SLC4A Transporters

Abstract

SLC4A gene family proteins include bicarbonate transporters that move HCO₃⁻ across the plasma membrane and regulate intracellular pH and transepithelial movement of acid–base equivalents. These transporters are Cl/HCO₃ exchangers, electrogenic Na/HCO₃ cotransporters, electroneutral Na/HCO₃ cotransporters, and Na⁺-driven Cl/HCO₃ exchanger. Studies of the bicarbonate transporters in vitro and in vivo have demonstrated their physiological importance for acid–base homeostasis at the cellular and systemic levels. Recent advances in structure/function analysis have also provided valuable information on domains or motifs critical for regulation, ion translocation, and protein topology. This chapter focuses on the molecular mechanisms of ion transport along with associated structural aspects from mutagenesis of particular residues and from chimeric constructs. Structure/function studies have helped to understand the mechanism by which ion substrates are moved via the transporters. This chapter also describes some insights into the structure of SLC4A1 (AE1) and SLC4A4 (NBCe1) transporters. Finally, as some SLC4A transporters exist in concert with other proteins in the cells, the structural features associated with protein–protein interactions are briefly discussed.

1. INTRODUCTION

In most cells, the internal ionic environment is regulated by membrane proteins that move ions across the cell membranes. The SLC4A bicarbonate transporters are a subset of these proteins that are specialized to move HCO₃⁻. HCO₃⁻ buffers H⁺, and therefore, HCO₃⁻ transport is equivalent to H⁺ transport in the opposite direction from a pH standpoint. The primary function of the bicarbonate transporters is to regulate intracellular H⁺. Some bicarbonate transporters move HCO₃⁻ into the cytosol and raise intracellular pH (acid extruders), whereas others remove HCO₃⁻ from the cytosol and lower intracellular pH (acid loaders). The activity of acid extruders and loaders affects steady-state pH in the cells. Cells also utilize these pH regulators to transfer large amounts of Na⁺ and HCO₃⁻ from one body compartment to another to maintain extracellular and total body acid–base balance. The bicarbonate transporters are also widely distributed in many epithelial tissues where they play a role in transepithelial absorption and secretion of HCO₃⁻. This transepithelial HCO₃⁻ movement contributes to tissue-specific physiological processes such as reabsorbing HCO₃⁻ or excreting acid into urine (kidney), neutralizing intestinal gastric acid (pancreatic ducts), or maintaining the cerebrospinal fluid pH (choroid plexus).

Many studies have been done to determine the functional and physiological significance of the SLC4A bicarbonate transporters in the body and to understand their pathological involvement in acid–base disorders. Cellular signals responsible for transport regulation have also been examined. In addition, the members of SLC4A transporters exhibit high homology in amino acid sequence, and this homology has enabled researchers to investigate domains or residues important for distinct molecular and functional properties of ion movement. Such efforts have produced valuable results to advance our knowledge of the underlying molecular mechanisms of ion transport. Furthermore, similar structural features among bicarbonate transporters have facilitated our understanding of transporter structure and the macromolecular complexes with other proteins. This chapter provides some insight into ion transport properties, structure, and protein–protein interactions of SLC4A bicarbonate transporters. This chapter focuses on anion exchangers AE1–3 and Na/HCO₃ cotransporters NBCe1 and NBCn1 because less has been done with other Na⁺-coupled transporters.

2. SLC4A TRANSPORTERS

The SLC4A transporters (Table 1) are Cl/HCO₃ exchangers, Na/HCO₃ cotransporters, Na⁺-driven Cl/HCO₃ exchanger, and Na/borate transporter. These are distinct from the SLC26A anion exchangers that transport a variety of monovalent and divalent anions (Sindic, Chang, Mount, & Romero, 2007). Some SLC26A proteins transport HCO₃⁻ in exchange for Cl⁻ or other anions.

Table 1.

Human SLC4A transporters

Human gene	Protein	Function	pHi regulation	Human chromosome
SLC4A1	AE1 Band 3	Cl−/HCO3− exchange	Acid loader	17q21–q22
SLC4A2	AE2	Cl−/HCO3− exchange	Acid loader	7q35–36
SLC4A3	AE3	Cl−/HCO3− exchange	Acid loader	2q36
SLC4A4	NBCe1 NBC1	Na+/HCO3− cotransport (electrogenic)	Acid extruder (except NBCe1-A)	4q21
SLC4A5	NBCe2 NBC4	Na+/HCO3− cotransport (electrogenic)	Acid extruder	2p13
SLC4A6	–			–
SLC4A7	NBCn1 NBC3	Na+/HCO3− cotransport	Acid extruder	3p22
SLC4A8	NDCBE	Na+/Cl−/2HCO3− cotransport/exchange*	Acid extruder	12q13
SLC4A9	AE4	Cl−/HCO3− exchange	Acid loader	5q31
SLC4A10	NBCn2 NBCE	Na+/HCO3− cotransport	Acid extruder	2q23–q24
SLC4A11	NaBC BTR1	Na+/B(OH)4−, OH− cotransport	Acid extruder for OH− transport	20p12

^*NDCBE cotransports external 1Na⁺ and 2HCO₃⁻ in exchange for internal Cl⁻.

2.1. Cl/HCO₃ Exchangers (AEs)

2.1.1. AE1 (SLC4A1), AE2 (SLC4A2), AE3 (SLC4A3)

The three Cl/HCO₃ exchangers (AE1–3) primarily move external HCO₃⁻ in exchange for internal Cl⁻ (Alper, 2009; Bonar & Casey, 2008). AE1 is the first cloned HCO₃⁻ transporter (Kopito & Lodish, 1985). The human AE1 contains 911 amino acids with the predicted structure of 13 transmembrane (TM) (The TM topology of AE1 is defined according to the 13-TM model (Bonar & Casey, 2008), in which 13 TMs and one unusual membrane-spanning segment (called extended structure) between TM11 and TM12 are present. The topology of NBCe1 is defined according to the 14-TM model (Zhu et al., 2010), in which 14 TMs are present.) segments and cytoplasmic N- and C-terminal domains (Zhu, Lee, & Casey, 2003). The large N-terminal domain (~400 amino acids) plays a regulatory role, while the transmembrane domain is responsible for anion exchange activity. The N-terminal domain interacts with other proteins such as ankyrin, glycolytic enzymes, and hemoglobin, whereas the C-terminal domain interacts with carbonic anhydrase II (CA II) (Vince & Reithmeier, 1998). AE1 exists as homodimers (Wang, Sarabia, Reithmeier, & Kuhlbrandt, 1994). AE1 is the most abundant membrane protein in red blood cells and plays a key role in CO₂ delivery from tissues to the lungs. AE1 is also localized to the basolateral membrane of type-A intercalated cells in the kidney collecting ducts (van Adelsberg, Edwards, & Al Awqati, 1993). The renal kAE1 lacks the first 65 amino acids because of an alternate promoter site. AE1 also transports other anions such as SO₄²⁻ at a very low rate (Milanick & Gunn, 1982).

AE2 was originally isolated from the kidney (Alper, Kopito, Libresco, & Lodish, 1988). AE2 is ~300 amino acids larger than AE1 and exists as multiple variants because of alternative promoter sites. Similar to AE1, the N- and C-terminal domains of AE2 play regulatory roles in its pH dependence, as well as response to NH₄⁺ and hypertonicity (Chernova et al., 2003; Stewart, Chernova, Shmukler, Wilhelm, & Alper, 2002), while the transmembrane domains are responsible for exchanger function (Stewart, Kurschat, Vaughan-Jones, Shmukler, & Alper, 2007). AE2 is widely distributed in the body, particularly in epithelial cells including the thick ascending limb and cortical collecting duct intercalated cells in the kidney (Brosius et al., 1995), and contributes to transepithelial movement of HCO₃⁻ and salt.

Human AE3 comprised the cardiac variant (cAE3) and the brain variant (bAE3) with different N-terminal domains (Kopito et al., 1989). Both variants are found in many different tissues in addition to the heart and brain. AE3 has a lower exchange activity than AE1 and AE2 because its membrane expression is weak (Fujinaga, Loiselle, & Casey, 2003).

2.1.2. AE4 (SLC4A9)

Human AE4 contains 983 amino acids and has three alternative splice variants according to the NCBI database. AE4 is more closely related to Na/ HCO₃ transporters than AEs in amino acid sequence. AE4 is functionally characterized as a Cl/HCO₃ exchanger (Tsuganezawa et al., 2001; Ko, Luo et al., 2002), but it is also suggested to be an electroneutral NBC without Cl/HCO₃ exchange activity (Parker et al., 2002). AE4 is localized to renal collecting duct type-A intercalated cells, but its membrane polarity appears to be species-specific. It is localized to the basolateral membrane of the intercalated cells in rats and mice, but to the apical and lateral membranes in rabbits (Ko, Luo et al., 2002).

2.2. Na⁺-Coupled Bicarbonate Transporters (NCBTs)

2.2.1. NBCe1 (SLC4A4), NBCe2 (SLC4A5)

Romero and Boron cloned the electrogenic NBCe1 from the salamander kidney (Romero, Hediger, Boulpaep, & Boron, 1997). On the basis of predicted topology and similarity to AE1, NBCe1 (1035 amino acids) was proposed to have a cytoplasmic N-terminal domain, 12–14 TM segments, and a cytoplasmic C-terminal domain. The most widely distributed variant is NBCe1-B containing the unique N-terminal 85 amino acids, which replace the first 41 amino acids of the renal variant NBCe1-A (Abuladze et al., 1998; Choi, Romero, Khandoudi, Bril, & Boron, 1999). The brain-specific NBCe1-C is identical to NBCe1-B except for the C-terminal 61 amino acids (Bevensee, Schmitt, Choi, Romero, & Boron, 2000). NBCe1-A heterologously expressed in Xenopus oocytes produces an electrogenic HCO₃⁻ current with a 1Na⁺ and 2HCO₃⁻ stoichiometry (Ducoudret, Diakov, Muller-Berger, Romero, & Frömter, 2001; Grichtchenko et al., 2001; Sciortino & Romero, 1999), mediating net Ducoudret, Diakov, Muller-Berger, Romero, & Frömter, 2001 influx. In contrast, the transporter in the proximal tubules moves 1Na⁺ and 2HCO₃⁻ across the basolateral membrane, mediating net HCO₃⁻ efflux. Studies show that the same transport molecule presumably changes its stoichiometry depending on the cellular environment and parameters such as intracellular Ca²⁺ levels (Muller-Berger, Ducoudret, Diakov, & Frömter, 2001) or phosphorylation activated by cAMP (Gross et al., 2003). Similar to AEs, NBCe1 forms a homodimer composed of two individually functional subunits (Gill & Boron, 2006a; Kao et al., 2008).

NBCe2 is the second type of the electrogenic Na/HCO₃ cotransporter (Pushkin et al., 2000; Virkki, Wilson, Vaughan-Jones, & Boron, 2002). NBCe2 is strongly expressed in hepatocytes and bile duct cholangiocytes in the liver and renal pelvis uroepithelial cells in the kidney (Abuladze et al., 2004). It is also found in the renal collecting ducts, choroid plexus epithelia (Bouzinova et al., 2005; Damkier, Nielsen, & Praetorius, 2007), and sarcolemmal membranes in skeletal muscles (Kristensen, Kristensen, & Juel, 2004). NBCe2 produces a HCO₃⁻-induced current with the 1Na⁺:2HCO₃⁻ stoichiometry when expressed in Xenopus oocytes (Virkki et al., 2002). However, it operates with the 1Na⁺:3HCO₃⁻ stoichiometry when expressed in the renal proximal tubule cell line (Sassani et al., 2002) and in intact cells prepared from choroid plexus epithelia (Millar & Brown, 2008).

2.2.2. NBCn1 (SLC4A7), NBCn2 (SLC4A10)

The electroneutral NBCn1 was isolated from rat vascular smooth muscle (Choi, Aalkjaer, Boulpaep, & Boron, 2000) and human skeletal muscle (Pushkin et al., 1999). Human NBCn1 contains 1214 amino acids with a predicted structure similar to that of NBCe1. NBCn1 exists as multiple variants because of N- and C-terminal splice events. The transporter is widely distributed in the body (Boedtkjer, Praetorius, Fuchtbauer, & Aalkjaer, 2008). NBCn1 is expressed in vascular smooth muscle cells and endothelial cells and modulates artery tone and blood pressure control (Boedtkjer et al., 2011). The variant containing the N-terminal cassette of 123 amino acids (cassette II) is predominantly found in the cardiovascular system (Cooper et al., 2006). Among SLC4A bicarbonate transporters, NBCn1 is least sensitive to the anion channel/transporter blocker 4,4′-diisothiocyanatostilbene disulfonate (DIDS) (Choi et al., 2000). In addition, NBCn1 has HCO₃⁻-independent, Na⁺ channel-like activity (Cooper et al., 2005).

NBCn2 is the second cloned electroneutral Na/HCO₃ cotransporter (Wang, Yano, Nagashima, & Seino, 2000). This transporter is predominantly expressed in neurons and choroid plexus epithelia (Damkier et al., 2007). NBCn2 operates with the 1Na⁺:1HCO₃⁻ stoichiometry with Cl⁻ self-exchange activity (Parker, Musa-Aziz et al., 2008). Other studies show that NBCn2 directly transports Cl⁻ in exchange for Na/HCO₃ cotransport and functions as a Na⁺-driven Cl/HCO₃ exchanger NCBE (Damkier, Aalkjaer, & Praetorius, 2010; Giffard, Lee, Ouyang, Murphy, & Monyer, 2003).

2.2.3. NDCBE1 (SLC4A8)

NDCBE1 is the Na⁺-driven Cl/HCO₃ exchanger (Grichtchenko et al., 2001; Parker, Bouyer, Daly, & Boron, 2008). This transporter is predominantly expressed in neurons (Chen, Haddad, & Boron, 2008; Sinning et al., 2011), where Na⁺-driven Cl/HCO₃ exchange has been known to primarily govern HCO₃⁻-dependent acid extrusion (Schwiening & Boron, 1994). NDCBE moves external 1Na⁺ and 2HCO₃⁻ in exchange for internal Cl⁻.

2.3. Sodium Borate (NaBC) Transporter

NaBC1 (SLC4A11) transports Na⁺ and borate (Park, Li, Shcheynikov, Zeng, & Muallem, 2004). This transporter has substantial sequence homology to SLC4A bicarbonate transporters and was assigned as SLC4A11 before its function was characterized. NaBC1 operates with the stoichiometry of 1Na⁺ and at least 2B(OH)₄⁻ (Park et al., 2004) The transporter moves Na⁺ and OH⁻ (H⁺) in the absence of borate.

3. ION TRANSPORT PROPERTIES

3.1. Ion Dependence and Selectivity

3.1.1. AEs

AEs are permissive to many anions in addition to HCO₃⁻ and Cl⁻. For example, AE1 has the capacity to move several monovalent and divalent anions in either direction or via anion self exchange (Knauf & Pal, 2003). AE2 also mediates the exchange of anions including NO₃⁻, Br⁻, and I⁻ with low affinity when expressed in Xenopus oocytes (Humphreys, Jiang, Chernova, & Alper, 1994). The amino acid residue best characterized for anion selectivity is Glu⁶⁸¹ (TM8) in erythrocyte AE1 and the corresponding Glu site in AE2. Chemical modification of Glu⁶⁸¹ to an alcohol leads to electrogenic SO₄²⁻/Cl⁻ exchange and electroneutral SO₄²⁻/SO₄²⁻ self exchange (Jennings, 1995; Jennings & Al Rhaiyel, 1988). Similar changes are also observed from the site-directed mutagenesis of the corresponding Glu in cloned AE1 and AE2 (Chernova et al., 1997; Sekler, Lo, & Kopito, 1995), indicating that the role of Glu⁶⁸¹ is conserved among AEs. The underlying mechanism involves an interaction of sulfate with H⁺ that is donated by the Glu residue lining the pore. Therefore, Glu⁶⁸¹ contributes to anion translocation, as discussed below. Other residues in AE are identified to serve as the anion selectivity sites in addition to Glu⁶⁸¹ of TM8. By scanning cysteine accessibility mutagenesis (SCAM) of AE1, Casey and his colleagues (Zhu & Casey, 2004) found that the region of Ser⁸⁵²–Leu⁸⁵⁷ in the extracellular loop between the last two TMs serves as an anion selectivity filter. Whether the residues Ser⁸⁵²–Leu⁸⁵⁷ are mechanistically associated with Glu⁶⁸¹ for function is unclear.

While it is generally accepted that the transmembrane domain plays an essential role in anion selectivity, the cytoplasmic domains may also contribute to maintaining HCO₃⁻ selectivity. Deletion and missense mutants without the C-terminal putative CAII-binding site lose Cl⁻/HCO₃⁻ exchange while retaining Cl⁻/Cl⁻ self-exchange (Dahl et al., 2003).

3.1.2. NCBTs

All NCBTs are dependent upon Na⁺ (Amlal, Wang, Burnham, & Soleimani, 1998; Sciortino & Romero, 1999). The current–voltage relationship for electrogenic HCO₃⁻ transport shows a positive and linear slope with outward currents at positive voltages. The current is highly selective to HCO₃⁻ and is not elicited by other anions (Fig. 1). The charge-conserved substitution of Glu⁵⁵⁵ with an Asp (i.e. D555E) produces dose-dependent Cl⁻ currents in the absence of HCO₃⁻ (Yang, Kim et al., 2009). Nonetheless, D555E produces a HCO₃⁻ current when both Cl⁻ and HCO₃⁻ are present. Anion selectivity experiments reveal that D555E is broadly permissive to other anions including NO₃⁻, SCN⁻, I⁻, and Br⁻, whereas wild-type NBCe1 shows a high selectivity to HCO₃⁻. D555 may act to distinguish HCO₃⁻ from other trigonal planar polyanions such as NO₃⁻, as well as provide steric hindrance for small anions and SCN⁻.The mechanism appears to involve a salt bridge between Asp⁵⁵⁵ and adjacent charge residues in TM5 (unpublished observation by Soojung Lee). NBCe1 may also transport OH⁻ at high extracellular pH (Amlal et al., 1998).

Figure 1.

Anion selectivity of NBCe1. A) NBCe1 expressed in Xenopus oocytes produces electrogenic outward currents in response to HCO₃⁻, but not other anions. Recording was done while oocytes were exposed to different anions (25 mM for each). B) Zero current voltage (E_rev) versus anion radius. Note that HCO₃⁻ has a large effective radius, but produces a current. The molecular or effective radius (in Å) is 1.81 Cl⁻, 1.89 NO₃⁻, 1.96 Br⁻, 2.13 SCN⁻, 2.20 I⁻ and 2.8 HCO₃⁻. n = 6 for each.

The apparent affinity of NBCe1-A for Na⁺ is about 30 mM, and independent of different voltages and external [HCO₃⁻] (Sciortino & Romero, 1999). Similar affinity values were also observed for all three NBCe1 variants, -A, -B, and -C (apparent K_m of 21–36 mM) (McAlear & Bevensee, 2006). The apparent K_m of NBCe1 for HCO₃⁻ is 4–6.5 mM (Grichtchenko, Romero, & Boron, 2000; McAlear & Bevensee, 2006).

3.2. Ion Translocation

3.2.1. AEs

In addition to anion dependence and selectivity, Glu⁶⁸¹ plays a role in ion translocation. Chemical modification or site-directed mutagenesis of this residue alters anion exchange kinetics (Chernova et al., 1997; Sekler et al., 1995), indicating that it is located in the anion translocation pathway. An SCAM study on TM8 (Tang, Kovacs, Sterling, & Casey, 1999) also identified residues that are predicted to line the translocation pore adjacent to Glu⁶⁸¹. These are Ala⁶⁶⁶, Ser⁶⁶⁷, Leu⁶⁶⁹, Leu⁶⁷³, Leu⁶⁷⁷, and Leu⁶⁸⁰ located in one face of α-helical TM8. Furthermore, cysteine mutants at Ile⁶⁸⁴ and Ile⁶⁸⁸ in the intracellular loop are also inhibited by sulfhydryl reagents and thus may be parts of the pore. Other TMs have been implicated as part of anion translocation. Lys⁵³⁹ in TM5 is the site where the anion channel/transporter blocker 4,4′-diisothiocyanostilbene-2,2′-disulfonate (DIDS) binds (Bartel, Hans, & Passow, 1989). Mutations of His⁷⁰³, Arg⁷³¹, His⁷³⁵, His⁸¹⁶, and His⁸³⁴ located in the region of TM10–13 result in inhibition of Cl⁻ flux (Muller-Berger, Karbach, Konig et al., 1995). The His residues, particularly His⁷³⁵, are proposed to interact with Glu⁶⁸¹ and form an access channel (Muller-Berger, Karbach, Kang et al., 1995). Several other TMs from 9 to 13 are required for transport function or contribute to ion translocation. Cysteine substitutions G714C, S725C, S731C, S762C, G790C, and F806C all reduced transporter activity to <40% of wild-type AE1 (Fujinaga, Tang, & Casey, 1999). The authors also found that the sulfhydryl reagent decreases the activity of AE constructs S852C (TM12) and A858C (TM13).

3.2.2. NCBTs

Regarding ion translocation of NBCe1, site-directed mutagenesis of amino acid residues in NBCe1 has been performed to identify the structural requirements for cotransport function (Abuladze et al., 2005). Many charged and polar residues, such as Glu, Asp and Arg, scattered throughout all the TMs were found critical for transport function. Nonetheless, while these residues contribute to charge interaction with ion substrates, it is unclear whether they are responsible for ion translocation or structural components of the protein. Furthermore, noncharged or nonpolar residues may play a more important role in ion translocation. An SCAM analysis of residues in TM8 identified Leu⁷⁵⁰ among others involved in ion translocation (McAlear & Bevensee, 2006). The sulfhydryl reagent pCMBS accessibility of L750C is strongly influenced by Na⁺ and HCO₃⁻, DIDS binding, and membrane potential. Kurtz and his colleagues used a similar approach to evaluate amino acid residues in TM1 and found that T442C, A435C, and A428C have substantial sensitivity to the methylsulfonate reagents (Zhu et al., 2009). The α-helical structure of TM1 suggests that these residues would be located on the same face of the TM and serve as binding sites for ion substrates or lining the pore. The functional significance of TM1 is also evident for S427L, which has a defect in either the voltage sensor or Na⁺-binding site and fails to mediate Na/HCO₃ efflux (Dinour et al., 2004). An SCAM analysis of residues in TM10–14 reveals that the C-terminal TMs are tightly folded with limited access to the extracellular aqueous environment (Zhu et al., 2010). D894C (in a loop between TM12 and TM13) and L937C (in TM14) appear to display a severe decrease in NBC function. Figure 2 summarizes amino acid residues that have been identified as residues critical for ion translocation via NBCe1. These residues are located in TM1, TM5, TM7, TM8, and TM14.

Figure 2.

Amino acid residues critical for ion translocation through NBCe1. The residues are located in TM1, TM5, TM7, TM8, and TM14. Identification and characterization of these residues are described in the text.

While structure/function studies continue to investigate the transmembrane domain for ion translocation, the cytoplasmic domain may also play a role. The naturally occurring mutant R298S is located in the cytoplasmic N-terminal domain of NBCe1 and has reduced transport activity. By modeling NBCe1 onto the crystal structure of the AE1 N-terminal domain, Romero and his colleagues hypothesized that Arg²⁹⁸ interacts with Glu⁹¹ or Glu²⁹⁵ via hydrogen bonds or charge–charge interactions (Chang, Dipiero, Sonnichsen, & Romero, 2008). The double mutation E91R/R298E rescues transport activity, indicating that the N-terminal domain contributes to HCO₃⁻ permeation.

3.3. Mechanisms of Ion Transport

The following two different mechanistic models are considered for SLC4A transporters: a sequential (or ping-pong) mechanism and a simultaneous mechanism (Fig. 3). In the sequential mechanism, one ion substrate binds to the transporter, causes a conformational change, and is released at the trans side of the membrane. Then, a second ion substrate binds to the transporter, regenerates the original conformation, and is released on the original cis side of the membrane. Therefore, the sequential mechanism involves a series of association–dissociation steps and conformational changes. However, the simultaneous mechanism proposes that the transporter forms a ternary complex with substrates before translocation occurs. The first and second ion substrates bind to the transporter and form a tertiary complex. This causes a conformational transition that makes the bound substrates face the trans side of the membrane. The two substrates are subsequently released.

Figure 3.

Sequential (ping pong) vs simultaneous transport mechanisms. A) Cl⁻/HCO₃ exchange via a sequential mechanism. In this mechanism, the transporter (T) accepts only one ion substrate at a time at each membrane surface. T_i and T_o represent the transporter in the inward- and outward-facing states, respectively. B) Na⁺/alanine cotransport via a simultaneous mechanism. Both Na⁺ and alanine (Ala) bind to the transporter before translocation occurs. The tertiary complex T·Na⁺·Ala undergoes a conformational change. Releasing the two substrates on the opposite side of the membrane causes the transporter to be returned to the original state.

The sequential mechanism has been proposed to account for the exchange kinetics of Cl⁻ and HCO₃⁻ by AEs (Frohlich & Gunn, 1986). H₂-DIDS can bind preferentially to the outward-facing state (Jennings, Whitlock, & Shinde, 1998). The exchange rates for HCO₃⁻ or Cl⁻ are different depending upon the inward facing or outward facing state (Knauf, Law, Leung, Gehret, & Perez, 2002). In addition, for the E699Q mutation of mouse AE1 (corresponding to E681Q in human AE1), elevating intracellular [SO₄²⁻] increases the apparent K_m for extracellular Cl⁻ and SO₄²⁻ (Chernova, Stewart, Barry, Jennings, & Alper, 2008). All these data are consistent with the sequential kinetics.

It is unclear if the mechanism used by NCBTs is simultaneous or sequential. The simultaneous mechanism can be distinguished from the sequential mechanism experimentally by studying the relationship of K_m and V_max as shown for the Na/alanine transporter (Jauch & Lauger, 1986). The simultaneous mechanism exhibits a variable K_m/V_max ratio for the variable substrate at different concentrations of the fixed substrate, whereas the sequential mechanism shows a constant K_m/V_max ratio. The apparent K_m and V_max of NBCe1 for Na⁺ are unaffected at 10 and 33 mM [HCO₃⁻] (Sciortino & Romero, 1999), implying that NBCe1-mediated transport may occur via a sequential mechanism although additional experiments are needed to confirm this. The ability of D555E to produce a Cl⁻ current (Yang, Kim et al., 2009) provides some insights into the binding order. The mutant transporter favorably selects HCO₃⁻ when HCO₃⁻ and Cl⁻ are present, but this HCO₃⁻ selection requires Na⁺. In the absence of Na⁺, D555E fails to select HCO₃⁻ and instead selects Cl⁻. This indicates that HCO₃⁻ binding requires the precondition of Na⁺ binding from the perspective of the ion binding order.

3.4. Electrogenicity and Stoichiometry

The electrogenicity of a given ion transporter determines whether it is affected by the membrane voltage in addition to the ion gradient across the cell membrane. Among SLC4A transporters, NBCe1 and NBCe2 are electrogenic despite their sequence similarity with other transporters. The molecular basis of electrogenicity is poorly understood. Studies on chimeric transporters constructed from NBCe1 and NBCn1 reveal that the NBCe1 electrogenicity is conferred by the conserved transmembrane domain, while the cytoplasmic N and C termini and the large third extracellular loop (EL3) between TM5 and TM6 are not critical (Choi, Yang, & Boron, 2007). Furthermore, electrogenicity is mediated via interactions between TM1–5 and TM6–14 because swapping these two regions with NBCe1 and NBCn1 results in electroneutral Na/HCO₃ transporters. The EL4 (32 amino acids) between TM7 and TM8 has been found to be critical for electrogenicity (Chen, Liu, & Boron, 2011). Chimeric transporters with NBCe1–EL4 exhibit electrogenic properties, while those with NBCn1– EL4 exhibit electroneutral properties. EL4 may directly constitute a part of the ion-binding vestibule or coordinate with TMs to modulate ion binding.

Related to the issue of electrogenicity is the ion transport stoichiometry. The Na⁺:HCO₃⁻ stoichiometry of NBCe1 is particularly essential for HCO₃ reabsorption in the proximal tubules of the kidney, where the basolateral NBCe1-A moves 1Na⁺ and 3HCO₃⁻ into the interstitium (Soleimani, Grassl, & Aronson, 1987). If the stoichiometry were 1Na⁺ and 2HCO₃⁻, then the transporter would instead move ions in the opposite direction (i.e. into the cell) because the equilibrium potential for 1Na⁺/2HCO₃⁻ transport is estimated to be more negative than the resting potential, typically −70 mV (Boron & Boulpaep, 1983). Indeed, NBCe1 in astrocytes has a 1Na⁺:2HCO₃⁻ stoichiometry and transports ions into the cells (Bevensee, Apkon, & Boron, 1997; Bevensee, Weed, & Boron, 1997). Similarly, NBCe1 in pancreatic ducts has a 1Na⁺:2HCO₃⁻ stoichiometry and transports ions into the cells (Ishiguro, Steward, Lindsay, & Case, 1996; Zhao, Star, & Muallem, 1994). Given the different N-terminal amino acids between NBCe1-A and NBCe1-B/C, it was initially expected that the stoichiometry would be determined by the N-terminal domain of the transporter. However, NBCe1-A expressed in Xenopus oocytes mediates Na/HCO₃ influx with a 1Na⁺:2HCO₃⁻ stoichiometry (Ducoudret et al., 2001; Grichtchenko et al., 2001; Sciortino & Romero, 1999). The stoichiometry changes from 1Na⁺:2HCO₃⁻ to 1Na⁺:3HCO₃⁻ when the cytosolic Ca²⁺ concentration increases (Muller-Berger et al., 2001). Both NBCe1-A and -B show a 1Na⁺:3HCO₃⁻ stoichiometry when expressed in a proximal tubule cell line and a 1Na⁺:2HCO₃⁻ stoichiometry when expressed in a mouse collecting tubule cell line (Gross et al., 2001). These findings provide a strong argument that the same transport molecule can express different stoichiometries, depending on the cellular environment. Gross and his colleagues demonstrate that the stoichiometry can be changed by a cAMP-dependent protein kinase A phosphorylation site (Ser⁹⁸² in NBCe1-A and Ser¹⁰²⁶ in NBCe1-B) in the C-terminal domain (Gross et al., 2003).

3.5. Pharmacological Inhibition of HCO₃⁻ Transport

All SLC4A HCO₃⁻ transporters except NBCn1 are inhibited by stilbene derivatives such as DIDS, 4-acetamide-4′-isothiocyanatostilbene-2,2′-disulfonic acid (SITS) and 4,4′-dinitrostilbene-2,2′ disulfonate (DNDS). These derivatives are broad inhibitors of anion transporters/channels (Cabantchik, Knauf, & Rothstein, 1978). DIDS binds to the pore entry of anion channels/transporters and occludes the pore vestibule (Tombola, Del Giudice, Papini, & Zoratti, 2000). The inhibition constant K_i of AE1 in erythrocytes for DIDS is as low as 0.04 µM (Knauf, Law, & Hahn, 1995), while the values of cloned AEs are much larger. For NBCe1, the inhibition constant is 36 µM (Liu, Williams, Sumpter, & Bevensee, 2007). DIDS reversibly and covalently binds to the first lysine (K) in the SKLIK motif in TM 5 of AE1 (Bartel et al., 1989). NBCe1 has the homologous KKMIK motif, and mutation analysis of this motif shows the second K (i.e. K⁵⁵⁹) with the most profound effect on the reversible DIDS binding (Lu & Boron, 2007). However, in the absence of DIDS, the mutation at K⁵⁵⁹ produces a HCO₃⁻ conductance similar to those of other mutations in the KKMIK motif. Thus, K⁵⁵⁹ does not appear to serve directly as a HCO₃⁻ translocation site. Other NCBTs contain similar DIDS motifs in the corresponding region of TM5, but their role in DIDS sensitivity has not been tested.

In addition to DIDS, fluorescent oxonol dyes such as WW-781, diBA(3) C4 and diBA(5)C4 are potent inhibitors for AEs (Alper et al., 1988). The inhibition potency of these oxonol dyes is higher than that of DIDS and varies among AE isoforms. The oxonol dyes diBA(3)C4 and diBA(5)C4 also inhibit NBCe1 (Liu et al., 2007). The oxonol dyes and DIDS compete with each other, and therefore appear to share at least part of the same external binding site. It is suggested that SLC4A transporters may contain a relatively large inner vestibule, in which a number of large anions bind and block Cl⁻ permeation. Other chemicals that inhibit AEs and/or NBCs include niflumic acid (Cousin & Motais, 1976; Liu et al., 2007), benzamil and tenidap (Ducoudret et al., 2001; Lu & Boron, 2007), and an N-cyanosulphonamide inhibitor S8059 (Ch’en, Villafuerte, Swietach, Cobden, & Vaughan-Jones, 2008).

3.6. HCO₃⁻ vs CO₃⁻ Transport

The possibility that HCO₃ transporters would carry CO₃²⁻ is based on an unexpected observation by Grichtchenko and Chesler (1994). They monitored extracellular pH in the brain slices and found a pH decrease that was caused by electrogenic Na/HCO₃ influx via NBCe1 into the astrocytes. They anticipated that inhibiting extracellular carbonic anhydrase (CA) would reduce the supply of extracellular HCO₃⁻ and thus blunt the pH decrease. However, CA inhibition had the opposite effect and stimulated the pH decrease. They hypothesized that NBCe1 transports CO₃²⁻ instead of HCO₃⁻. At physiological pH (i.e. pH 7.4 in the extracellular space), HCO₃⁻ is in the millimolar range, while CO₃²⁻ is in the micromolar range and H⁺ is in the nanomolar range. If CO₃²⁻ is consumed via NBCe1, the reaction HCO₃⁻ → CO₃²⁻ + H⁺ would produce a large increase in [H⁺]. However, [HCO₃⁻] would remain relatively unchanged as it is in the millimolar range. From the reaction CO₂ + H₂O ↔ HCO₃⁻ + H⁺, one can expect that an increase in [H⁺] makes the reaction run leftward. Therefore, inhibiting extracellular carbonic anhydrase would blunt this leftward reaction and consequently enhance a decrease in extracellular pH. Boron and his colleagues (Grichtchenko & Boron, 2002; Lee, Grichtchenko, & Boron, 2011) tested this model by coexpressing NBCe1 and GPI-linked CAIV in Xenopus oocytes and measuring surface pH (pH_s). They found that stimulating NBCe1 caused pHs to change as predicted from CO₃²⁻ transport and inhibiting CAIV augmented all pH_S changes. These results are consistent with CO₃²⁻ transport rather than HCO₃⁻ by NBCe1.

3.7. Channel-Like Activity

Historically, ion transporters have been considered mechanistically and structurally different from ion channels (Hille, 2001). Ion transporters move ions slowly and exhibit distinct substrate coupling (i.e. stoichiometry) and kinetics, whereas ion channels move ions rapidly (>10⁶ ions per second) and exhibit gating and permeation. A typical view is that ion transporters structurally undergo rather marked conformational changes between inward-facing and outward-facing states, whereas ion channels undergo more subtle gating changes that open and close the pore. However, this distinction between transporters and channels became increasingly unclear as some membrane proteins share functional characteristics of both groups. Neurotransmitter transporters, including serotonin transporters, norepinephrine transporters, glutamate transporters and dopamine transporters, have imbedded channel properties (DeFelice & Blakely, 1996; Torres & Amara, 2007). Similarly, NBCn1 has robust Na⁺ channel-like activity (Na⁺ ≫ K⁺ = Cs⁺ = NMDG⁺) (Choi et al., 2000; Cooper et al., 2005). This activity can cause a positive shift in the membrane potential, and raise intracellular Na⁺ when NBCn1 is expressed in HEK 293 cells or Xenopus oocytes. The channel-like activity is not coupled to Na/HCO₃ cotransport. The domain responsible for the channel-like activity of NBCn1 is the region between TM6 and TM14 (Choi et al., 2007). NBCn1 is not the only HCO₃⁻ transporter exhibiting this activity, other SLC4A proteins also have similar activities. Erythrocyte AE1 mediates an anion flux that is stimulated by cell hyperpolarization, implicating a conductive pathway (Freedman & Novak, 1997). Trout AE1 expressed in Xenopus oocytes induces Cl⁻ currents (Fievet, Gabillat, Borgese, & Motais, 1995). The electroneutral NDAE1 in Drosophila has a small inward current (Romero et al., 2000). NBCe1 variants display a Na⁺-independent HCO₃⁻ current (McAlear, Liu, Williams, McNicholas-Bevensee, & Bevensee, 2006). Furthermore, the Cl/HCO₃ exchangers Slc26a3 and Slc26a6 produce large NO₃⁻ and SCN⁻ currents (Ko, Shcheynikov et al., 2002; Shcheynikov et al., 2008).

The channel-like activity is more than just an oddity of the transporter because it affects cellular membrane potential and controls entry of ions at least in heterologous expression systems. For example, the Na⁺ channel activity of NBCn1 lowers an electrochemical Na⁺ gradient across the cell membrane. The lowered Na⁺ gradient may alter activities of other Na⁺-conducting ion channels if such channels and NBCn1 are present together. The significance of this effect has yet to be explored.

4. STRUCTURE OF SLC4A TRANSPORTERS

AE1 is predominantly composed of dimers although both monomers and tetramers are known to exist (Reithmeier, Chan, & Popov, 1996). A structural analysis by electron microscopy and three-dimensional image reconstruction at 20 Å resolution reveal the dimeric membrane domain in the U-shaped structure (Dolder, Walz, Hefti, & Engel, 1993; Wang et al., 1994). The two monomers are in contact at the lower part (called the basal domain) embedded in the membrane, while forming a pair of protrusions at the upper part that faces the cytosol. The protrusions form the sides of a canyon, which surrounds a wide space and then converges into a depression in the basal domain. This depression may represent the pore entry or the opening to a transporter vestibule located at the dimer interface.

In support of the above model, the three-dimensional structure of the AE1 membrane domain shows V-shaped densities near the center of the dimer at 7.5 Å resolution (Hirai, Hamasaki, Yamaguchi, & Ikeda, 2011; Yamaguchi et al., 2010). Interestingly, similar V-shaped densities exist in the previously reported structure of a prokaryotic chloride channel (ClC) protein. The ClC protein contains 14 TMs with two internal repeats of 7 TMs inserted into the membrane in the transverse orientation (Dutzler, Campbell, Cadene, Chait, & MacKinnon, 2002). A V-shaped pair of helical repeats (TM1 +2 and TM8 +9) in AE1 are similar to corresponding repeats in the ClC. This similarity in the density and projection maps coincides with a sequence identity of 20%. Nonetheless, there are difficulties with this model. All the TMs cannot be clearly accounted for at 7.5 Å. The beginning of TM1 predicted from NMR studies (Chambers, Bloomberg, Ring, & Tanner, 1999) does not align with the beginning of that for ClC.

The cytoplasmic N-terminal domain of NBCe1-A has been crystallized at 3 Å (Gill & Boron, 2006b). The domain forms two monomers related by a 2-fold axis. The dimer architecture is similar to that observed for the N-terminal crystals of AE1 (Zhang, Kiyatkin, Bolin, & Low, 2000).

5. INTERACTION WITH OTHER PROTEINS

Similar to many other membrane proteins, some SLC4A transporters exist in concert with other proteins and form macromolecular complexes. It would be worth discussing the structural features associated with protein– protein interactions, which mainly occur in the cytoplasmic N- and C-terminal domains of the transporters.

5.1. AEs

The N-terminal domain in AE1 has an ankyrin-binding motif in residues 175–185 (Chang & Low, 2003). The AE1/ankyrin interaction tethers the cytoskeletal protein spectrin to the microdomains near the plasma membrane. Tyrosine phosphorylation promotes dissociation of AE1 from the spectrin-actin skeleton and may enable erythrocytes to undergo adaptive changes in response to stimuli such as malaria parasite invasion, cell shrinkage, and aging (Ferru et al., 2011). AE1 also binds to glycolytic enzymes, hemoglobin and protein 4.1 and 4.2.

The cytoplasmic C-terminal domain of AE1 interacts with CAII (McMurtrie et al., 2004; Vince & Reithmeier, 2000). The consensus sequence for the CAII binding site is DADD, where the first Asp makes an electrostatic interaction with His or Lys residues in the N-terminal region of CAII. Other AEs and different CA isoforms also interact with each other (Casey, Sly, Shah, & Alvarez, 2009). The AE1/CAII interaction significantly increases the rate of HCO₃⁻ transport (Sterling, Reithmeier, & Casey, 2001). Cl⁻/HCO₃⁻ exchange via AE1 in erythrocytes is a rate-limiting step for CO₂ delivery from tissues to the lungs (Wieth, Andersen, Brahm, Bjerrum, & Borders, 1982). Maximizing the HCO₃⁻ transport rate helps fully utilize the CO₂-transport capacity of the blood.

5.2. NCBTs

IP₃ receptor binding protein released with IP₃ (IRBIT) is a signaling protein that is released from IP₃ receptors and regulates downstream target molecules. IRBIT binds to the N terminus of NBCe1-B, but not NBCe1-A, and increases HCO₃⁻ transport activity (Shirakabe et al., 2006). IRBIT-mediated NBCe1-B stimulation has been demonstrated in pancreatic ducts (Yang, Shcheynikov et al., 2009a). The N-terminal 87 amino acids of NBCe1-B is an autoinhibitory domain that, when deleted, stimulates transport activity (McAlear et al., 2006). Deleting the N-terminal residues 2–16 of NBCe1-B abolishes the IRBIT-stimulated effect while maintaining autoinhibition (Lee, Boron, & Parker, 2012). Thus, autoinhibitory and IRBIT-binding determinants are not identical.

Some SLC4A transporters contain a motif in their C terminus to which PDZ proteins bind (PDZ stands for Post synaptic density protein PSD95, Drosophila disc large tumor suppressor Dlg1, and Zonula occludens-1 protein zo-1). PDZ proteins are proteins that interact with a variety of membrane and cytosolic proteins, modulate activities of their binding partners, associate the partners with cytoskeletal protein complexes near the membrane, and recruit signaling proteins (Kim & Sheng, 2004). NBCn1 contains the PDZ-binding motif (amino acid sequence ETSL) and interacts with the 56 kDa subunit of H-ATPase (Pushkin et al., 2003). The two proteins colocalize to the collecting duct intercalated cells of the kidney. NBCn1 also interacts with ezrin-binding protein 50 (EBP50; alias Na/H exchanger regulatory factor 1 NHERF-1) in the apical membrane of the pancreatic ducts and salivary glands (Park et al., 2002). The NBCn1/EBP50 interaction helps scavenge luminal HCO₃⁻ at rest but facilitates Cl⁻-mediated HCO₃⁻ release during stimulation. NBCn1 also interacts with the postsynaptic density protein PSD-95 at synapses in the hippocampus (Park et al., 2010). The NBCn1/PSD-95 interaction appears to increase channel-like activity of the transporter. NBCn1 also interacts with harmonin to form a protein complex in the ear (Reiners et al., 2005). Other SLC4A bicarbonate transporters capable of interacting with PDZ proteins include NBCn2-C, which contains ETCL at its C terminus and interacts with the cytoskeleton and EBP5 (Lee, Ouyang, & Giffard, 2006).

Similar to AEs, NCBTs are also reported to interact with CAII isoforms. For example, the cytoplasmic C-terminal domain of NBCe1 contains D⁹⁸⁶NDD for the CA II binding and the interaction enhances transport function (Becker & Deitmer, 2007; Pushkin et al., 2004). In contrast, other studies reveal no interaction between NBCe1 and CA II (Lu et al., 2006) or normal transport activity of a mutant lacking the CAII-binding domain (Yamada, Horita, Suzuki, Fujita, & Seki, 2011).

6. CONCLUSION

SLC4A transporters regulate intracellular pH in a variety of cells. Almost all cells in the body possess at least one of these transporters, and changes in pH influence activities of other physiologically important proteins. For example, pH alters ion permeation, sensitivity, and response to agonists of ion channel receptors in the nervous system. Slc4a-knockout mice exhibit aberrant neuronal activities that are caused by acid–base disturbance. In addition, the bicarbonate transporters are essential for transepithelial movement of acid–base equivalents in epithelial tissues such as the kidney where the transporters contribute to regulating plasma pH. Studying the molecular mechanism of how ions are selected and carried via the transporters helps better understand acid–base homeostasis at the cellular and whole-body levels. Studying the transporter structure is valuable for drug designs for transporter-related pathophysiology. Future molecular studies of SLC4A bicarbonate transporters will certainly advance our understanding of transporter physiology.