Research ArticleION TRANSPORT

Modulation of Cl signaling and ion transport by recruitment of kinases and phosphatases mediated by the regulatory protein IRBIT

See allHide authors and affiliations

Science Signaling  30 Oct 2018:
Vol. 11, Issue 554, eaat5018
DOI: 10.1126/scisignal.aat5018

Combinations to regulate Cl signaling

The secretion of HCO3-containing fluids is vital to the function of all epithelia and is enabled in part by the activity of the Na+-coupled HCO3 transporter NBCe1-B. Vachel et al. identified five serine residues in NBCe1-B whose phosphorylation status was controlled by the regulatory protein IRBIT. The phosphorylation status of Ser12 and Ser65 affected the Cl sensitivity of two intracellular Cl-sensing motifs. Moreover, IRBIT recruited a distinct kinase/phosphatase pair for each serine residue. The three remaining phosphorylation sites were phosphorylated in distinct combinations that determined the relative basal activity level of NBCe1-B and the potential for further activation by IRBIT. These results demonstrate how distinct phosphorylation patterns may enable epithelial cells to fine-tune the HCO3 transport activity of NBCe1-B in response to varying conditions in different parts of the organ.

Abstract

IRBIT is a multifunctional protein that controls the activity of various epithelial ion transporters including NBCe1-B. Interaction with IRBIT increases NBCe1-B activity and exposes two cryptic Cl-sensing GXXXP sites that enable regulation of NBCe1-B by intracellular Cl (Clin). Here, phosphoproteomic analysis revealed that IRBIT controlled five phosphorylation sites in NBCe1-B that determined both the active conformation of the transporter and its regulation by Clin. Mutational analysis suggested that the phosphorylation status of Ser232, Ser233, and Ser235 was regulated by IRBIT and determined whether NBCe1 transporters are in active or inactive conformations. The absence of phosphorylation at Ser232, Ser233, or Ser235 produced NBCe1-B in the conformations pSer233/pSer235, pSer232/pSer235, or pSer232/pSer233, respectively. The activity of the pSer233/pSer235 form was similar to that of IRBIT-activated NBCe1-B, but it was insensitive to inhibition by Clin. The properties of the pSer232/pSer235 form were similar to those of wild-type NBCe1-B, whereas the pSer232/pSer233 form was partially active, further activated by IRBIT, but retained inhibition by Clin. Furthermore, IRBIT recruited the phosphatase PP1 and the kinase SPAK to control phosphorylation of Ser65, which affected Clin sensing by the 32GXXXP36 motif. IRBIT also recruited the phosphatase calcineurin and the kinase CaMKII to control phosphorylation of Ser12, which affected Clin sensing by the 194GXXXP198 motif. Ser232, Ser233, and Ser235 are conserved in all NBCe1 variants and affect their activity. These findings reveal how multiple kinase and phosphatase pathways use phosphorylation sites to fine-tune a transporter, which have important implications for epithelial fluid and HCO3 secretion.

INTRODUCTION

Cl is the principal extracellular and intracellular anion for all vertebrate cells. Extracellular Cl (Clout) maintains bodily fluid and volume homeostasis and blood pressure and is thus fairly constant, except in disease states mainly associated with metabolic acidosis and alkalosis (1). Conversely, intracellular Cl (Clin) varies considerably between cells and changes markedly in response to cell stimulation. The resting Clin concentration in epithelial cells is as high as 60 mM (24), whereas in neurons and skeletal muscle, it is between 5 to 20 mM (5, 6). When Cl-absorbing and HCO3-secreting pancreatic and salivary gland ducts are stimulated, ductal Clin is reduced from ~40 to 4 mM by basolateral membrane Cl extrusion to allow the high HCO3 concentration (up to 140 mM) in the secreted fluids. These changes are accomplished by enhancement of basolateral NBCe1-B–dependent electrogenic 2HCO3out/Na+out cotransport and luminal cystic fibrosis transmembrane conductance regulator (CFTR) Cl channel and electrogenic Slc26a6-mediated 2HCO3in/Clout exchange (4, 7).

Clin is regulated by ion channels and transporters coupled to other ions, mainly Na+, K+, and HCO3. In turn, Clin regulates the cellular activity of several ions and intracellular pH to coordinate transepithelial fluid and electrolyte secretion (7). In addition, Clin regulates transporter function by modulating the activity of several protein kinases, such as WNK1 (with no lysine 1) (8), WNK4 (9), WNK3 (10), SPAK (Ste20p-related proline alanine-rich kinase) (11), JNK (c-Jun N-terminal kinase), and p38 (12), by binding directly to the kinases and modifying their activities. Thus, Clin regulates the activity of several transporters and channels in at least two ways. The activity of Cl-coupled transporters is determined through changes in driving force. In addition, reduction in Clin activates NKCC1 (Na+/K+/2Cl cotransporter 1) (13), NKCC2 (14), and NCC (15) and several Na+-HCO3 cotransporters (NBCs) (16) by causing phosphorylation of the transporters by an unknown mechanism.

Clin may directly interact with transporters to modulate their activity. However, the mechanisms by which Clin is sensed to modulate the activity of ion transporters are not well understood. Analysis of the ClC family of Cl channels and transporters has identified GXXXP motifs as Cl-sensing sites that participate in Cl-mediated regulation of SLC26A2 (17), NBCe1-B (also known as SLC4A4), and NBCe2-C (also known as SLC4A5) (16, 1820). In NBCe1-B, Clin is sensed by two GXXXP motifs, and both are required for inhibition of the NBCs by Clin between 5 and 60 mM. The first motif is 32GXXXP36, which is located in the NBCe1-B autoinhibitory domain (AID; residues 1 to 40), and the second motif is 194GXXXP198.

NBCe1-B and other epithelial transporters that are functionally coupled to NBCe1-B to mediate epithelial fluid and electrolyte secretion, including CFTR, Slc26a6, KCC, NCC, and NKCC1, are regulated by multiple kinases and phosphatases (2124). In secretory gland ducts, NBCe1-B, CFTR, and Slc26a6 are activated by IRBIT [inositol 1,4,5-trisphosphate (IP3) receptor-binding protein released with IP3] (16, 2527). IRBIT interacts with the AID in the N terminus of NBCe1-B to relieve autoinhibition by an unknown mechanism (25, 26). NBCe1-B, CFTR, and Slc26a6 are regulated by the kinases WNK1 and WNK4, which phosphorylate and activate the kinase SPAK (26, 27). Phosphorylation of NBCe1-B Ser65 by SPAK inhibits the transporter, an effect that is reversed by the phosphatase PP1 (protein phosphatase 1) (28). Key questions in understanding signaling by Clin are how Clin sensing by the GXXXP motifs is achieved, how IRBIT modulates regulation by Clin, and whether the kinase and phosphatase pathways interface with Clin and IRBIT to form an integrated signaling pathway. Moreover, it is not clear how interaction of IRBIT with the AID activates the transporters. Here, we used NBCe1-B as a model to address these questions by identifying the phosphorylation sites controlled by IRBIT, the kinases and phosphatases recruited by IRBIT to modify these sites, the effect of specific phosphorylation sites on Clin signaling, and the role of these sites in IRBIT-mediated relief of autoinhibition.

We report that IRBIT exposed the two cryptic N-terminal GXXXP motifs and recruited the kinases SPAK and calmodulin-dependent protein kinase II (CaMKII) or the phosphatases PP1 and calcineurin (CaN) to the AID of NBCe1-B. Liquid chromatography–tandem mass spectrometry (LC-MS/MS) analysis of NBCe1-B serine and threonine phosphorylation sites identified 25 sites in total. IRBIT regulated the phosphorylation status of Ser232, Ser233, and Ser235 generating three possible conformers (pSer233/pSer235, pSer232/pSer235, or pSer232/pSer233). IRBIT also regulated the phosphorylation status of Ser65 and Ser12 in NBCe1-B. The phosphorylation of Ser232, Ser233, and Ser235 controlled the active state of NBCe1-B and its regulation by IRBIT and Clin. The active states of (Δ1-95)NBCe1-B and NBCe1-A, which are not regulated by IRBIT, were regulated in the same manner, suggesting that the conserved Ser232, Ser233, and Ser235 sites determined the active state of these NBC transporters. Functional analysis of NBCe1-B GXXXP motifs showed that SPAK and PP1 acted on Ser65 to regulate Clin sensing by the 32GXXXP36 motif, whereas CaN dephosphorylated pSer12 to regulate Clin sensing by the 194GXXXP198 motif. These findings uncover an intricate mechanism by which Clin signals through IRBIT-mediated recruitment of kinases and phosphatases to regulate key transporters involved in epithelial fluid and electrolyte secretion. These findings have implications for various epithelial diseases, including cystic fibrosis, pancreatitis, and Sjögren’s syndrome.

RESULTS

IRBIT interacts with and recruits PP1, CaN, CaMKII, and SPAK to NBCe1-B N terminus

IRBIT interacts with PP1 (26, 29) and CaMKII (30). We showed here that IRBIT coimmunoprecipitated with PP1, CaMKII, the phosphatase CaN, and the kinase SPAK (Fig. 1, A and B). Moreover, IRBIT recruited the phosphatases and kinases to full-length NBCe1-B (Fig. 1, C and D) and, more specifically, to the N terminus of NBCe1-B(1-95) (Fig. 1, E and F), which includes the AID of NBCe1-B. The entire NBCe1-B N-terminal domain (1 to 429) also showed IRBIT-dependent interaction of the phosphatases and kinases (Fig. 1, E and F).

Fig. 1 IRBIT interacts with and recruits kinases and phosphatases to the AID of NBCe1-B.

(A and B) Immunoprecipitates (IP) prepared from lysates from human embryonic kidney (HEK) cells transfected with IRBIT and CaN, CaMKII, PP1, or SPAK were immunoblotted (IB) for the indicated proteins. (A) IRBIT immunoprecipitates were probed for the indicated kinases or phosphatases. Representative example of three to four similar experiments. (B) Kinase or phosphatase immunoprecipitates were probed for IRBIT. Representative example of three to four similar experiments. (C) The indicated immunoprecipitates were prepared from lysates from HEK cells expressing NBCe1-B with or without IRBIT and were immunoblotted for the indicated kinases and phosphatases. In the controls (Con), precipitating antibodies were not added. (D) The means ± SEM of IRBIT-enhanced coimmunoprecipitates were determined from six similar experiments and normalized to coimmunoprecipitation in the absence of IRBIT. (E) Lysates from HEK cells transfected with IRBIT and CaN, CaMKII, PP1, or SPAK were incubated with partially purified His-tagged NBCe1-B(1-95) or NBCe1-B(1-429) fragments, and pulldown experiments were performed. (F) The means ± SEM of the IRBIT-enhanced pulldown determined from three to four independent experiments and normalized to pulldown in the absence of IRBIT. AU, arbitrary units.

PP1 and SPAK modulate Clin sensing by the 32GXXXP36 motif

We measured NBCe1-B activity in the presence or absence of IRBIT by recording HCO3-induced currents at pipette Cl concentration (Clin) of 5 and 140 mM (Fig. 2A). PP1 activates and SPAK inhibits NBCe1-B in an IRBIT-independent manner (26), and the presence of IRBIT did not further activate NBCe1-B activated by PP1 (Fig. 2B). NBCe1-B is inhibited by Clin, which in the absence of IRBIT occurred at Clin between 40 and 140 mM (Fig. 2C). In the presence of IRBIT, the inhibition occurred at Clin between 5 and 40 mM. Inhibition by the lower Clin depends on IRBIT-mediated exposure of two cryptic Clin-sensing 32GXXXP36 and 194GXXXP198 motifs (16). Mutation of either motif increased the Clin required to inhibit NBCe1-B, and mutation of both motifs is required to eliminate inhibition by Clin (16).

Fig. 2 PP1 acts on NBCe1-B(Ser65) to control Clin sensing by the 32GXXXP36 site.

(A) Example of current time course traces in response to adding bath HCO3 and the current/voltage (I/V) at peak current. (B) HEK cells transfected with NBCe1-B and with or without IRBIT and PP1 were used to measure NBCe1-B–mediated current due to Na+-2HCO3 cotransport at 5 mM Clin. The columns show the means ± SEM. NS, not significant. (C) Current was measured in HEK cells transfected with NBCe1-B alone or with PP1, IRBIT, or IRBIT and PP1. Pipette solutions containing the indicated Cl concentrations are shown. Normalized current is shown to illustrate the effect of IRBIT on Cl sensing. (D) Current was measured in HEK cells transfected with the mutants NBCe1-B(32GP/AA36) or NBCe1-B(194GP/AA198) and with IRBIT and PP1 at pipette solutions containing 5, 40, or 140 mM Cl. The models in (C) and (D) depict the effect of the indicated condition on the state of Cl sensing. Closed GXXXP motifs indicate Cl sensing by both motifs and inhibition by Clin at high affinity, partially opened GXXXP motifs indicate loss of Cl sensing by one GXXXP motif and Cl sensing by the available motif at low affinity, and fully opened GXXXP motifs indicate complete loss of Clin sensing. The numbers in the columns and next to the symbols indicate the number of current measurements obtained from at least three independent experiments. The results are given as means ± SEM.

It is not known whether or how signaling by Clin is regulated by kinases and phosphatases and the phosphorylation sites involved. To address these questions, we examined the effect of PP1 on Clin-mediated inhibition of NBCe1-B. PP1 coexpression in the absence of IRBIT resulted in activation of NBCe1-B (Fig. 2B), which could be mediated by interaction of PP1 with a potential PP1-binding site on NBCe1-B (949RVHLF953). Similar to NBCe1-B expressed alone, NBCe1-B expressed with PP1 showed low sensitivity for inhibition by Clin (Fig. 2C, black and green traces), indicating that PP1 does not expose the cryptic GXXXP motifs and does not affect inhibition by Clin when the GXXXP motifs are not first exposed by IRBIT. Coexpressing NBCe1-B with PP1 and IRBIT induced inhibition of NBCe1-B, but only by Clin at 40 and 140 mM (Fig. 2C, red trace). Thus, PP1 eliminated the inhibition of NBCe1-B by 5 to 20 mM Clin induced by activation of NBCe1-B by IRBIT in the absence of PP1 (Fig. 2C, blue trace). This finding indicates that PP1 acting on IRBIT-activated NBCe1-B must have affected Clin sensing by one of the GXXXP motifs. In the NBCe1-B(32GP/AA36) mutant, the 194GXXXP198 motif is available for Clin sensing, and PP1 coexpression did not affect the inhibition of this mutant by 140 mM Clin (Fig. 2D). By contrast, coexpression of the NBCe1-B(195GP/AA199) mutant (in which the 32GXXXP36 motif is available) with PP1 eliminated inhibition by Clin, indicating that PP1 affects Clin sensing by the 32GXXXP36 motif and not by the 194GXXXP198 motif (Fig. 2D).

SPAK phosphorylates Ser65 in NBCe1-B to inhibit its activity, and PP1 dephosphorylates Ser65 to reverse the inhibitory effect of SPAK (28). In this case, phosphorylation of Ser65 by SPAK may affect the same GXXXP motif affected by PP1, with SPAK and PP1 having opposing effects on inhibition by Clin. To test this prediction, we compared the effects of coexpressing PP1 or dominant-negative (DN) SPAK on the inhibition of NBCe1-B by Clin because expression of wild-type SPAK markedly reduces NBCe1-B surface abundance (28). Similar to PP1, expression of DN-SPAK prevented low Clin from inhibiting IRBIT-activated NBCe1-B (Fig. 3A) and eliminated inhibition of NBCe1-B(194GP/AA198) by 140 mM Clin, but not of NBCe1-B(32GP/AA36) (Fig. 3B), indicating that DN-SPAK interfered with Cl sensing by the 32GXXXP36 motif. Because SPAK phosphorylates Ser65, NBCe1-B(S65A) should behave like NBCe1-B in the presence of DN-SPAK. IRBIT-activated NBCe1-B(S65A) was not inhibited by low Clin, whereas the phosphomimetic NBCe1-B(S65E) mutant showed the expected strong inhibition by 10 to 40 mM Clin (Fig. 3C). In addition, the NBCe1-B(S65A) mutation eliminated Cl sensing by the 32GXXXP36 motif (Fig. 3D). Last, because PP1 dephosphorylates Ser65, PP1 did not affect the inhibition of NBCe1-B(S65A) by 140 mM Clin (fig. S1). As illustrated in the GXXXP models (Fig. 3, B to D), these findings indicate that SPAK acts on Ser65 to modulate Clin sensing by the 32GXXXP36 motif, similar to PP1.

Fig. 3 SPAK acts on Ser65 in NBCe1-B to control Clin sensing by the 32GXXXP36 site.

(A to D) Current was measured in HEK cells transfected with wild-type NBCe1-B and IRBIT (A and C), NBCe1-B(32GP/AA36) or NBCe1-B(194GP/AA198) and IRBIT (B and D), or NBCe1-B(S65E) and IRBIT (C), NBCe1-B(S65A) alone or with IRBIT (C and D), and NBCe1-B, IRBIT, and DNSPAK (A and B) and with pipette solutions containing the indicated Cl concentrations. Example of current traces at the indicated Clin concentrations and the indicated NBCe1-B mutants are shown in (A) and (D). The models in (B) to (D) depict the effect of the indicated conditions on the state of Cl sensing by the GXXXP motifs. The numbers in the columns and next to the symbols indicate the number of current measurements obtained from three to five independent transfections, and the results are given as means ± SEM.

CaN and CaMKII reciprocally regulate NBCe1-B

IRBIT recruited CaN and CaMKII to NBCe1-B (Fig. 1, A to F). In oocytes, coexpression of constitutively active (CA) CaN (Fig. 4A) and CaMKII (Fig. 4B) did not affect NBCe1-B activity in the absence of IRBIT. However, in the presence of low levels of IRBIT, coexpression of CA-CaN markedly activated NBCe1-B to the level observed with maximal stimulation with 9 ng of IRBIT (the highest amount of IRBIT that we transfected; Fig. 4A). 271LxVP274 is a CaN-interacting motif in IRBIT, and the LCVP/AAAA mutant did not activate NBCe1-B whether expressed with or without CA-CaN (Fig. 4A). CaMKII inhibited NBCe1-B activated by intermediate concentrations of IRBIT (Fig. 4B). Thus, these findings suggest that the IRBIT/CaN and IRBIT/CaMKII reciprocally regulate the activity of NBCe1-B.

Fig. 4 CaN and CaMKII regulate NBCe1-B activity, and CaN controls Clin sensing by the 194GXXXP198 motif.

(A) Current was measured in Xenopus oocytes injected with complementary RNAs (cRNAs) encoding NBCe1-B, after NBCe1-B, H2O, CA-CaN(R392X), the indicated amount of IRBIT, and the IRBIT (LCVP/AAAA) mutant alone or together with CA-CaN. (B) Current was measured in Xenopus oocytes injected with cRNAs encoding NBCe1-B and CaMKII, IRBIT, or both. The numbers in the columns indicate the number of current measurements obtained from four independent batches of oocytes from four frogs, and the results are given as means ± SEM. (C to E) Current was measured in HEK cells transfected with NBCe1-B (C and D), NBCe1-B (32GP/AA36) or NBCe1-B(194GP/AA198) (E), and CA-CaN (C to E) or CA-CaN + PP1 (C and D). Normalized current is shown in (D). The models in (C) and (E) depict the effect of the indicated condition on the state of Cl sensing by the GXXXP motifs. The numbers in the columns and next to the symbols indicate the number of current measurements obtained from three to four independent transfections, and the results are given as means ± SEM.

CaN and CaMKII modulate Clin sensing by the 194GXXXP198 motif

Similar to results obtained in oocytes, coexpression of CaN modestly activated NBCe1-B in the absence of IRBIT in HEK cells (Fig. 4C). In the presence of IRBIT, CaN coexpression increased the Clin required for inhibition of NBCe1-B (Fig. 4, C and D). Use of the GXXXP mutants showed that CaN inhibited Clin sensing by the 194GXXXP198 motif (Fig. 4E), rather than the 32GXXXP36 motif affected by PP1. Accordingly, coexpressing both PP1 and CaN with IRBIT-activated NBCe1-B eliminated inhibition by Clin (Fig. 4, C and D). Hence, the two GXXXP motifs appear to cooperate in determining regulation by Clin. Elimination of either motif reduced the affinity for Clin to a similar extent, and when both motifs are not available, NBCe1-B was no longer inhibited by Clin (Fig. 4, C and D). Similar studies in HEK cells with CaMKII were not possible because CaMKII coexpression modestly activated NBCe1-B and altered inhibition by Clin (fig. S2, A and B). It is not clear why CaMKII expression had such different effects in oocytes and mammalian cells. However, this is not without precedent. Regulation of NBCe1-B by PP1 is quite different in HEK cells (26) and oocytes (31).

NBCe1-B regulates the phosphorylation status of Ser232, Ser233, and Ser235 in IRBIT

To identify NBCe1-B phosphorylation sites affected by IRBIT, we defined the NBCe1-B phosphorylation sites in the presence and absence of IRBIT using LC-MS/MS–based phosphoproteomic (32) and identified 20 phosphorylation sites at the long N-terminal domain and 5 sites in the C terminus (fig. S3A and data file S1). All previously reported phosphorylation sites—Thr49, Ser1026 (33), and Ser65 (28)—were identified by the phosphoproteomic analysis. A similar analysis performed in cells expressing IRBIT (data file S2) revealed that IRBIT promoted the phosphorylation of Ser12 (fig. S3A, red) and restricted the phosphorylation of Ser232, Ser233, and Ser235 (fig. S3A, magenta). For Ser12, only one phosphorylated peptide was found (fig. S3B). For Ser232, Ser233, and Ser235, two separate phosphorylated peptides were identified in this region (fig. S3C). For peptides in which both Ser232 and Ser233 were phosphorylated, Ser235 was not phosphorylated. For peptides in which Ser235 was phosphorylated, either Ser232 or Ser233 was phosphorylated, but not both. These findings indicate that IRBIT had a combinatorial effect on the phosphorylation status of these sites, to generate NBCe1-B phosphorylated at Ser232 and Ser233, NBCe1-B phosphorylated at Ser232and Ser235, and NBCe1-B phosphorylated at Ser233 and Ser235. Localization of the phosphorylation sites in the N terminus of NBCe1-B and the orientation of these sites with respect to the GXXXP motifs and the ion-transporting sector of NBCe1-B are shown in the structural model of NBCe1-B (fig. S4). All three phosphorylation sites are conserved in the kidney-specific isoform NBCe1-A, and the first phosphorylation site (equivalent to Ser232 in NBCe1-B) is conserved in all members of the NBC family (fig. S5).

Ser12 in NBCe1-B mediates the effect of CaN on Cl sensing by the 194GXXXP198 motif

Mutating Ser12 to a nonphosphorylatable residue [NBCe1-B(S12A)] or a phosphomimetic residue [NBCe1-B(S12D)] did not affect the surface expression of NBCe1-B or its interaction with IRBIT (Fig. 5A). At 5 mM Clin, the S12A mutation did not affect NBCe1-B activity in the absence of IRBIT (Fig. 5B) and had a small stimulatory effect in the presence of IRBIT (Fig. 5C), whereas the S12D mutation increased NBCe1-B activity in the presence or absence of IRBIT by about 60% (Fig. 5, B and C). By contrast, the S12D mutation did not affect the inhibition of IRBIT-activated NBCe1-B by Clin, whereas the S12A mutation increased the Clin required for inhibition at the high millimolar range (Fig. 5, C and D). Coexpression of PP1 or DN-SPAK, but not of CA-CaN, reduced the inhibition of the S12A mutant by 140 Clin (Fig. 5, E and F). These findings suggest that Ser12 regulates Clin sensing by the 194GXXXP198 motif and that pSer12 is dephosphorylated by CaN. Accordingly, the S12A mutation eliminated inhibition by 140 mM Cl of the NBCe1-B(32GP/AA36) mutant (in which the 194GXXXP198 motif is available), but not of the NBCe1-B(194GP/AA198) mutant (in which the 32GXXXP36 motif is available) (Fig. 5, E and F).

Fig. 5 CaN acts on Ser12 in NBCe1-B to control Clin sensing by the 194GXXXP198 motif.

(A) Effect of the S12A and S12D mutations on NBCe1-B surface expression and interaction with IRBIT (n = 3 independent experiments). (B to D) Effect of S12D and S12A mutations on the activity of NBCe1-B measured at 5 mM Clin (B) and increasing concentrations of Clin (C and D). The current in (C) is shown as normalized current in (D). (E) Example of current traces measured in cells expressing NBCe1-B(S12A) or NBCe1-B(S12D) and the indicated Clin. (F) NBCe1-B(S12A) current was measured in the presence of 5 or 140 mM pipette solution in HEK cells expressing NBCe1-B(S12A) alone or together with PP1, DN-SPAK, or CaN. Current was measured with the double mutants NBCe1-B(S12A/32GP/AA36) or NBCe1-B(S12A/194GP/AA198). The models in (C) and (F) depict the effect of the indicated condition on the state of Cl sensing by the GXXXP motifs. The numbers in the columns and next to the symbols indicate the number of current measurements obtained from three to four independent transfections, and the results are given as means ± SEM.

Ser232, Ser233, and Ser235 regulate the active state of NBCe1-B and activation by IRBIT

Because IRBIT regulates the phosphorylation state of Ser232, Ser233, and Ser235 in a combinatorial manner, the role of each serine residue was studied by mutating them individually and in combination to alanine or to aspartate residues. The combined and individual mutations had no obvious effect on NBCe1-B total and surface expression or interaction with IRBIT (Fig. 6, A and B). However, mutating all three serine residues to alanine residues [NBCe1-B(AAA)] resulted in fully activated NBCe1-B that was not activated further by IRBIT (Fig. 6C). On the other hand, mutating all three serine residues to aspartate residues [NBCe1-B(DDD)] nearly completely inhibited NBCe1-B activity, which could not be rescued or activated by IRBIT (Fig. 6C). To further confirm that NBCe1-B(AAA) and NBCe1-B(DDD) function was independent of IRBIT, we mutated the three serine residues in NBCe1-B(Δ1-95), which lacks the autoinhibitory IRBIT-binding domain (28, 34). The AAA mutation also maximized the activity of NBCe1-B(Δ1-95), whereas the DDD mutation markedly inhibited NBCe1-B(Δ1-95) activity (Fig. 6C). In addition, we mutated the analogous three residues in NBCe1-A, which functions independently of IRBIT. NBCe1-A activity was increased by the AAA mutation and inhibited by the DDD mutation (Fig. 6C).

Fig. 6 IRBIT-modulated phosphorylation of Ser232, Ser233, and/or Ser235 affects NBCe1-B activity and activation by IRBIT.

(A and B) Effect of the Ser232, Ser233, and/or Ser235 mutations on NBCe1-B surface expression and interaction with IRBIT (n = 4 independent experiments). (C and D) Current was measured with pipette solution containing 5 mM Cl and in the presence (+) or absence of IRBIT, as indicated. The cells were transfected with (C) NBCe1-B , NBCe1-B(S232A/S233A/S235A), NBCe1-B(S232D/S233D/S235D), NBCe1-A and its 3S/3A and 3S/3D mutants, NBCe1-B(Δ1-95) and its 3S/3A and 3S/3D mutants, (D) S232A or S232D, S233A or S233D, and S235A or S235D. The models in (C) and (D) depict the phosphorylation state of Ser232, Ser233, and Ser235 for the indicated mutants. Asterisk denotes P < 0.01 relative to no IRBIT. The numbers in the columns indicate the number of current measurements obtained from three to four independent transfections, and the results are given as means ± SEM.

Because IRBIT affects the phosphorylation of Ser232, Ser233, or Ser235, we measured the activity of the individual mutants first at 5 mM Clin (Fig. 6D). The S232A mutation (conformer A, in which Ser233 and Ser235 are phosphorylated) resulted in activation of NBCe1-B with minimal further activation by IRBIT. The S233A mutation (conformer B, in which Ser232 and Ser235 are phosphorylated) did not activate NBCe1-B, which was then fully activated by IRBIT. The S235A mutation (conformer C, in which Ser232 and Ser233 are phosphorylated) resulted in activation of NBCe1-B, which was further increased by IRBIT (Fig. 6D). S232D and S233D minimally affected basal NBCe1-B activity but eliminated activation by IRBIT. By contrast, the S235D mutation substantially inhibited basal NBCe1-B activity, but the residual activity was activated by IRBIT to a similar extent as wild-type NBCe1-B (Fig. 6, C and D). Thus, inhibition of NBCe1-B by the phosphomimetic aspartate mutations was independent of IRBIT. Mutation of these residues in NBCe1-B(Δ1-95) did not affect surface expression (fig. S6A) but caused inhibition of NBCe1-B(Δ1-95) current similar to that of full-length NBCe1-B (fig. S6B).

The decoupling of NBCe1-B activity from IRBIT caused by mutating Ser232 raised the question of the role played by Ser232 and this region of NBCe1-B in the activation of the cotransporter. We addressed this question by mutating Ser232 to various residues. All mutations tested were tolerated and did not affect NBCe1-B total and surface expression or interaction with IRBIT (fig. S7A). Measurement of NBCe1-B activity showed that adding a negative charge at Ser232 (Asp and Glu) resulted in inhibition of transport and loss (S232D) or reduced (S232E) activation by IRBIT (fig. S7B). Replacing Ser232 with residues that can be phosphorylated by kinases (Thr and Tyr) did not restore the wild-type phenotype but resulted in a phenotype similar to S232A. A positive charge (Lys), a polar residue (Gln), a small residue (Leu), and a bulky (Trp) hydrophobic residue resulted in a phenotype similar to that of S232A. Residues that introduce kinks (Gly and Pro) were less effective in activating NBCe1-B but resulted in maximal activation by IRBIT. These findings suggest that Ser232 is in a conformation-sensitive region of the protein and that IRBIT limits the phosphorylation of this site to stabilize an active conformation. The equivalent Ser232 is conserved in NBCe2-C (fig. S5), and introducing the equivalent mutation NBCe2-C(S236D) reduced transport activity (fig. S8, A and B). Together, the findings suggest that the phosphorylation state of Ser232 determines whether NBCe1-B is in an active or inactive conformation and that by binding to the AID, IRBIT activates NBCe1-B by limiting the dephosphorylation of this site.

Ser232, but not Ser235, modulates Clin sensing

Current measurement of NBCe1-B(AAA) showed that activation of the current did not affect the reversal potential, and thus ion selectivity, of NBCe1-B (Fig. 7A). However, the Ser232, Ser233, and Ser235 mutants affected signaling by Clin. NBCe1-B(AAA) and NBCe1-B(S232A) were not inhibited by Clin in the presence or absence of IRBIT (Fig. 7B). Hence, when Ser233 and Ser235 are phosphorylated (in the S232A mutant), NBCe1-B is fully active, with the Clin-sensing motifs occluded within the N-terminal domain. In the presence of IRBIT, the S233A mutant was inhibited by 40 and 140 mM Clin (Fig. 7C). The actual and normalized currents of NBCe1-B(S235A) (Fig. 7D) showed that in the absence of IRBIT, the increased activity of NBCe1-B(S235A) was inhibited by 140 mM, but not by 40 mM Clin, similar to the inhibition of NBCe1-B alone. IRBIT maximized the activity of NBCe1-B(S235A) and conferred inhibition by low Clin (Fig. 7D). The inhibition followed a shallower curve than that observed with NBCe1-B activated by transfection with IRBIT complementary DNA (cDNA) (1 μg/ml; for example, Figs. 2C and 5C). We reason that this effect may be due to the maximal activity observed with NBCe1-B(S235A) activated by IRBIT. Transfecting the cells with wild-type NBCe1-B and a large amount of IRBIT resulted in a Clin inhibitory curve similar to that observed with NBCe1-B(S235A) stimulated by IRBIT (Fig. 7D). These findings suggest that inhibition by Clin depends on the extent of activation of NBCe1-B by IRBIT and that when Ser232 and Ser233 are phosphorylated (as in the S235A mutant), NBCe1-B has higher basal activity. Binding of IRBIT to this NBCe1-B form fully activates it and exposes its cryptic GXXXP Clin-sensing motifs.

Fig. 7 Ser232 affects Clin sensing by NBCe1-B.

(A) Example of current time course traces in response to adding bath HCO3 measured in HEK cells expressing NBCe1-B, NBCe1-B + IRBIT, or the 3S/3A mutant, 3S/3A + IRBIT, and 3S/3D with pipette solution containing 5 or 140 mM Cl and the I/V at peak current. (B) Current density for NBCe1-B, the 3S/3A, and the S232A mutants in the presence (+) or absence of IRBIT was measured at 5 and 140 mM Clin. (C) Current of NBCe1-B(S233A) + IRBIT at 5, 40, and 140 mM Clin. (D) Effect of Clin on current measured with NBCe1-B + 5 μg of IRBIT, NBCe1-B(S235A) alone, or with 1 μg of IRBIT. The right plot shows the normalized current in the left plot. (E) A model of the three IRBIT-dependent NBCe1-B conformations and their properties with respect to autoinhibition, Clin sensing, and regulation of Clin sensing by the PP1/SPAK and CaN/CaMKII. The models in (B), (D), and (E) depict the phosphorylation state of Ser232, Ser233, and Ser235 and Clin sensing by the GXXXP motifs of each mutant. The numbers in the columns indicate the number of current measurements obtained from three to four independent transfections, and the results are given as means ± SEM.

DISCUSSION

Cl is the major anion and an important osmolyte in most living cells and has a prominent role in cellular homeostasis by affecting cytoplasmic ionic composition, pH, cell volume, and the membrane potential. These roles are related mostly to the transport function of Cl. Another form of regulation by Cl involves direct interaction of Cl with the transporters. For example, Clin regulates mammalian neurotransmitter transporters (35, 36), the anion exchanger AE1 (37), and the SLC26 transporters Slc26a5 (also known as prestin) (38), Slc26a1, and Slc26a2 (17, 39). Cl is sensed by a GXXXP motif in Slc26a2 (17). The family of cotransporters and exchangers has one or more GXXXP motifs, and Clin sensing by the GXXXP motifs inhibits the transporters at Clin between 5 and 40 mM (16), well within the physiological range of Clin changes. In the present study, we used regulation of NBCe1-B by Clin as a model to understand how Cl regulates the transporters and how Cl sensing by the transporters is regulated.

In the basal state, NBCe1-B has low activity due to inhibition by its AID, a conformation in which the two Clin-sensing GXXXP motifs are not accessible. In the resting state, IRBIT is sequestered by binding to IP3 receptors (IP3Rs) (27, 40). Upon cell stimulation, IRBIT is released from the IP3Rs and interacts with and activates NBCe1-B and other epithelial transporters (27) and exposes their Cl-sensing motifs (16).

Here, we found that IRBIT controlled autoinhibition by controlling the phosphorylation status of Ser232, Ser233, and Ser235 in a combinatorial manner to induce various NBCe1-B conformations, each of which has different properties (Fig. 7E). Ser232, Ser233, and Ser235 appear to reside in a conformation-sensitive region of NBCe1-B. Ser232 is conserved in all members of the NBC family, and lack of phosphorylation at this site results in NBCe1-B phosphorylated at Ser233 and Ser235 (Fig. 7E, conformation A). This conformation is active independently of autoinhibition and Clin-dependent inhibition, as demonstrated by the findings that mutation of Ser232 to any residue, except for the phosphomimetic Asp and Glu, resulted in partial or full activation of NBCe1-B independent of IRBIT, and that similar activation was observed with the CA NBCe1-B mutant lacking the AID [NBCe1-B(Δ1-95)] and NBCe1-A (Fig. 6, C and D, and fig. S7B). Moreover, mutation of Ser232 to Asp or Glu inhibited NBCe1-B, NBCe1-B(Δ1-95), and NBCe1-A without affecting the interaction with IRBIT or the surface expression of the transporter [Fig. 6 (A to D) and figs. S6 (A and B) and S7 (A and B)]. Together, these findings suggest that Ser232 controls the communication between the large cytoplasmic N terminus and the transmembrane sector of NBCe1-B to affect opening of the permeation pathway.

The second IRBIT-induced NBCe1-B conformation is phosphorylated at Ser232 and Ser235 (Fig. 7E, conformation B). The properties of conformation B are similar to those of wild-type NBCe1-B. It is inactive in the absence of IRBIT with cryptic Clin-sensing motifs. IRBIT activates NBCe1-B(pSer232/pSer235) and exposes the Clin-sensing motifs that are, in turn, regulated by the PP1/SPAK and CaN/CaMKII affecting Clin sensing by the respective GXXXP motifs. The third IRBIT-induced NBCe1-B conformation is phosphorylated at Ser232 and Ser233 (Fig. 7E, conformation C). The properties of conformation C are intermediate between those of conformations A and B. In conformation C, basal NBCe1-B activity is increased independently of the AID, as suggested by the low sensitivity to Clin. However, IRBIT further activates conformation C and exposes the cryptic GXXXP motifs that are regulated by PP1/SPAK and CaN/CaMKII. Hence, regulation by IRBIT is more complicated than originally thought, with IRBIT able to tune NBCe1-B activity to match physiological demands.

Clin sensing is conferred by two GXXXP motifs that are selectively affected by the action of SPAK and PP1 on Ser65 and those of CaN and CaMKII on Ser12. The two GXXXP motifs appear to cooperate to determine regulation by Clin at one site or two associated Clin sites. This is suggested by the findings that elimination of any of the motifs, either by SPAK/PP1 and CaMKII/CaN or by mutation, similarly reduces the affinity for inhibition by Clin to shift the inhibition to a higher Clin. Only when both motifs are not available for Clin sensing is inhibition by Clin eliminated.

The present and previous works suggest that the kinase and phosphatase pathways affect NBCe1-B beyond regulation of Clin sensing. Hence, SPAK controls transporter activity by two mechanisms, modifying transporter surface expression and sensing of Cl by the 32GXXXP36 motif. In the absence of IRBIT, NBCe1-B activity is low, and SPAK inhibits NBCe1-B activity by reducing surface expression (28). Although IRBIT antagonizes the effect of SPAK on NBCe1-B surface expression (26), SPAK can still phosphorylate Ser65 in the presence of IRBIT, which affects Clin-mediated regulation of NBCe1-B at the 32GXXXP36 motif. Cl may regulate SPAK activity directly or indirectly through WNK4 (24) to inhibit SPAK activity. However, the effect that Cl between 5 and 20 mM has on SPAK activity is unclear, and the concentrations at which SPAK affects Cl regulation of NBCe1-B are not known.

Our findings have broad physiological implications by demonstrating that multiple phosphorylation sites are used to fine-tune transporter activity to meet variable demands. A good example is fluid and HCO3 secretion by stimulated pancreatic and salivary ducts, which secrete fluid containing 140 mM HCO3 (21). As fluid and HCO3 secretion progresses from the intercalated duct to the intralobular duct and to the main duct, secretion occurs at increased luminal HCO3 concentration and thus requires enhanced basolateral HCO3 influx. The duct cells likely use combinations of phosphorylation sites to tune HCO3 influx by NBCe1-B, depending on their position in the duct. It is likely that in the intercalated duct, Ser65, Ser232, and Ser235 in NBCe1-B are phosphorylated, whereas Ser12 and Ser233 in NBCe1-B are not phosphorylated, and together with the high Clin, this phosphorylation combination maintains low NBCe1-B activity that is sufficient for the initial HCO3 secretion. In the main duct, Ser12 in NBCe1-B is phosphorylated, and Ser65 and Ser232 are not phosphorylated, which, together with the low Clin, maximize NBCe1-B activity and basolateral HCO3 uptake to support HCO3 secretion into a lumen that already contains 120 mM HCO3. Combinations of these extremes operate along the ducts to tune NBCe1-B activity and ductal HCO3 secretion. Thus, phosphorylation and dephosphorylation of Ser12, Ser65, Ser232, Ser233, and Ser235, together with changes in Clin between 5 and 40 mM, function as a rheostat to tune NBCe1-B activity and functions.

MATERIALS AND METHODS

Plasmids

The p3xFLAG-CMV-7.1/IRBIT, pCMV-HA-IRBIT, pEGFP-C1/NBCe1-B, pEGFP-C1/NBCe2-C, and pCDNA 3.1+/NBCe1-A constructs were described previously (28). pCDNA3-Flag CaN was a gift from R. Bassel-Duby and E. Olson (University of Texas Southwestern Medical Center), pCDNA-CaMKIIα was a gift from L. Nutt (St. Jude Children’s Research Hospital, Memphis), and pCDNA3/mPP1 was a gift from E. Delpire (Vanderbilt University Medical Center, Nashville). The cDNA encoding SPAK (a gift from M. Cobb, University of Texas Southwestern Medical Center) was excised from the original vectors and transferred to pCMV-myc. Point mutations were generated by site-directed mutagenesis with a Quick light chain mutagenesis kit (Agilent), using primers to change specific regions. All constructs were verified by sequencing of the entire open reading frames.

Surface expression and coimmunoprecipitation

Biotinylation of surface proteins was performed by incubating transfected HEK cells with EZ-LINK Sulfo-NHS-LC-biotin (0.5 mg/ml; Thermo Fisher Scientific) for 30 min on ice. The cells were then treated with 100 mM glycine for 10 min to terminate the biotinylation reaction, washed with phosphate-buffered saline, and lysed with ice-cold lysis buffer [10 mM NaPO43−, 137 mM NaCl, 2.7 mM KCl, 50 mM NaF, 1% Triton X-100, and protease inhibitor cocktail (Roche)] for 20 min on ice. After brief sonication, lysates were collected by centrifugation. Biotinylated proteins were captured with avidin beads (Thermo Fisher Scientific) by incubation for 1 hour at 4°C. The precipitated proteins were analyzed by SDS–polyacrylamide gel electrophoresis (PAGE). Coimmunoprecipitation was performed by incubation overnight of 300 μg of proteins with corresponding antibodies for each protein: NBCe1-B [anti-GFP (green fluorescent protein), A11122, Invitrogen], IRBIT [anti-HA (hemagglutinin), 2367S, Cell Signaling], CaN (anti-FLAG, F3165, Sigma) anti-MYC (2276S, Cell Signaling), and CAMKIIα (13141, Santa Cruz). The next day, Protein G Sepharose beads (GE Healthcare) are added and incubated for 2 hours at 4°C. Beads were collected by centrifugation and washed three times with lysis buffer, and proteins were recovered by 30-min heating at 56°C in SDS sample buffer. Proteins were subjected to SDS-PAGE, and the blots probed for the indicated proteins.

Pulldown experiments

The N-terminal 95 (1 to 95) amino acids and 429 amino acids (1 to 429) of NBCe1-B were subcloned in the pQE-TriSystem His-Strep vector (Qiagen). The final plasmid constructs were transformed into the Escherichia coli strain Rosetta (DE3) competent cells. Colonies were amplified in 10 ml of LB medium containing ampicillin (100 ug/ml) and incubated at 37°C overnight under shaking (250 rpm). Aliquots were then used to inoculate larger volume cultures. Protein expression was induced until the OD600 (optical density at 600 nm) of the culture reached 0.6 by the addition of isopropyl-β-d-thiogalactopyranoside to a final concentration of 0.1 mM. His-tagged proteins were extracted by lysis and sonication, and extracts were clarified by centrifugation at 20,000g for 20 min. Fusion proteins were purified with Ni-NTA superflow beads (Qiagen). Purity of proteins was better than 80%, as determined by Coomassie Blue staining. HEK cell lysates expressing NBCe1-B and the kinases or phosphatases as indicated were incubated with Strep-tagged fusion proteins bound to the beads for 4 hours at 4°C. The beads were washed with lysis buffer, and proteins were released by incubation in SDS sample buffer at 56°C for 20 min and analyzed by SDS-PAGE and Western blotting.

Determination of NBCe1-B phosphorylation site

Cells expressing NBCe1-B alone or NBCe1-B and IRBIT were lysed with 8 M urea and sonicated. The cell lysates were reduced by the addition of 20 mM dithiothreitol, alkylated with 40 mM iodoacetamide, and diluted with 20 mM triethylammonium bicarbonate buffer (pH 8.5) to reduce urea to 1 M. The lysates were then digested with trypsin (enzyme, Promega) at a lysate/trypsin ratio of 1:20. Peptides were desalted using hydrophilic-lipophilic–balanced extraction cartridges (Oasis) and then subjected to phosphopeptide enrichment using either Fe-NTA or TiO2 columns as per the manufacturer’s protocol (Thermo Fisher Scientific). The enriched peptides were desalted using graphite columns, vacuum-dried, and stored at −80°C. Peptides were resuspended with 0.1% formic acid for mass spectrometry (Orbitrap Fusion ETD mass spectrometer, Thermo Fisher Scientific) analysis. Mass Spec spectra were analyzed using Proteome Discoverer 1.4 software (Thermo Fisher Scientific). Peptide-spectra matching (false discovery rate, <0.01; peptide rank, 1) was analyzed by both Mascot and Sequest HT software and the human Swiss-prot (18 November 2016) protein database. The probabilities of the phosphorylation site localizations were calculated using the phosphoRS 3.1 Module within Proteome Discoverer 1.4. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD010598.

Current measurement in HEK-293 cells

The effect of Clin on current mediated by NBCe1-B, NBCe1-A, and NBCe2-C activity was analyzed in transiently transfected HEK-293 cells using whole-cell current recording at room temperature, exactly as detailed previously (16), and varying the pipette Cl concentration between 5 and 140 mM. Patch pipettes had a resistance of 5 to 7 megohms when filled with CsCl-based pipette solution. The cell capacitance was between 10 and 18 pF. The pipette solutions contained 2 mM MgSO4, 1 mM adenosine 5′-triphosphate, 0.5 mM EGTA, 10 mM Hepes, and a mixture of CsCl and Cs-gluconate to yield Cl concentrations of 5, 10, 20, 30, 40, and 140 mM. pH was adjusted to 7.3 with CsOH. The Hepes-buffered bath solution contained 140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 10 mM glucose, and 10 mM Hepes (pH 7.4 with NaOH). The HCO3-buffered solution was prepared by replacing 25 mM NaCl with equimolar amount of NaHCO3, and the solution was equilibrated with 5% CO2/95% O2. The current was recorded by 400-ms rapid alteration of membrane potential from −60 to +60 mV every 2 s from a holding potential of 0 mV. Before current recording, the junction potential at each pipetted solution Cl concentration was offset to 0 using the Axopatch 200B amplifier. The current recorded at +60 mV was used to calculate current density as picoampere/picofarad. Axopatch 200B patch-clamp amplifier, Digidata -1440A, and pClamp 10 software (Molecular Devices) were used for data acquisition and analysis. The currents were filtered at 1 kHz and sampled at 10 kHz.

Preparation of cRNA

The human NBCe1-B/pcDNA3.1 and human IRBIT/pcDNA3.1 plasmids were used as templates for preparation of the corresponding cRNAs. DNA was linearized with SmaI (NBCe1-B) or Xba1 (IRBIT) (New England Biolabs, Ipswich, MA). cRNA was then transcribed in vitro using a T7 PNA polymerase kit (mMessage mMachine T7 Kit; Ambion Inc., Austin, TX). cRNA was purified by phenol/chloroform extraction, precipitated with isopropanol, and then dissolved in nuclease-free water (Ambion Inc., Austin, TX).

Preparation and injection of Xenopus oocytes

Oocytes were obtained by partial ovariectomy of female Xenopus laevis (Xenopus Express, Beverly Hills, FL), anesthetized with methanesulfonate salt of 3-aminobenzoic acid ethyl ester (2.0 g/liter; Sigma, St. Louis, MO). Follicular cells were removed in OR-2 Ca2+-free medium [82.5 mM NaCl, 2.4 mM KCl, 1.0 mM MgCl2, and 5.0 mM Hepes-Na (pH 7.5)] with the addition of collagenase B (1 mg/ml; Boehringer Mannheim, Indianapolis, IN). Defolliculated oocytes were washed four to five times with Ca2+-free OR-2. Healthy oocytes in stages V to VI were collected under a microscope and maintained at 18°C overnight in ND96 solution [96.0 mM NaCl, 2.0 mM KCl, 1.8 mM CaCl2, 1.0 mM MgCl2, 2.5 mM sodium pyruvate, 5.0 mM Hepes-Na (pH 7.5), streptomycin (100 μg/ml), and penicillin (100 U/ml)]. Oocytes were injected with the indicated amount of cRNA in a final volume of 50 nl using a Nanoliter 2000 injector (World Precision Instruments Inc., Sarasota, FL). Oocytes were incubated at 18°C in ND96 solution with pyruvate and antibiotics. The medium was changed everyday, and oocytes were tested 48 to 96 hours after cRNA injection.

Measurement of membrane current and voltage in oocytes

Electrophysiological recordings were performed at room temperature with two-electrode voltage clamp or current clamp using an OC-725C Oocyte Clamp System (Warner Instruments Corp., Hamden, CT). The microelectrodes were filled with 3 M KCl (resistance, 0.5 to 2 megohms). During recording, the oocytes were superfused with ND96 solution. The Cl and HCO3-containing solutions were of 71.0 mM NaCl, 25.0 mM NaCl or NaHCO3, 2.0 mM KCl, 1.8 mM CaCl2, 1.0 mM MgCl2, and 5.0 mM Hepes-Na (pH 7.5). Solutions were constantly gassed with 5% CO2/95% O2. Current and voltage were digitized via a Digidata 1322A A/D converter (Axon Instrument, Foster City, CA) and analyzed using Clampex 8.1 system.

Structure prediction and protein modeling

A segment of mouse NBCe1-B, consisting of 1000 amino acids (NP_061230.2), was submitted to both ROBETTA and hhPRED. The ROBETTA software generated different models for three segments of the submitted sequence. No homology was identified for segment 1-106. Segment 107-413 was homologous to Putative Substrate Access Tunnel in the Cytosolic Domain of Human Anion Exchanger 1 (Protein Data Bank ID: 4KY9), which is similar to human erythrocyte band 3 cytoplasmic domain (Protein Data Bank ID: 1HYN) that we previously used to predict NBCe1-B N terminus (16), and the same structure is used here. The transmembrane segment spanning residues 414 to 1000 was found to be homologous to the structure of the anion exchanger domain of human erythrocyte band 3 (Protein Data Bank ID: 4YZF). The highest transmembrane domain (TMD) homology score provided by the hhPRED software was for the Electrogenic sodium bicarbonate cotransporter 1, NBCe1 (PDB_ID 6CAA). (Probability, 100; E value, 2.5 × 10−164; SS,108.9; Cols,1003). The NBCe1 TMD structure has been recently solved by cryo–electron microscopy (CryoEM) (41). The final model was generated with PyMOL using the TMD obtained of the CryoEM structure, and the intracellular domains were predicted by hhPRED and ROBETTA based on the N-terminal domain of AE1. The orientation of the N terminus with respect to the TMD is not known with certainty and may not be accurate. However, the overall structure is reasonable because the software identified the N terminus of AE1 as the best template, which, like NBCe1-B, is a member of the SLC4 transporters superfamily. In addition, we have previously predicted the structure of NBCe1-B to be similar to the crystal structure of UraA (16), which is similar to the crystal structure of AE1. This indicates that the software predicted the same structure twice and independently on the basis of several similar, yet different structures.

Statistics

Results are given as means ± SEM, and statistical significance was analyzed by Student’s t test or two-way analysis of variance (ANOVA), as appropriate.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/11/554/eaat5018/DC1

Fig. S1. PP1 and SPAK, but not CaN, act through NBCe1-B Ser65.

Fig. S2. CaMKII activates NBCe1-B when expressed in HEK cells.

Fig. S3. NBCe1-B phosphorylation sites and the sites affected by IRBIT.

Fig. S4. Structural model of NBCe1-B.

Fig. S5. Sequence alignment of the NBC family.

Fig. S6. Effect of S232D, S233D, and S235D mutations on NBCe1-B(Δ1-95) expression and activity.

Fig. S7. Effect of S232X mutation on NBCe1-B expression and activity.

Fig. S8. Effect of the S236D mutation in NBCe2-C on transport activity.

Data file S1. MS data of HEK cells phosphoproteins in the absence of IRBIT.

Data file S2. MS data of HEK cells phosphoproteins in the presence of IRBIT.

REFERENCES AND NOTES

Acknowledgments: We thank D. M. Schwartz (NIADS) for fruitful discussion and suggestions and R. Bassel-Duby and E. Olson for providing the CA-CaNR392X plasmid. Funding: These studies were funded by the Division of Intramural Research National Institute of Dental and Craniofacial Research (NIDCR), intramural grants DE000735-07 (to S.M.), and the National Heart, Lung, and Blood Institute (NHLBI) (projects ZIA-HL001285 and ZIA-HL006129 to M.A.K.). The NHLBI Proteomics Core Facility (M. Gucek, Director) was used. O.Y. was supported by fellowships from the Nakatomi, Sumitomo Life Welfare, Culture and the Mochida Memorial Foundations, and JSPS KAKENHI grant number JP16K18992; and A.Y.-N. was supported by the Mitsukoshi Health and Welfare Foundation. Funds were also provided by grant no. 2015003 from the United States–Israel Binational Science Foundation to S.M. and E.O., ISF grants 271/16 and 2164/16 by the Israel Science Foundation to E.O., and by a Basic Science Research Program through the National Research Foundation of Korea (NRF-2016R1A5A2008630 to D.M.S.). Author contributions: L.V., N.S., O.Y., A.S., A.Y.-N., and E.O. performed experiments and interpreted results. M.F. and E.O. performed structural prediction. C.-R.Y. and M.A.K. performed LC-MS/MS–based phosphoproteomics and interpreted results. D.M.S. interpreted results. S.M. conceived the study, interpreted results, and drafted the manuscript with contribution from all authors. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All plasmids generated in the Muallem laboratory and used in this study require a material transfer agreement from NIDCR, NIH. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD010598. All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article