Mercury and zinc differentially inhibit shark and human CFTR orthologues: involvement of shark cysteine 102

Gerhard J. Weber, Ali Poyan Mehr, Jeffrey C. Sirota, Stephen G. Aller, Sarah E. Decker, David C. Dawson, John N. Forrest Jr.


The apical membrane is an important site of mercury toxicity in shark rectal gland tubular cells. We compared the effects of mercury and other thiol-reacting agents on shark CFTR (sCFTR) and human CFTR (hCFTR) chloride channels using two-electrode voltage clamping of cRNA microinjected Xenopus laevis oocytes. Chloride conductance was stimulated by perfusing with 10 μM forskolin (FOR) and 1 mM IBMX, and then thio-reactive species were added. In oocytes expressing sCFTR, FOR + IBMX mean stimulated Cl conductance was inhibited 69% by 1 μM mercuric chloride and 78% by 5 μM mercuric chloride (IC50 of 0.8 μM). Despite comparable stimulation of conductance, hCFTR was insensitive to 1 μM HgCl2 and maximum inhibition was 15% at the highest concentration used (5 μM). Subsequent exposure to glutathione (GSH) did not reverse the inhibition of sCFTR by mercury, but dithiothreitol (DTT) completely reversed this inhibition. Zinc (50–200 μM) also reversibly inhibited sCFTR (40–75%) but did not significantly inhibit hCFTR. Similar inhibition of sCFTR but not hCFTR was observed with an organic mercurial, p-chloromercuriphenylsulfonic acid (pCMBS). The first membrane spanning domain (MSD1) of sCFTR contains two unique cysteines, C102 and C303. A chimeric construct replacing MSD1 of hCFTR with the corresponding sequence of sCFTR was highly sensitive to mercury. Site-specific mutations introducing the first but not the second shark unique cysteine in hCFTR MSD1 resulted in full sensitivity to mercury. These experiments demonstrate a profound difference in the sensitivity of shark vs. human CFTR to inhibition by three thiol-reactive substances, an effect that involves C102 in the shark orthologue.

  • chloride transport
  • Xenopus laevis oocytes
  • dithiothreitol
  • glutathione
  • p-chloromercuriphenylsulfonic acid
  • cystic fibrosis transmembrane regulator

salt secretion by epithelial cells has been the focus of intensive investigation (7). Given its anatomic simplicity and functional specialization, the rectal gland of the spiny dogfish shark, Squalus acanthias, is a powerful model to examine transepithelial chloride secretion. Study of this elasmobranch tissue has led to the characterization of numerous proteins involved in epithelial chloride transport (3, 2022, 27, 42, 51).

The heavy metal mercury is a potent inhibitor of secretion in the perfused rectal gland (58). With the use of in vitro assays and heterologous expression, several proteins expressed on the basolateral membrane, including the Na+-K+-2Cl cotransporter, Na+-K+-ATPase, and membrane adenylate cyclase, all of which mediate specific steps of the secretory pathway, were proposed as targets of mercury in the rectal gland (5, 32, 34, 59, 62).

Mercurials have broad cellular toxicity, especially in neurons and renal tubular cells (11, 25, 26, 43, 45, 47, 54, 64). Their capability to impair epithelial ion transport was exploited in mercurial diuretics acting on the loop of Henle of the human kidney (9, 10, 18, 25, 30). At a molecular level, the effects of mercuric substances are thought to be mediated by their high affinity for the thiolate anion (23, 43). Eventually, these reactions induce conformational changes or disrupt protein structure, leading to malfunction (32, 41, 43, 62).

Recent work from our laboratory demonstrated that the inhibition of Cl secretion in the shark rectal gland by HgCl2 occurs in a highly polarized fashion. When applied to the apical membrane of rectal gland cell monolayers in primary culture, HgCl2 (1–10 μM) effected a profound inhibition of Cl transport. In contrast, when HgCl2 was applied on the basolateral side, inhibition of Cl transport was five fold lower (50). The apically expressed protein that likely accounts for this differential effect is the shark orthologue of CFTR (sCFTR), the predominant conductive pathway for chloride exit in the shark rectal gland cell (17, 22, 40).

We sought to examine two hypotheses: 1) that sCFTR is a primary target of mercury toxicity in the rectal gland, and 2) that there are significant differences in the response to thiol-reactive agents between shark and human CFTR, likely mediated by differences in cysteine residues. We used the Xenopus laevis oocyte heterologous expression system to assay the effects of thiol reactants, HgCl2, ZnAc, and pCMBS on sCFTR and hCFTR. Our data provide strong evidence that sCFTR chloride channels are inhibited by HgCl2, ZnAc, and pCMBS, whereas the hCFTR orthologue is relatively insensitive to these three agents. Our data also suggest that the sensitivity of sCFTR to HgCl2 is mediated by the amino acid residue C102 in the first membrane-spanning segment.


Preparation of sCFTR and hCFTR cRNA.

sCFTR contains several cryptic bacterial promoter sequences that result in expression of toxic proteins when transfected into bacteria, making the wild-type gene impossible to clone without site-directed mutagenesis. To prepare wild-type, nonmutated sCFTR cRNA, we synthesized a sense oligonucleotide primer containing the T7 RNA polymerase promoter, a Kozak initiation sequence, and sCFTR sequence, beginning at the start codon. An antisense oligo was designed from sequence immediately flanking the stop codon of sCFTR. These primers were used in a PCR reaction on shark rectal gland cDNA, using a mixture of Taq and Pwo thermostable polymerases (Stratagene, La Jolla, CA). The reaction mixture was heated to 95°C for 1 min before adding the enzyme, and 28 cycles were performed at 95°C for 30 s, 55°C for 30 s, and 68°C for 4 min. cDNA was prepared from three shark rectal glands. Each batch of cDNA was then subjected to 3–5 separate thermocycling reactions to produce a total of 13 different PCR amplifications of shark CFTR. The resulting PCR products were pooled and verified by bidirectional sequencing to be wild-type sCFTR without mutations. The resulting 4,523-bp product was extracted with phenol/chloroform and served as the template for cRNA synthesis. Capped messenger RNA was synthesized from the PCR product using T7 in vitro transcription (Ambion, Austin, TX). hCFTR plasmid cDNA flanked by Xenopus laevis β-globin UTRs (kindly provided by Dr. Carol Semrad, New York) was purified by affinity columns (Qiagen) and linearized with KpnI. Capped hCFTR cRNA was synthesized using T7 RNA polymerase and in vitro transcription.

Oocyte preparation and expression of sCFTR and hCFTR.

Mature female Xenopus laevis were anesthetized in a 0.15% cold solution of tricaine for 20 min, and several ovarian lobules were removed under sterile conditions through an abdominal incision per a protocol approved by the Yale University Institutional Animal Care and Use Committee. Oocytes were incubated in a 2.5 mg/ml solution of type I collagenase for 2.5 h and subsequently manually defolliculated. Mature stage V and VI oocytes were selected and stored. After 12–24 h, the oocytes were injected with 10–20 ng sCFTR cRNA/50 nl, 4–10 ng of hCFTR, or an equivalent volume of water and then stored for 1–3 days in MBSH at 18°C. The amount of RNA was adjusted to achieve a level of expression that resulted in stimulated conductances ∼100 μS. Oocytes were stored in modified Barth solution holding (MBSH) medium containing (in mM) 88 NaCl, 1 KCl, 2.4 NaHCO3, 0.82 MgSO4, 0.33 Ca(NO3)2·4H2O, and 10 HEPES (5 Na+ salt, 5 acid), buffered to pH 7.4, and 150 mg/l gentamicin sulfate.

Creation of chimeric and mutant CFTR constructs.

The human-shark chimeric protein (h-s MSD1) was prepared using PCR and restriction enzyme digestion. Membrane spanning domain 1 (MSD1) of shark was amplified using a sense primer with an overlapping sequence homologous to the NH2 terminal of hCFTR and an antisense primer with an introduced restriction site for BsiWI (sense, AAAATCCTAAACTCATTAATGCACTTCGCCGATG; antisense, CACGTTCGTACGGCAGATGGAAATTGTCTG). The NH2 terminal end of hCFTR was then amplified in a separate PCR (sense, ATACGACTCACTATAGG; antisense, GCATTAATGAGTTTAGG), and both fragments were combined in a third reaction. This product was joined to the rest of hCFTR using NotI and BsiWI restriction sites, and cloned into the pBlueskript KS(−) vector (Invitrogen). Mutant hCFTRs were constructed using the nested PCR method. For each mutant, both fragments carrying the desired mutation were amplified in the first PCR step using flanking and mutation-carrying primers (mutation-carrying primer: L101C-sense AGTACAGCCTCTCTGCCTGGGAAGAA; L101C-antisense TTCTTCCCAGGCAGAGAGGCTGTACT; V302C-sense GCCTATTGCAGATACTTCAATAGC; V302C-antisense AAGTATCTGCAATAGGCTGCCTTCCG; flanking primer: SacII TGTAAAACGACGGCCAGTGAG; BstZ17 I GTATACTGCTCTTGCTAAAGA). In the second PCR step, the two fragments were combined using only the flanking primers, and subcloned into the vector containing hCFTR from which the native fragment had been removed with SacII and BstZ17I.

Two-electrode voltage clamping.

Electrophysiological recordings were performed 2–4 days after injection. Electrodes were pulled on a micropipette puller (Sutter Instruments, Novato, CA) and had input resistances between 0.6 and 0.9 MΩ. For electrophysiological measurements, electrodes were filled with 3 M KCl. Membrane potential was recorded continuously on a chart recorder (model SR6253, Datamark). During two-electrode voltage-clamping recording, current-voltage (I-V) curves were obtained by clamping the voltage over a ramp from −120 to +60 mV at a rate of 100 mV/s with the use of a two-electrode voltage clamp (TEV-200, Dagan Instruments, Foster City, CA). After being corrected for capacity currents, reversal potentials were determined and conductances were calculated over a range of the corrected reversal potential ± 10 mV.

During a typical experiment, oocytes were perifused with frog Ringer solution containing (in mM) 98 NaCl, 2 KCl, 1.8 CaCl2, 1 MgCl2, 10 HEPES (5 Na+ salt, 5 acid), buffered to pH 7.4 (reagents from Sigma, St. Louis, MO). I-V ramps were taken under basal conditions, during steady-state stimulation by forskolin (10 μM) and IBMX (1 mM) and after addition of HgCl2, pCMBS, or ZnAc to the forskolin-IBMX solution. In all experiments inhibition was determined by the ratio of conductances measured immediately before and 30 min after addition of the sulfhydryl reactants. Data were analyzed with pCLAMP software (Axon Instruments). All data are means ± SE.


Conductances and reversal potentials of Xenopus oocytes expressing sCFTR-, hCFTR-, or water-injected controls.

Conductance and reversal potential of oocytes were measured under basal and stimulated conditions in 140 oocytes injected with either sCFTR (n = 84), hCFTR (n = 43), or water (n = 13). The mean time intervals between cRNA injection and two-electrode voltage-clamping experiments were similar in all groups (2.8 ± 0.1, 2.5 ± 0.2, and 3.0 ± 0.4 days for sCFTR-, hCFTR-, and water-injected oocytes, respectively).

sCFTR- and hCFTR-injected oocytes had basal conductances of 16.1 ± 1.4 and 7.5 ± 1.5 μS, respectively. Water-injected control oocytes had a significantly lower baseline conductance of 3.4 ± 1.5 μS (P < 0.02 compared with both sCFTR and hCFTR) (Fig. 1, top). Baseline reversal potentials were −35.0 ± 1.1 mV, −38.0 ± 2.6 mV, and −40.2 ± 6.0 mV for sCFTR-, hCFTR-, and water-injected oocytes, respectively (Fig. 1, bottom).

Fig. 1.

Conductances and reversal potentials of Xenopus oocytes expressing either the shark orthologue of cystic fibrosis transmembrane conductance regulator (sCFTR; n = 84) or human CFTR (hCFTR; n = 43) compared with water-injected controls (n = 13). Oocytes were perifused with frog Ringer for 10–15 min until the conductance reached a steady state (baseline). Oocytes were then stimulated with foskolin (FOR; 10 μM) and IBMX (1 mM; stimulated). Current-voltage (I-V) ramps were taken before and after 50 min of stimulation to determine conductance and reversal potential under baseline and stimulated conditions, respectively. After stimulation with FOR and IBMX, oocytes expressing sCFTR- and hCFTR had nearly identical conductances (112 ± 3.8 and 117 ± 6.7 μS) and reversal potentials (−28.4 ± 0.8 and −25.1 ± 1.0 mV). In both sCFTR- and hCFTR-expressing oocytes, changes in conductance (P < 0.001) and reversal potential (P < 0.01) after stimulation were highly significant compared with baseline values. In contrast to CFTR- expressing oocytes, conductance and reversal potential of water-injected oocytes did not change significantly after stimulation with FOR and IBMX. Water-injected oocytes had no response to mercury. All values are mean ± SE. *P < 0.01; **P < 0.001.

Because we determined the effects of mercuric chloride after activation of CFTR, it was important to establish comparable steady-state conductances for hCFTR- and sCFTR-injected oocytes upon activation. After addition of 10 μM forskolin (FOR) and 1 mM IBMX to the perifusate, sCFTR- and hCFTR-injected oocytes had nearly identical steady-state conductances (112 ± 3.8 and 117 ± 6.7 μS, respectively; Fig. 1, top). Simultaneously, both sets of oocytes depolarized (−28.4 ± 0.8 mV and −25.1 ± 1.0 mV for sCFTR and hCFTR, respectively) toward the chloride equilibrium potential. In contrast, water-injected oocytes remained hyperpolarized and their conductance did not change significantly upon stimulation (Fig. 1, bottom).

Effects of mercuric chloride on sCFTR- and hCFTR-injected oocytes in representative experiments: time-course and I-V plots.

The time-course and I-V relationships in representative experiments with our protocol are shown in Fig. 2. In both sCFTR- and hCFTR-injected oocytes, stimulation with FOR + IBMX resulted in a prompt increase in conductance, which reached a steady state after 30–50 min (Fig. 2, A and B). Addition of 1μM HgCl2 to the perifusate rapidly decreased conductance in the sCFTR-injected oocyte (75.8% inhibition after 30 min of HgCl2) (Fig. 2A). In contrast, in an hCFTR- injected oocyte, the effect of HgCl2 on conductance was substantially less: addition of 1 μM HgCl2 decreased conductance by only 26.8% in this representative experiment (Fig. 2B). I-V relationships for these oocytes are shown in Fig. 2, C and D. Throughout the experiments, both sCFTR- and hCFTR- injected oocytes exhibited nearly linear I-V relationships. After the addition of HgCl2, the reversal potential repolarized toward basal values in the sCFTR-injected oocyte (−43.0, −35.2, and −41.2 mV under basal, stimulated and HgCl2 conditions, respectively) (Fig. 2C). In the hCFTR-injected oocyte, mercury had an insignificant effect on the reversal potential (Fig. 2D).

Fig. 2.

Inhibition of sCFTR and hCFTR by HgCl2. Time course of the effects of 1 μM HgCl2 on FOR + IBMX stimulated conductances in sCFTR (A) and hCFTR (B). Representative I-V plots from oocytes expressing sCFTR (C) and hCFTR (D). Oocytes were perifused with frog Ringer solution for 10–15 min until the conductance reached a baseline steady state and were then stimulated with FOR (10 μM) and IBMX (1 mM) for 50–65 min. HgCl2 (1 μM) was then added to the perifusate for 30 min. I-V ramps were taken after 15, 65, and 105 min, at the end of each condition.

Dose response to HgCl2 in sCFTR- vs. hCFTR-injected oocytes.

A dose response to HgCl2 (0.1 to 5 μM) was determined for both sets of oocytes. In sCFTR-injected oocytes, the addition of HgCl2 resulted in a sigmoidal, dose-dependent inhibition of FOR+IBMX stimulated conductance when plotted on a logarithmic scale (see Fig. 3). Mercuric chloride (1 μM) inhibited the stimulated conductance in sCFTR-expressing oocytes by a mean of 68.9 ± 5.2%, with an IC50 of 0.8 μM. Mean maximal inhibition of the stimulated conductance was 77.7 ± 4.3% at 5 μM HgCl2. In contrast, in hCFTR-expressing oocytes, mean maximal inhibition (at 5 μM HgCl2) was only 15.0 ± 2.1% of the stimulated conductance. At concentrations >10 μM HgCl2, irreversible depolarization of oocytes was consistently seen with an accompanying large increase in conductance (data not shown).

Fig. 3.

Logarithmic dose response demonstrating relative sensitivity of sCFTR vs. hCFTR to inhibition by HgCl2. Oocytes expressing either sCFTR or hCFTR were perifused with frog Ringer solution for 10–15 min and then stimulated with FOR (10 μM) and IBMX (1 mM) for 50–70 min. HgCl2 (0.1–5 μM) was then added to the perifusate. Conductances were measured after 30-min exposure to each concentration of HgCl2. Mean maximum inhibition by HgCl2 was 77.7 ± 4.3% for sCFTR and 14.0 ± 2.1% for hCFTR. The calculated IC50 for sCFTR was 0.8 μM HgCl2 (n = 3–10 observations per concentration of HgCl2 in 38 sCFTR and 24 hCFTR injected oocytes). All values are means ± SE.

Reversal of the inhibitory effect of mercuric chloride by dithiothreitol but not by glutathione.

The effects of HgCl2 were not reversed after being washed with Ringer solution for up to 60 min (data not shown). We next examined the effects of dithiothreitol (DTT), a permeant small molecule, and the relatively impermeant, larger tripeptide, glutathione (GSH) (19, 39). Differential effects of these agents have been used to localize the site of action of Hg in several proteins (2, 6, 35). Although both are commonly used to reduce disulfide bonds, both are also well known to be high affinity metal ligands (8, 36, 48). In these experiments, DTT (5 mM) reversed the effects of HgCl2 in both sCFTR (80 ± 11% reversal, n = 8)- and hCFTR-injected oocytes (101 ± 29% reversal, n = 4). A representative tracing depicting the effects of DTT is shown in Fig. 4. After inhibition of conductance by 2 μM HgCl2, DTT (5 mM) promptly reversed the effects of HgCl2. This figure also shows the sensitivity of sCFTR to the sulfonylurea glibenclamide (600 μM), an inhibitor of CFTR chloride channels.

Fig. 4.

Reversal of HgCl2 inhibition of sCFTR conductance by dithiothreitol (DTT). After stimulation by 10 μM FOR and 1 mM IBMX, conductance was inhibited to near baseline values by 2 μM HgCl2; 5 mM DTT was then added in the presence of HgCl2 and entirely reversed the inhibition by HgCl2. The stimulated conductance was also sensitive to glibenclamide (Glib; 600 μM). This figure is representative of 8 experiments where the reversal by DTT was 80 ± 11%.

A separate series of experiments was carried out to compare directly the effects of GSH and DTT in the same sCFTR- injected oocytes (Fig. 5). The time course is depicted in a representative experiment in Fig. 5A. Whereas GSH (10 mM) showed minimal reversal of inhibition by mercury, DTT (5 mM) promptly reversed the effect of the metal (Fig. 5B). The mean percent reversal in these experiments is shown in Fig. 5C [GSH 8 ± 5%, P = not significant (NS) compared with HgCl2, and DTT 91 ± 8%, P < 0.001 compared with HgCl2]. Higher concentrations of GSH (up to 100 mM, n = 4) showed no further reversal of inhibition. In control experiments, DTT (5 mM) had no effect on either baseline or FOR+IBMX-stimulated conductance in sCFTR-expressing oocytes (n = 4).

Fig. 5.

Comparison of the effects of glutathione (GSH) and DTT to reverse the effects of HgCl2 in sCFTR-injected oocytes. A: representative experiment indicating the time course of the protocol. Whereas GSH showed minimal reversal of inhibition by mercury, DTT promptly reversed these effects. B: percent conductance after stimulation by FOR + IBMX, inhibition by HgCl2 and reversal of inhibited conductance by GSH and DTT in 5 experiments. Percent conductance was calculated as (value after condition/stimulated conductance)·100. FOR/IBMX stimulated conductance values ranged from 140 to 340 μS. The value for each subsequent condition was then divided by this FOR/IBMX value to give a conductance value. **P < 0.001. C: mean percent reversal by GSH (8 ± 5%) and DTT (91 ± 8%) in 5 experiments. Percent reversal was calculated as (value after reversal − value of inhibition)/(value of stimulation − value of inhibition)·100.

Inhibition of shark CFTR by zinc acetate.

We also compared the sensitivity of sCFTR and hCFTR to zinc acetate (ZnAc), a nonmercuric compound that binds to cysteine residues (12, 24, 28, 29, 46, 53) but with a lower affinity than Hg. ZnAc inhibited chloride conductance in sCFTR but not in hCFTR injected oocytes. Figure 6A illustrates a representative experiment with ZnAc (50 μM) in an oocyte expressing sCFTR. Inhibition by ZnAc was readily reversible on washout. After the same protocol used for testing mercurial compounds, ZnAc (50 μM) inhibited the stimulated conductance of the sCFTR- expressing oocytes by 41% in four experiments. Over a range from 1 μM to 200 μM, ZnAc inhibited sCFTR induced Cl conductances in a dose-dependent manner (Fig. 6B). From a sigmoidal plot of the ZnAc dose response curve in sCFTR- expressing oocytes, we calculated an IC50 of 63 μM and a mean maximal inhibition of 72 ± 4.7% at the highest concentration employed (200 μM). In contrast, in hCFTR-injected oocytes, ZnAc inhibition of the Cl conductance was substantially less (6.5 ± 1.8% at 200 μM ZnAc) (Fig. 6B).

Fig. 6.

Inhibition of sCFTR by the thiol-reactive agent zinc acetate (ZnAc). A: representative experiment with an oocyte expressing sCFTR. After stimulation with FOR (10 μM) and IBMX (1 mM), ZnAc (50 μM) inhibited sCFTR in the presence of FOR and IBMX. Unlike the inhibition by HgCl2, the effect of ZnAc was reversible on washout. B: dose response and relative sensitivity of sCFTR vs. hCFTR to inhibition by ZnAc. Oocytes expressing either sCFTR or hCFTR were perifused with frog Ringer solution for 10–15 min and then stimulated with FOR (10 μM) and IBMX (1 mM). ZnAc (1–200 μM) was then added to the perifusate for 30 min. Conductances were measured after 30-min exposure to each concentration of ZnAc. Maximum inhibition by ZnAc was 72 ± 4.7% for sCFTR and 6.5 ± 1.8% for hCFTR. The calculated IC50 for sCFTR was 63 μM ZnAc (n = 3–4 observations per concentration of ZnAc). Water-injected oocytes had no response to zinc.

Inhibition of sCFTR by the organic mercurial para-chloromercuribenzenesulfonate.

Finally, we investigated the effect of the organic mercurial para-chloromercuribenzenesulfonate (pCMBS; see Fig. 7). This compound is expected to be less permeant than the inorganic mercury complexes present in aqueous solution due to the presence of the sulfonic acid group. Similar to HgCl2 and ZnAc, the efficacy of pCMBS (100 μM) was much greater in oocytes expressing sCFTR. Conductance was inhibited by 54 ± 5% in sCFTR- vs. 4 ± 2% in hCFTR-injected oocytes (P < 0.001; Fig. 7A). DTT (5 mM) reversed the inhibition of pCMBS by 84.5 ± 6% in sCFTR-injected oocytes. A dose response (0.1 μM to 200 μM pCMBS) revealed an IC50 of 53 μM and a maximal inhibition of 90.3% in oocytes expressing sCFTR (Fig. 7B). In contrast, in hCFTR- injected oocytes, maximal inhibition by pCMBS (at 200 μM) was only 18%. pCMBS had no effect on conductance in water-injected oocytes.

Fig. 7.

Inhibition of CFTR by the membrane impermeant mercury compound pCMBS. A: oocytes expressing either sCFTR or hCFTR were perfused with frog Ringer solution for 10–15 min and then stimulated with FOR (10 μM) and IBMX (1 mM). p-Chloromercuriphenylsulfonic acid (pCMBS; 100 μM) was then added to the perifusate for 30 min. Conductances were measured after 30 min of exposure. Similar to HgCl2, pCMBS had a greater inhibitory effect on sCFTR than on hCFTR (54 ± 5%, n = 6, vs. 4 ± 2%, n = 4, **P < 0.001). B: dose response of sCFTR to inhibition by pCMBS. pCMBS (0.1–200 μM) was added and conductances were measured after 30-min exposure to each concentration. Maximum inhibition by pCMBS was 90.3%, and the calculated IC50 was 53 μM.

Effects of mercury on chimeric h-s CFTR and hCFTR site-specific mutants.

Numerous studies (1, 4, 15, 5557, 60) have suggested that components of MSD1 form the channel pore and determine the anion selectivity of CFTR. Given the presence of two unique cysteines in MSD1 of sCFTR, we first tested the hypothesis that a chimeric channel, in which human MSD1 was replaced by shark MSD1, would be highly sensitive to HgCl2. This chimera demonstrated striking sensitivity to mercury (89 ± 3% inhibition at 1 μM HgCl2, n = 9, P < 0.001 compared with hCFTR) (Fig. 8).

Fig. 8.

Inhibition of forskolin + IBMX stimulated conductances by 1 μM HgCl2 in sCFTR, hCFTR, human-shark (h-s) MSD1, hL101, and V302C. sCFTR, h-s MSD1, and hL101C show similar sensitivity to mercury (68.9 ± 5.2%, 89 ± 3%, and 78 ± 8% inhibition at 1 μM HgCl2, each P < 0.001 compared with hCFTR). The inhibition of hV302C (20 ± 5%) was not significantly different from hCFTR. The basal conductance and the conductance after stimulation were 10.2 ± 2 μS and 135.6 ± 5.5 μS for sCFTR (n = 10); 6.3 ± 1.1 μS and 150.5 ± 12.6 μS for hCFTR (n = 10); 7.4 ± 1.0 μS and 155.4 ± 27.3 μS for h-s MSD1 (n = 9) 4.2 ± 1.7 μS and 207.8 ± 42.1 μS for hL101C (n = 7); and 13.8 ± 5.9 μS and 123.7 ± 43.9 μS for hV302C (n = 4).

Because the h-sMSD1 chimera fully reproduced the mercury sensitivity of sCFTR, we next mutated single residues in hCFTR (L101C and V302C) to the corresponding cysteines (C102 and C303) present in MSD1 of sCFTR. Mutant hL101C was markedly inhibited by 1 μM HgCl2 (79 ± 8%, n = 7; P < 0.001 compared with hCFTR), whereas mutant hV302C remained insensitive to 1 μM HgCl2 (20 ± 5% inhibition, n = 4, P = NS compared with hCFTR) (Fig. 8).


Previous studies in the shark rectal gland have examined the mercury sensitivity of membrane Na+-K+-ATPase (49), adenylate cyclase (59), and the Na+-K+-2Cl cotransporter (32, 49). Work from our laboratory, however, has indicated that a protein located in the apical membrane of shark rectal gland cells is a primary site of mercury toxicity (50). When the membrane polarity of mercuric chloride toxicity was determined by short-circuit current measurements in cultured monolayers of rectal gland cells, we demonstrated a markedly higher inhibition of short-circuit current when Hg2+ was added on the apical compared with the basolateral side (50). Because sCFTR is apically expressed (40, 42), whereas Na+-K+-ATPase and the Na+-K+-2Cl cotransporter are localized on the basolateral membrane (5, 32), the membrane polarity of the HgCl2 effect suggests a model in which sCFTR is a major target of mercury toxicity.

In the present study, we provide direct support for the hypothesis that CFTR chloride channels are a primary site of mercury toxicity in the shark rectal gland and provide the first evidence that shark and human orthologues of CFTR Cl channels have markedly different sensitivity to inhibition by mercuric chloride, pCMBS, and other thiol reactive substances.

Speculation on the site of action of mercury must take into account the mercury species that are most abundant in a physiological salt solution (14). Mercury, like many metals, exists in solution in anionic complexes and the nature of the anionic ligand is a primary determinant of reactivity. In the presence of 126 mM chloride, divalent mercury is present in small amounts and is unlikely to be the species involved in reactions with protein thiols (63). Under these conditions the majority of mercury exists in three forms, in roughly equal proportions, HgCl2, HgClMath and HgClMath. Reaction with thiols occurs via a ligand substitution whereby one or more of the chloride ligands is exchanged for the thiolate anion. In view of the negative charge of the reactive thiolate, electrostatic considerations render the neutral species, HgCl2, most likely to participate in a ligand substitution reaction, i.e., protein-S + HgCl2 ↔ protein-S-HgCl.

The inhibition of sCFTR by mercury was irreversible upon washout. One explanation is that mercury caused irreversible conformational change or disruption of the protein that prevented restoration of conductance. However, our findings show that mercury inhibition could be immediately and completely restored by applying DTT to the bath solution. The failure of GSH, a larger molecule, to reverse the reaction of mercury with sCFTR, even at concentrations 20-fold higher than DTT, likely reflects the relative impermeability of the tripeptide (19). Our data suggest a model in which DTT competes for the thiolate-liganded mercury. This mechanism is supported by the Zn2+ inhibition, which mimics the inhibition by HgCl2 and pCMBS, but is readily reversed on washout. Wang et al. (61) recently reported that hCFTR with substituted cysteines was inhibited by zinc and that this effect depended on interactions with the engineered cysteines. Therefore, we suggest ligand substitution of mercurials and Zn2+ on specific thiolate groups in sCFTR, with the structural integrity of the protein remaining intact, to be the underlying mechanism of inhibition.

Nondestructive binding with reversibility has been proposed as the mechanism for mercury toxicity in other epithelial proteins, including the Na+-K+-2Cl-cotransporter, the Na+-K+-ATPase and aquaporins, in which mercury binds to specific cysteine residues (13, 32, 37, 38, 41, 44, 62). In the shark Na+-K+-2Cl-cotransporter, for example, Hg2+ interacts with cysteine residue C697 in the transmembrane domain 11, causing inhibition of the cotransporter activity that is restored upon addition of DTT (32).

Once we established a high sensitivity of sCFTR to thiol reactive agents, we investigated the effect of these compounds on hCFTR. We found a pronounced difference in the sensitivity to mercurials between sCFTR and hCFTR: the human protein was inhibited to a much lesser extent than the shark homologue (15.0 ± 2.1% vs. 77.7 ± 4.3% inhibition at 5 μM HgCl2 of hCFTR vs. sCFTR). Furthermore, not only HgCl2 but also the thiol reactive agents pCMBS and Zn2+ had similar patterns of inhibition: both compounds inhibited hCFTR either weakly (pCMBS), or insignificantly (ZnAc). In other membrane protein systems, pCMBS sensitivity has correlated with the presence of exofacial cysteine residues (31, 65). Cysteine residues in the R-domain, sensitive to 10- to 100-fold higher concentrations of mercury, were recently suggested as sites influencing channel gating in human CFTR Cl channels (33). In agreement with our findings was the lack of significant inhibition of hCFTR by pCMBS (33). The striking differences in sensitivity to HgCl2, zinc, and pCMBS in oocytes expressing sCFTR and hCFTR, and the presence of unique cysteines not found in the human protein, reinforce the idea that cysteine thiolates are the site(s) of action of mercury as well as pCMBS and zinc.

Comparison of the amino acid sequences of the two proteins reveals that sCFTR shares 72% of its amino acid sequence with the human protein (42, 52). Of the 18 cysteine residues in sCFTR, six are unique to the shark orthologue. Two of these cysteines (C102 and C303) are located in the MSD1, including one (C102) present on an exofacial site. Both cysteines are not present in other species, where CFTR sequence is known, including zebrafish, pufferfish, and salmon (GenBank). Because MSD1 has been proposed to be involved in the pore region of CFTR (1, 16, 57), we focused our chimeric and mutational experiments on this segment. Replacing MSD1 of human with the corresponding segment of sCFTR created a chimeric protein that, like sCFTR, was highly sensitive to HgCl2. The single mutation to cysteine (hL101C) (corresponding to wild-type C102 in shark) resulted in complete sensitivity to mercury, whereas the single mutation to cysteine (hV302C) (corresponding to wild-type C303 in shark) showed minimal sensitivity to HgCl2, and was not different from hCFTR (Fig. 8). This finding suggests that L101 contributes either directly or indirectly to the formation of the channel pore.

Taken together with our previous work in cultured monolayers, the present experiments establish that CFTR is a major target of mercury in epithelial cells of the shark rectal gland. Furthermore, we describe marked differences to three thiol reactive agents between human and shark CFTR chloride channels and identify the cysteine residue likely responsible for these effects. These findings establish a novel property of the sCFTR protein that advances our understanding of the structure and function of CFTR.


This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-34208 (to J. N. Forrest, Jr.) and DK-45880 (to D. C. Dawson), the Center for Membrane Toxicology Studies Grant P30-ES-3828, and The Cystic Fibrosis Foundation Grant 0210 (to D. C. Dawson), and American Heart Association Grant 9607741S (to J. N. Forrest, Jr).


We thank Marie Bewley, Peter Burrage, and Ryan Martin for excellent technical assistance.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract