A One-Domain Voltage-Gated Sodium Channel in Bacteria

See allHide authors and affiliations

Science  14 Dec 2001:
Vol. 294, Issue 5550, pp. 2306-2308
DOI: 10.1126/science.1067417

Learning, memory, movement, sensation, and other complex processes are all coordinated by electrical signals (action potentials) conducted along the long fibers of nerve cells. In both multicellular animals and complex single-cell organisms such as the eukaryote Paramecium, electrical signals are generated locally and action potentials are conducted globally through the activity of a family of voltage-gated sodium channels. This ion channel family was not thought to exist in bacteria, the jellyfish being the simplest organism in which it had been found (1). On page 2372 of this issue, Ren et al. (2) surprise us with their discovery of the first voltage-gated ion channel in a prokaryote. They describe the structure and function of a sodium channel in the salt-loving bacterium Bacillus halodurans.

The voltage-gated sodium (Nav) channels of eukaryotes are complex proteins composed of more than 2000 amino acid residues. The large pore-forming α subunit comprises four homologous domains containing six transmembrane α helices (see the figure) (3). It is bell-shaped (4) and associates with smaller β subunits (3). As Ren et al. report (2), the bacterial sodium channel (NaChBac) is much simpler, consisting of a single domain with 274 amino acid residues and six transmembrane α helices (see the figure). This structure resembles voltage-gated potassium (Kv) channels, which are homotetramers (composed of four identical subunits) (5). The primary structure of NaChBac raises provocative questions about the three most important tasks of sodium channels—selective ion conductance, voltage-dependent activation, and fast inactivation.

Comparing Nav and NaChBac channels.

(A) Arrangement of the transmembrane α subunit of the mammalian Nav1.2 voltage-gated sodium channel. Note the four-domain structure with six α-helical transmembrane segments in each domain. The S5 and S6 segments and the pore loop region between them are highlighted in green. The positively charged S4 gating segments are highlighted in red. The inactivation gate and the isoleucine-phenylalanine-methionine (IFM) motif that is crucial for inactivation are highlighted in yellow. (B) The three-dimensional structure of the Nav channel α subunit at 20 Å resolution, compiled from electron micrograph reconstructions (4). Note the bell shape with most of the mass on the intracellular side of the membrane and a cap over the extracellular opening of the pore. (C) NaChBac, the bacterial voltage-gated sodium channel. Note the similarity to one domain of the Nav channel α subunit.

In vertebrates, an ion selectivity filter enables voltage-gated sodium channels to conduct sodium ions 15 to 50 times as rapidly as potassium and calcium ions, the other prevalent cations in physiological fluids (6). The transmembrane pore of these ion channels is formed by the S5 and S6 segments of the α subunit and the pore loop segment between them [reviewed in (3)]. Ion selectivity is thought to be determined by a small number of amino acid residues in the second half of the pore loop, which forms a narrow opening into the extracellular end of the pore (7). As discussed by Ren et al. (2), preferential sodium selectivity in Nav channels is thought to require an asymmetric pore structure with different amino acid residues in key positions in the pore loops of each of the four domains. At one critical position, the amino acid residues in the pore loops of the four domains of Nav channels are aspartate, glutamate, lysine, and alanine (DEKA). Evidence from mutagenesis experiments indicates that these four residues must be different from each other so that they can form an asymmetric ion selectivity filter, which selects sodium ions over calcium ions (7, 8). In light of these previous results, it is surprising that the NaChBac channel, expressed in bacteria as a homotetramer with glutamate residues in the key position in all four pore loops (EEEE), has such high sodium selectivity. This finding implies either that the homotetramer can form an asymmetric pore structure (by arranging identical subunits in an asymmetric assembly) or that asymmetry of the pore structure is not really required for sodium selectivity. Experiments to test this question should help to reveal the mechanism of sodium channel ion selectivity.

The pharmacology of NaChBac's pore region is also unexpected, as it more closely resembles that of voltage-gated calcium (Cav) channels. Although NaChBac is not blocked by the traditional sodium channel blocker tetrodotoxin, it is blocked by divalent cations and two different types of calcium channel blockers (2). Cav channels have four glutamate residues (the EEEE motif) in the key position in the pore loop. These channels conduct sodium ions well in the absence of divalent cations, but sodium conductance is prevented by the presence of calcium ions or other divalent cations (9, 10). Perhaps the EEEE motif of NaChBac, like the similar motif in Cav channels, is permissive for sodium conductance, whereas other unidentified molecular features of the NaChBac pore prevent divalent cation binding and conductance and thereby confer sodium selectivity.

The second hallmark property of the voltage-gated ion channel family is steep voltage-dependent opening in response to changes in the electrical potential across cell membranes. Current molecular models of the activation gating process suggest that the S4 transmembrane α helices in each domain serve as voltage sensors. They contain four to seven positive charges from cationic amino acids (called gating charges), which are positioned at intervals of three residues. Depolarization of the membrane exerts an electrostatic force, pushing the positive charges in the S4 segments in the outward direction across the transmembrane electric field. The S4 segments are thought to respond to depolarization by a combination of outward movement and rotation, exposing their outermost gating charges to the extracellular medium and eventually resulting in opening of the pore (11-15). Surprisingly, the extracellular loop connecting the S3 and S4 transmembrane domains of NaChBac is only two or three amino acid residues in length, preventing large outward movements of the S4 segment independently of the S3 segment. A short S3-S4 loop created by mutagenesis of a potassium channel also allowed activation gating (16). This short-loop structure implies a different gating mechanism from the most widely accepted models. Either the S3 and S4 segments move outward and rotate together, or NaChBac achieves transmembrane movement of its S4 gating charges primarily by combining rotation and rearrangement of the surrounding protein to expose the positive charges to the extracellular medium with only small outward movement (14).

Fast inactivation during prolonged depolarization is the third key property of Nav channels, allowing them to close and return to the resting state within 1 ms, as required for generation of action potentials at high frequency. Impairment of this process by mutation causes inherited diseases of nerve hyperexcitability, such as periodic paralysis, cardiac arrhythmia and epilepsy (17). Inactivation may involve closure of an intracellular inactivation gate formed by the peptide loop connecting homologous domains III and IV, which occludes the intracellular end of the pore (3, 18-21). NaChBac is inactivated like a Nav channel, but more slowly by a factor of 100. Because it is a homotetramer, it has no equivalent of the intracellular loop connecting domains III and IV of the Nav channels.

How is NaChBac inactivated? The single-domain Kv channels may provide the answer. These channels are inactivated in two ways at a slow rate comparable to that of NaChBac inactivation. N-type inactivation involves the amino-terminal segment of the Kv channel, which folds like a ball and chain into the channel structure and blocks the pore from the inside (like Nav channel inactivation) (22, 23). In contrast, C-type inactivation of the Kv channel results from closure of the pore by a mechanism involving the pore loops in all four domains, closure resembling the annular movements of a camera shutter (24, 25). The related slow inactivation of Nav channels may work similarly (26, 27). The amino-terminal domain of NaChBac is far shorter than that of Kv channels, and it seems too short to serve as a tethered ball and chain for inactivation. Therefore, it seems most likely that NaChBac uses C-type inactivation to yield concerted closure of the pore by the pore loops in all four domains. The new NaChBac channel now provides an opportunity to explore this poorly understood form of inactivation.

Perhaps the most exciting aspect of NaChBac's discovery is the possibility that it can be expressed in large amounts, crystallized, and analyzed by x-ray crystallography to determine its three-dimensional structure. To date, only the pore region of a related bacterial potassium channel (KcsA) has been analyzed at the structural level (28). It has yielded a wealth of information about how the pore works and maintains potassium ion selectivity. The KcsA channel, however, does not have voltage-dependent gating, thus no information about this crucial mechanism of excitation can be gleaned from its structure. With NaChBac—a primordial member of the ion channel gene family—in hand, we can look forward to resolving questions about ion selectivity, voltage-dependent activation, and inactivation at the structural level.


View Abstract

Navigate This Article