Effect of acute acid-base disturbances on ErbB1/2 tyrosine phosphorylation in rabbit renal proximal tubules

Abstract

The renal proximal tubule (PT) is a major site for maintaining whole body pH homeostasis and is responsible for reabsorbing ∼80% of filtered HCO₃⁻, the major plasma buffer, into the blood. The PT adapts its rate of HCO₃⁻ reabsorption (J_HCO3⁻) in response to acute acid-base disturbances. Our laboratory previously showed that single isolated perfused PTs adapt J_HCO3⁻ in response to isolated changes in basolateral (i.e., blood side) CO₂ and HCO₃⁻ concentrations but, surprisingly, not to pH. The response to CO₂ concentration can be blocked by the ErbB family tyrosine kinase inhibitor PD-168393. In the present study, we exposed enriched rabbit PT suspensions to five acute acid-base disturbances for 5 and 20 min using a panel of phosphotyrosine (pY)-specific antibodies to determine the influence of each disturbance on pan-pY, ErbB1-specific pY (four sites), and ErbB2-specific pY (two sites). We found that each acid-base treatment generated a distinct temporal pY pattern. For example, the summated responses of the individual ErbB1/2-pY sites to each disturbance showed that metabolic acidosis (normal CO₂ concentration and reduced HCO₃⁻ concentration) produced a transient summated pY decrease (5 vs. 20 min), whereas metabolic alkalosis produced a transient increase. Respiratory acidosis (normal HCO₃⁻ concentration and elevated CO₂ concentration) had little effect on summated pY at 5 min but produced an elevation at 20 min, whereas respiratory alkalosis produced a reduction at 20 min. Our data show that ErbB1 and ErbB2 in the PT respond to acute acid-base disturbances, consistent with the hypothesis that they are part of the signaling cascade.

the maintenance of ph inside cells [intracellular pH (pH_i)] and in the extracellular fluid [extracellular pH (pH_o)] is vital for almost every biological process. Thus, uncompensated pH disturbances affect a wide range of processes (21, 30, 58, 74, 91) and such parameters as neuronal excitability (37, 40, 61), blood pressure regulation (6, 7), apoptosis (10, 20, 23), and ventilation (53, 66). Moreover, solid tumors create local acid-base disturbances to provide a selective advantage over their normal competitors (85, 86).

In mammals, CO₂/HCO₃⁻ is by far the most important buffer system. The two major components interact via the reactions CO₂ + H₂O ↔ H₂CO₃ ↔ HCO₃⁻ + H⁺, and thus blood pH depends on the ratio of the CO₂ concentration ([CO₂]) to the HCO₃⁻ concentration ([HCO₃⁻]). Controlling blood [CO₂] is the respiratory system, which responds to elevations in [CO₂] or decreases in pH by increasing ventilation and thereby promoting the excretion of metabolically generated CO₂ (8). Controlling blood [HCO₃⁻] is the urinary system, which responds to acidosis by secreting H⁺ into the lumen of renal tubules (28). The majority of this secreted H⁺ titrates the HCO₃⁻ filtered by the glomeruli and converts it to CO₂; the equivalent amount of HCO₃⁻ moves into the blood, representing “reabsorbed HCO₃⁻.” The kidney also secretes an amount of H⁺ equivalent to the H⁺ absorbed in the diet, the OH⁻ excreted in the feces, and the nonvolatile acids produced by metabolism; the equivalent amount of HCO₃⁻ transferred into the blood represents “new HCO₃⁻.”

The renal proximal tubule (PT) is responsible for ∼80% of total H⁺ secretion. The extrusion of H⁺ into the PT lumen occurs via Na⁺/H⁺ exchanger 3 (NHE3) and the vacuolar-type H⁺ pump, both in the brush-border membrane (9, 29, 97). Carbonic anhydrase (CA) IV converts luminal H⁺ and HCO₃⁻ to H₂O and CO₂. H₂O enters the PT cell via the water channel [aquaporin 1 (AQP1)] in the apical membrane (54, 64, 76). CO₂ also enters the PT cell, probably largely via AQP1 (9, 19, 48, 52). Inside the PT cell, CA II converts CO₂ and H₂O to HCO₃⁻ and H⁺. The aforementioned acid extruders transfer H⁺ to the lumen, and electrogenic Na⁺-HCO₃⁻ cotransporter (NBC)e1-A completes the process by transferring Na⁺ and what appears to be 3 HCO₃⁻ across the basolateral (BL) membrane.

Three landmark studies (11, 24, 69) have described the effects of respiratory acidosis (RAc; a fall in arterial pH caused by a rise in [CO₂]) on dogs, showing that the kidney responds rapidly by increasing acid secretion. Metabolic acidosis (MAc; a fall in arterial [HCO₃⁻] at constant [CO₂]) elicits short-term and long-term adaptations by the kidney. For example, when rabbits are challenged with 2-h MAc, the subsequently isolated apical/brush-border PT membrane vesicles exhibit an increase in NHE3, whereas BL vesicles exhibit an increase in NBC activity (80). Because RAc and MAc in a living animal involve changes in [CO₂], [HCO₃⁻], and H⁺ concentration, a remaining puzzle was how to distinguish the extent to which tubule cells sense changes in each of these three parameters.

In 1995, our laboratory developed a rapid mixing technique for generating temporarily out-of-equilibrium (OOE) solutions, enabling one to make isolated changes in [CO₂], [HCO₃⁻], or pH (104). The method can be easily adapted for BL or bath exposure of single, isolated perfused PTs (105). A surprising observation was that varying BL pH (pH_BL) at fixed BL [CO₂] ([CO₂]_BL) and BL [HCO₃⁻] ([HCO₃⁻]_BL) had no effect on the rate of HCO₃⁻ reabsorption (J_HCO3⁻). However, either lowering [HCO₃⁻]_BL at constant [CO₂]_BL or raising [CO₂]_BL at constant [HCO₃⁻]_BL caused a substantial increase in J_HCO3⁻ (107). The opposite changes in [HCO₃⁻]_BL or [CO₂]_BL caused J_HCO3⁻ to decrease (107). This study demonstrated that the key parameters sensed by the PT, at least over a period of ∼20 min, are [CO₂]_BL and [HCO₃⁻]_BL but not pH_BL. Further work showed that the irreversible ErbB family inhibitor PD-168393 (35 nM) or the reversible ErbB inhibitor BPIQ-1 (10 nM) each prevent the PT from responding to isolated changes in [CO₂]_BL (106).

The ErbB family comprises ErbB1/HER1/EGF receptor (EGFR), ErbB2/HER2/Neu, ErbB3/HER3, and ErbB4/HER4, which form homo- or heterodimers. Ligand-induced conformational changes within the extracellular ligand-binding domain activate cytosolic tyrosine kinase domains (26, 57, 100, 103). ErbB2 is an orphan receptor with no known ligand; its extracellular domain adopts an extended conformation, similar to that of invertebrate EGFR and ligand-bound human ErbB1 (2, 17). However, ErbB2 has an autoinhibition mechanism (25). ErbB3 lacks key residues required for appreciable catalytic activity but can elicit downstream signaling through heterodimerization with other ErbB family members (38, 78). Kinase activation initiates transphosphorylation at tyrosine residues along the COOH-terminal portions of each receptor, forming docking platforms for Src homology 2 (SH2) and phosphotyrosine (pY)-binding domains of target signaling molecules (45, 84, 101). The ErbB family regulates diverse cellular processes, including proliferation, differentiation, adhesion, and migration (1, 56). Deregulated expression and mutation of ErbB receptors occurs frequently in cancerous tissues (71, 99). The PTs of normal adult kidneys express ErbB1 and ErbB2, with a strong BL localization of ErbB1 and a weaker, more diffuse pattern for ErbB2 (49, 63, 102). Low levels of ErbB3 expression have been reported in the adult PT (65, 102). ErbB4, which regulates cell polarity and tubulogenesis during development, maintains moderate expression in the adult PT (81, 93, 102).

The purpose of the present work was to test the hypothesis that changes in [CO₂]/[HCO₃⁻] signal in part via ErbB1 and ErbB2. Our approach was to prepare suspensions of the rabbit renal cortex enriched in PTs and to expose them to a series of in-equilibrium CO₂/HCO₃⁻ solutions mimicking RAc, MAc, respiratory alkalosis (RAlk), metabolic alkalosis (MAlk), and fully compensated respiratory acidosis (cRAc), which is the same as fully compensated MAlk. We found that each acid-base disturbance produced a characteristic set of time courses (5 and 20 min) for four ErbB1 pY sites (pY845, pY992, pY1068, and pY1173) and two ErbB2 pY sites (pY1221 and pY1222).

MATERIALS AND METHODS

Solutions

For the isolation of tubules, we used a variant of DMEM-Ham's F-12 medium (DMEM-F-12, 50:50), obtained as a powder without l-glutamine or NaHCO₃ (catalog no. 90-091-PB, Corning, Cellgro Mediatech, Manassas, VA). After diluting the powder with Milli-Q water, we lowered the temperature to 0°C in an ice bath and supplemented the medium with 15 mM HEPES, 2 mM l-glutamine, 2 mM l-(+)-lactic acid, and 2 mM heptanoic acid (catalog no. 75190, Sigma-Aldrich, St. Louis, MO). After titrating the pH to 7.40 at 0°C, we added 15 mM NaHCO₃ and gassed the solution with 3% CO₂-balance air. Finally, we adjusted the osmolality to 310 ± 5 mosM with H₂O or NaCl. For the Percoll gradient, 10× PBS was made from tablets without calcium or magnesium (catalog no. 2810305, MP Biomedicals, Solon, OH).

Table 1 shows the compositions of our test solutions. CO₂-containing solutions contained the indicated percentages of CO₂, 21% O₂, and balance N₂. We created the gas mixtures in a computer-controlled gas-mixing system (Series 4000, Environics, Tolland, CT) and bubbled the solutions for 30–40 min. Nominally CO₂-free solutions (for which CO₂ is not specified in Table 1) were equilibrated with room air.

Table 1.

Compositions of test solutions

Solution	HEPES	HEPES + 20 mM Butyrate	HEPES + 20 mM Butyrate	HEPES + 1 ng/ml EGF	2.5% CO2/22 mM HCO3−	5% CO2/11 mM HCO3−	5% CO2/22 mM HCO3−	5% CO2/44 mM HCO3−	10% CO2/22 mM HCO3−	10% CO2/44 mM HCO3−
Solutes	HEPES	Butyrate (pH 7.1)	Butyrate (pH 7.4)	EGF	Respiratory alkalosis	Metabolic acidosis	Control	Metabolic alkalosis	Respiratory acidosis	Fully compensated respiratory acidosis
Nominal CO2	∼0	∼0	∼0	∼0	0.6	1.2	1.2	1.2	2.4	2.4
NaHCO3	∼0	∼0	∼0	∼0	22	11	22	44	22	44
pH*	7.4	7.1	7.4	7.4	7.7	7.4	7.4	7.7	7.1	7.4
NaCl	133	115	115	133	111	123	111	89	111	89
KCl	5	5	5	5	5	5	5	5	5	5
H3PO4	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2
CaCl2	1	1	1	1	1	1	1	1	1	1
MgSO4·7H2O	1.2	1.2	1.2	1.2	1.2	1.2	1.2	1.2	1.2	1.2
Glucose	10.5	10.5	10.5	10.5	10.5	10.5	10.5	10.5	10.5	10.5
l-Lactic acid	2	2	2	2	2	2	2	2	2	2
l-Glutamine	2	2	2	2	2	2	2	2	2	2
HEPES	15	15	15	15	15	15	15	15	15	15
EGF	0	0	0	1 ng/ml	0	0	0	0	0	0
N-butyric acid (i.e., butyrate)	0	20	20	0	0	0	0	0	0	0

Values are in mM (except for pH and where indicated). In some experiments, we added 50 nM PD-168393 to the solutions.

^*We titrated all solutions to the indicated pH at 37°C and adjusted the osmolality to 310 ± 5 mosM.

In butyrate-containing solutions, we titrated 20 mM N-butyric acid with NaOH to produce butyrate. Our laboratory has previously used 10 mM (73) or 16 mM (18, 31) butyrate/butyric acid as a surrogate for the pH_i-lowering effects of 1.5% CO₂ in Xenopus oocytes and 30 mM butyrate/butyric acid for 5% CO₂ (94). However, in PTs, BL CO₂/HCO₃⁻ produces only a transient fall in pH_i (due to CO₂ entry) followed by a sustained alkalinization (51). Also in PTs (based on results with 10 mM acetate), Na⁺-monocarboxylate cotransport at the apical membrane (which is of unknown accessibility in PT suspensions) is expected to alkalinize the cell (50). We chose 20 mM as the concentration of butyrate/butyric acid that (acting at the exposed BL membrane) is likely to reproduce at least a transient fall in pH_i created by CO₂/HCO₃⁻ in PT suspensions, perhaps (acting at the apical membrane) followed by a sustained pH_i increase.

We purchased l-(+)-lactic acid solution (30% in water by weight; L1875) and butyric acid (+99%) from Sigma-Aldrich (B103500), lyophilized purified mouse EGF1 from AbD Serotec (MorphoSys, Martinsried/Planegg, Germany), and PD-168393 (catalog no. 513033) from Calbiochem (La Jolla, CA).

Harvesting of tissue.

Female New Zealand White rabbits were used as the source of kidneys for renal PT suspensions using a method adapted from Doctor et al. (22) and others (33, 44, 72). Protocols for the housing and handling of New Zealand White rabbits were approved by the Institutional Animal Care and Use Committee of Case Western Reserve University. Rabbits under anesthesia [∼10 ml iv (5%) pentobarbital sodium in H₂O] were euthanized by exsanguination. The kidneys were removed, decapsulated, sliced into 5- to 6-mm coronal sections, and placed into ice-cold DMEM-F-12. The renal cortex was trimmed and finely diced. We digested the tissue using 0.5 mg/ml collagenase type IV (catalog no. C5138, Sigma-Aldrich) in DMEM-F-12, shaking at 225 rpm for 10 min at 37°C, harvested by centrifugation at 150 g for 1 min, and washed three times with DMEM-F-12 to remove collagenase.

Isolation of PTs

Digested cortical tubules were placed on ice and separated from one another by gentle pressure with a flat glass pestle followed by trituration through a 10-ml pipette. Once dispersed, cortical tubules were washed three times in DMEM-F-12, with harvesting each time at 150 g for 2 min. Cortical tubules were shaken gently in freshly gassed DMEM-F-12 at 4°C for 1 h, resuspended in 4 × 25-ml aliquots of ice-cold 45% Percoll solution (45 ml Percoll, 50 ml freshly gassed DMEM-F-12, and 4.5 ml of 10× PBS), and separated on a self-forming gradient by centrifugation at 25,000 g for 35 min at 4°C. The material migrated into strata (see Fig. 1A). As shown in Fig. 1, B and C, immunoblots of tubule fractions F1 to F5 isolated from individual strata showed gradual enrichment of NHE3 (present in PTs) and NBCe1 (PT marker) and depletion of AQP2 (collecting duct), podocin (podocytes in glomerulus), Na⁺-K⁺-Cl⁻ cotransporter 2 (NKCC2; thick ascending limb), and Na⁺-Cl⁻ cotransporter (NCC; distal convoluted tubule) from tubule fraction F1 (highest stratum) to tubule fraction F5 (lowest stratum). It was the two most dense, PT-enriched fractions (tubule fractions F4 and F5) that we used for experimentation. Enriched PTs were washed several times in freshly gassed, ice-cold DMEM-F-12.

Fig. 1.

Characteristics of our renal cortical tubule preparation. A: a mixture of renal tubules from the rabbit kidney cortex was applied to a 45% self-forming Percoll gradient and subjected to centrifugation at 25,000 × g for 35 min at 4°C. Tubules typically migrated into tubule fractions F1–F5, with proximal tubules (PTs) migrating in lower tubule fractions F4 and F5. B and C: characterization of tubule fractions F1–F5 by Western blot. Lysates were prepared from material isolated from tubule fractions F1–F5, resolved by SDS-PAGE, transferred to polyvinylidene difluoride (PVDF) membranes, and probed with the indicated antibodies. From top to bottom in B: aquaporin 2 (AQP2; collecting duct marker), Na⁺/H⁺ exchanger 3 (NHE3; brush-border membrane), α-subunit of Na⁺-K⁺-ATPase (basolateral transporter, renal tubule epithelium), β-actin (loading control), podocin (podocyte marker), β-actin (loading control), Na⁺-HCO₃⁻ cotransporter (NBC)e1-A (PT marker), and β-actin (loading control). From top to bottom in C: Na⁺-K⁺-Cl⁻ cotransporter 2 (NKCC2; thick ascending limb marker), β-actin (loading control), Na⁺-Cl⁻ cotransporter (NCC; distal convoluted tubule marker), and β-actin (loading control).

Challenging PTs With Acid-Base Disturbances

The enriched PTs-fractions, tubule fraction F4 and F5, from four Percoll gradients (from two kidneys) were pooled and mixed end over end to create a uniform suspension (typically of 70–100 ml), divided into aliquots of ∼5 ml each containing ∼3–8 mg protein (see Preparation of Lysate below), placed in a 50-ml, conical-bottom polypropylene tube with a screw cap (catalog no. 352070, BD Biosciences, Durham, NC), diluted to a total volume of ∼10 ml in freshly gassed DMEM-F-12, and allowed to rest at room temperature for 30 min before experimentation. Each PT aliquot was briefly pelleted at 150 g for 1 min, resuspended (after decanting) in 50 ml (virtually to the plastic screw cap) of one of our prewarmed, pregassed test solutions (Table 1), and then incubated for 5 or 20 min at 37°C in a water bath.¹ After the incubation, PTs were harvested by centrifugation at 1,000 g for 30 s, test solutions were decanted, and PT pellets were flash frozen in liquid nitrogen and stored at −80°C, pending analysis by Western blot.

Antibodies

ErbB1.

We used rabbit polyclonal anti-EGFR (no. 2232, Cell Signaling Technology, Danvers, MA).

ErbB2.

We used rabbit polyclonal anti-ErbB2 (no. 18299-1-AP, ProteinTech Group, Chicago, IL) and goat polyclonal anti-Neu C-14 (sc-31154, Santa Cruz Biotechnology, Santa Cruz, CA).

Pan-pY.

We used two antibodies that target generic pY motifs: mouse monoclonal anti-pY phospho-Tyr¹⁰⁰ (no. 9411, Cell Signaling Technology) and mouse monoclonal cocktail anti-pY 4G10 Platinum (no. 05–1050, EMD Millipore, Billerica, MA). These produced similar staining patterns and, because of our normalization procedure (see Data Analysis below), were used interchangeably.

Phosphospecific ErbB1-pY.

We used two antibodies that target pY845 of ErbB1. In early experiments, we used mouse monoclonal anti-ErbB1 pY-845 clone 12A3 (no. 04-283, EMD Millipore). Later, we found greater sensitivity, although a similar staining pattern, with rabbit polyclonal anti-ErbB1 pY-845 (no. sc-101669, Santa Cruz Biotechnology). To target pY992, we used a single rabbit polyclonal antibody (no. 40-8250, Invitrogen, Grand Island, NY). To target pY1068, we used a single rabbit polyclonal antibody (no. 324867, Calbiochem). To target pY1173, we used a single rabbit polyclonal antibody (XBP-4088, ProSci, Poway, CA).

Phosphospecific ErbB2-pY.

To target the tandem site pY1221/pY1222, we used a single rabbit polyclonal antibody (sc-101694, Santa Cruz Biotechnology).

Actin.

To normalize sample loading on protein gels, we used mouse monoclonal anti-β-actin antibody clone AC-15 (no. A1978, Sigma-Aldrich).

Tubule fraction markers.

We used AQP2 rabbit polyclonal antibody (sc-28629, Santa Cruz Biotechnology), NHE3 mouse monoclonal antibody (3H3) to opossum NHE3 (kindly provided by Dr. Daniel Biemesderfer, Yale University), Na⁺-K⁺-ATPase mouse monoclonal antibody (464.6) to the α₁-subunit (ab7671, Abcam, Cambridge, MA), podocin rabbit polyclonal antibody (sc-21009, Santa Cruz Biotechnology), NBCe1 rabbit polyclonal antibody [NBC-3 (75)], NKCC2 rabbit polyclonal antibody, NKCC2 rabbit polyclonal antibody (sc-133823, Santa Cruz Biotechnology), and NCC rabbit polyclonal antibody (AB3553, Millipore).

Secondary antibodies.

We used a goat affinity-purified antibody to rabbit IgG, horseradish peroxidase (HRP) conjugated (AP132P, Millipore), and goat affinity-purified antibody to mouse IgG, HRP conjugated (no. 55563, MP Biomedicals).

Preparation of Lysate

PT pellets were thawed and rapidly placed on ice before the addition of ice-cold lysis buffer of the following composition (in mM): 25 HEPES (pH 7.50), 100 NaCl, 50 NaF, 10 Na-pyrophosphate (Na₄P₂O₇·10 H₂O), 1 EDTA, and 1% Nonidet P-40 (AB01425, American Bioanalytical, Natick, MA), to which we added protease inhibitor cocktail (Sigma-Aldrich) at 1:25 (vol/vol). Using a 21-gauge needle and syringe, we sheared and rapidly lysed the PT pellet. After thorough homogenization, we added 2 mM Na₃VO₄ to each sample to stabilize pY sites, and the cell debris was removed by centrifugation. Finally, a Pierce BCA protein assay (catalog no. 23227, Thermo Scientific) was performed to determine total protein.

Protein Gels and Western Blot Analysis

Equal amounts of protein were loaded onto NuPAGE Novex 4–12% Bis-Tris gels (NP0322 and NP0366, Invitrogen), resolved by electrophoresis, electrophoretically transferred to a polyvinylidene difluoride membrane, and blocked in Tris-buffered saline plus Tween 20 [containing 10 mM Tris·HCl (pH 7.50), 150 mM NaCl, and 0.01% Tween 20] supplemented with 5% milk and 2 mM Na₃VO₄ at room temperature for 1 h. We probed overnight with an anti-pY primary antibody (see above). After washing well to remove unbound primary antibodies, we incubated blots with secondary antibodies linked to HRP (see above), washed, and detected immunoreactive bands with Pierce ECL2 reagent (catalog no. 80196, Thermo Scientific) on X-ray film or using a FluorChem E gel-documentation device (Protein Simple, San Jose, CA). Before reprobing (e.g., with antibodies to actin, ErbB1, or ErbB2), we stripped blots by an incubation in Pierce Restore Plus Western blot stripping buffer (catalog no. 46430, Thermo Scientific) at 37°C for 30 min.

Image Acquisition

For densitometry, we scanned films as 16-bit grayscale images using an Epson Perfection V500 photo scanner at a positive film setting at high resolution (600–800 dpi). The densitometry of β-actin on each blot was used as a loading control with which to normalize pY immunoreactivity. We performed all densitometry analyses, both after scanning film or imaging with the FluorChem E, using ImageJ software (National Institutes of Health).

Data Analysis

After densitometry analysis, we normalized the intensity of pY immunoreactivity to the intensity of actin immunoreactivity in the same lane. In the analyses shown in Figs. 4A, 5A, 7A, 8A, 10A, 11A, and 12, A and C, we divided the normalized pY-to-actin ratio for any “solution treatment” by the normalized pY-to-actin ration after nominally CO₂/HCO₃⁻-free HEPES treatment. In the analyses shown in Figs. 4B, 5B, 7B, 8B, 10B, 11B, and 12, B and D, we divided the normalized pY-to-actin ratio for each solution treatment by the normalized pY-to-actin ratio after our control (Ctrl) 5% CO₂/22 mM HCO₃⁻ treatment. In the analyses shown in Figs. 4C, 5C, 7C, 8C, 10C, and 11C, we divided the normalized pY-to-actin ratio for each solution treatment in the presence of PD-168393 by the normalized pY-to-actin ratio after treatment in the same solution without PD-168393. For statistical analysis of each data set, we log₂ transformed each point (so that that a twofold increase is treated the same as a twofold decrease) and discarded the ∼5% of data points that were >2 SD away from the mean of each set, as previously described by others (67). We performed one-tailed, paired t-tests using Excel 2010.

Fig. 4.

Quantification of pY immunoreactivity after solution treatments of 5 min. Shown are the results of densitometry of blots similar to those shown in Fig. 3. For each treatment, pan-pY was first normalized to actin. A: ratios of pan-pY to actin (pan-pY/actin) for nominally CO₂/HCO₃⁻-free treatments, normalized to pan-pY/actin for the nominally CO₂/HCO₃⁻-free HEPES treatment. B: pan-pY/actin for nominally CO₂/HCO₃⁻-free HEPES and CO₂/HCO₃⁻ solutions, normalized to pan-pY/actin for our Ctrl treatment (5% CO₂/22 mM HCO₃⁻). C: pan-pY/actin for treatments like those in B but with and without the ErbB inhibitor PD-168393, normalized to pan-pY/actin for the equivalent treatment without PD-168393. Sample sizes (n) are shown beneath each graph. †Bar for which all constituent data points are <0 (or >0) but not statistically different from 0. #Bar for which P < 0.05 (compared with 0 by a one-tailed, paired t-test); in the text, we state that the parameter “tends or tended” to rise/fall. *Bar that is significantly different from 0, even after the very conservative Bonferroni correction (P < 0.05 divided by the number of treatment solutions considered); in the text, we refer to the change as “significant.” ‡Pair of bars for which P < 0.05 (compared with each other by a one-tailed, paired t-test).

Fig. 5.

Quantification of pY immunoreactivity after solution treatments of 20 min. Except for the duration of the treatments, this figure is identical to Fig. 4. Sample sizes (n) are shown beneath each graph. #Bar for which P < 0.05 (compared with 0 by a one-tailed, paired t-test); in the text, we state that the parameter “tends or tended” to rise/fall. *Bar that is significantly different from 0, even after the very conservative Bonferroni correction (P < 0.05 divided by the number of treatment solutions considered); in the text, we refer to the change as “significant.” ‡Pair of bars for which P < 0.05 (compared with each other by a one-tailed, paired t-test).

Fig. 7.

Quantification of ErbB1 immunoreactivity at individual pY sites after solution treatments of 5 min. These are the results of densitometry of blots similar to those shown in Fig. 6. For each treatment, pY-specific immunoreactivity at each ErbB1 site was first normalized to actin. A: ratios of ErbB1-pY to actin (ErbB1-pY/actin) for nominally CO₂/HCO₃⁻-free treatments, normalized to ErbB1-pY/actin for the nominally CO₂/HCO₃⁻-free HEPES treatment. B: ErbB1-pY/actin for nominally CO₂/HCO₃⁻-free HEPES and CO₂/HCO₃⁻ solutions, normalized to ErbB1-pY/actin for our Ctrl treatment (5% CO₂/22 mM HCO₃⁻). C: ErbB1-pY/actin for treatments like those in B but with and without the ErbB inhibitor PD-168393, normalized to ErbB1-pY/actin for the equivalent treatment without PD-168393. Sample sizes (n) are shown beneath each graph. †Bar for which all constituent data points are <0 (or >0) but not statistically different from 0. #Bar for which P < 0.05 (compared with 0 by a one-tailed, paired t-test); in the text, we state that the parameter “tends or tended” to rise/fall. *Bar that is significantly different from 0, even after the very conservative Bonferroni correction (P < 0.05 divided by the number of treatment solutions considered); in the text, we refer to the change as “significant.”

Fig. 8.

Quantification of ErbB1 immunoreactivity at individual pY sites after solution treatments of 20 min. Except for the duration of the treatments, this figure is identical to Fig. 7. Sample sizes (n) are shown beneath each graph. †Bar for which all constituent data points are <0 (or >0) but not statistically different from 0. #Bar for which P < 0.05 (compared with 0 by a one-tailed, paired t-test); in the text, we state that the parameter “tends or tended” to rise/fall. *Bar that is significantly different from 0, even after the very conservative Bonferroni correction (P < 0.05 divided by the number of treatment solutions considered); in the text, we refer to the change as “significant.”

Fig. 10.

Quantification of ErbB2 immunoreactivity at pY1221/pY1222 after solution treatments of 5 min. These are the results of densitometry of blots similar to those shown in Fig. 9. For each treatment, pY-specific immunoreactivity at ErbB2 pY1221/pY1222 was first normalized to actin. A: ratios of ErbB2-pY1221/pY1222 to actin [(ErbB2-pY1221/2)/actin] for nominally CO₂/HCO₃⁻-free treatments normalized to (ErbB2-pY1221/2)/actin for the nominally CO₂/HCO₃⁻-free HEPES treatment. B: ratios of pY1221/pY1222 to actin [(pY1221/2)/actin] for nominally CO₂/HCO₃⁻-free HEPES and CO₂/HCO₃⁻ solutions, normalized to (pY1221/2)/actin for our Ctrl treatment (5% CO₂/22 mM HCO₃⁻). C: (ErbB2-pY1221/2)/actin for treatments like those in B but with and without the ErbB inhibitor PD-168393, normalized to (pY1221/2)/actin for the equivalent treatment without PD-168393. Sample sizes (n) are shown beneath each graph. †Bar for which all constituent data points are <0 (or >0) but not statistically different from 0. #Bar for which P < 0.05 (compared with 0 by a one-tailed, paired t-test); in the text, we state that the parameter “tends or tended” to rise/fall. *Bar that is significantly different from 0, even after the very conservative Bonferroni correction (P < 0.05 divided by the number of treatment solutions considered); in the text, we refer to the change as “significant.”

Fig. 11.

Quantification of ErbB2 immunoreactivity at pY1221/pY1222 after solution treatments of 20 min. Except for the duration of the treatments, this figure is identical to Fig. 10. Sample sizes (n) are shown beneath each graph. †Bar for which all constituent data points are <0 (or >0) but not statistically different from 0. #Bar for which P < 0.05 (compared with 0 by a one-tailed, paired t-test); in the text, we state that the parameter “tends or tended” to rise/fall. *Bar that is significantly different from 0, even after the very conservative Bonferroni correction (P < 0.05 divided by the number of treatment solutions considered); in the text, we refer to the change as “significant.”

Fig. 12.

Quantification of pseudo-pan-ErbB1/2-pY (ErbB-pY′) after solution treatments of 5 and 20 min. The responses of ErbB1-pY845, ErbB1-pY992, ErbB1-pY1068, ErbB1-pY1173 (Figs. 7 and 8), and ErbB2-pY1221/pY1222 (Figs. 10 and 11) were summed to provide an approximation of the pan-ErbB1/2-pY response to the various acid-base disturbances. A: ErbB-pY′ at 5 min for nominally CO₂/HCO₃⁻-free treatments, normalized to ErbB-pY′ for the nominally CO₂/HCO₃⁻-free HEPES treatment at 5 min. B: ErbB-pY′ at 5 min for nominally CO₂/HCO₃⁻-free HEPES and CO₂/HCO₃⁻ solutions, normalized to ErbB-pY′ for our Ctrl treatment (5% CO₂/22 mM HCO₃⁻). C: ErbB-pY′ at 20 min for nominally CO₂/HCO₃⁻-free treatments, normalized to ErbB-pY′ for the HEPES treatment at 20 min. D: ErbB-pY′ at 20 min for HEPES and CO₂/HCO₃⁻ solutions, normalized to ErbB-pY′ for our control treatment (5% CO₂/22 mM HCO₃⁻). Sample sizes (n) are shown beneath each graph. #Bar for which P < 0.05 (compared with 0 by a one-tailed, paired t-test); in the text, we state that the parameter “tends or tended” to rise/fall. *Bar that is significantly different from 0, even after the very conservative Bonferroni correction (P < 0.05 divided by the number of treatment solutions considered); in the text, we refer to the change as “significant.”

RESULTS

Detection of Major pY Bands

Figure 2A, left, shows an anti-pan-pY Western blot of protein prepared from an aliquot of a PT suspension that had been transferred from DMEM-F-12 and then incubated in our Ctrl 5% CO₂/22 mM HCO₃⁻/pH 7.4 solution for 5 min. In this sample, we detected prominent pan-pY bands with molecular weights (MWs) of ∼180, ∼160, ∼130, and ∼120 kDa. We also saw a dim band at ∼65 kDa and an inconsistent band (present in this blot but not all others) at ∼55 kDa. The bands below 150 kDa do not appear to be related to ErbB1 or ErbB2 because, unlike the ∼180 and ∼160-kDa bands, as shown below, the pan-pY signals for bands < 150 kDa seldom responded to acid-base disturbances or the inhibitor PD-168393. For this reason, we did not further pursue these lower MW bands.

Fig. 2.

Western blots of rabbit renal PT preparations. Tubules were exposed to our control (Ctrl) 5% CO₂/22 mM HCO₃⁻ solution for 5 min, and aliquots of extracted protein were resolved by SDS-PAGE, transferred to PVDF membranes, probed with the indicated phosphotyrosine (pY) antibodies, stripped, and then reprobed with an antibody to ErbB1 or ErbB2. A: blots were probed first with anti-pan-pY (left) and reprobed with anti-ErbB1 (right). B: blots were probed first with anti-ErbB1/pY845. C: blots were probed first with anti-ErbB1/pY992. D: blots were probed first with anti-ErbB1/pY1068. E: blots were probed first with anti-ErbB1/pY1173. F: blots were probed first with anti-ErbB2/pY1221/pY1222.

The same blot, reprobed with an anti-ErbB1 antibody, Fig. 2A, right, shows that ErbB1 immunoreactivity was located in the vicinity of both the 180- and 160-kDa pan-pY bands shown in Fig. 2A, left; thus, ErbB1-pY likely contributes to these two bands. Indeed, similar blots probed with antibodies directed against specific pY sites in the COOH-terminal tail of ErbB1 (pY845, pY992, pY1068, and pY1173 in Fig. 2, B–E) showed ErbB1-pY immunoreactivity in this MW range. The ErbB2-pY1221/2 immunoreactivity was also in the 160- to 180-kDa range (Fig. 2F).

In addition to the bands at ∼180 and ∼160 kDa, our phosphospecific antibodies sometimes showed reactivity between 100 and 120 kDa, often with ErbB1-pY992 (Fig. 2C), rarely with ErbB1-pY1068 (not present in Fig. 2D), often with ErbB1-pY1173 (Fig. 2E), and often with ErbB2-pY1221/2 (not present in Fig. 2F). The ErbB1 antibody, raised against an epitope surrounding Y1068, never reacted with any of these bands at 100–120 kDa. Because the appearance of these bands was inconsistent, their identity is unknown, and because these fragments cannot be part of an intact ErbB receptor, we excluded them from further analysis. We also excluded from further analysis high MW (>250 kDa) material detected with ErbB2-pY1221/2 (Fig. 2F, left) but that was less detectable when probing total ErbB2 (Fig. 2F, right). Thus, we focused our attention on pY immunoreactivity at 180 kDa (pY_180kDa) and 160 kDa (pY_160kDa).

The ratio of pY_180kDa-to-pY_160kDa intensities varied among rabbits and PT treatments. However, in the majority of cases, as evidenced by the example blots shown in Figs. 2A and 3, pY_180kDa was greater in intensity than pY_160kDa. For example, the ratio pY_180kDa to pY_160kDa was 1.7 ± 0.2 after a 5-min exposure to our Ctrl solution (5% CO₂/22 mM HCO₃⁻/pH 7.4, n = 12 rabbits).

Fig. 3.

Pan-pY immunoreactivity at 180 and 160 kDa in rabbit PT preparations subjected to acid-base disturbances (5 and 20 min). Tubules were exposed to one of several treatment solutions, and protein was subjected to Western blot analysis. For each lane, we show (from top to bottom) the pan-pY immunoreactivity from tubules that were exposed to treatments for 5 min and a reprobe of the same blot with an anti-actin antibody (top), a summary of the treatment conditions (table in middle), and the pan-pY immunoreactivity from tubules that were exposed to treatments for 20 min and a reprobe of the same blot with an anti-actin antibody (bottom). MAc, metabolic acidosis; MAlk, metabolic alkalosis; RAc, respiratory acidosis; RAlk, respiratory alkalosis; cRAc, fully compensated respiratory acidosis.

The Pan-pY Profile of PT Suspensions Varies with Acid-Base Disturbances and PD-168393 Treatment

General approach.

We divided a PT suspension into multiple equivalent aliquots and exposed each to one of a series of solutions designed to mimic diverse acid-base disturbances (see Table 1) for either 5 or 20 min in the presence or absence of the ErbB inhibitor PD-168393. Because of physical limitations on the amount of PT material that we could isolate from a single rabbit and on the number of samples that we could process simultaneously, we performed our study in a series of overlapping assays. Figure 3 shows Western blots of material isolated from two such assays that, between them, covered the full array of solutions shown in Table 1.

Note that each assay set included one sample treated with 1) our nominally CO₂/HCO₃⁻-free/pH 7.4 solution (HEPES) and another sample treated with 2) our Ctrl 5% CO₂/22 mM HCO₃⁻/pH 7.4 solution.

Five-minute treatment.

Figure 4A shows the pan-pY signal intensity at 180 and 160 kDa (measured from blots such as those shown in Fig. 3) for nominally CO₂/HCO₃⁻-free treatments, normalized first to the intensity of actin immunoreactivity in the same sample and normalized second to the pan-pY signal of the HEPES-treated sample (i.e., HEPES = 1.0). Thus, these data represent the CO₂/HCO₃⁻-independent effects of nominally CO₂/HCO₃⁻-free EGF or butyrate challenges on pY_180kDa and pY_160kDa signals.

The left bar of each pair in Fig. 4A represents the normalized pY_180kDa signal and the right bar of each pair represents the normalized pY_160kDa signal. Analysis of the data shown in Fig. 4A revealed the following:

1. Each treatment affects the intensities of pY_180kDa and pY_160kDa to similar extents (results of a one-tailed, paired t-test showed no differences between pY_180kDa and pY_160kDa in each pair, P > 0.05).

2. After 5 min of nominally CO₂/HCO₃⁻-free EGF treatment (which ought to activate ErbB1 and ErbB2), pY_180kDa and pY_160kDa intensities tend to increase, consistent with the hypothesis that ErbB activation contributes to pY_180kDa and pY_160kDa intensities in nominally CO₂/HCO₃⁻-free solutions (HEPES vs. EGF in Fig. 4A).²

3. After 5 min of nominally CO₂/HCO₃⁻-free butyrate treatment (which ought to cause pH_i to fall) accompanied by a decrease in pH_o to 7.1, pY_180kDa tends to fall, mirrored by pY_160kDa, which is consistently reduced but with a high variance (HEPES vs. butyrate, pH 7.1, in Fig. 4A).

Figure 4B shows pan-pY signal intensities (measured from blots such as those shown in Fig. 3) for HEPES and Ctrl CO₂/HCO₃⁻ conditions as well as five acid-base disturbances imposed in the presence of CO₂/HCO₃⁻. In each case, we normalized to the intensity of actin immunoreactivity from each sample and then normalized to the pan-pY signal of the Ctrl-treated sample (i.e., Ctrl = 1.0). Analysis of the data shown in Fig. 4B revealed that:

1. Nominally CO₂/HCO₃⁻-free HEPES treatment, which tends to reduce pY_180kDa, is the only condition that influences the intensity of pY_180kDa at 5 min (i.e., HEPES is less than Ctrl in Fig. 4B).

2. MAlk is the only treatment that significantly influences pY_160kDa at 5 min (i.e., MAlk/160 is significantly greater than Ctrl/160 in Fig. 4B).²

3. MAlk enhances pY_160kDa more than pY_180kDa (i.e., MAlk/180 vs. MAlk/160 in Fig. 4B).

Figure 4C shows pan-pY signal intensities (measured from blots such as those shown in Fig. 3) for a subset of experiments that we performed with and without the ErbB inhibitor PD-168393. For each sample, we normalized to the intensity of actin immunoreactivity and then normalized to the PD-168393-free signal (i.e., PD-168393 untreated = 1.0 for each solution type). Thus, these data represent the PD-168393 sensitivity of each pY signal. Analysis of the data shown in Fig. 4C revealed that:

1. PD-168393 significantly reduces the intensity of pY_160kDa but not pY_180kDa in HEPES.

2. PD-168393 reduces both pY_180kDa and pY_160kDa in RAlk, MAc, Ctrl, MAlk, and RAc. This observation is consistent with the hypothesis that ErbB1/2 activation can contribute to pY_180kDa and pY_160kDa intensity in the presence of CO₂/HCO₃⁻.²

3. PD-168393 does not significantly influence either pY_180kDa or pY_160kDa in cRAc.

4. With the exception of RAlk, the inhibitory effect of PD-168393 is greater for pY_160kDa than for pY_180kDa. For RAlk, pY_180kDa and pY_160kDa are equally sensitive to inhibition.

In summary, a 5-min treatment with nominally CO₂/HCO₃⁻-free EGF tends to raise both pY_180kDa and pY_160kDa (Fig. 4A), whereas after a 5-min treatment with nominally CO₂/HCO₃⁻-free butyrate (pH 7.1), the simultaneous decrease in pH_o and pH_i, pY_180kDa tends to fall. MAlk raises pY_160kDa (Fig. 4B). Finally, PD-168393 generally lowers both pY_180kDa and pY_160kDa (Fig. 4C).

Twenty-minute treatment.

Figure 5, A–C, shows data equivalent to those shown in Fig. 4, A–C but for a 20-min exposure rather than a 5-min exposure to each treatment solution. We provide a comparison of the 5- and 20-min data in the discussion (see Fig. 14).

Fig. 14.

Time-dependent changes of ErbB-pY′ and pan-pY signals at ∼180 and ∼160 kDa for each acid-base disturbance. Each line represents the time course of the log₂-transformed parameter for one of the five acid-base disturbances. A: ErbB-pY′. B: pan-pY180kDa. C: pan-pY160kDa. In these analyses, we assume that the parameter value at 0 min (i.e., before imposition of one of the acid-base disturbances) is equivalent to the parameter value for our Ctrl treatment (5% CO₂/22 mM HCO₃⁻) at 5 or 20 min, where, by definition, log₂(pY/actin) = 0. Sample sizes (n) are shown in Figs. 4B, 5B, 12B and 12D. The upward-sloping or downward-sloping arrows indicate a statistically significant difference between the corresponding values at 5 and 20 min. #Bar for which P < 0.05 (compared with 0 by a one-tailed, paired t-test); in the text, we state that the parameter “tends or tended” to rise/fall. *Bar that is significantly different from 0, even after the very conservative Bonferroni correction (P < 0.05 divided by the number of treatment solutions considered); in the text, we refer to the change as “significant.”

Figure 5A shows pY_180kDa and pY_160kDa signal intensities (measured from blots such as those shown in Fig. 3) after PT exposure to nominally CO₂/HCO₃⁻-free solutions normalized to the HEPES (pH 7.4) solution. Analysis of these data revealed that:

1. As with the 5-min treatment (Fig. 4A), nominally CO₂/HCO₃⁻-free EGF enhances the intensity of both pY_180kDa and pY_160kDa (i.e., EGF is significantly greater than HEPES in Fig. 5A); however, different from the effect of a 5-min exposure, at 20 min, pY_160kDa is enhanced to a greater extent than pY_180kDa (i.e., EGF/180 vs. EGF/160 in Fig. 5A).²

2. In contrast to the 5-min treatment (Fig. 4A), nominally CO₂/HCO₃⁻-free butyrate at pH_o 7.1 reduces pY_180kDa but not pY_160kDa (i.e., at pY_180kDa, butyrate at pH_o 7.1 is significantly less than HEPES in Fig. 5A).

Figure 5B shows pY_180kDa and pY_160kDa signal intensities (measured from blots such as those shown in Fig. 3) after a 20-min exposure of PT preparations to nominally CO₂/HCO₃⁻-free HEPES or various CO₂/HCO₃⁻ solutions. In each case, we normalized to the Ctrl solution. Analysis of these data revealed that:

1. MAlk after 20 min, in contrast to the situation at 5 min (Fig. 4B), where MAlk is the only CO₂/HCO₃⁻ treatment that elicits an effect (an increase in pY_160kDa), tends to decrease pY_180kDa and has no effect on pY_160kDa.²

2. RAlk reduces pY_180kDa and also tends to reduce pY_160kDa (i.e., RAlk/180 vs. Ctrl/180 and RAlk/160 vs. Ctrl/160 in Fig. 5B).²

3. In contrast, neither MAc nor RAc (the two acidosis treatments) influences either pY_180kDa or pY_160kDa (i.e., MAc/180 or RAc/180 vs. Ctrl/180 and MAc/160 or RAc/160 vs. Ctrl/160 in Fig. 5B).

4. cRAc tends to enhance both pY_180kDa and pY_160kDa (i.e., cRAc/180 vs. Ctrl/180 and cRAc/160 vs. Ctrl/160 in Fig. 5B).

5. Nominally CO₂/HCO₃⁻-free HEPES after 20 min, in contrast to the 5-min exposure (Fig. 4B), reduces not only pY_180kDa but also pY_160kDa (i.e., HEPES/180 vs. Ctrl/180 and HEPES/160 vs. Ctrl/160 in Fig. 5B).

Figure 5C shows, for a subset of the data shown in Fig. 5B, the PD-168393 sensitivities of pY_180kDa and pY_160kDa after 20 min of treatment. Analysis of these data revealed that:

1. PD-168393 reduces both pY_180kDa and pY_160kDa for all treatments.²

2. As was the trend at 5 min (Fig. 4C), the inhibition by PD-168393 at 20 min tends to be greater for pY_160kDa than for pY_180kDa.

3. The inhibition by PD-168393 is generally more robust at 20 min (Fig. 5C) than at 5 min (Fig. 4C).

In summary, a 20-min treatment with nominally CO₂/HCO₃⁻-free butyrate (pH 7.1) reduces pY_180kDa (Fig. 5A, butyrate at pH_o 7.1). Nominally CO₂/HCO₃⁻-free EGF enhances both pY_180kDa and pY_160kDa (Fig. 5A). Moreover, PD-168393 generally reduces the two pY signals across the treatments (Fig. 5C). These trends at 20 min are similar to (but tend to be more robust than) those at 5 min. The 20-min data differ most strikingly from the 5-min data for acid-base disturbances in the presence of CO₂/HCO₃⁻. At 5 min (Fig. 4B), only MAlk/160 shows a response. At 20 min (Fig. 5B), pY_180kDa and pY_160kDa exhibit a significant decrease or tend to decrease at low percent CO₂ (HEPES and RAlk) and tend to increase at elevated percent CO₂ and [HCO₃⁻] (HEPES vs. Ctrl vs. cRAc) independent of pH_o.

For ErbB1, the Time Course of pY at Specific Sites Depends on the Acid-Base Status and Inhibition With PD-168393

General approach.

We hypothesized that the pY180kDa and pY160kDa signals shown in Figs. 4 and 5 reflect, in large measure, changes in the pY status of sites on ErbB1 and ErbB2. In parallel with the pan-pY analyses shown in Figs. 4 and 5, we also analyzed tyrosine phosphorylation on four specific tyrosine residues in ErbB1-Y845 in the activation loop of the kinase domain as well as Y992, Y1068, and Y1173 in the COOH-terminal tail of ErbB1. Figure 6 shows representative examples of Western blots probed with phosphospecific antibodies at 5 and 20 min. (Below, we present a similar analysis for the Y1221/Y1222 doublet in the COOH-terminal tail of ErbB2.)

Fig. 6.

ErbB1 immunoreactivity at individual pY sites in rabbit PT preparations subjected to acid-base disturbances (5 and 20 min). Tubules were exposed to one of several treatment solutions, and protein was subjected to Western blot analysis. For each lane, we show (from top to bottom) pY845 at 5 and 20 min, pY992 at 5 and 20 min, pY1068 at 5 and 20 min, pY1173 at 5 and 20 min, and a summary of the treatment conditions. Note that the blots on the right repeat some of the conditions on the left. We do not show the reprobes of the blots for actin.

Five-minute treatment.

Figure 7A shows site-specific pY signal intensities (measured from blots such as those shown in Fig. 6) for nominally CO₂/HCO₃⁻-free treatments. We normalized these first to the intensity of actin immunoreactivity in the same lane and then normalized again to the corresponding normalized signal from the HEPES-treated sample (i.e., HEPES = 1.0). Thus, these data represent the effect of nominally CO₂/HCO₃⁻-independent acid-base disturbances on the pY status of each of the four ErbB1 tyrosine residues. Analysis of the data shown in Fig. 7A revealed that:

1. After 5 min of nominally CO₂/HCO₃⁻-free EGF treatment, pY immunoreactivity at all four sites tends to increase, as expected (HEPES vs. EGF in Fig. 7A); pY845 exhibits the most robust increase.²

2. Nominally CO₂/HCO₃⁻-free butyrate treatment at pH_o 7.4 tends to increase pY1173 [HEPES/Y1173 vs. butyrate (7.4/1173) in Fig. 7A] but does not affect pY845, pY992, or pY1068.

3. Nominally CO₂/HCO₃⁻-free butyrate at pH_o 7.1 does not alter pY845 or pY992 [HEPES vs. butyrate (pH 7.1) in Fig. 7A]. However, we note that this treatment consistently reduced pY1068 and pY1173, although with too great a variance for statistical significance.

Figure 7B shows site-specific pY signal intensities (measured from blots such as those shown in Fig. 6) for nominally CO₂/HCO₃⁻-free HEPES plus acid-base disturbances in the presence of CO₂/HCO₃⁻. We normalized values to actin and the Ctrl sample, as shown in Fig. 5B. Analysis of the data shown in Fig. 7B revealed that the four ErbB1 sites responded differently to each acid-base disturbances at 5 min:

1. The nominally CO₂/HCO₃⁻-free HEPES treatment tends to reduce only pY1173 (HEPES/Y1173 vs. Ctrl/Y1173 in Fig. 7B).

2. RAlk tends to reduce only pY845 (RAlk/Y845 vs. Ctrl/Y845 in Fig. 7B).

3. MAc tends to reduce only pY1068 (MAc/Y1068 vs. Ctrl/Y1068 in Fig. 7B).

4. MAlk tends to increase only pY845 (HEPES/Y845 vs. Ctrl/Y845 in Fig. 7B).

5. RAc did not cause an alteration in the pY status of any of the four sites (RAc vs. Ctrl in Fig. 7B).²

6. cRAc tends to reduce only pY992 (HEPES/Y992 vs. Ctrl/Y992 in Fig. 7B).

In summary, acid-base disturbances at 5 min tend to produce modest pY effects at specific sites.

Figure 7C shows, for a subset of the data shown in Fig. 7B, the PD-168393 sensitivity of pY at the four ErbB1 residues after 5-min treatments. The inhibition by the drug is reflected by the extent to which the values are <1.0. Analysis of these data revealed that:

1. For the nominally CO₂/HCO₃⁻-free HEPES treatment, the only significant effect of PD-168393 is a reduction in pY1068.

2. For RAlk, MAc, Ctrl, MAlk, and RAc, PD-168393 reduces each of the four pY signals.²

3. For cRAc, the only significant effect of the drug is a reduction in pY845.

In summary, and not surprisingly, for nearly all of the acid-base conditions, PD-168393 reduces the phosphorylation at all four pY sites. In other words, over a range of acid-base disturbances, PD-168393-sensitive kinases are active and required to maintain pY levels. However, under conditions of cRAc, a PD-168393-sensitive kinase is less critical to maintain the phosphorylation of Y992; cRAc at 5 min has eliminated that kinase activity.

Twenty-minute treatment.

Figure 8, A–C, shows equivalent data to those shown in Fig. 7, A–C, but for a 20-min exposure rather than a 5-min exposure to each treatment solution. We provide a comparison of the 5- and 20-min data in the discussion (see ErbB1 data in Fig. 13).

Fig. 13.

Time-dependent changes of ErbB1/2 immunoreactivity at individual pY sites. A–E: log₂-transformed data for the four individual ErbB1 sites and one compound ErbB2 site, normalized to our Ctrl treatment (5% CO₂/22 mM HCO₃⁻), were plotted as a function of time and grouped according to acid-base disturbance. A: MAc. B: MAlk. C: RAc. D: RAlk. E: cRAc. F–J: the same data grouped according to pY site. F: ErbB1-pY845. G: ErbB1-pY992. H: ErbB1-pY1068. I: ErbB1-pY1173. J: ErbB2-pY1221/pY1222. In these analyses, we assume that the pY status at 0 min (i.e., before incubation with each treatment solution) is equivalent to the pY status of that site at 5 or 20 min under Ctrl conditions, where, by definition, log₂(pY/actin) = 0. Sample sizes (n) are shown in Figs. 7B, 8B, 10B, and 11B. The upward-sloping or downward-sloping arrows indicate a statistically significant difference between the corresponding values at 5 and 20 min. #Bar for which P < 0.05 (compared with 0 by a one-tailed, paired t-test); in the text, we state that the parameter “tends or tended” to rise/fall. *Bar that is significantly different from 0, even after the very conservative Bonferroni correction (P < 0.05 divided by the number of treatment solutions considered); in the text, we refer to the change as “significant.”

Figure 8A shows the effect of nominally CO₂/HCO₃⁻-free treatments on each of the four ErbB1 tyrosine residues (measured from blots such as those shown in Fig. 6). Analysis of these data revealed trends similar to those found in the 5-min data shown in Fig. 7A:

1. Nominally CO₂/HCO₃⁻-free EGF tends to increase pY immunoreactivity at pY1173 and significantly at pY992 and pY1068 (HEPES vs. EGF in Fig. 8A).²

2. Nominally CO₂/HCO₃⁻-free butyrate treatment at pH_o 7.4 does not significantly affect the pY status of any of the four sites.

3. Nominally CO₂/HCO₃⁻-free butyrate treatment at pH_o 7.1 significantly reduces pY1068 and tends to reduce pY1173 (HEPES vs. butyrate at pH_o 7.1 in Fig. 8A).

Figure 8B shows site-specific pY signal intensities (measured from blots such as those shown in Fig. 6) after 20-min treatments in nominally CO₂/HCO₃⁻-free HEPES or CO₂/HCO₃⁻. Analysis of the data shown in Fig. 8B revealed that the pY status evolved since the 5-min mark (see Fig. 7B):

1. The nominally CO₂/HCO₃⁻-free HEPES treatment significantly reduces pY845, and pY992 and tends to reduce pY1068 but not pY1173 (HEPES vs. Ctrl in Fig. 8B), in contrast to 5 min, where pY1173 tends to be reduced (Fig. 7B).

2. RAlk tends to reduce only pY992 (RAlk/Y992 vs. Ctrl/Y992 in Fig. 8B).

3. MAc does not alter the pY status of any of the four sites (MAc vs. Ctrl in Fig. 8B), although it had tended to reduce pY1068 at 5 min.

4. MAlk treatment tends to reduce only pY1068 (MAlk/1068 vs. Ctrl/1068 in Fig. 8B), a pattern different from 5 min.

5. RAc and cRAc now increase or tend to increase pY845 and pY1173, with cRAc-pY845 being the most robust change (cRAc/Y845 vs. Ctrl/Y845 in Fig. 8B).²

In summary, acid-base disturbances at 20 min produce specific effects on pY levels that are different from those at 5 min and are generally more robust.

Fig. 8C shows, for a subset of the data shown in Fig. 8B, the PD-168393 sensitivity of pY at the four ErbB1 residues after 20-min treatments. Analysis of these data revealed that:

1. For almost all conditions, PD-168393 robustly reduces pY levels at all four sites; HEPES-pY1173 and RAlk-pY845 may be special cases, either because of a small number of successful assays or a large variance.²

2. HEPES-pY992 is not significantly altered by PD-168393. Thus, the PD-168393-sensitive kinases responsible for phosphorylating Y992 have become relatively inactive after 20 min in HEPES.

In summary, as at 5 min, the four pY sites exhibit largely independent responses to acid-base disturbances, although the patterns are different from those at 5 min. The ability of PD-168393 to reduce pY levels becomes more robust by 20 min, especially for cRAc (where the effect of PD-168393 was less robust at 5 min).

For ErbB2, the Time Course of pY1221/2 Depends on the Acid-Base Status and Inhibition With PD-168393

General approach.

In parallel with the pan-pY and ErbB1-pY analyses shown in Figs. 4–8, we assessed tyrosine phosphorylation for the Y1221/2 compound site of ErbB2. Figure 9 shows representative Western blots probed at 5 and 20 min with a pY1221/2 phosphospecific antibody.

Fig. 9.

ErbB2 immunoreactivity at pY1221/pY1222 in rabbit PT preparations subjected to acid-base disturbances (5 and 20 min). Tubules were exposed to one of several treatment solutions, and protein was subjected to Western blot analysis. For each lane, we show (from top to bottom) the compound pY1221/pY1222 site at 5 and 20 min and a summary of the treatment conditions. Note that the blots on the right repeat some of the conditions on the left. We do not show the reprobes of the blots for actin.

Five-minute treatment.

Figure 10, A–C, shows, for ErbB2 pY1221/pY1222, data (measured from blots such as those shown in Fig. 9) similar to those shown for ErbB1-specific pY sites in Fig. 7, A–C. Note that pY1221/pY1222 significantly rose with nominally CO₂/HCO₃⁻-free EGF treatment (Fig. 10A, HEPES vs. EGF), tended to fall in nominally CO₂/HCO₃⁻-free HEPES (Fig. 10B, HEPES vs. Ctrl), rose in MAlk, tended to rise in RAc (Fig. 10B, MAlk and RAc vs. Ctrl), and fell in all conditions with PD-168393 (Fig. 10C).

Twenty-minute treatment.

Figure 11, A–C, shows, for ErbB2 pY1221/2, data (measured from blots such as those shown in Fig. 9) similar to those presented for ErbB1-specific pY sites in Fig. 8, A–C. Nominally CO₂/HCO₃⁻-free EGF treatment appeared to increase ErbB2 pY1221/pY1222, although with too high a variance to reach significance. Note that, as at 5 min (Fig. 10, B and C), pY1221/pY1222 fell in nominally CO₂/HCO₃⁻-free HEPES (Fig. 11B, HEPES vs. Ctrl) and generally fell with exposure to PD-168393 (Fig. 11C). In contrast to the 5-min data, where MAlk increased pY1221/pY1222, after 20 min, MAlk tended to decrease pY1221/2. We provide a comparison of the 5- and 20-min data in the discussion (see ErbB2 data in Fig. 13).

In summary, the pY status of ErbB2 pY1221/pY1222 is responsive to acid-base disturbances, and, as is the case for the ErbB1-specific sites, the pattern evolves between 5 min and 20 min. However, the response pattern of ErbB2 pY1221/pY1222 differs from that of any of the ErbB1-specific sites.

Comparison of CTRL Data at 5 and 20 min

To test whether the PT suspensions were stable between the 5- and 20-min incubations, we performed Western blots on PT aliquots incubated under Ctrl conditions for 5 and 20 min. In tissue from three rabbits, we found no noticeable differences (not shown) between the 5- and 20-min samples, when run side by side, for any of the five pY sites (i.e., 845, 992, 1068, and 1173 on ErbB1 and 1221/1222 on ErbB2). Thus, even 5 min at 37°C was sufficient for PTs to establish a steady state with respect to these pY sites. Indeed, work on isolated perfused rabbit PTs has indicated that 2 min is sufficient for pH_i to fully recover from a sizeable acid load (15).

The “Total” ErbB1/2 pY (Pseudo-Pan-ErbB1/2-pY) Profile of PT Suspensions Varies With Acid-Base Disturbance and PD-168393 Treatment

Motivation.

Recall that in the experiments with the pan-pY antibody (shown in Figs. 4 and 5), we found that acid-base disturbances produced characteristic changes in pY levels of proteins near the MWs of ErbB1 and ErbB2. One way of evaluating whether these pan-pY signals do indeed reflect ErbB1/2 is to compare them with the sum of the five site-specific pY signals for ErbB1 and ErbB2 (shown in Figs. 7, 8, 10, and 11). Moreover, if the pan-pY data match up well with the summation data, it would suggest that our analysis of the five pY sites covers a representative portion of the pY space modulated by acid-base disturbances.

General approach.

Adding the responses of ErbB1-pY845, ErbB1-pY992, ErbB1-pY1068, ErbB1-pY1173 (Figs. 7 and 8), and ErbB2-pY1221/2 (Figs. 10 and 11) provides a pseudo-pan-ErbB1/2-pY (ErbB-pY′) response to acid-base disturbances. Figure 12, A–D, shows the results of this analysis. Note that, because of the normalization procedures that we used, the following analysis weights all site-specific antibodies equally.

Five-minute treatment.

The response of ErbB-pY′ to nominally CO₂/HCO₃⁻-free EGF and butyrate treatments (Fig. 12A) recapitulated the pan-pY response (Fig. 4A), inasmuch as EGF caused pY to rise (Fig. 4A, HEPES vs. EGF), whereas nominally CO₂/HCO₃⁻-free butyrate caused pY to fall at pH_o 7.1 but not at pH_o 7.4 (Fig. 4A, HEPES vs. butyrate). Similarly, the response of the summated ErbB-pY′ signal to CO₂/HCO₃⁻-dependent acid-base disturbances (Fig. 12B) largely recapitulated the pan-pY response, but with greater statistical strength, to those same treatments (Fig. 4B). Specifically:

1. Nominally CO₂/HCO₃⁻-free HEPES treatment significantly reduces pY/ErbB (Fig. 12B, HEPES vs. Ctrl).

2. MAc causes ErbB-pY′ to fall (Fig. 12B, MAc vs. Ctrl).²

3. MAlk causes ErbB-pY′ to rise (Fig. 12B, MAlk vs. Ctrl).²

4. RAlk, RAc, and cRAc do not significantly affect ErbB-pY′ (Fig. 12B, RAlk, RAc, or cRAc vs. Ctrl).²

In summary, at 5 min, ErbB-pY′ falls significantly in nominally CO₂/HCO₃⁻-free HEPES and MAc but rises in MAlk, consistent with the pan-pY signal at 5 min (Fig. 4).

Twenty-minute treatment.

As at 5 min, the response of ErbB-pY′ at 20 min to nominally CO₂/HCO₃⁻-free EGF (Fig. 12C) generally recapitulated that of pan-pY (Fig. 5A). However, different from the pan-pY response, ErbB-pY′ was not significantly affected by nominally CO₂/HCO₃⁻-free butyrate treatment at pH_o 7.1 (Figs. 12C vs. 5A). The response of ErbB-pY′ signal to CO₂/HCO₃⁻-dependent acid-base disturbances (Fig. 12D) generally matched well with the pan-pY response to those same treatments (Fig. 5B). Specifically:

1. Nominally CO₂/HCO₃⁻-free HEPES treatment significantly reduces ErbB-pY′ (Fig. 12D, HEPES vs. Ctrl).

2. RAlk tends to decrease ErbB-pY′ (Fig. 12D, RAlk vs. Ctrl).

3. MAc causes no significant alteration of ErbB-pY′ (Fig. 12D, MAc vs. Ctrl).²

4. MAlk tends to decrease ErbB-pY′ (Fig. 12D, MAlk vs. Ctrl).

5. RAc and cRAc treatments significantly increase ErbB-pY′ (Fig. 12D, RAc or cRAc vs. Ctrl).²

One inconsistency with the pan-pY data is the significant increase in ErbB-pY′ with RAc (Fig. 12D, RAc vs. Ctrl), which contrasts with the lack of response to the same treatment in the pan-pY study (Fig. 5B, RAc vs. Ctrl). If our pan-pY and site-specific-pY data and analyses can be taken at face value, we could conclude that RAc must cause a substantial fall in pY at some site other than the five sites that we examined on ErbB1, ErbB2, or in pY on another protein that runs at the same MW. We provide a comparison of the 5- and 20-min data in the discussion (see Fig. 14A).

In summary, at 20 min, ErbB-pY′ falls or tends to fall with nominally CO₂/HCO₃⁻-free HEPES, RAlk, and MAlk but rises with RAc and cRAc. We also note that ErbB-pY′ exhibits a different pattern of responses at 5 and 20 min, as we detailed in the discussion.

DISCUSSION

Zhou et al. (107), who flowed OOE CO₂/HCO₃⁻ solutions over the BL surface of isolated perfused PTs, found that the PT responds to independent changes in [CO₂]_BL and [HCO₃⁻]_BL with compensatory alterations in J_HCO3⁻ that would restore normal arterial pH. However, J_HCO3⁻ is not responsive to alterations in pH_BL per se (107). Later, Zhou et al. (106) found that the ErbB inhibitors BPIQ-I and PD-168393 eliminate J_HCO3⁻ responses to alterations in [CO₂]_BL. These authors proposed that unidentified BL sensors for [CO₂]_BL and [HCO₃⁻]_BL trigger signaling cascades (dependent on receptor tyrosine kinases) that rapidly modulate both brush-border NHE3 and BL NBCe1-A and thereby regulate J_HCO3⁻. Indeed, Quigley and Baum (68) had previously demonstrated that two ErbB1 receptor ligands, EGF and transforming growth factor-α, can each increase J_HCO3⁻ (68).

Based on the above discussion, it is reasonable to hypothesize that changes in [CO₂]_BL or [HCO₃⁻]_BL result in time-dependent changes in the pY status of one or more sites (the Y/pY fingerprint) on ErbB family members in the PT. In the present study, we chose to examine ErbB1 and ErbB2 because these are the two most abundant ErbB family members in the adult PT.

Scope of the Study

We chose to work on suspensions of native PTs because these most closely approximate the isolated perfused PTs that we have used in physiological studies. A disadvantage of native tissue is that changes in pY signals will not be as robust as in artificial systems in which ErbB1/2 might be heavily overexpressed. On the other hand, one might question the biological relevance of observations made in such artificial systems.

We worked on total ErbB1/2. It is possible that acid-base disturbances affect the Y/pY status of only a subset of ErbB1/2. If so, our results would underestimate the magnitude of the changes in the Y/pY status in the relevant subpopulation of ErbB1/2.

Of the well-characterized ErbB1 autophosphorylation sites, we focused on three sites involved with major signal transduction pathways and for which reliable antibodies were available: pY992 [which recruits phospolipase C-γ (58, 59)], pY1068 [which recruits growth factor receptor-bound protein 2 (4)], and pY1173 [which recruits Shc and Shp1 (35, 39, 98)]. We unsuccessfully attempted to study pY1045 [which recruits the E3 ubiquitin ligase Cbl (32)] with two different antibodies. We also studied pY845 [which is phosphorylated by Src (5, 88)].

Of the well-characterized sites on ErbB2, we focused on the compound 1221/1222 pY site [which recruits Shc (70)].

We chose to examine our samples at 5 and 20 min because 5 min was the shortest practical time for resuspending PTs and 20 min approximates the duration of experiments on isolated, perfused PTs (105–107).

Although it would have been ideal to use OOE solutions, which allow one to alter [CO₂]_BL, [HCO₃⁻]_BL, or pH_BL one at a time, this approach is impractical with PT suspensions. Instead, we examined a series of classical acid-base disturbances, RAlk, MAc, MAlk, RAc, and cRAc, to examine how changes in acid-base parameters affect the five chosen pY sites on ErbB1 and ErbB2.

Effect of Acid-Base Disturbances on the pY Time Course for Individual ErbB1/2

Here, we consider the differences in pY at specific sites as the system moves from a state of having its strongest acid-base correlation with HCO₃⁻ (at 5 min) to a state of having its strongest correlation with CO₂ (at 20 min).

To examine the dynamic response of a portion of the ErbB1/2 Y/pY fingerprint to changes in the acid-base status, we plotted the log₂-transformed data shown in Figs. 7B, 8B, 10B, and 11B versus time in Fig. 13, A–E and F–J. Under Ctrl conditions, log₂(pY/actin) = 0. In Fig. 13, A–E, we grouped all pY sites together for each of the acid-base disturbances.

MAc.

The results shown in Fig. 13A (the time course for MAc) revealed that pY1068 tended to be less than Ctrl at 5 min and to rise significantly between 5 and 20 min (indicated by the upward arrowhead).

MAlk.

The results shown in Fig. 13B (the time course for MAlk) revealed a pattern that generally tended to be opposite to that of MAc. After 5 min of MAlk, pY845 tended to be greater than Ctrl, and pY1221/pY1222 was significantly greater. Between 5 and 20 min, pY845, pY1068, and pY1221/2 all fell significantly (downward arrowheads). Finally, at 20 min, pY1068 and pY1221/2 both tended to be less than Ctrl.

RAc.

The results shown in Fig. 13C (the time course of RAc) were distinct from those of either MAc or MAlk. After 5 min of RAc, pY1221/pY1222 tended to be greater than Ctrl. Between 5 and 20 min, pY1173 rose significantly. Finally, at 20 min, pY845 tended to be greater than Ctrl, and pY1173 was significantly greater.

RAlk.

The results shown in Fig. 13D (the time course of RAlk) were distinct from RAc as well as MAc and MAlk. After 5 min of RAlk, pY845 tended to be less than Ctrl, and, after 20 min, pY992 tended to be less than Ctrl.

cRAc.

Figure 13E shows the time course of cRAc, which, in terms of chemistry, has the elevated [CO₂] of RAc and the elevated [HCO₃⁻] of MAlk but without the pH changes. The pY response of cRAc was unique but most resembled that of RAc. After 5 min of cRAc, pY992 tended to be less than Ctrl. Between 5 and 20 min, pY992 significantly increased. Finally, after 20 min, pY845 was significantly greater than Ctrl, and pY1173 tended to be greater.

Thus, Fig. 13, A–E, shows that each acid-base disturbance causes a unique pattern of pY changes between 5 and 20 min. In Fig. 13, F–J, we plotted the inverse analysis, grouping all acid-base disturbances (including HEPES) together for each of the pY sites. We found that each pY site has a unique response to each of the acid-base disturbances.

Effect of Acid-Base Disturbances on Summated and Pan-pY Time Courses

As noted in the results, the responses of the summated pY status of the individual ErbB1/2 sites to various acid-base disturbances (ErbB-pY′ in Fig. 12) largely reproduce, in those same PT preparations, the responses of pY_180kDa and pY_160kDa (Figs. 4 and 5). This result implies that the responses of the ErbB1-pY and ErbB-2 pY sites that we did not examine in the present study generally were either similar to the sites that we did examine or were sufficiently weak so as to not interfere with the overall pattern. The lone exception was the response to RAc, where ErbB-pY′ rose at 20 min (Fig. 12D), whereas pan-pY was unchanged (Fig. 5B). Presumably, RAc led to the dephosphorylation of one or more of the unexplored ErbB1/2 pY sites or of a pY site on another protein of similar MW.

ErbB-pY′.

To assess the overall trends in the cumulative pY status of ErbB1/2 over time, we replotted the log₂-transformed ErbB-pY′ data for each acid-base treatment, as shown in Fig. 14A. After 5 min, MAc and MAlk (metabolic-type disturbances) exerted significant but opposite effects on ErbB-pY′. Between 5 and 20 min, we found that ErbB-pY′ significantly rose with MAc, RAc, and cRAc (acidosis-related disturbances) but significantly fell with MAlk and RAlk (alkalosis-related disturbances).

Pan-pY.

Figure 14, B and C, show analyses similar to that shown in A except that we examined pan-pY signals at ∼180 and ∼160 kDa, respectively. The major difference was the lack of a MAlk response at 5 min in the case of pY_180kDa.

In summary, the time courses of the summated ErbB-pY′ signals (Fig. 14A) and of the pY_180kDa signals (Fig. 14B) and pY_160kDa signals (Fig. 14C) are unique for each disturbance. As shown in Fig. 13, A–E, we found that the Y/pY fingerprints for ErbB1/2 (i.e., the collection of pY properties), as far as we have been able to determine them, evolved over time in a unique fashion for each of the five acid-base disturbances. Viewed from the opposite perspective (Fig. 13, F–J), only pY845 exhibited a unique response to each of the acid-base disturbances. For example, the pY992 response was flat for MAc, RAc, and MAlk. In other words, aside from pY845, it takes a combination of pY sites to encode the acid-base disturbances.

PD-168393

PD168393 is a highly specific, potent, and irreversible inhibitor of ErbB1 and ErbB2 (27) that blocks the increase (or decrease) in J_HCO3⁻ that normally occurs when isolated perfused PTs are subjected to an increase (or decrease) in [CO₂]_BL (106). Several interesting insights arose from our results with PD-168393.

First, from our experiments with pan-pY antibodies, the largest PD-168393-induced changes in pY levels occur in the MW range of the ErbB family (i.e., pY_180kDa and pY_160kDa), and here the drug generally reduces pY_160kDa more than pY_180kDa (Figs. 4C and 5C). Second, in our experiments with site-specific pY-ErbB antibodies, PD-168393 nearly always produces a significant reduction in pY (Figs. 7C and 8C). From these two sets of observations, we can conclude that a PD-168393-dependent process must maintain the ErbB1/2 pY status across a wide range of acid-base states.

Third, at 5 min, cRAc tends to reduce pY992 (Fig. 7B), but PD-168393 has no effect on cRAc/pY992 (Fig. 7C). At 20 min, nominally CO₂/HCO₃⁻-free HEPES reduces pY992 (Fig. 8B), but PD-168393 has no effect on HEPES/pY992 (Fig. 8C). These results are consistent with the hypothesis that the PD-168393-sensitive kinases responsible for the phosphorylation of Y992 have a relatively low activity under conditions of cRAc at 5 min and nominally CO₂/HCO₃⁻-free HEPES at 20 min.

Acute Acid-Base Feedback in the PT

The present study may be the first to examine cell signaling in a broad matrix of acid-base conditions (six conditions total: HEPES, RAlk, MAc, Ctrl, MAlk, RAc, and cRAc). Moreover, as far as we are aware, the present data are the first evidence showing that any acute acid-base disturbance can produce a change in any phosphorylation level (specifically pY and specifically on ErbB1/2) in a native tissue. Our work with ErbB1/2 could be an important step in understanding how PT cells transduce acute acid-base disturbances to compensatory changes in acid-base transport.

Any feedback control system requires a sensor, an integrator, and an effector. In the case of the PT, the acid-base effectors include NHE3 at the apical membrane and NBCe1-A at the BL membrane; together, the two transporters produce HCO₃⁻ reabsorption. But what are the sensors and integrators?

Acid-base sensors.

Vertebrates have a wide range of mechanisms for sensing changes in acid-base status, and most of these are present in the kidney (for reviews, see Refs. 12, 79, and 89). First, G protein-coupled receptors (GPCRs) [including GPR4 (43, 83), OGR1 (43, 47), and TDAG8 (36)] can sense a decrease in pH_o. GPR4 is expressed predominantly in collecting ducts and thick ascending limbs and mediates the pH-sensitive accumulation of cAMP (83). OGR1, which is widely distributed in the mouse nephron, including the PT, acts through phospholipase C (47). Second, several ion channels [including TWIK1 (14, 34), TASK2 (34, 41), ASIC (34, 96), and TRPV1 (34, 95)] sense decreases in pH_o. Many are present in the kidney (12, 79, 89), including the PT (12, 79). Third, soluble adenylyl cyclase, which is present in intercalated cells of the collecting duct of the kidney (60), binds cytosolic HCO₃⁻ and responds by generating cAMP (16). Fourth, soluble guanylyl cyclases in the olfactory system (GC-D and GC-G) respond to HCO₃⁻ by generating cGMP (13, 77, 82). Fifth, the tyrosine kinase Pyk2, which is present in the PT (46), responds to in vitro decreases in pH_i with increases in kinase activity (42, 62). Finally, receptor protein tyrosine phosphatase (RPTP)-γ has a ligand-binding domain that is homologous to canonical CAs (3) and is a candidate CO₂/HCO₃⁻ sensor.

Of the six general mechanisms outlined above, we can rule out, as sensors of acute acid-base challenges in the PT, all those that sense pH_o or pH_i per se (i.e., GPCRs, ion channels, and Pyk2) because work with OOE solutions has revealed that PTs are sensitive not to pH_BL or pH_i but to changes in [CO₂]_BL and [HCO₃⁻]_BL per se (107). We can rule out soluble adenylyl cyclase because increases in the intracellular concentration would inhibit, not stimulate, J_HCO3⁻ (59). There is no evidence showing that HCO₃⁻-dependent guanylyl cyclases are present in the PT.³ RPTP-γ is an attractive candidate because its mRNA is present in the PT³, and RPTP-γ could potentially influence the ErbB1/2 Y/pY fingerprint and thereby produce rapid changes in J_HCO3⁻.

Integrators.

The present study shows that acid-base disturbances modulate the pY fingerprint of ErbB1/2. Thus, ErbB1/2 and their downstream signaling cascades are in a position to integrate acid-base signals in the PT. Work on OKP cells has revealed that an acute exposure to acidic media (i.e., MAc) rapidly activates Pyk2, which is required for the activation, after a short delay, of Src (42). Independently, acidic media activate the MEK-ERK-c-fos pathway in OKP cells (90). If the work on OKP cells translates to native PTs, which appear not to be intrinsically pH sensitive in the short term, then it is possible that CO₂ and HCO₃⁻ sensors could lead to the activation of Pyk2 in the PT.

The major observation of the present study is that the four classical acid-base disturbances (RAlk, MAc, MAlk, and RAc) as well as nominally CO₂/HCO₃⁻-free HEPES (an extreme cRAlk) and cRAc each produce a distinctive ErbB1/2 Y/pY fingerprint, a pattern of Y/pY at specific sites that evolves over time. This conclusion is corroborated by the time courses of summated ErbB-pY′ as well as pan-pY_180kDa and pan-pY_160kDa for each of the conditions. At present, we do not know which of these changes in Y/pY status, and at which times, are important for transducing increases in extracellular [CO₂] to increases in J_HCO3⁻ versus transducing increases in extracellular [HCO₃⁻] to decreases in J_HCO3⁻ and so on. However, the PD-168393 sensitivity of changes in both J_HCO3⁻ and Y/pY are consistent with the hypothesis that at least some of the Y/pY changes that we report here are likely to be physiologically important. We also note that ErbB1 localizes strongly to the BL membrane, which is in direct contact with the OOE solutions in a physiological study (107) of J_HCO3⁻ in the PT.

We caution that, in addition to the five sites already examined, other ErbB1/2 Y/pY sites may be important in responses to acid-base disturbances. A major challenge will be to examine the importance of each site for each disturbance. Given the complexity of Y/pY responses to each acid-base disturbance, we think it unlikely that future research will point to a simple relationship among acid-base disturbances, Y/pY sites, and J_HCO3⁻ responses (i.e., that any one Y/pY site would be solely responsible for mediating the response to any one disturbance). Instead, with each disturbance, the time-dependent Y/pY pattern on ErbB1/2 would determine the time-dependent recruitment of diverse signaling proteins containing SH2 and pY-binding domains. We hypothesize that each disturbance, with its unique Y/pY fingerprint, initiates a unique signaling cascade, altering Y/pY, S/pS, and T/pT patterns on other PT proteins, and ultimately changes the activity of NBCe1-A, NHE3, and other transport-related proteins, which leads to the appropriate and observed modification in J_HCO3⁻ while stabilizing volume reabsorption (107).

The present work provides the first insights into how ErbB1/2 tyrosine residues are affected by specific acid-base disturbances at 5 or 20 min. We are in a position to begin to dissect the downstream effectors that lead to the HCO₃⁻-reabsorbing machinery itself. In addition, it will be important to identify acid-base sensors upstream to ErbB1/2. Finally, although a major technical challenge, it would be especially powerful to develop approaches for using OOE solutions to address questions of evolving pY status.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases grant DK81567 and a National Kidney Foundation Postdoctoral Fellowship (to L. A. Skelton).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: L.A.S. and W.F.B. conception and design of research; L.A.S. performed experiments; L.A.S. analyzed data; L.A.S. and W.F.B. interpreted results of experiments; L.A.S. prepared figures; L.A.S. drafted manuscript; L.A.S. and W.F.B. edited and revised manuscript; L.A.S. and W.F.B. approved final version of manuscript.