+/-2 SD from the mean) in the baseline (timepoint “0”) data set only, were excluded from analysis; which equated to 1 animal from the plethysmography data set, 1 animal from the pCO₂ data set, and 2 animals from the plasma pH data set being excluded. Conversely, assessment of plasma electrolytes required sacrificing mice after 1-, 2-, or 3-days of control or MAc-challenged conditions; thus each day represents a different cohort of mice. However, in these same experiments, metabolic and urinary acid-excretion parameters were recorded daily, leading to sequentially larger sample sizes for day 1, then day 2 or day 3. Thus, comparison between WT and KO_b/c metabolic and urinary acid-parameters over the 3-day experimental time course were assessed by 2-way ANOVA, using genotype and time as independent variables, with the main effect of genotype specifically reported (“G”, Table 2 and Figure 7). For urinary parameters in Figure 7, the genotype x time (“G*T”) interaction effect is also reported and comparison of WT and KO_b/c responses were assessed at each timepoint using Student’s unpaired two-tailed t-test. Both male and female mice were used non-discriminately throughout the study based on availability at the time of each experiment. Inclusion of sex as an independent factor in ANOVA was used to assess for a significant effect of sex; if a significant sex interaction was found this is reported specifically. All analyses were performed in Prism GraphPad version 9 or IBM SPSS. Figures were prepared using Microsoft PowerPoint, Microsoft Excel, BioRender.com, and Prism GraphPad version 9.

Protein homogenates prepared from NBCe1-A-EGFP cRNA-, NBCe1-B-EGFP cRNA-, or water- injected oocytes were loaded on a single gel, in triplicate. After protein transfer, the PVDF membrane was cut into thirds, and antibodies (Abs) specific to Total-NBCe1 (affinity for all NBCe1 variants), NBCe1-A, or NBCe1-B/C were used to individually probe the three membrane sections (Figure 1A). The anti-Total-NBCe1-Ab-probed blots reported bands that are consistent with the molecular weights of NBCe1-EGFP monomers and dimers (Parker et al., 2012) in both the NBCe1-A-EGFP and NBCe1-B-EGFP-loaded lanes (Figure 1A, left). The anti-NBCe1-A and anti-NBCe1-B/C-Ab-probed blots only reported bands in the lanes loaded with NBCe1-A-EGFP or NBCe1-B-EGFP protein, respectively (Figure 1A, center and right).

To assess the validity of the NBCe1-B/C specific antibody in mouse kidney preparations (where NBCe1-A and NBCe1-B are expressed, but NBCe1-C is not), we probed for NBCe1-B in WT and KO_b/c kidney protein homogenates treated with 0.1M dithiothreitol (DTT; see ‘Methods–Western blotting’ for rationale; Figure 1B). We detect immunoreactivity consistent with monomeric NBCe1-B in homogenates from WT but not from KO_b/c mice. Thus the antibody is appropriate for specific detection of NBCe1-B in the kidney.

Before testing the hypothesis that NBCe1-B/C has a role in the respiratory response to acidosis we first confirmed the loss of NBCe1-B/C in the brainstem medulla of KO_b/c mice. Sections were cut from the rostral medulla of the brainstem, starting from the dotted line illustrated in Figure 2A and moving rostrally, which is suggested to be the optimal sampling location for sampling the RTN (Lavezzi et al., 2019). The images in Figure 2B demonstrate Phox2B expression (yellow arrowheads), a transcription factor expressed in RTN neurons (Stornetta et al., 2006; Dubreuil et al., 2008; Ruffault et al., 2015), in both WT (top) and KO_b/c (bottom) medullary brainstem sections. Thus, Phox2B- expressing neurons appear intact in the brainstem medulla of KO_b/c mice. Figure 2C shows representative images of medullary brainstem sections from WT (top) and KO_b/c (bottom) mice probed with the NBCe1-B/C antibody, showing absence of NBCe1-B/C immunoreactivity in brainstem medulla of KO_b/c mice. Figure 2D summarizes the results of three experimental replicates in which the intensity of NBCe1-B/C immunolabel was quantified in identically treated/imaged WT and KO_b/c sections, with the percent of NBCe1-B/C expression in KO_b/c sections calculated relative to WT. Similar calculations were made for negative control sections in which either the primary antibody or the secondary antibody was excluded (−1° and −2°, respectively; images not shown). The chromogenic signal intensity in KO_b/c sections treated with the NBCe1-B/C antibody was significantly less than in WT sections (p = 0.015, Bonferroni corrected), and was not significantly different from either of the negative controls, confirming the absence of majority NBCe1-B/C expression in the brainstem of KO_b/c mice.

The original characterization of NBCe1-B expression in the kidney was in the context of the congenitally acidemic NBCe1-A KO mouse, and since NBCe1-B expression is controlled by an acid-sensitive promoter (Snead et al., 2011), the NBCe1-B expression described in NBCe1-A KO mice may not be representative of expression in WT mice. Therefore, we aimed to determine the expression pattern and abundance response of NBCe1-B in WT mice at baseline and during MAc. We also investigated the abundance response of NBCe1-A during MAc in both WT and KO_b/c mice.

Using the NBCe1-A and NBCe1-B/C specific antibodies described in Figure 1, we first assessed the abundance of NBCe1-A and NBCe1-B in WT cortical (cor) and medullary (med) protein preparations by Western blot (Figures 3A, B). Although the NBCe1-B/C specific antibody could also recognize NBCe1-C, only NBCe1-B is expressed in the kidney (Fang et al., 2018). Since these preparations contained less protein compared to whole kidney lysates, in this experiment two µg of protein was loaded per lane. The Memcode total protein stain was used to normalize protein loading/transfer among lanes (data not shown). We found, as expected, that NBCe1-A abundance is significantly greater in the cortex than medulla (p = 0.002; Figure 3C, left). On the other hand, NBCe1-B abundance was not significantly different between the cortex and medulla (Figure 3C, right). To determine the specific location of NBCe1-B expression in WT kidneys we used immunohistochemistry (Figures 3D–F, n = 3 replicates for each). In kidney sections from control WT mice, we observed NBCe1-B immunoreactivity on the basolateral membrane of PTs located in cortical medullary rays, as well as in PTs of the OSOM (Figure 3D; yellow outlined arrowheads indicate representative tubules that have positive basolateral immunoreactivity). After 3-days of MAc, we qualitatively observed a greater intensity of basolateral NBCe1-B staining in the PTs of the cortical medullary ray and OSOM (Figure 3E). As a control, basolateral NBCe1-B immunoreactivity was absent in KO_b/c mice even after 3 days of MAc (Figure 3F, n = 3 replicates). Together, these data indicate that NBCe1-B in WT mice is expressed primarily in the PTs of the cortical medullary ray and OSOM, with no obvious expression detectable in PTs of the cortical labyrinth, even after 3-days of MAc-challenged conditions.

Next, we assessed the change in protein abundance of both NBCe1-A and NBCe1-B variants in the kidneys of WT mice, as well as the change in NBCe1-A kidney abundance in KO_b/c mice, during MAc. In these experiments, sex-matched littermates were paired, with one subjected to control conditions (con) and the other to MAc-challenged conditions for 1 or 3 days. For Western blotting, 10 µg of kidney lysate protein was loaded per lane. Prior to antibody application each blot was treated with the reversible Memcode total-protein stain in order to provide an index for normalizing protein loading/transfer among lanes (data not shown). The WT membrane was then cut, and probed with either the NBCe1-A or NBCe1-B specific antibody, while the KO_b/c membrane was probed with just the NBCe1-A specific antibody. Figure 4A shows a representative Western blot from a 3-day experiment demonstrating NBCe1 monomer immunoreactivity. NBCe1 abundance (from Western blot) was normalized to total-protein abundance (from Memcode stain) and the ratio of this normalized abundance was calculated between paired MAc-challenged and control littermates; thus, each point in Figure 4B represents a single MAc/control pair (i.e., control = 1, represented by the dotted line in Figure 4B). In WT mice, after 1 or 3 days of MAc, NBCe1-A abundance was not significantly different from control (Figure 4B). On the other hand, in the same WT mice, NBCe1-B abundance was significantly increased after both 1- and 3-days of MAc (1-day ratio, p = 0.033; 3-day ratio, p < 0.001; Figure 4B). Finally, similar to WT mice, there was no significant change in NBCe1-A abundance in KO_b/c mice after 1- or 3-days of MAc (Figure 4B). Overall, this data indicates that NBCe1-B is expressed in the PTs of the medullary and OSOM in WT mice and that the abundance of NBCe1-B, but not NBCe1-A, increases during MAc.

Next, we assessed the ventilation of WT and KO_b/c mice prior to, and during each day of, a 3-day MAc-challenge using unrestrained whole-body plethysmography. Ventilation parameters (minute volume, tidal volume, and frequency) of WT and KO_b/c mice were assessed for 2 days under unchallenged conditions and averaged to establish a baseline (Figures 5A–F, timepoint “0”), followed by repeated measures after each 24-h period of a 3-day MAc-challenge (Figures 5A–F, timepoints “1–3”). Minute volume and tidal volume were corrected for body-weight in individual mice. Figures 5A,C,E show the average values for WT and KO_b/c groups at each timepoint. At baseline, KO_b/c mice had a significantly higher minute volume than WT (p = 0.019; Figure 5A, timepoint 0) due to a significantly greater baseline tidal volume (p = 0.038; Figure 5C, timepoint 0), while there was no significant difference in baseline frequency (Figure 5E, timepoint 0). There were no significant differences in the values of these parameters for the duration of the 3-day MAc-challenge (Figures 3A,C,E, timepoints 1–3).

Since there were significant baseline differences in ventilation parameters between WT and KO_b/c mice, we assessed the respiratory response of WT and KO_b/c mice to MAc by calculating the percent change from baseline of each ventilatory parameter during the 3-day MAc-challenge (Figures 3B,D,F). MAc had a significant effect on WT minute volume (p < 0.001 by RM-ANOVA; Figure 3B) but had no significant effect on KO_b/c minute volume (p = 0.227 by RM-ANOVA; Figure 3B). Furthermore, after the first day of MAc, pairwise comparisons demonstrate a significantly greater percent increase in minute volume in WT mice compared to KO_b/c mice (p = 0.026; Figure 5B, timepoint 1). MAc had a significant effect on tidal volume in both WT and KO_b/c mice (WT: p < 0.001, KO_b/c: p = 0.007 by RM-ANOVA; Figure 5D), however WT mice had a significantly greater percent increase in tidal volume compared to KO_b/c mice after the first day of MAc (p = 0.010; Figure 5D, timepoint 1). Lastly, MAc had a significant effect on respiratory frequency in both WT and KO_b/c mice (WT: p < 0.001, KO_b/c: p = 0.007 by RM-ANOVA; Figure 5F), but there were no significant differences in the percent change in frequency between WT and KO_b/c mice at any timepoint (Figure 5F). Overall, the absence of an increase in minute volume and diminished increase in tidal volume after 1 day of MAc in KO_b/c mice is consistent with the hypothesis that the respiratory response to MAc is impaired in KO_b/c mice.

To determine the effect of NBCe1-B/C loss on the defense of acid-base status during MAc, we assessed the pCO₂, pH, and [HCO₃ ⁻] of WT and KO_b/c mice at baseline (Figures 6A–F, timepoint “0”) followed by repeated measures after each 24-h period of a 3-day MAc-challenge (Figures 6A–F, timepoints “1–3”). Figures 6A,C,E show the average values for WT and KO_b/c groups at each timepoint. At baseline, KO_b/c mice had a significantly lower pCO₂ than WT mice (p = 0.046; Figure 6A, timepoint 0), but had no significant difference from WT in baseline pH or baseline [HCO₃ ⁻] (Figures 6C, E; timepoint 0). During the 3-day MAc-challenge, there were no significant differences between WT and KO_b/c pCO₂ or [HCO₃ ⁻] at any timepoint (Figures 6A–E, timepoints 1–3), however KO_b/c pH was significantly lower than WT after the first day of MAc (p = 0.038; Figure 6C, timepoint 1).

Since there was a baseline difference in pCO₂ between WT and KO_b/c mice, we assessed the recovery of acid-base status in WT and KO_b/c mice during MAc by calculating the percent change from baseline in each parameter during the 3-day MAc-challenge (Figures 6B,D,F). For plasma pH, we transformed all values to hydrogen ion concentration ([H⁺]) in order to directly calculate the percent change in acidity from baseline (Figure 6D). MAc had a significant effect on pCO₂ in WT mice (p = 0.011 by RM-ANOVA; Figure 6B) but had no effect on pCO₂ in KO_b/c mice (p = 0.889 by RM-ANOVA; Figure 6B). Furthermore, after 1-day of MAc, pairwise comparisons demonstrate a significantly greater percent decrease in pCO₂ in WT mice compared to KO_b/c mice (p = 0.029; Figure 6B, timepoint 1). MAc had a significant effect on [H⁺] in both WT and KO_b/c mice (WT: p = 0.039, KO_b/c: p < 0.001 by RM-ANOVA; Figure 6D), however the increase in [H⁺] after the first day of MAc was significantly greater in KO_b/c mice than in WT mice (p = 0.005; Figure 6D, timepoint 1). Lastly, MAc had a significant effect on [HCO₃ ⁻] in both WT and KO_b/c mice (WT: p < 0.001, KO_b/c: p < 0.001 by RM-ANOVA; Figure 6F), and there were no significant differences in the percent change from baseline [HCO₃ ⁻] between WT and KO_b/c mice at any timepoint (Figure 6F). In summary, the lack of a decrease in pCO₂ in KO_b/c mice appears to underlie a greater fall in plasma pH after 1-day of MAc but does not inhibit the recovery of pH after 2- and 3-days of MAc.

Since we observed a recovery of plasma pH in KO_b/c mice to a level similar to that in WT mice, despite KO_b/c mice having an impaired respiratory response (Figure 5) and more severe acidemia (Figure 6) after 1 day of MAc, we hypothesized that renal acid-excretion is enhanced in KO_b/c mice during the MAc-challenge. To test this hypothesis, we subjected WT and KO_b/c mice to control or MAc-challenged conditions for 1, 2, or 3 days while housed in metabolic cages and assessed daily 24-h urine collections for NH₄ ⁺ excretion, titratable acid (TA) excretion, and pH. Additionally, mice were sacrificed after each timepoint via cardiac puncture for assessment of electrolytes. As expected, mice subjected to MAc-challenged conditions exhibited significantly greater NH₄ ⁺ and TA excretion, and significantly lower urinary pH than animals of the same genotype under control conditions (p < 0.001 for all three parameters in both genotypes by ANOVA). Acid-excretion parameters were compared between WT and KO_b/c groups using 2-way ANOVA with genotype and time as independent variables. A significant main effect of genotype (“G”) indicates a significant difference between overall group means (i.e., average daily NH₄ ⁺ excretion between WT and KO_b/c mice) and a significant genotype x time (“G*T”) interaction effect indicates that the difference between genotypes depends on time. Thus, pairwise comparisons between WT and KO_b/c groups were assessed at each timepoint; for transparency, the significance of pairwise comparisons are reported for all parameters regardless of G*T significance.

Under control conditions, there was no significant difference between WT and KO_b/c daily average NH₄ ⁺ excretion (Figure 7A, control). Under MAc-conditions, there was a significant genotype x time (G*T) interaction effect (p = 0.049 ANOVA), indicating a possible difference in the time course of WT and KO_b/c NH₄ ⁺ excretion. Indeed, on day 2 of MAc, KO_b/c mice excreted significantly more NH₄ ⁺ than WT mice (p = 0.018; Figure 7A, MAc, timepoint 2). To determine if this peak in KO_b/c NH₄ ⁺ excretion over day 2 was a result of increased NH₄ ⁺ excretion in individual mice, we calculated the change in NH₄ ⁺ excretion that occurred in each animal between days 1 and 2 (ΔNH₄ ⁺ day_1-2, Figure 7B). Under control conditions there was no significant difference in ΔNH₄ ⁺ day_1-2 (Figure 7B, control). However, under MAc conditions, KO_b/c mice had a significantly greater ΔNH₄ ⁺ day_1-2 than WT mice (p = 0.046; Figure 7B, MAc).

Along with increased NH₄ ⁺ excretion, increased TA excretion and urine acidification are also expected during MAc; thus, we similarly assessed TA excretion and urine pH in WT and KO_b/c mice (Figures 7C–F). Under control and MAc conditions, the daily average TA excretion was significantly less in KO_b/c mice than WT mice (G: p < 0.001 and p = 0.026; Figure 7C). However, the change in TA excretion between days 1 and 2 (ΔTA day_1-2) was not significantly different between WT and KO_b/c mice under control conditions or MAc-conditions (Figure 7D, control and MAc). Average daily urine pH was significantly greater in KO_b/c mice than WT mice while under control conditions (G: p < 0.001; Figure 7E, control). However, during MAc, there was no significant difference between the average urine pH of WT and KO_b/c mice. Lastly, the change in urine pH between days 1 and 2 (Δurine pH day_1-2) was not significantly different between WT and KO_b/c mice under control conditions or MAc-conditions (Figure 7F, control and MAc).

Table 2 displays relevant metabolic and electrolyte daily averages measured in WT and KO_b/c mice during the 1–3-days of control (0.5% sucrose added to drinking water) or MAc (0.28M NH₄Cl + 0.5% sucrose added to drinking water) challenged conditions. To summarize these results, KO_b/c mice were significantly smaller by body-weight, and ate and drank less compared to their WT counterparts. However, when food and fluid intake were normalized to bodyweight these differences in intake became non-significant. Under control conditions, KO_b/c mice had a significantly lower [Na⁺] [Cl⁻], and BUN. During MAc [Cl⁻] and BUN both remained significantly lower in KO_b/c mice compared to WT, while there was no significant difference in [Na⁺]. Overall, these results make it unlikely that differences in fluid intake, food intake, hydration status, or metabolism of NH₄Cl, account for the observed differences in renal acid-excretion.

Since both male and female mice were used in these experiments, we assessed for differences in all parameters by including sex as an independent factor in ANOVA. Notably, only NH₄ ⁺ excretion under the MAc condition was found to have a significant sex interaction (genotype x time x sex: p = 0.047). Analysis of the sex-split data set suggests that male KO_b/c mice are the primary drivers of the increase in NH₄ ⁺ excretion observed in KO_b/c mice on day 2 of MAc (Day 2 [males]–WT: 216 ± 30 µmol/day, KO_b/c: 411 ± 65 µmol/day, p = 0.001, n = 11 per group; Day 2 [females]–WT: 157 ± 22 µmol/day, KO_b/c: 120 ± 29 µmol/day, p = 0.460, n = 13 and 7, respectively). Possible explanations for this sex difference are discussed below (see ‘Discussion’ section); however, since characterization of sex-specific differences was not a primary aim of this study and we did not observe any significant sex-interactions in any other parameters, data from male and female mice were kept pooled for reporting in Figure 7. Altogether, these data suggest that NH₄ ⁺ excretion, but not TA excretion or urine acidification, is enhanced in KO_b/c mice over day 2 of MAc.

Results of recent studies in NBCe1-A KO mice and combined NBCe1-A/B kidney specific KO mice have indicated that both NBCe1-A and NBCe1-B are important for the normal upregulation of ammoniagenesis during MAc (Lee et al., 2018; 2022). Thus, we were surprised to observe an increase (rather than decrease) in NH₄ ⁺ excretion in KO_b/c mice during MAc (Figure 7). To this end, we hypothesized that in KO_b/c mice, ammoniagenesis is enhanced in PTs of the cortical labyrinth, where only NBCe1-A and not NBCe1-B is expressed (Figure 3). To test this hypothesis, we compared the expression of PEPCK and GS between WT and KO_b/c PTs located in the cortical labyrinth and OSOM after 2-days of MAc-challenged conditions. MAc normally stimulates an increase in PEPCK expression and a decrease in GS expression. We found no significant difference between WT and KO_b/c PEPCK expression in the cortex or in the OSOM (cortex, p = 0.743; OSOM, p = 0.112; Figures 8A–C), indicating PEPCK expression is stimulated to a similar extent in WT and KO_b/c mice. Conversely, GS expression in the cortex was significantly lower in KO_b/c mice than WT mice (p = 0.016), while there was no significant difference between WT and KO_b/c GS expression in the OSOM (p = 0.550; Figures 9A–C). This result indicates that the decrease in GS expression expected during MAc is enhanced in PTs of the cortical labyrinth of KO_b/c mice, overall supporting the hypothesis that ammoniagenesis is stimulated to a greater extent in KO_b/c mice.