A Cysteine-scanning Mutagenesis Study of Transmembrane Domain 8 of the Electrogenic Sodium/Bicarbonate Cotransporter NBCe1*

Na/HCO3 cotransporters (NBCs) such as NBCe1 are members of a superfamily of bicarbonate transporters that includes anion exchangers. Residues within putative transmembrane domain 8 (TMD8) of anion exchanger 1 are involved in ion translocation (Tang, X. B., Kovacs, M., Sterling, D., and Casey, J. R. (1999) J. Biol. Chem. 274, 3557–3564), and the corresponding domain in NBCe1 variants is highly homologous. We performed cysteine-scanning mutagenesis to examine the role of TMD8 residues in ion translocation by rat NBCe1-A. We accessed function and/or sulfhydryl sensitivity and p-chloromercuribenzene sulfonate (pCMBS) accessibility of 21 cysteine-substituted NBC mutants expressed in Xenopus oocytes using the two-electrode, voltage clamp technique. Five NBC mutants displayed <10% wild-type activity: P743C, A744C, L746C, D754C, and T758C. For the remaining 16 mutants, we compared transporter-mediated inward currents elicited by removing external Na+ before and after exposing oocytes to either 2-aminoethylmethane thiosulfonate (MTSEA) or pCMBS. MTSEA inhibited NBC mutants T748C, I749C, I751C, F752C, M753C, and Q756C by 9–19% and stimulated mutants A739C, A741C, L745C, V747C, Q755C, and I757C by 11–21%. pCMBS mildly inhibited mutants A739C, A740, V747C, and Q756C by 5 or 8%, and stimulated I749C by 10%. However, both sulfhydryl reagents strongly inhibited the L750C mutant by ≥85%. Using the substituted cysteine accessibility method, we examined the accessibility of the NBC mutant L750C under different transporter conditions. pCMBS accessibility is (i) reduced when the transporter is active in the presence of both Na+ and \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{HCO}_{3}^{-}\) \end{document}, likely due to substrate competition with pCMBS; (ii) reduced in the presence of a stilbene inhibitor; and (iii) stimulated at more positive membrane potentials. In summary, TMD8 residues of NBCe1, particularly L750, are involved in ion translocation, and accessibility is influenced by the state of transporter activity.

Sodium/bicarbonate cotransporters (NBCs) 2 are proteins that cotransport HCO 3 Ϫ and/or CO 3 2Ϫ with Na ϩ across plasma membranes, thereby contributing to the regulation of intracellular pH (pH i ) and ion homeostasis in many tissues, including kidney, heart, and brain. Electrogenic NBCs, including NBCe1 and NBCe2, as well as electroneutral NBCs such as NBCn1 have been cloned, characterized, and localized in many tissues and cell types (1,2). Splice variants exist for the different NBCs. For example, NBCe1 contains three splice variants that differ at their amino and/or carboxyl termini: NBCe1-A, -B, and -C. Based on sequence homology, NBCs in conjunction with anion exchangers (AEs) and sodium-driven chloride-bicarbonate exchangers are members of a superfamily of bicarbonate transporters.
Over the last several years, considerable molecular advances have been made in understanding both the function and regulation of cloned NBCs, particularly NBCe1 variants. When expressed in Xenopus oocytes, all three NBCe1 variants have similar ion and voltage dependences, although the A variant is ϳ4-fold more active than the B and C variants due to its unique amino terminus (3). However, functional properties of the transporters (e.g. stoichiometry) appear to be cell-type-dependent (4). NBC activity was first characterized in the salamander proximal tubule (5), and the cDNA encoding the protein was later cloned by Romero et al. (6). Native NBCe1-A has a 1:3 Na ϩ :HCO 3 Ϫ stoichiometry in kidney (7) but a 1:2 stoichiometry when expressed in oocytes (3,8). Regarding regulation, protein kinase A-mediated phosphorylation of a carboxyl-terminal serine of human NBCe1 can decrease the apparent Na ϩ :HCO 3 Ϫ stoichiometry from 1:3 to 1:2 in transfected mammalian cells (9,10). In addition, cAMP-mediated stimulation of human NBCe1-B activity requires a threonine at the amino terminus (10). Working on NBCe1-A expressed in oocytes, Perry et al. (11) recently reported that phorbol 12-myristate 13-acetateinduced inhibition of NBCe1-A involves Ca 2ϩ -dependent protein kinase C ␣␤␥ , whereas angiotensin II-induced inhibition involves Ca 2ϩ -independent protein kinase C ⑀ .
Although the results from molecular studies have revealed regions important for NBC activity and regulation, we still lack information on the mechanisms by which NBCs bind and translocate ions. In a recent large scale mutagenesis study, Abuladze et al. (12) identified numerous residues in the transmembrane segments and intracellular/extracellular loops of NBCe1-A that are required for transporter activity. Many of these residues may be involved in ion selectivity or translocation. Considerably more work addressing mechanisms of ion binding and translocation has been done on AE proteins. AEs and NBCe1 are ϳ30% identical at the amino acid level, and their predicted membrane topologies are quite similar (1). Results from structure-function and cysteine-scanning mutagenesis studies on AE1 provide the foundation for targeting regions that may be involved in ion binding and translocation of other bicarbonate transporters.
Using cysteine-scanning mutagenesis and sulfhydryl chemistry, the Casey laboratory provided evidence that TMDs 8,9,13, and 14 are involved in ion translocation by AE1 (13)(14)(15). Particularly compelling data came from work on TMD8, which appears to line the translocation pathway. After assigning residues Met 664 -Gln 683 to TMD8 of AE1, Tang et al. (14) proposed that preceding residues Arg 656 -Met 663 form a "vestibule" that may draw substrates toward the translocation region based on accessibility studies with a cysteine-directed reagent. In a further cysteine-scanning study, the Casey group identified several residues, particularly leucines and isoleucines, of TMD8 that appear to line the transmembrane pore of AE1 (13). When these residues were mutated to cysteines, Cl-HCO 3 exchanger activity was inhibited by the sulfhydryl reagents pCMBS and/or MTSEA. At the near physiologic pH of 7.5, pCMBS (pK ϳ 1.5) is anionic and MTSEA (pK ϳ 8.5) exists in cationic and neutral forms in an MTSEA:MTSEA ϩ ratio of ϳ1:10. These reagentsensitive residues of AE1 were predicted to comprise one side of an ␣-helical model of TMD8 that threads through the membrane (see Fig. 8).
Further support for ion translocation involving TMD8 of AE1 comes from work done on the nearby glutamate residue at position 681 (Glu 681 ), which is thought to be involved in the transport process near the permeability barrier. Applying Woodward's reagent K plus BH 4 converts the carboxyl group of Glu 681 to an alcohol. Consequently, Cl Ϫ transport is inhibited but sulfate transport is stimulated (16 -18). Glu 681 is only three residues away from assigned TMD8. Although AE1 and NBCe1 share an overall identity of ϳ30%, their TMD8s are 73% homologous (see Fig. 1A). We therefore hypothesized that putative TMD8 of NBCe1 also contributes to ion translocation.
In the present study, we performed cysteine-scanning mutagenesis on rat NBCe1-A expressed in Xenopus oocytes to examine 21 residues in putative TMD8 and their potential involvement in ion translocation. Of the 21 cysteine-substituted mutants, five were non-functional (Ͻ10% NBC activity), 13 were mildly to moderately sensitive to cationic/neutral MTSEA and/or anionic pCMBS, and one (L750C) was strongly inhibited by both reagents. The sulfhydryl data are consistent with residues in TMD8 lining the translocation pathway. In subsequent experiments using the substituted cysteine accessibility method or SCAM (19,20) with the L750C mutant, we characterized the accessibility of the translocation pathway by comparing rates of pCMBS inhibition under different transport conditions.

EXPERIMENTAL PROCEDURES
Generating NBC Constructs-We used the previously described wild-type rat NBCe1-A tagged with an extracellular hemagglutinin epitope and subcloned into the oocyte expression vector pTLNII (3). The hemagglutinin epitope inserted at residue 647 in the extracellular loop between TMDs 5 and 6 does not alter transporter activity. All NBCe1 mutants were generated from hemagglutinin-tagged constructs using the QuikChange kit (Stratagene, La Jolla, CA). Vector NTI Advance 9.0 software (InfoMax, Invitrogen) was used for analyzing DNA sequences and ordering PCR primers. All constructs were verified by bidirectional sequencing (DNA Sequencing Core, Center for AIDS Research and the Genomics Core Facility, Heflin Center for Human Genetics; both at the University of Alabama at Birmingham).
Oocytes and cRNA-Oocytes from female Xenopus laevis frogs were harvested, dissociated, and washed as previously described (3). Healthy stage V/VI oocytes were isolated and then incubated (before and after injection) at 18°C in sterile ND96 supplemented with 10 mM sodium/pyruvate and 10 mg ml Ϫ1 gentamycin (Mediatech Inc., Herndon, VA). With a "Nanoject II" microinjector (Drummond Scientific, Broomall, PA), oocytes were injected with 48 nl of either RNase-free water or a stock solution containing 500 ng l Ϫ1 of NBC cRNA. Experiments were performed 3-7 days after injection.
NBCe1-A constructs subcloned into pTLNII (21,22) were first linearized with MluI and then transcribed using an SP6 transcription kit (Ambion, Austin, TX). The cRNA was purified with the RNeasy kit (Qiagen, Santa Clarita, CA) and stored at Ϫ80°C.
Two-electrode Voltage Clamping-Our method for measuring NBC currents with the two-electrode voltage clamp technique is described in detail in McAlear et al. (3). Oocytes were placed in a flow-through chamber (ϳ4 ml min Ϫ1 ), and solution changes were made using a custom-designed, solution-delivery system. To avoid contamination of our solution-delivery system, sulfhydryl-containing solutions were delivered to the chamber via a separate line that was shorter. In addition, the flow rate through this separate line was reduced to minimize excessive consumption of the reagents.
Microelectrodes were pulled from borosilicate glass capillaries, filled with saturated KCl, and attached to two channels of an OC-725C voltage clamp apparatus (Warner Instruments, New Haven, CT). Microelectrode resistances were typically 1.0 -2.0 M⍀ for the voltage electrode and 0.2-0.8 M⍀ for the current electrode. Current signals were digitized with a 1322A interface (Axon Instruments, Molecular Devices, San Jose, CA). In continual current recordings, data were obtained at a filtering frequency of 10 Hz with an 8-pole Bessel filter (LFP-8, Warner Instruments) and at a sampling frequency of 30 Hz. For current-voltage (I-V) recordings, oocytes were subjected to 12 sweeps in which the voltage was held at Ϫ60 mV for 60 ms, then stepped to 1 of 12 voltages (Ϫ200 to 20 mV, in 20-mV steps) for 20 ms, and then returned to Ϫ60 mV for 20 ms before the next sweep. Data were filtered at 800 Hz and sampled at 2 kHz. Axon Instruments' pClamp 8.2 ClampEx was used for data acquisition and ClampFit for analysis.
Solutions-The ND96 solution at pH 7.5 contained (in mM): 96 NaCl, 2 KCl, 1 MgCl 2 , 1.8 CaCl 2 , 5 HEPES, and 2.5 NaOH. For the 5% CO 2 /33 mM HCO 3 Ϫ solution, 33 mM NaCl was replaced with an equimolar amount of NaHCO 3 , and the solution was equilibrated with 5% CO 2 /95% O 2 to pH 7.5. For Na ϩfree solutions, Na ϩ was replaced with an equimolar amount of N-methyl-D-glucammonium. MTSEA and pCMBS powders were stored at Ϫ20°C. Due to the temperature sensitivity of MTS reagents, stock solutions of 1 M MTSEA in H 2 O were rapidly prepared on ice and in a 4°C cold room, and then immediately stored at Ϫ20°C. Stock solutions of the reagent were thawed on ice (for a maximum of 2 h), and periodically diluted to the 5 mM working solution during each experiment. Solutions containing light-sensitive pCMBS were protected from light at all times.
MTSEA and pCMBS were obtained from Toronto Research Chemicals, Inc. (North York, Ontario, Canada). All other chemicals were obtained from Sigma-Aldrich.
Immunoblotting-The techniques for evaluating NBC expression either in a microsomal membrane fraction by immunoblotting, or at the plasma membrane by single-oocyte chemiluminescence (SOC) have been previously described (3). For immunoblotting, individual oocytes were homogenized in buffer containing protease inhibitors, and suspensions were centrifuged to pellet cell debris and nuclei. The proteins were then separated by SDS-7.5% PAGE, transferred to an Immobilon-P polyvinylidene difluoride membrane (Millipore, Bedford, MA), and then probed first with a rabbit polyclonal antibody (Rab3A, 1:100 or 1:200) to the amino terminus of NBCe1 (23), and then with a secondary goat ␣-rabbit-IgG antibody (1:10,000) conjugated to horseradish peroxidase (Jackson ImmunoResearch Laboratories, West Grove, PA). Labeling was detected with the SuperSignal WestPico chemiluminescence kit (Pierce).
SOC-Individual oocytes were fixed with 4% paraformaldehyde for 15 min in ND96, rinsed three times in ND96, and incubated for 30 min in a 1% bovine serum albumin-ND96 blocking solution. Oocytes were then probed first with a rat monoclonal ␣-hemagglutinin antibody (1:100, Roche Applied Science, Indianapolis, IN), and then with a secondary goat ␣-rat-IgG antibody (1:400) conjugated to horseradish peroxidase (Jackson ImmunoResearch Laboratories). Labeled oocytes were incubated in 50 l of SuperSignal Elisa Femto substrate (Pierce) and immediately placed in a TD-20/20 luminometer (Turner Designs, Inc., Sunnyvale, CA). Luminescence was measured 10 s later.
Statistics-Data are reported as means Ϯ S.E. For the data on transporter function and expression, as well as sulfhydryl sensitivity, significance was determined using one-way analysis of variance (ANOVA) using the Tukey criterion, which is one of the more stringent criterions (24). p Ͻ 0.05 was considered significant. Significance was also determined using the unpaired Student's t test (one-tailed). Plots of normalized NBC activity versus cumulative time of exposure to pCMBS were well fit with a first degree exponential decay ( y ϭ y o ϩ Ae Ϫx/t ) using Origin 7.5 software (OriginLab, Northampton, MA). Portions of this work have been reported in abstract form (25). 3

Sensitivity of Introduced Cysteines to Sulfhydryl Reagents
Replacing the Native Cysteine in TMD8-As described in the Introduction, the TMD8s of AE1 and NBCe1 are highly homologous (Fig. 1A), and previous data are consistent with TMD8 of AE1 being involved in ion translocation. Therefore, we used cysteine-scanning mutagenesis with the cationic/neutral MTSEA and the anionic pCMBS to examine the potential involvement of TMD8 in ion translocation by NBCe1-A expressed in Xenopus oocytes. Rat NBCe1-A has 15 native cysteines (22). However, based on results from preliminary studies, short exposures to sulfhydryl reagents such as MTSEA did not appreciably alter the activity of NBCe1-A expressed in oocytes (not shown). This observation was confirmed in our sulfhydryl sensitivity studies associated with Figs. 3 and 4 (see below) where neither MTSEA nor pCMBS irreversibly inhibited NBC activity. Either the native cysteines are not accessible to the sulfhydryl reagents, or their modification by the reagents does not alter NBC function. In our mutagenesis studies, we therefore retained all but one (see below) of the native cysteines in NBCe1-A. Results from a cysteine-scanning mutagenesis study on a protein containing native cysteines (27,28) are more likely to yield information relevant to the native conformation of the protein.
We did replace the cysteine at position 737 located at the beginning of TMD8 (  other introduced cysteines nearby. We then compared the voltage dependences of NBCe1-A and NBCe1-A C737S . HCO 3 Ϫ -dependent I-V plots were computed from the differences between I-V plots from oocytes first in nominally HCO 3 Ϫ -free ND96, and then after being exposed to 5% CO 2 /33 mM HCO 3 Ϫ for 5 min (3). The corresponding mean HCO 3 Ϫ -dependent plot from H 2 Oinjected oocytes (not shown, n ϭ 3) was subtracted from each NBC-dependent I-V plot. As shown in Fig. 1B, the mean voltage dependences of NBCe1-A (open squares, n ϭ 4) and NBCe1-A C737S (closed diamonds, n ϭ 4) were similar.
From the same batch of oocytes, we also used SOC to compare the plasma-membrane expression levels of NBCe1-A and NBCe1-A C737S . Luminescence from each oocyte (including those injected with H 2 O) was normalized to the mean luminescence from NBCe1-A-injected oocytes. As shown in the inset of Fig. 1B, the mean normalized luminescence values (Norm. Lum.) were not different (p ϭ 0.08) from oocytes injected with NBCe1-A or NBCe1-A C737S and ϳ20-fold higher than the value obtained from H 2 O-injected oocytes. In summary, the C737S substitution in NBCe1-A does not appreciably alter the function or surface expression of NBCe1-A expressed in oocytes. NBCe1-A C737S was used in the following cysteine-scanning mutagenesis studies, and will be referred to as wt*.
Assessing Transport Function and Surface Expression of Cysteine-substituted NBCe1-A Mutants-Each of the 21 residues of putative TMD8 (see Fig. 1A) was individually replaced with a cysteine. After injecting each cRNA construct into oocytes, we assessed both transporter function using the two-electrode voltage-clamp technique and plasma-membrane expression using SOC. In our functional assay, we measured the NBCmediated outward current elicited by exposing a voltageclamped (V h ϭ Ϫ60 mV) oocyte to 5% CO 2 /33 mM HCO 3 Ϫ (3). Representative traces from voltage-clamped oocytes injected with H 2 O or cRNA encoding either wt* or one of the cysteinesubstituted mutants (P743C) are shown in Fig. 2A. At the start of the experiments, the oocytes were bathed in ND96. Subsequently exposing the wt*-expressing oocyte to 33 mM HCO 3 Ϫ elicited a rapid outward current of ϳ1.6 A that decayed slowly as previously reported for oocytes expressing wild-type rat NBCe1-A (3). In contrast, the HCO 3 Ϫ -induced current was nearly absent in oocytes injected with H 2 O or expressing the P743C mutant.
From experiments similar to those shown in Fig. 2A, we determined the mean HCO 3 Ϫ -induced outward current from oocytes injected with each cysteine-substituted NBC construct as a percentage of the corresponding mean current obtained from batch-matched, wt*-injected oocytes (Fig. 2B). Small HCO 3 Ϫ -dependent currents obtained from batch-matched, H 2 O-injected oocytes were subtracted. Although a majority of the NBC mutants displayed a mean decrease in transporter activity, most retained at least 40% of wt* activity. However, the following five substitutions lead to Ͼ90% loss of wt* activity: P743C, A744C, L746C, D754C, and T758C. These five cysteine-substituted NBC mutants were not used in subsequent sulfhydryl studies due to their low activities.
We used the SOC technique to examine oocyte surface expression of the five low active NBC mutants. Luminescence readings were normalized to the mean value from batch-matched, wt*-injected oocytes. Of the five mutants, only those with substitutions P743C and A744C displayed a markedly lower mean Norm. Lum. compared to that of wt* (Fig. 2C). Based on immunoblot data, these two NBC mutants were present in total microsomal protein from injected oocytes (insets above bar graphs, Fig. 2C). Immunoblot data from corresponding wt*-expressing oocytes are also shown. The reduced surface expression of the P743C mutant may be partially due to reduced total mutant protein expression. In summary, low NBC activity with substitutions P743C and A744C is predominantly due to reduced surface expression, whereas the low activity with substitutions L746C, D754C, and T758C is due to reduced function of the transporter at the plasma membrane. SOC analysis of surface expression was performed on all the cysteine-substituted NBC mutants, and the summary data are reported in supplemental Fig. S1. Reduced NBC activity was associated with no/little change in surface expression for 38% of the mutants, or a decrease in surface expression for 29% of the mutants.
Sensitivity to MTSEA-The 16 of 21 cysteine-substituted NBC mutants that displayed Ͼ10% of wt* activity were used in subsequent sulfhydryl sensitivity studies. We used the following assay to test for sensitivity of NBC mutants to either MTSEA or pCMBS. After being incubated in a solution containing 5% CO 2 /33 mM HCO 3 Ϫ , a voltage-clamped oocyte at Ϫ -dependent outward currents from oocytes expressing each cysteine-substituted NBC mutant. The mean currents are reported as the percentage of the mean current from batchmatched wt*-expressing oocytes, and corresponding currents from H 2 O-injected oocytes are subtracted. n Ն 4 for each bar. †, reduced NBC function; *, increased NBC function ( p Յ 0.05, ANOVA). C, mean normalized luminescence (Norm. Lum.) of oocytes expressing one of five low-active NBC mutants or wt*, or injected with H 2 O. n Ն 4 for each bar. *, The mean Norm. Lum. of oocytes expressing P743C and A744C NBC mutants was similar to that of H 2 O-injected oocytes. The mean Norm. Lum. of oocytes expressing the other NBC mutants compared to wt* was no different ( p Ͼ 0.07, ANOVA). Immunoblots of total microsomal protein from oocytes expressing either P743C or A744C NBC mutants and batch-matched wt* are displayed above bars 3 and 4. The lower wt* NBC band likely represents a breakdown product of the full-length protein. For each blot, 40% of total protein from a single oocyte was examined.
Ϫ60 mV was transiently exposed to a Na ϩ -free, HCO 3 Ϫ solution both before and after being incubated in the HCO 3 Ϫ solution containing the sulfhydryl reagent (pH ϳ7.5). Removing external Na ϩ elicits an inward current due to NBC-mediated transport of Na ϩ , HCO 3 Ϫ , and net-negative charge out of the oocyte (6,8). The simplest interpretation of a smaller inward current after sulfhydryl exposure is NBC inhibition due to an irreversible, covalent interaction of the reagent with the introduced cysteine.
An example of a sensitivity experiment on a voltage-clamped oocyte (Ϫ60 mV) expressing wt* is shown in Fig. 3A. The 5% CO 2 /33 mM HCO 3 Ϫ solution elicited an outward current of ϳ1.5 A similar to that shown in Fig. 2A. The oocyte was incubated in the HCO 3 Ϫ solution for 5 min to allow for intracellular equilibration of the physiologic buffer before being subjected to the Na ϩ -removal protocol. Removing external Na ϩ elicited a robust inward current of ϳ1.2 A that was reversible when the oocyte was returned to the Na ϩ -containing HCO 3 Ϫ solution. Exposing the oocyte to the HCO 3 Ϫ solution containing 5 mM MTSEA for 4 min also generated an inward current, apparently due to NBC inhibition. However, this inhibition was non-covalent because it was readily reversible. Removing external Na ϩ a second time after MTSEA exposure again elicited an inward current (ϳ1.2 A) of similar magnitude to that before MTSEA exposure. Thus, MTSEA in our assay does not alter wt* activity through an irreversible, covalent interaction.
The same experimental protocol yielded markedly different results with an oocyte expressing NBCe1-A with the substitution L750C (Fig. 3B). As expected from the Fig. 2B summary data, both the HCO 3 Ϫ -stimulated outward current and the 0 Na ϩ -induced inward current with the L750C construct were smaller than the corresponding currents with wt*. Importantly however, the inward current elicited by 5 mM MTSEA was not readily reversible with the L750C mutant. Furthermore, the inward current elicited by removing external Na ϩ was inhibited ϳ92% after the oocyte was exposed to MTSEA. The inhibition is consistent with MTSEA covalently interacting with C750 in TMD8 and sterically blocking ion translocation. The experimental maneuvers shown in Fig. 3 (A and B) elicited only small nanoamp currents in an H 2 O-injected oocyte (Fig. 3C).
From experiments similar to those shown in Fig. 3 (A and B), the percent inhibition by MTSEA of each of the 16 functional cysteine-substituted NBC mutants is plotted in Fig. 3D. Data for each mutant were obtained from at least two batches of oocytes, and the means of small currents from the Na ϩ -removal assay on batch-matched, H 2 O-injected oocytes were subtracted. 4 Seven NBC mutants with an introduced cysteine were inhibited ( p Ͻ 0.03) by MTSEA compared with the slight 3% stimulation seen with wt*. MTSEA modestly inhibited mutants T748C, I749C, I751C, F752C, M753C, and Q756C by 9 -19% and strongly inhibited the L750C mutant by 85%. The data are consistent with these seven positions, which are clustered together from position 748 to 756, contributing to ion translocation by NBCe1. We also found that MTSEA stimulated ( p Ͻ 0.04) the following six mutants by 11-21%: A739C, A741C, L745C, V747C, Q755C, and I757C. Such stimulation also supports the involvement of TMD8 in ion translocation.
Sensitivity to pCMBS-Similar sulfhydryl experiments were performed on the 16 functional cysteine-substituted NBC mutants using 1 mM pCMBS for 2 min (Fig. 4) rather than 5 mM MTSEA for 4 min. Our sulfhydryl-sensitivity assay with pCMBS yielded results qualitatively similar to those with FIGURE 3. MTSEA sensitivity of cysteine-substituted NBC mutants. A, a voltage-clamped oocyte (V h ϭ Ϫ60 mV) expressing wt* and incubated in 5% CO 2 /33 mM HCO 3 Ϫ was subjected to a Na ϩ -removal assay both before and after exposure to 5 mM MTSEA for 4 min. Removing external Na ϩ elicited inward currents (due to reverse NBC activity) of similar magnitude before and after MTSEA exposure. B, a similar experiment was preformed on an oocyte expressing the L750C NBC mutant. After the oocyte was exposed to MTSEA, removing Na ϩ elicited an inward current that was 14-fold smaller than before the MTSEA exposure. C, a H 2 O-injected oocyte subjected to the same protocol displayed only small currents when exposed to the Na ϩ -removal protocol before and after MTSEA exposure. D, summary data of % NBC inhibition by MTSEA from oocytes expressing each of the cysteine-substituted NBC mutants subjected to the experimental protocol shown in A and B. n Ն 4 for each bar, and oocytes were from at least two batches. †, inhibited by MTSEA ( p Ͻ 0.03, ANOVA) compared to wt*. *, stimulated by MTSEA ( p Ͻ 0.04, ANOVA) compared to wt*.
MTSEA for oocytes expressing wt* (Fig. 4A) or the L750C mutant (Fig. 4B), or injected with H 2 O (Fig. 4C). From experiments similar to those shown in Fig. 4 (A and B), the percent inhibition by pCMBS of each of the 16 functional cysteine-substituted NBC mutants is plotted in Fig. 4D after subtracting the small currents from H 2 O-injected oocytes exposed to the Na ϩremoval assay. Data for each mutant were obtained from at least two batches of oocytes. Four substitutions created NBC mutants that were mildly, but significantly ( p Ͻ 0.05) inhibited by pCMBS compared to the slight 1% stimulation seen with wt*.

Accessibility of L750C to pCMBS
Accessibility Assay-According to the results from our sulfhydryl sensitivity studies, the L750C construct is particularly sensitive to inhibition by both MTSEA and pCMBS. Therefore, position 750 is a critical part of the ion translocation pathway of NBCe1. To characterize this pathway further under different FIGURE 4. pCMBS sensitivity of cysteine-substituted NBC mutants. A, a voltage-clamped oocyte (V h ϭ Ϫ60 mV) expressing wt* and incubated in CO 2 /HCO 3 Ϫ was subjected to the Na ϩ -removal assay before and after exposure to 1 mM pCMBS for 2 min. Removing external Na ϩ elicited inward currents of similar magnitude before and after pCMBS exposure. B, a similar experiment was performed on an oocyte expressing the L750C mutant, and no appreciable inward current upon removing Na ϩ was observed after pCMBS exposure. C, a H 2 O-injected oocyte subjected to the same protocol displayed only small currents when exposed to the Na ϩ -removal protocol before and after pCMBS exposure. *, the magnitude of the transient current spike due to changing solutions was truncated. D, summary data of % NBC inhibition by pCMBS from oocytes expressing each of the cysteine-substituted NBC mutants subjected to the experimental protocol shown in A and B. n Ն 4 for each bar, and oocytes were from at least two batches. † inhibited by pCMBS ( p Ͻ 0.05, ANOVA) compared to wt*. *, stimulated by pCMBS ( p ϭ 0.02, ANOVA) compared to wt*. FIGURE 5. pCMBS accessibility assay. A, a voltage-clamped oocyte (V h ϭ Ϫ60 mV) expressing wt* was exposed to 1 mM pCMBS in CO 2 /HCO 3 Ϫ (CB) for 15 s (c-e trace, arrow) between control exposures to CB without pCMBS (e.g. a-c and e-g traces). The magnitudes of the NBC-mediated outward currents elicited by CB did not change following repeated exposures to pCMBS. B, a similar experiment was performed on an oocyte expressing the L750C mutant. The magnitudes of the NBC-mediated outward currents decreased after repeated exposures to pCMBS. C, plot of normalized NBC current (Norm. I NBC ) versus cumulative time of pCMBS exposure from the experiments shown in A and B. The L750C data were fit with an exponential decay with rate constant k. experimental conditions, we quantitated the accessibility of L750C from the rate of transport inhibition by pCMBS.
The details of our experimental protocol are presented in Fig.  5A on a wt*-expressing oocyte voltage clamped at Ϫ60 mV and initially bathed in ND96. NBC activity was determined from the magnitude of the outward current elicited by exposing the oocyte briefly to a solution containing 5% CO 2 /33 mM HCO 3 Ϫ (Fig. 5A, CB and trace a-c). The oocyte was then transiently exposed to the HCO 3 Ϫ solution containing 1 mM pCMBS for 15 s (arrows). As expected from the Fig. 4A results, the HCO 3 Ϫinduced NBC current was smaller in the presence (trace c-e) versus absence (trace a-c) of pCMBS, at least partially due to reversible inhibition by the sulfhydryl reagent. NBC activity after pCMBS exposure was again assessed by transiently exposing the oocyte to the HCO 3 Ϫ solution without pCMBS (trace e-g). The oocyte was then alternately exposed to the HCO 3 Ϫ solution with or without pCMBS several additional times. The HCO 3 Ϫ -induced current after each 15-s exposure to pCMBS was normalized to the first HCO 3 Ϫ -induced current prior to pCMBS exposure (trace a-c). The normalized currents were then plotted as a function of cumulative exposure to pCMBS (Fig. 5c, wt*). For wt*, the NBC-mediated outward currents did not change appreciably with increasing cumulative exposure to pCMBS, a finding that corroborates the pCMBS insensitivity of wt* shown in Fig. 4.
A similar experiment performed on an oocyte expressing the L750C mutant yielded different results (Fig. 5B). After the oocyte was exposed to pCMBS for only 15 s (trace c-e), the following HCO 3 Ϫ -induced current (trace e-g) was ϳ30% smaller than the current before exposure to the sulfhydryl reagent (trace a-c). Furthermore, the currents became progressively smaller following subsequent 15-s exposures to pCMBS. The smaller currents reflect NBC inhibition due to covalent binding of pCMBS. The progressively smaller currents required the presence of the sulfhydryl reagent. For example, oocytes expressing the L750C construct displayed a mean decrease in NBC current of only 4 Ϯ 7% (n ϭ 3) after five 15-s exposures to the HCO 3 Ϫ solution without pCMBS. As shown in Fig. 5C  (L750C), inhibition of the L750C mutant as a function of cumulative exposure to pCMBS was well fit by an exponential decay with a rate constant (k) of 0.027 s Ϫ1 . NBC inhibition of ϳ85% at the 105-s time point corroborates the high pCMBS sensitivity of the L750C mutant shown in Fig. 4.
Using the aforementioned accessibility assay, we next examined how pCMBS accessibility of L750C was influenced by different states of transporter activity. We compared rates of pCMBS inhibition at high and low concentrations of the transported substrates (i.e. Na ϩ and HCO 3 Ϫ ), in the presence of the reversible inhibitor DNDS, and at different membrane potentials. For these experiments, oocytes were subjected to these different conditions only during pCMBS exposure. In a series of control experiments on wt*, the pCMBS-pulse protocol under all the different conditions did not elicit a progressive decrease in NBC currents (e.g. see Fig. 5, A and C). In fact, there was a mild mean stimulation that ranged from 7 to 32% by the end of the experiments. However, such stimulations appear to be pCMBS-independent because similar responses were observed with the Na ϩ -and HCO 3 Ϫ -containing solution without pCMBS in oocytes at Ϫ60 or ϩ60 mV.
Effect of Transported Substrates on Accessibility-We performed experiments and analyses similar to those shown in Fig.  5 on the L750C construct, except that pCMBS was applied to oocytes (V h ϭ Ϫ60 mV) in the presence or nominal absence of 33 mM HCO 3 Ϫ , and in the presence of either 98.5 mM or 1 mM Na ϩ . In the low Na ϩ experiments, we first exposed oocytes to a Na ϩ -free solution for ϳ20 s to lower external Na ϩ in the bath solution prior to applying pCMBS. In control experiments on the L750C mutant, the experimental conditions had little inhibitory effect on the NBC currents in the absence of pCMBS. For example, oocytes displayed a mean decrease in NBC current of only 9 Ϯ 4% (n ϭ 4) after five 15-s exposures to the 1 mM Na ϩ , HCO 3 Ϫ solution without pCMBS. However, the mean percent decrease in NBC current was approximately an order of magnitude greater for all conditions in the presence of pCMBS. More specifically, the mean current decreased by 83 Ϯ 1% (n ϭ 12) with 98.5 mM Na ϩ /33 mM HCO 3 Ϫ , 91 Ϯ 1% (n ϭ 7) with 98.5 mM Na ϩ /"0" mM HCO 3 Ϫ , 91 Ϯ 2% (n ϭ 4) with 1 mM Na ϩ /"0" mM HCO 3 Ϫ , and 90 Ϯ 1% (n ϭ 5) with 1 mM Na ϩ /33 mM HCO 3 Ϫ . Rate constants of pCMBS inhibition under the different conditions (Fig. 6A) were normalized to the mean rate constant with 98.5 mM Na ϩ /33 mM HCO 3 Ϫ obtained from batchmatched experiments. Mean normalized rate constants (Norm. k) were 1.7-to 2.2-fold greater ( p Յ 0.013) with low levels of one or both of the substrates (bars 2-4) than with 98.5 mM Na ϩ and 33 mM HCO 3 Ϫ (bar 1). Thus, position 750 is accessible to pCMBS in the absence of transporter activity (e.g. in the nominal absence of HCO 3 Ϫ ). In addition, physiologic levels of Na ϩ and HCO 3 Ϫ together reduce pCMBS accessibility, probably due to competition of the transported anion (either HCO 3 Ϫ , CO 3 2Ϫ , or NaCO 3 Ϫ ) with anionic pCMBS through the active transporter.
Effect of DNDS Binding on Accessibility-In further experiments similar to those described above, we examined the effect of stilbene binding on L750C accessibility to pCMBS. We used the stilbene derivative DNDS because it is a reversible inhibitor of HCO 3 Ϫ transporters (29) and inhibits rat NBCe1-B expressed in oocytes (3). In the present study, we found that 2 mM DNDS partially and reversibly inhibited the HCO 3 Ϫ -induced outward current (V h ϭ Ϫ60 mV) by 36 Ϯ 2% (n ϭ 3) in oocytes expressing wt* (Fig. 6B, left pair of bars). DNDS also inhibited the HCO 3 Ϫ -induced outward current to a slightly greater extent (54 Ϯ 0.3%, n ϭ 3) in oocytes expressing the L750C mutant (Fig.  6B, right pair of bars). In our accessibility assay, we performed experiments similar to those shown in Fig. 5, except that pCMBS was applied with or without DNDS. Because DNDS is reversible, the time course of inhibition of the HCO 3 Ϫ -induced outward currents after exposures to pCMBS plus DNDS is mainly due to covalent pCMBS inhibition and the effect of DNDS on pCMBS accessibility.
In control experiments on the L750C mutant, the experimental protocol with DNDS had a minimal effect on the HCO 3 Ϫ -induced currents in the absence of pCMBS. Expectedly, DNDS did not irreversibly inhibit NBC activity. For example, two oocytes expressing the L750C mutant displayed a mean increase in NBC current of only 3% (n ϭ 2) after five 15-s expo- Ϫ . This pCMBS-dependent decrease in NBC current was clearly less than that observed in the absence of DNDS reported above (83 Ϯ 1%). Therefore, DNDS reduced the degree of pCMBS inhibition after these five pCMBS exposures. DNDS also reduced the rate of pCMBS inhibition. As summarized in Fig. 6C, the mean Norm. k of sulfhydryl inhibition was ϳ45% less in the presence (bar 2) versus absence (bar 1) of DNDS. Thus, the binding of DNDS reduced the accessibility of L750C to pCMBS, likely due to a steric or allosteric effect of the stilbene.

Effect of Membrane Potential on Accessibility-Because
NBCe1 is electrogenic, we next tested the hypothesis that accessibility of L750C to pCMBS is influenced by membrane potential. The experimental protocol in Fig. 7A is the same as that shown in Fig. 5B, except that the holding potential (V h ) was stepped to Ϫ120 mV a few seconds before and after the oocyte was exposed to the pCMBS solution (arrows). The HCO 3 Ϫ -induced currents progressively decreased after oocytes were repetitively exposed to 33 mM HCO 3 Ϫ plus pCMBS at Ϫ120 mV. Similar experiments were performed at three or four V h values between Ϫ120 mV and ϩ60 mV, and with pCMBS either in the ND96 or 33 mM HCO 3 Ϫ solution. In control experiments on the L750C mutant, the experimental HCO 3 Ϫ protocols at the different V h values had little inhibitory effect (4 -6%, n ϭ 2 or 3 for each protocol) on NBC current in the absence of pCMBS. However, for experiments on the L750C mutant with pCMBS, the mean NBC current decreased by 76 -91% after four or five 15-s exposures to the pCMBS-containing solutions Ϯ 33 mM HCO 3 Ϫ at the different V h values (n Ն 3 for all conditions). Rate constants of pCMBS inhibition obtained from the experiments were normalized to the mean value from batch-matched exper-FIGURE 6. Effect of Na ؉ , HCO 3 ؊ , and DNDS on the accessibility of L750C to pCMBS. A, mean normalized rate constants (Norm. k) were obtained from experiments similar to the one shown in Fig. 5B. pCMBS solutions contained either high or low concentrations of Na ϩ with or without 5% CO 2 /33 mM HCO 3 Ϫ . †, the mean Norm. k in the presence of physiologic levels of both Na ϩ and HCO 3 Ϫ (bar 1) was significantly less ( p Յ 0.001, ANOVA) than the mean Norm. k with low levels of one or both substrates (bars 2-4). n Ն 4 for each bar, and oocytes were from two or more batches. The mean Norm. k in "0" HCO 3 Ϫ /1 mM Na ϩ (bar 3) was slightly less ( p ϭ 0.04, ANOVA) than the mean Norm. k in "0" HCO 3 Ϫ /98.5 mM Na ϩ (bar 2). B, mean HCO 3 Ϫ -induced outward currents from oocytes (V h ϭ Ϫ60 mV) expressing either wt* (left pair of bars) or the L750C NBC mutant (right pair of bars) in the absence and presence of 2 mM DNDS (paired data). n ϭ 3 for each bar. C, the mean Norm. k from experiments similar to those described in panel A in which the pCMBS solution contained 33 mM HCO 3 Ϫ Ϯ 2 mM DNDS. The mean Norm. k was 45% less in the presence versus absence of DNDS ( p Ͻ 0.001, unpaired Student's t test). n Ն 4 for each bar, and oocytes were from two batches. FIGURE 7. Effect of membrane potential on the accessibility of L750C to pCMBS. A, a voltage-clamped oocyte (V h ϭ Ϫ60 mV) expressing the L750C mutant was exposed to 1 mM pCMBS in 5% CO 2 /33 mM HCO 3 Ϫ (CB) for 15 s (arrows) between control exposures to CB without pCMBS. Immediately before and after pCMBS application, V h was changed to Ϫ120 mV (top trace). Similar to that shown in Fig. 5B, the magnitudes of the NBC-mediated outward currents elicited by CB decreased following repeated exposures to pCMBS. B, mean normalized rate constants (Norm. k) from experiments similar to the one shown in panel A. Data were normalized to batch-matched mean values from experiments with HCO 3 Ϫ at Ϫ60 mV (dotted lines). Norm. k were obtained with pCMBS pulse protocols at different V h values and with or without HCO 3 Ϫ . n Ն 4 for each data point, and oocytes were from two or more batches. Error bars smaller than symbols are not shown.
iments performed with HCO 3 Ϫ and at V h ϭ Ϫ60 mV. Mean Norm. k in both the presence and absence of HCO 3 Ϫ increased at more positive V h values in a linear fashion (Fig. 7B). Similar to the Ϫ60-mV data presented in Fig. 6A, the rate constants were 2.0-to 2.5-fold greater in the nominal absence (open squares) versus presence (closed squares) of HCO 3 Ϫ at all V h values (Fig. 7B).

DISCUSSION
We report, for the first time, a region of a cation-coupled HCO 3 Ϫ transporter that is involved in ion translocation across the plasma membrane. In a cysteine-scanning mutagenesis study of putative TMD8 of rat NBCe1-A, 13 cysteine-substituted NBC mutants were mildly to moderately sensitive to cationic/neutral MTSEA and/or anionic pCMBS, and the L750C mutant was strongly inhibited by both reagents. Inhibition by anionic pCMBS is consistent with TMD8 residues lining the anion pathway. The L750C mutant was used with SCAM to characterize pCMBS accessibility of the pathway under different conditions of transport. Accessibility of TMD8 is influenced by Na ϩ and HCO 3 Ϫ , stilbene inhibition, and membrane potential.
Sulfhydryl Sensitivity of TMD8 of NBCe1-In our cysteinescanning mutagenesis studies, we used the sulfhydryl reagents MTSEA and pCMBS because both have been used in a similar study on TMD8 of AE1 (13). MTSEA is a logical reagent to use in an initial screen because it exhibits the widest reactivity compared with other sulfhydryl reagents. Furthermore, both MTSEA and pCMBS are relatively small compared to other sulfhydryl reagents (20).
As described in the Introduction, pCMBS is anionic, whereas MTSEA exists in both cationic and uncharged forms (MTSEA: MTSEA ϩ ratio of ϳ1:10) in our solutions at pH ϳ7.5. The pCMBS data are consistent with residues of TMD8 lining the anion pathway of NBCe1, similar to that observed in AE1 (13). Based on the size and configuration of pCMBS, the diameter of the pathway is predicted to have a lower limit of ϳ6 Å (20). The transported species through the pathway could be HCO 3 Ϫ , CO 3 2Ϫ , or NaCO 3 Ϫ . Compared to the pCMBS data, the MTSEA data are less conclusive because MTSEA effects could result from either the cationic or neutral form traveling through the pathway, or the membrane permeant, neutral form bypassing the pathway and diffusing into the membrane. It is therefore not surprising that more of our NBC mutants are sensitive to MTSEA than pCMBS. According to preliminary data, the MTSEA-inhibited L750C mutant is relatively insensitive to the purely cationic sulfhydryl reagent MTSET. 5 This finding supports the possibility that MTSEA inhibition is due to the neutral form of the reagent. On the other hand, MTSET accessibility may be less because it is larger than MTSEA. Further studies with cationic MTS reagents are necessary to test for TMD8-mediated cation translocation.
In our studies, we found that sulfhydryl reagents stimulated the activity of some cysteine-substituted NBC mutants. Six mutants were stimulated 11-21% by MTSEA, and one was stimulated 10% by pCMBS. Such stimulation may result from bound reagent altering the conformation of the protein (or pathway), or the bound cationic form of MTSEA drawing transported anions into the pathway. In cysteine-scanning studies on ion channels, current potentiation can result from a sulfhydrylinduced decrease in EC 50 for an agonist or a large change in gating kinetics (20). The sulfhydryl-induced stimulation of our NBC mutants may therefore reflect an increase in the activation kinetics of the transporter.
Comparison of Cysteine-scanning Results on TMD8s of NBCe1 and AE1-There are similarities and differences between the sulfhydryl data on NBCe1 (this study) and AE1 (previously reported by Tang et al. (13)). We mapped the TMD8 sequence of NBCe1 (Fig. 8A, right) on the ␣-helical topology model of AE1 (Fig. 8A, left) presented by Tang et al. (13) and then compared the pCMBS-sensitive sites between the two domains. The majority of the sites in both TMD8s lie on one side of the ␣ helix. This positioning in NBCe1 is clear from the helical wheel model (Fig. 8B) in which all but one of the pCMBS-inhibited sites are within a 100°arc of the helix as it spirals through the membrane. Three of the pCMBS-sensitive sites in NBCe1 (Ala 739 , Ala 740 , and Leu 750 ) are at identical positions in AE1 (Fig. 8A). The Leu 750 position is particularly interesting because the L750C mutant is completely inhibited by pCMBS and 85% inhibited by MTSEA. In a similar fashion, the homologous AE1 mutant (L677C) is one of only two residues sensitive to both reagents (13).
Leu 750 also contributes to function and ion translocation of other NBCe1 variants based on the following two observations (not shown). First, replacing the leucine in the homologous position (794) of rat NBCe1-C with either a threonine, isoleucine, or cysteine (NBCe1-C L794C ) caused a decrease of ϳ85% in transporter function in oocytes without affecting surface expression. Second, the residual activity of NBCe1-C L794C was inhibited ϳ75% after a 4-min preincubation in MTSEA. In summary, Leu 750 in the A variant (and Leu 794 in the B and C variants) appears to be a critical residue in ion translocation involving TMD8.
We could not test all TMD8 residues for sulfhydryl sensitivity because five NBC mutants had very low (Ͻ10%) activity. One or more of these five positions may also be involved in ion translocation, especially because three of these residues (Pro 743 , Leu 746 , and Asp 754 ) are interspersed with pCMBS-sensitive sites on the same side of the transmembrane helix (Fig. 8). Leu 746 is homologous to AE1's Leu 673 , which is one of two TMD8 sites sensitive to both MTSEA and pCMBS. Asp 754 is homologous to AE1's Glu 681 , which is sensitive to Woodward's reagent K and involved in the transport process near the permeability barrier as described in the Introduction. In a large scale mutagenesis study on human NBCe1-A, Abuladze et al. (12) also observed a dramatic decrease in transporter activity when Asp 754 was mutated to one of three other residues.
The most notable difference between the NBCe1 and AE1 sulfhydryl data on TMD8 is the relatively high number of MTSEA-sensitive NBC mutants. 13 cysteine-substituted NBC mutants are sensitive to MTSEA compared to only three AE1 mutants (13). For NBCe1, the introduced cysteines that lead to MTSEA-induced NBC inhibition do not dominate one side of the helix, but instead, cluster at positions 748 -756 near the predicted inner surface of the membrane. Three pCMBS-sensitive sites described above are also found in this location. The cytoplasmic versus extracellular half of TMD8 may therefore contribute more to the structure of the translocation pathway.
In light of the aforementioned data, we suspect that TMD8mediated ion translocation is different for NBCs and AEs. To explore this possibility in more detail, we created an NBCe1-AE2 chimera in which TMD8 of NBCe1 was replaced with TMD8 of AE2 (not shown). Unfortunately, the chimera expressed in oocytes did not express at the plasma membrane based on SOC analysis. Swapping the TMD8s apparently alters protein conformation enough to inhibit proper trafficking to the oocyte surface. Further studies are required to determine if differences in TMD8-mediated ion translocation contribute to the fundamental differences in ion selectivities (Na ϩ versus Cl Ϫ ) and modes of transport (cotransport versus exchange) of AEs and NBCs.
Substrate Dependence of pCMBS Accessibility-We were somewhat surprised by the finding that pCMBS accessibility of L750C is greater in the absence versus presence of physiologic levels of Na ϩ and HCO 3 Ϫ . Although one might expect accessibility to be higher in the presence of transported substrates, the reduced accessibility is likely due to competition of pCMBS with the anionic substrate for the translocation pathway. The fact that both Na ϩ and HCO 3 Ϫ are required for the reduced accessibility suggests that the competition requires an active transporter, or simply the binding of both substrates. The requirement for the presence of both Na ϩ and HCO 3 Ϫ to reduce accessibility is also consistent with the transported anionic species being NaCO 3 Ϫ . Indeed, NBCe1 expressed in oocytes has been reported to transport CO 3 2Ϫ based on results from measurements of surface pH (30).
DNDS Inhibition of Accessibility-In studies on AE1 and NBCe1, mutating lysine residues involved in inhibition by stilbenes does not inhibit transporter function (31)(32)(33)(34). However, two stilbene-reactive lysines in AE1 appear to have allosteric effects on substrate binding (31). More recently, Salhany (35) has presented kinetic evidence that stilbenes compete allosterically rather than directly with substrate binding. Although the stilbene and substrate binding sites appear to be distinct, stilbene binding may interfere with accessibility of the translocation pathway. Indeed, we found that DNDS reduced accessibility of L750C to pCMBS by 45%. In rat NBCe1-A, two stilbene binding sites with the amino acid sequence KL(X)K (where X ϭ I, V, or Y) are found near the extracellular end of TMD5 (resi-  (13). Very low active residues of NBCe1 displayed Ͻ10% function in our study, whereas very low active residues of AE1 displayed Յ25% function (13). B, TMD8 of NBCe1 spiraling through the membrane is modeled as a helical wheel (36) viewed from the extracellular solution above the membrane.
dues 558 -561), as well as within the intracellular loop between TMD8 and -9 (residues 768 -771) (22). Because anionic stilbenes have limited membrane permeability (29), extracellular DNDS likely inhibits NBCe1 predominantly at the TMD5 site (see Ref. 33). The reduced pCMBS accessibility by DNDS is therefore consistent with the extracellular end of TMD5 being in close contact with TMD8 and DNDS sterically reducing access to the translocation path. Alternatively, DNDS binding may reduce accessibility by altering the transporter's conformation.
Effect of Membrane Potential on Accessibility-We found that the accessibility of L750C to pCMBS is voltage-dependent. More specifically, pCMBS accessibility was progressively greater at more positive potentials for oocytes either in the nominal absence or presence of CO 2 /HCO 3 Ϫ . The effect in the absence of HCO 3 Ϫ indicates that the voltage dependence of accessibility does not require transporter activity. The voltage dependence of pCMBS accessibility could reflect voltage-dependent conformational changes of TMD8. Alternatively, electrostatic forces may influence the translocation pathway, and anionic pCMBS may be drawn into the pathway at more positive potentials.
In summary, we have used cysteine-scanning mutagenesis and sulfhydryl reagents MTSEA and pCMBS to identify residues within TMD8 of rat NBCe1-A involved in ion translocation. According to pCMBS accessibility studies on the most sulfhydryl-sensitive NBC mutant (L750C), TMD8-mediated ion translocation can be influenced by the transported substrates, stilbene binding, and membrane potential. As described in the Introduction, TMDs other than 8 also appear to be involved in AE1-mediated ion translocation. Cysteine-scanning mutagenesis and SCAM will help in characterizing the homologous regions of NBCe1. The results will be germane to other members of the transporter superfamily and will divulge similarities and differences in ion translocation among HCO 3 Ϫdependent exchangers and cotransporters.