Na⁺/HCO₃^– Cotransporter NBCn2 Mediates HCO₃⁻ Reclamation in the Apical Membrane of Renal Proximal Tubules

Abstract

The kidney maintains systemic acid-base balance by reclaiming from the renal tubule lumen virtually all HCO₃⁻ filtered in glomeruli and by secreting additional H⁺ to titrate luminal buffers. For proximal tubules, which are responsible for about 80% of this activity, it is believed that HCO₃⁻ reclamation depends solely on H⁺ secretion, mediated by the apical Na⁺/H⁺ exchanger NHE₃ and the vacuolar proton pump. However, _NHE3 and the proton pump cannot account for all HCO₃⁻ reclamation. Here, we investigated the potential contribution of two variants of the electroneutral Na⁺/HCO₃^– cotransporter NBCn2, the amino termini of which start with the amino acids MCDL (MCDL-NBCn2) and MEIK (MEIK-NBCn2). Western blot analysis and immunocytochemistry revealed that MEIK-NBCn2 predominantly localizes at the basolateral membrane of medullary thick ascending limbs in the rat kidney, whereas MCDL-NBCn2 localizes at the apical membrane of proximal tubules. Notably, NH₄Cl-induced systemic metabolic acidosis or hypokalemic alkalosis downregulated the abundance of MCDL-NBCn2 and reciprocally upregulated NHE₃. Conversely, NaHCO₃-induced metabolic alkalosis upregulated MCDL-NBCn2 and reciprocally downregulated NHE₃. We propose that the apical membrane of the proximal tubules has two distinct strategies for HCO₃⁻ reclamation: the conventional indirect pathway, in which NHE₃ and the proton pump secrete H⁺ to titrate luminal HCO₃⁻, and the novel direct pathway, in which NBCn2 removes HCO₃⁻ from the lumen. The reciprocal regulation of NBCn2 and NHE₃ under different physiologic conditions is consistent with our mathematical simulations, which suggest that HCO₃⁻ uptake and H⁺ secretion have reciprocal efficiencies for HCO₃⁻ reclamation versus titration of luminal buffers.

The kidney plays a central role in regulating systemic acid-base balance by controlling the movement of HCO₃⁻ from renal tubule cells into the blood. For well over a half century, the consensus has been that this HCO₃⁻ traffic—about 80% of which normally occurs in the proximal tubule (PT)—is coupled exclusively with the secretion of H⁺ into the tubule lumen.^1–4 As shown in Figure 1, this H⁺ secretion is because of the action of the apical Na⁺/H⁺ exchanger NHE3 (Slc9a3) and vacuolar-type H⁺ pump. Under normal conditions, the vast majority of the H⁺ secreted into the PT lumen follows pathway 1 in Figure 1, converting filtered HCO₃⁻ to CO₂ and H₂O, a reaction catalyzed by carbonic anhydrase 4. The CO₂ then diffuses into the PT epithelia to recreate HCO₃⁻ and H⁺ under the influence of carbonic anhydrase 2. The HCO₃⁻ exits the cell via the basolateral Na⁺/HCO₃⁻ cotransporter NBCe1 (Slc4a4).^2,5–7 Pathway 1 (i.e., HCO₃⁻ reabsorption) merely reclaims filtered HCO₃⁻, and adds no net HCO₃⁻ to the blood.

Figure 1.

Conventional model, in which proximal-tubule HCO₃⁻ reclamation depends solely on H⁺ secretion. In this model, the acid-base traffic across the apical membrane of the proximal tubule occurs via the Na⁺/H⁺ exchanger NHE3 and a vacuolar-type proton pump. The secreted protons have three major fates: (1) titration of luminal HCO₃⁻ to generate CO₂, which enters the epithelial cells to recreate HCO₃⁻ and H⁺ (pathway 1); (2) titration of NH₃ to form NH₄⁺ (pathway 2); and (3) titration of HPO₄⁼, creatine, and other luminal buffers. In the cytosol of the proximal-tubule epithelium, the protons to be secreted across the apical membrane arise from three sources: (1) hydration of CO₂ reclaimed from the lumen (pathway 1); (2) metabolism of glutamine (pathway 2); and (3) hydration of CO₂ that enters from the blood (pathway 3). According to this model, HCO₃⁻ reclamation and the creation of new HCO₃⁻ rely entirely on H⁺ secretion. All reclaimed HCO₃⁻ (pathway 1) and all newly created HCO₃⁻ (pathways 2 and 3) move via the basolateral NBCe1 into the interstitial space and finally appear in the blood. For details, see the Introduction. The color of the arrows denotes the flows of carbon in the three different pathways in the proximal tubule.

A tiny fraction of the secreted H⁺ follows pathways 2 and 3 in Figure 1 to titrate buffers other than HCO₃⁻. In pathway 2, glutamine generates NH₃, which enters the lumen and combines secreted H⁺ to form NH₄⁺, which ultimately appears in the urine. The glutamine metabolism also generates “new HCO₃⁻,” which appears in the blood. In pathway 3, CO₂ enters the cell across the basolateral membrane (BLM) to generate (1) more H⁺, which is secreted into the lumen to titrate other buffers (e.g., HPO₄⁼ and creatine); and (2) more new HCO₃⁻, which eventually moves into the blood. It is the new HCO₃⁻ generated in pathways 2 and 3 that titrates the H⁺ generated by body metabolism and traffic across the gastrointestinal tract.

The apical NHE3 and proton pump in Figure 1 can account for most, but not all, HCO₃⁻ reclamation by PT. Genetic disruption of Slc9a3 or application of NHE inhibitor N-ethyl-N-isopropyl amiloride (EIPA) reduces HCO₃⁻ reclamation (J_HCO3) by 50%−60% in the PT.^8–11 Blocking the proton pump with bafilomycin (BAF) reduces J_HCO3 by 22%–35% in the PT.^9,10 Application of BAF in NHE3-null mice reduces J_HCO3 by approximately 80% in the PT, compared with J_HCO3 in uninhibited wild-type mouse PTs.¹⁰ In rats, application of EIPA together with BAF reduces J_HCO3 by 76%.⁹ Consistent with the above perfusion studies, genetic disruption in Slc9a3 reduces plasma [HCO₃^–] by only 2.5–5 mM, causing mild metabolic acidosis (MAc).^10–12 Note that mutations in Slc4a4 often reduce plasma [HCO₃^–] by more than half, causing severe MAc (for review, see Parker and Boron⁷ and Romero et al.¹³). Together, it appears that the apical NHE3 and proton pump are responsible for approximately 80% of J_HCO3 by PT. The remaining approximately 20% is likely mediated by apical mechanisms other than NHE3 and proton pump.

Here, we investigate a candidate mechanism potentially underlying the unaccounted-for approximately 20% of J_HCO3 in the PT. Previous studies have shown that the rat kidney expresses two groups of full-length NBCn2 (Slc4a10) variants, differing in the amino termini (Nts), one starting with residue MCDL (mainly NBCn2-G), and the other with MEIK (mainly NBCn2-C).^14,15 Here, we report that MCDL-NBCn2 is expressed at the apical membrane of the PT, whereas MEIK-NBCn2 is expressed at the BLM of the medullary thick ascending limb (mTAL) and inner medullary collecting ducts (IMCDs). Systemic acid-base disturbances reciprocally regulate the abundances of MCDL-NBCn2 and NHE3. Our data are consistent with the hypothesis that PT apical membranes reabsorb HCO₃⁻ by two strategies. The major one is the classic indirect CO₂ pathway, depending on H⁺ secretion by the apical NHE3 and proton pumps. The minor one is the direct HCO₃^– pathway, in which NBCn2 transfers HCO₃⁻ across the apical membrane.

Results

Distribution and Localization of NBCn2 in the Rat Kidney

Consistent with previous observations,¹⁴ anti-MCDL recognizes a band of apparent molecular mass approximately 150 kD from the rat kidney, but not the brain (Figure 2A). Anti-MEIK recognizes bands of apparent molecular mass approximately 180 kD from the brain and approximately 190 kD from the kidney (Figure 2B). The difference between the apparent and predicted molecular masses (approximately 125 kD) reflects N-glycosylation of NBCn2 proteins (see Liu et al.¹⁴ and Chen et al.¹⁶, and Supplemental Material). The difference in the apparent molecular mass of MEIK-NBCn2 in the brain and kidney presumably reflects differences in MEIK-NBCn2 variants and glycosylation in these tissues (Supplemental Material).

Figure 2.

Distinct distribution of MCDL-NBCn2 and MEIK-NBCn2 in rat brain and kidney, and in subcellular fractions of kidney. (A and B) Expression of MCDL-NBCn2 and MEIK-NBCn2 in the rat brain and kidney. The numbers above each lane indicate the amount of total membrane proteins loaded into the lane for SDS-PAGE. (C) Distribution of NBCn2 variants in various tissues of rat kidney. We loaded 40 μg of total membrane proteins into each lane. β-actin was probed to verify equal loading in each lane. (D) Distribution of NBCn2 variants in basolateral versus BBMs. In (D), equal volumes of the total, BLM, and BBM preparations (see Concise Methods for details) were loaded into each lane for SDS-PAGE and Western blotting in such a way that, in theory, the amount of a given protein in total equals the sum of the amounts of this protein in BBM and BLM. Although this is true for MCDL, MEIK, and α1, the exclusively apical NHE3 paradoxically generates a stronger signal in BBM than in total. AQP1, aquaporin 1.

MCDL-NBCn2 is primarily expressed in the cortex (CX), outer stripe of outer medulla (OSOM), and inner stripe of outer medulla (ISOM), but not detectable in the inner medulla (IM) in rat kidneys (Figure 2C). In contrast, MEIK-NBCn2 is predominantly expressed in the ISOM and IM. Consistent with previous studies,^17–19,40 we show that AQP1 is predominantly expressed in the CX, OSOM, and ISOM; NHE3 is predominantly expressed in the CX; and the α1 subunit of Na⁺-K⁺ pump (α1) is expressed throughout the kidney (Figure 2C). Thus, MCDL-NBCn2 has a tissue distribution similar to that of AQP1 in rat kidneys.

In isolated BLMs and brush border membranes (BBMs) from rat kidneys, MCDL-NBCn2 is mainly distributed in the BBM, whereas MEIK-NBCn2 is predominantly distributed in the BLM (Figure 2D). The observations that α1 (basolateral marker) is primarily basolateral but NHE3 (apical marker) is predominantly apical confirm our separation of the BLM and BBM. Together, our data indicate that MCDL-NBCn2 and MEIK-NBCn2 have distinct tissue and subcellular distributions in the rat kidney. MCDL-NBCn2 is predominantly expressed in the apical membranes in the CX, OSOM, and ISOM, whereas MEIK-NBCn2 is predominantly expressed in the BLMs of the ISOM and IM.

Consistent with the data in Figure 2, C and D, immunocytochemistry shows that MCDL-NBCn2 is predominantly expressed in the CX and OSOM of rat kidney (Figures 3 and 4). Immunodepletion with the Nt of MCDL-NBCn2 eliminates the fluorescence (Figure 3B, inset), suggesting that the immunofluorescence signal is specific. A higher magnification view shows that MCDL-NBCn2 colocalizes with AQP1 in the apical membrane of tubules in the CX (Figure 5, A–C). Moreover, in the CX and OSOM, tubules enriched in MCDL-NBCn2 exhibit relatively low expression of α1, whereas those lacking MCDL-NBCn2 exhibit very strong expression of α1 (Figure 5, D–I). Others have shown that α1 is highly enriched in all segments, from the mTAL through to the cortical collecting tubule, but is much lower in the PTs of the rat kidney.^17,20 Indeed, in the CX, the tubules enriched in α1 account for only a small fraction of total tubules (Figures 4 and 5, D–F), consistent with our conclusion that these α1-enriched cortical structures are not PTs. In the OSOM, tubules lacking MCDL-NBCn2 but highly enriched in α1 are presumably mTALs. Considering all of these data, we conclude that, in the rat kidney, MCDL-NBCn2 is expressed in the apical membrane of PTs but not in the mTAL through to the cortical collecting tubule.

Figure 3.

Overview showing colocalization of AQP1 and MCDL-NBCn2 in the rat kidney. A section of rat kidney was costained with (A) anti-AQP1 (a specific marker for the PT and thin descending limb of Henle), (see refs. 17, 19, 40), and (B) anti-MCDL directed against NBCn2, and counterstained with DAPI. (C) shows the merge of (A and B) plus DAPI. The figure was generated from images of five contiguous regions. In the CX and OSOM, MCDL-NBCn2 staining is seen in the tubules enriched in AQP1 expression. The inset in (B) shows a section stained with anti-NBCn2 that has been immunodepleted with GST-NBCn2-Nt.

Figure 4.

Overview showing differential localization of the α1 subunit of the Na⁺-K⁺ pump and MCDL-NBCn2 in rat kidney tubules. A section of rat kidney was costained with (A) anti-α1 (a specific marker for the BLM of renal tubules), (see refs. 17, 20), and (B) anti-MCDL. The section was counterstained with DAPI. (C) shows the merge of (A and B) plus DAPI. The figure was generated from images of five contiguous regions. The distribution profile of MCDL-NBCn2 is distinct from that of the α1 subunit of Na⁺-K⁺ pump, with the tubules that are positive for MCDL-NBCn2 expression having low expression of α1.

Figure 5.

High-magnification view showing apical localization of MCDL-NBCn2 in proximal tubules of rat kidney. (A–C) Section of renal CX costained with anti-AQP1 and anti–MCDL-NBCn2. (D–F) Section of renal CX costained with anti-α1 subunit of Na⁺-K⁺ pump and anti–MCDL-NBCn2. (G–I) Section of the OSOM and ISOM costained with anti-α1 and anti–MCDL-NBCn2. The star indicates a tubule a high level of expression of the Na⁺-K⁺ pump; the triangle, a tubule with less expression; and the arrow, a tubule with no visible expression of Na⁺-K⁺ pump. The dashed lines in (G–I) represent the demarcation between OSOM and ISOM. S1, the s1 segment of proximal tubule; G, glomerulus.

In the ISOM (Figure 6, A–C), MEIK-NBCn2 is localized at the BLMs of structures highly enriched in α1, characteristic of the BLM of the mTAL.^17,20 Figure 6, D–F shows the basolateral localization of α1 and the apical localization of NHE3 in the mTAL. In the IM (Figure 6, G–I), MEIK-NBCn2 also colocalizes with α1 at the BLM of structures that are presumably IMCDs. The basolateral localization of MEIK-NBCn2 is consistent with Figure 2D. Thus, it is likely that MEIK-NBCn2 is localized at the BLMs of mTALs in the ISOM, and of IMCDs in the IM.

Figure 6.

High magnification view showing basolateral localization of MEIK-NBCn2 in the mTAL of the ISOM, as well as in the IMCD of rat kidney. (A–C) A section of renal ISOM, costained with anti-α1 subunit of Na⁺-K⁺ pump and anti-MEIK. (D–F) A section of the renal ISOM, costained with anti-α1 and anti-NHE3. (G–I) A section of the renal IM, costained with anti-α1 and anti–MEIK-NBCn2. Strong labeling of MEIK-NBCn2 was observed at the BLM of mTALs. The labeling of MEIK-NBCn2 was also observed, to a weaker level, at the BLM of IMCDs. The white arrows in (C) indicate the tubules that were cut through the tubule wall mostly parallel to the long axis of the lumen; therefore, the nucleus is surrounded by BLMs. The arrowheads in (C and F) indicate the approximate locations of BBMs (on the side of the nuclei abutting the tubule lumen). Note that the BBMs indicated by arrowheads in (C) are invisible because of the lack of any staining.

Effect of NH₄Cl-Induced MAc on NBCn2 Expression in the Kidney

We were surprised to find MCDL-NBCn2 in the apical membrane of the PT. Here, NBCn2 would mediate Na⁺/HCO₃⁻ influx from the tubule lumen to the cytosol. To address the physiologic relevance of the apical MCDL-NBCn2 in PTs, we investigated the effects of systemic acid-base disturbances on the expression of MCDL-NBCn2 and NHE3 in the rat kidney.

As shown in Figure 7A, NH₄Cl-induced MAc for 1 week lowers the abundance of MCDL-NBCn2 in the rat kidney by approximately 20% compared with the control condition (no MAc), although this difference is not statistically significant (see Supplemental Table 1 for detailed MCDL-NBCn2 data). NH₄Cl treatment for 2 weeks decreases MCDL-NBCn2 abundance by approximately 31% (P=0.03; Figure 7B). As shown in Figure 7C, compared with control, NH₄Cl treatment increases NHE3 abundance by approximately 127% in rat kidneys after 1 week (P<0.01; see Supplemental Table 2 for detailed NHE3 data), and by approximately 97% after 2 weeks (P=0.003; data not shown), consistent with previous reports.^21–25 Thus, the expression of MCDL-NBCn2 and NHE3 are inversely regulated by NH₄Cl-induced MAc.

Figure 7.

Inverse regulation of the expression of MCDL-NBCn2 and NHE3 in rat kidney by systemic acid-base disturbances. (A) Effect of 1 week of NH₄Cl overload on MCDL-NBCn2. Here, 1.5% NH₄Cl was in drinking water only. (B) Effect of 2 weeks of NH₄Cl overload on MCDL-NBCn2. Here, 1.5% NH₄Cl was in the drinking water and 2% in the chow (with the rats being fed normal chow every third day). (C) Effect of 1 week of NH₄Cl overload on NHE₃. (D) Effect of 1 week of NaHCO₃ overload on MCDL-NBCn2. (F) Effect of 2 weeks of NaHCO₃ overload on MCDL-NBCn2. (F) Effect of 1 week of NaHCO₃ overload on NHE₃. (G) Effect of 2 weeks of potassium depletion on MCDL-NBCn2. (H) Effect of 2 weeks of potassium depletion on NHE₃. The top of each panel shows representative Western blots of the corresponding transporters and β-actin for the normalization of loading. The bar graph at the bottom of each panel shows the summary of densitometry data (presented as mean±SEM) from the blots like those shown on the top. Two-tailed paired t tests were performed for statistical analysis.

Effect of NaHCO₃-Induced Metabolic Alkalosis on NBCn2 Expression in the Kidney

NaHCO₃ overload induces metabolic alkalosis (MAlk). In this study, treatment with NaHCO₃ for 1 week raises the mean abundance of MCDL-NBCn2 by approximately 21% in the kidney compared with controls, although this difference is not significant (Figure 7D). However, extending the NaHCO₃ treatment to 2 weeks increases MCDL-NBCn2 abundance by approximately 52% (P=0.002; Figure 7E). Conversely, we find that, compared with control, NaHCO₃ overload decreases NHE3 abundance by approximately 50% in rat kidneys after 1 week (P=0.02; Figure 7F), and approximately 46% after 2 weeks (P<0.01; data not shown). This last observation is consistent with previous studies showing that NaHCO₃-induced MAlk substantially downregulates renal NHE3 expression.^21,26 NaCl overload does not significantly change the abundance of MCDL-NBCn2 in rat kidney compared with controls (Supplemental Figure 1), ruling out an effect of Na⁺ overload per se on the interpretation of our NaHCO₃ data. Thus, our NaHCO₃-induced MAlk data provide a second example of an inverse relationship between the regulation of renal MCDL-NBCn2 and NHE3 expression.

Effect of Hypokalemic Alkalosis on NBCn2 Expression in the Kidney

Potassium depletion causes hypokalemic alkalosis, presumably by stimulating the production and secretion of NH₄⁺ in PTs.^27,28 Indeed, potassium depletion substantially increases the abundance of NHE3 in rat kidneys.²⁹

Here, we find that potassium depletion for 2 weeks decreases MCDL-NBCn2 abundance by approximately 45% (P<0.03; Figure 7G) but increases NHE3 expression by approximately 121% in the rat kidney (P<0.01; Figure 7H). Our hypokalemic data provide a third example of an inverse relationship between the regulation of MCDL-NBCn2 and NHE3 expression.

Discussion

In rats, the Slc4a10 gene can give rise to multiple NBCn2 variants that fall into two major groups. Depending on the usage of different promoters,¹⁵ the extreme Nt of these variants can begin either with MCDL (NBCn2-E/F/G/H) or MEIK (NBCn2-A/B/C/D). In the MCDL group, the rat kidney mainly expresses NBCn2-G, and in the MEIK group, it expresses mainly NBCn2-C.^14,15 Here, we find that MCDL-NBCn2 and MEIK-NBCn2 exhibit distinct tissue and subcellular distributions in rat kidney, with MCDL-NBCn2 being localized in the apical membrane of PTs, and MEIK-NBCn2 localized in the BLM of mTALs and IMCDs.

The basolateral localization of MEIK-NBCn2 in the mTAL parallels that of NBCn1.^24,30 Because NBCn1 contributes to NH₄⁺ reabsorption in the mTAL, MEIK-NBCn2 likely does the same.

The apical localization of MCDL-NBCn2 in PTs suggests that the cell can directly reclaim HCO₃⁻ per se or a related species (e.g., CO₃⁼ or NaCO₃^–)—collectively referred to as “HCO₃⁻” hereafter—across the apical membrane of the PT. Figure 8 summarizes our new model for the molecular mechanisms underlying HCO₃⁻ reclamation in the PT. We propose that the apical membrane of PTs reclaims HCO₃⁻ by two distinct strategies: the conventional indirect CO₂ pathway 1 and the novel HCO₃⁻ pathway 4 (Figure 8). The conventional CO₂ pathway couples HCO₃⁻ reclamation to proton secretion by apical NHE3 and vacuolar H⁺ pump, thereby generating CO₂ and H₂O, which then diffuse across the apical membrane for eventual reconstruction of the original HCO₃⁻ in the cytosol of the PT cell. In contrast, the HCO₃⁻ pathway moves luminal HCO₃⁻ directly into the cytosol of the epithelia. Although not shown in Figure 8, the direct uptake of HCO₃⁻ from the lumen would also indirectly generate luminal H⁺ for the titration of other luminal buffers (Supplemental Figure 2). In a steady state, the sum of apical fluxes of CO₂ (J_CO2^AP) and HCO₃⁻ (J_HCO3^AP) represents the total disappearance of HCO₃⁻ from PT lumen. As noted in the Introduction, the PT retains approximately 20% of its J_HCO3 even after disruption/inhibition of both NHE3 and proton pump. We propose that the apical NBCn2 accounts for this remaining J_HCO3.

Figure 8.

New model, in which proximal-tubule HCO₃^– reclamation depends on both H⁺ secretion (conventional indirect pathways #1, 2, and 3) and HCO₃^– reabsorption (novel direct pathway #4). In the PT, HCO₃⁻ is reclaimed from the lumen via two distinct mechanisms: (1) pathway 1 as described in Figure 1; (2) the novel pathway 4, in which HCO₃⁻ (or a HCO₃⁻-related species like CO₃⁼ or NaCO₃^–) is directly reabsorbed by the apical Na⁺/HCO₃⁻ cotransporter NBCn2. The total flux of CO₂ across the apical membrane (J_CO2^AP) via pathway 1 and the total flux of HCO₃⁻ across the apical membrane (J_HCO3^AP) via pathway 4 represent the total HCO₃⁻ reclaimed by the PT cells, which accounts for approximately 80% of total HCO₃⁻ reclamation by the kidney. Of the 80% of HCO₃⁻ reclaimed by the PT, pathway 1 is responsible for approximately 80%, thereby accounting for approximately 65% of the total HCO₃⁻ reclamation by the kidney. We propose that pathway 4 is responsible for the remaining approximately 20% of the approximately 80% of the total filtered HCO₃⁻ that is reclaimed by the PT, thereby accounting for approximately 15% of the total HCO₃⁻ reclamation by the kidney. Inside the PT cells, the CO₂ that entered across the apical membrane (pathway 1) is converted to HCO₃⁻. This HCO₃⁻ from pathways 1 and 4 diffuses to the BLM. In addition to these two routes of HCO₃⁻ reclamation, the PT cell generates new HCO₃⁻ in pathway 2 (glutamine metabolism) and pathway 3 (hydration of CO₂ derived from the blood). The HCO₃^– efflux across the BLM (J_HCO3^BL, presumably all via NBCe1) equals the sum of the reclaimed and new HCO₃⁻ (i.e., J_HCO3^BL=J_CO2^AP+J_HCO3^AP+J_HCO3^new). The color of the arrows denotes the flows of carbon in the four different pathways in the PT.

The apical MCDL-NBCn2 that we demonstrate here is consistent with a previous study identifying 4,4′-diisothiocyanostilbene-2,2′-disulfonate (DIDS)–sensitive acid-base transport at the apical membrane of rat PTs.³¹ Luminal DIDS reduces fluid reabsorption (J_V) by approximately 19% in rat PTs,³¹ a fraction consistent with the unaccounted-for approximately 20% of J_HCO3 (see Introduction). Moreover, the inhibitory effect of luminal DIDS on J_V requires the presence of HCO₃⁻.³² It is true that, in rabbit PTs, luminal 4-acetamido-4′-isothiocyanostilbene-2,2′-disulfonate—administered at the same time luminal Na⁺ is removed—has no effect on the intracellular acidification induced by Na⁺ removal.³³ This apparent inconsistency with our hypothesis could reflect a species specificity or, more likely, insufficient time for luminal 4-acetamido-4′-isothiocyanostilbene-2,2′-disulfonate to block NBCn2 before the reversal of Na⁺-coupled transporters was well underway.

One question that arises is whether the direct HCO₃⁻ pathway via apical NBCn2 serves merely as a back-up for the indirect CO₂ pathway, which depends on the apical NHE3 and vacuolar proton pump. If so, then we would expect systemic acid-base disturbances to affect all of these transporters in parallel. However, we find that the regulation of the protein abundances of NBCn2 and NHE3 are strikingly reciprocal. Under chronic NH₄Cl-induced MAc or hypokalemic alkalosis for 2 weeks, the protein abundance of NHE3 is substantially upregulated, as is well established by previous work^21–25,29 and confirmed in this study. Conversely, the protein abundance of MCDL-NBCn2 is markedly downregulated. In contrast, under chronic NaHCO₃-induced MAlk, MCDL-NBCn2 is upregulated, whereas NHE3 is downregulated. Thus, we conclude that NBCn2 cannot be merely a back-up pathway for HCO₃^– reclamation.

A second question that arises is whether the reciprocal regulation of MCDL-NBCn2 and NHE3 in the PT confers a functional advantage for the animal. Is it possible that the two pathways have different efficiencies for new HCO₃⁻ creation (accounting for 1%–2% of total PT acid-base traffic under physiologic conditions) on the one hand and HCO₃⁻ reclamation (accounting for the remaining 98%–99%) on the other? One approach for addressing this question is to simulate mathematically the effects of apical H⁺ extrusion versus apical HCO₃⁻ uptake. Because a reaction–diffusion model for CO₂/HCO₃⁻ and various buffers does not exist for PTs, we adapt an existing Xenopus oocyte model,^34,35 with the goal of providing a proof of principle. As outlined in Supplemental Figure 2, we simulate three models: (1) extrusion of H⁺ across a cell membrane, (2) uptake of the same amount of HCO₃⁻, or (3) uptake of half that amount of CO₃⁼ (i.e., same number of acid-base equivalents). Note that both NHE3 and NBCn2 are energized by the coupled influx of one Na⁺, and carry the same number of acid-base equivalents. For each model, we compute the fraction of acid-base transport across the membrane used to (1) titrate the weak base A^– (a surrogate for non-HCO₃ buffers like NH₃ and HPO₄²⁻) to its conjugate weak acid HA (e.g., H₂PO₄⁻) in a thin layer of extracellular fluid near the membrane, and (2) reclaim HCO₃⁻ from this same thin layer. Also, for each model, we perform simulations at two values of bulk extracellular pH (pH_o), 7.50 (standard oocyte pH_o, mimicking luminal pH near the beginning of the PT) and 7.10 (mimicking luminal pH further down the PT), and at two values of the extracellular surface CA acceleration factor (A_os), 25,000 (extremely high A_os that approximates erythrocyte cytoplasm) and 150 (approximating the surface of the control Xenopus oocyte).

As summarized in Table 1, our simulations predict that the H⁺ extrusion model is most efficient for A^– titration but least efficient for HCO₃⁻ reclamation, whereas the HCO₃⁻ uptake model is just the opposite. The CO₃⁼ model makes intermediate predictions. The discrepancies among models (i.e., efficiency of A^– titration versus HCO₃⁻ reclamation) increase as pH_o increases (7.10 → 7.50), and A_os decreases (25,000 → 150). For example, at pH_o 7.50, H⁺ extrusion is more efficient than HCO₃^– uptake for A^– titration (a process essential for the formation of new HCO₃⁻ that titrates the net acid generated in the body) by approximately 26% at A_os=25,000, and by approximately 750% at A_os=150. The PT probably lies between these two values. During MAc, the upregulation of NHE3 (at the expense of MCDL-NBCn2) is energetically advantageous for the critical need of increasing new HCO₃⁻ generation (which requires increasing luminal A^– titration) to return blood pH toward normal. Upregulation of NHE3 is specifically not required for HCO₃⁻ reclamation because the HCO₃^– load to the PT naturally falls during MAc. Of course, the preexisting levels of NHE3 and NBCn2 were already adequate to reclaim larger HCO₃^– loads under normal acid-base conditions. Thus, during MAc it is energetically advantageous to decrease HCO₃⁻-reclamation capacity by downregulating MCDL-NBCn2. Conversely, in situations like MAlk, where the kidney must decrease new HCO₃⁻ generation to return blood pH toward normal, it is energetically advantageous to reduce A^– titration by downregulating NHE3. Because MAlk simultaneously increases the HCO₃⁻ filtered load, the kidney must increase HCO₃⁻ reclamation, which it achieves most efficiently by upregulating MCDL-NBCn2.

Table 1.

Results of mathematical simulations

Surface CA Acceleration Factor	Model	pHo=7.50			pHo=7.10
		2 H+, %	2 HCO3–, %	1 CO3=, %	2 H+, %	2 HCO3–, %	1 CO3=, %
25,000	A− titration	2.25	1.79	2.02	4.47	3.76	4.11
	HCO3– reclamation	97.76	98.21	97.98	95.54	96.25	95.89
150	A− titration	11.95	1.40	6.66	18.54	3.0	10.83
	HCO3– reclamation	88.04	98.60	93.33	81.41	97.01	89.12

We use three different reaction–diffusion models for a spherical cell (extrusion of two H⁺ ions, uptake of two HCO₃⁻ ions, or uptake of 1 CO₃= ion) to simulate at two different bulk extracellular pH (pH_o) values and two different sets of CA activities—the fraction of total traffic of acid-base equivalents across the cell membrane that results in titration of extracellular A^– versus reclamation of extracellular HCO₃⁻. For details, see Supplemental Material. Note that we captured these data 30 seconds into a simulation in which we initiate acid-base transport at 0 seconds. Data summaries for simulations at 10 and 100 seconds are shown in Supplemental Table 4.

MCDL-NBCn2 and NHE3 are also under reciprocal regulation during hypokalemic alkalosis, an observation that raises the possibility that one set of controls reciprocally modulates the expression of these two proteins in the PT. Note that we have not simulated the titration of NH₃ to NH₄⁺ in the extracellular fluid. Also, we have not considered the possibility, raised by Kinsella and Aronson,³⁶ that NHE3 may, to some extent, engage in Na⁺-NH₄⁺ exchange.

In summary, we find that the Na⁺/HCO₃^– cotransporter MCDL-NBCn2 is expressed at the apical membrane of renal PTs. We suggest that the apical PT membrane uses two strategies for HCO₃^– reclamation, with different acid-base consequences: the indirect CO₂ pathway depending on the H⁺ secretion by the apical NHE3, and the direct HCO₃^– pathway depending on the apical MCDL-NBCn2. The PT reciprocally regulates MCDL-NBCn2 and NHE3 under conditions that require reciprocal modulation of HCO₃^– reclamation versus new HCO₃^– creation. Indeed, our mathematical simulations show that these two pathways, appropriately, have reciprocal efficiencies for HCO₃⁻ reclamation versus creation of new HCO₃⁻. These predictions call for testing in future studies. Nevertheless, our findings about the direct HCO₃^– pathway for HCO₃^– reclamation in the apical membrane of PT are of fundamental importance for understanding the molecular mechanisms underlying the regulation of renal acid-base balance in health and disease.

Concise Methods

Antibodies and Fusion Protein

Rabbit polyclonal antibodies against NBCn2, including anti-MCDL and anti-MEIK, were generated by Genscript (Nanjing, China). Anti-MCDL directed against SGNRKVMQPGTCEHC in the unique Nt of MCDL-NBCn2 has been described previously,¹⁴ and is further characterized in this study (see Supplemental Figures 1, A–C and 2, A–C). Anti-MEIK was directed against MEIKDQGAQMEPLLPTC, an epitope in the unique portion of the Nt of MEIK-NBCn2 that has been well characterized previously.³⁷ The underlined Cys residues were artificially introduced for conjugating the peptide to keyhole limpet hemocyanin for immunizing rabbits. All these polyclonal antibodies were affinity-purified with the corresponding immunizing peptides. Goat anti-AQP1 was purchased from Santa Cruz Biotechnology (catalog no. sc-9878). Mouse anti-α1 (catalog no. ab-7671) against the α1 subunit of Na⁺-K⁺ pump and mouse anti-NHE3 (catalog no. ab-95299) were purchased from Abcam (Hong Kong). Mouse anti-actin was purchased from Beyotime (catalog no. AA128; Jiangsu, China).

Secondary antibodies conjugated with HRP for Western blotting were goat anti-rabbit (catalog no. A0208) and goat anti-mouse (catalog no. A0216) from Beyotime, and donkey anti-goat (catalog no. SA00001–3) from ProteinTech Group, Inc. (Chicago, IL). Secondary antibodies used for immunocytochemistry were Alexa Fluor 488 AffiniPure goat anti-mouse (catalog no. 115–545–003) and Alexa Fluor 488 AffiniPure donkey anti-goat (catalog no. 705–545–003) from Jackson ImmunoResearch Laboratories (West Grove, PA), and DyLight 549 goat anti-rabbit (catalog no. E032320–01) from EarthOx (Millbrae, CA).

The construct for recombinantly expressing fusion protein GST-MCDL-Nt, consisting of glutathione-S-transferase and the Nt of MCDL-NBCn2, was described previously.¹⁴ The fusion protein was expressed in bacteria and affinity-purified with glutathione-S-transferase resin (Genscript) according to the manufacturer’s instructions.

Animals

Adult Sprague–Dawley (SD) rats with 200–300 g body wt were used to examine the effects of systemic acid-base disturbances on the expression of renal acid-base transporters. For NH₄Cl overload, rats from the same batch were randomly assigned to one of two groups. One group was supplied with drinking water containing 1.5% NH₄Cl plus 0.5% sucrose for 1 week. The other (control) was supplied with water containing 0.5% sucrose only. The rats had free access to drinking water and were supplied with regular rodent chow. In other experiments, to maximize the acid load, the rats were treated with NH₄Cl both in drinking water and chow as described previously,^22,38 with slight modifications. Briefly, the acidotic rats had drinking water containing 1.5% NH₄Cl plus 0.5% sucrose, and chow containing 2% NH₄Cl (replaced with normal chow every third day). The NH₄Cl chow was prepared by mixing powdered regular chow and 3% NH₄Cl solution with a ratio of 4:3 (wt/vol). The control rats were pair-fed with water containing 0.5% sucrose and regular chow. After 1 or 2 weeks, the rats were anesthetized by subcutaneous injection of pentobarbital sodium and euthanized for tissue collection.

For NaHCO₃ overload, adult SD rats were randomly assigned to one of two groups. One was supplied with water containing 0.28 M NaHCO₃ for 1 or 2 weeks. The control group was supplied with plain water. After the treatment, the rats were then euthanized for tissue collection.

For hypokalemic alkalosis, the rats were fed with either potassium-depleted chow (catalog no. TP0901–01M, potassium level of 0.01%) or standard AIN93 chow (catalog no. LAD3001M, potassium level of 0.36%) purchased from Trophic Animal Feed High-Tech Co. Ltd. (Jiangsu, China). The control rats had a limited supply of chow, so that they had similar caloric intakes compared with the potassium-depletion group. All rats had free access to ultrapure water. The rats were treated for 2 weeks and then euthanized for tissue collection.

Membrane Protein Preparation and Western Blotting

Rat kidney tissue (approximately 500 mg) was placed in 6 ml of precooled protein isolation buffer (7.5 mM NaH₂PO₄, 250 mM sucrose, 5 mM EDTA, 5 mM EGTA, pH 7.0) containing 1% protease inhibitor cocktail (catalog no. P8340; Sigma-Aldrich, St. Louis, MO) and homogenized by 15 strokes with a Glas-Col Teflon glass homogenizer (Glas-Col, Terre Haute, IN). The crude homogenate was centrifuged at 3000×g for 10 minutes at 4°C to remove the cell debris. The supernatant was then centrifuged at 100,000×g for 60 minutes at 4°C. The pellet containing the plasma membranes was resuspended in buffer containing 20 mM Tris, 5 mM EDTA, 5% SDS, pH 8.0. The protein concentration was determined with Enhanced BCA Protein Assay Kit (catalog no. P0010; Beyotime). The protein preparations were stored in aliquots at −80°C until usage.

BBMs and BLMs were prepared from whole rat kidneys by MgCl₂ precipitation, as described previously,³⁹ with slight modifications. Briefly, 250 mg of rat kidney tissue was homogenized in 6 ml of precooled protein isolation buffer and centrifuged at 3000×g for 10 minutes at 4°C. The supernatant was then ultracentrifuged at 48,000×g for 1 hour. The resultant pellet was resuspended in 500 μl of protein isolation buffer and 100 μl of this resuspension was saved as the “total” membrane protein. The remaining 400 μl was diluted with 2.5 ml of protein isolation buffer and 45 μl of 1 M MgCl₂, and then incubated at 4°C for 20 minutes. The samples were centrifuged at 3000×g for 10 minutes at 4°C to collect the precipitated membrane proteins. The supernatant was subjected to two more rounds of MgCl₂ precipitation. The pellets from all three rounds of centrifugation were pooled, resuspended in 400 μl of protein isolation buffer, and used as BLMs. The supernatant from the last MgCl₂ precipitation was centrifuged at 48,000×g at 4°C for 30 minutes. The resultant pellet was resuspended in 400 μl protein isolation buffer and used as BBMs.

For Western blotting, the membrane proteins were separated by SDS-PAGE and then blotted onto a PVDF membrane. The blot was blocked with 5% milk in TBST (1 mM Tris, 150 mM NaCl, 0.1% Tween 20, pH 7.4) for 1 hour at room temperature (RT). The blot was then probed with primary antibody in TBST containing 1% milk at 4°C overnight. The blot was washed five times with TBST, incubated with HRP-conjugated secondary antibody in TBST containing 1% milk for 1 hour at RT, and then washed five times with TBST. The blot was treated with SuperSignal West Pico chemiluminescent reagent (Thermo Scientific, Rockford, IL) before x-ray exposure.

Immunocytochemistry

Adult SD rats were fixed by transcardial perfusion with 4% paraformaldehyde in PBS buffer (77.4 mM Na₂HPO₄, 22.6 mM NaH₂PO₄, pH 7.4). The kidneys were collected and stored in PBS containing 0.1% PFA plus 0.02% NaN₃ until usage for cryo-section. For immunocytochemical staining, a section of 5 μm was incubated in 60°C overnight and then rehydrated in TBS for 60 minutes. The section was treated at 98°C for 20 minutes in Improved Citrate Antigen Retrieval Solution (catalog no. P0083; Beyotime). After washing five times with TBS, the section was blocked with Immunol Staining Blocking Buffer (catalog no. P0102; Beyotime) for 30 minutes, and then incubated with primary antibody at 4°C overnight. The section was washed five times with TBS and incubated with secondary antibody at RT for 60 minutes. After washing three times with TBS, the section was counterstained with 4′,6′-diamidino-2-phenylindole at RT for 5 minutes, washed three times with TBS, and mounted with Antifade Polyvinylpyrrolidone Medium (catalog no. P0123; Beyotime). Images were acquired on a FluoView FV1000 confocal microscope (Olympus, Tokyo, Japan).

Mathematical Simulations

A time-dependent reaction–diffusion model of transmembrane CO₂ fluxes in a Xenopus oocyte was adapted to simulate the effects of apical H⁺ extrusion versus apical HCO₃⁻ uptake in the PT.³⁴ Here, we briefly summarize the main features of these models. For details on the theory behind the models and on the numerical implementation, refer to the first-generation model described previously.³⁵ The second-generation model includes refinements that lead to more realistic simulations of an oocyte exposed to a solution containing equilibrated CO₂/HCO₃⁻.³⁴

The starting point for the simulations in this study is the second-generation model of a CAIV oocyte (i.e., an oocyte expressing carbonic anhydrase 4). The model assumes that the oocyte is a symmetric spherical cell surrounded by a thin spherical layer of extracellular unconvected fluid (EUF) of thickness d=10 μm. The EUF is surrounded by the bulk extracellular fluid, which is an infinite reservoir of pre-equilibrated solution. In the intracellular fluid and EUF, diffusion of solutes (CO₂, H₂CO₃, HCO₃⁻, H⁺, HA, and A⁻) and reactions among solutes (CO₂ hydration/dehydration, carbonic acid dissociation, and HA dissociation) occur. In this study, we extend the second-generation model to include the uptake of one CO₃⁼ ion; we therefore include the new solute CO₃⁼ (assumed to have the same mobility as HCO₃⁻) and model the reaction HCO₃⁻⇄CO₃⁼+H⁺, with pK=10.2258 at 22°C. Moreover, in this study, in order to mimic inorganic phosphate in the PT lumen, we assume that the extracellular HA/A⁻ buffer has a pK=6.8 and [HA]+[A⁻]=0.7 mM. Finally, we assume that the acceleration factor for extracellular surface CA activity is 25,000 (or 150) and that the acceleration factor for cytosolic CA activity is 20 (or 5). All other parameter values are the same as described previously (see Supplemental Material for further details).^34,35

Disclosures

None.