Ion channels are a special class of proteins that conduct small ions such as Na⁺, K⁺, Ca²⁺ or Cl^-. The opening and closing of ion channels may depend on membrane voltage or the binding of molecules (the ligand). Its mechanism remains unclear. However, the mystery of their selectivity has been solved.

Ion Selectivity

Potassium Channels

By definition, a potassium channel conducts mainly K⁺ ions while excluding Na⁺ and Ca²⁺  ions. We note that a Na⁺ ion is smaller than a K⁺ ion. How could a channel conduct K⁺ ions while excluding smaller ions? The answer lies in the ion's hydration energy.

The diameter of the potassium channel's selectivity filter (the narrowest region of its channel pore) should be about the same as the size of a Cs⁺ ion (diameter = 3.3Å) because Cs⁺ is the largest ion which can barely pass through the potassium channel. In a solution, ions are surrounded by water molecules. The first hydration shell of a K⁺ or Na⁺ ion contains six water molecules. The diameters of unhydrated Na⁺ and K⁺ ions are 1.96Å and 2.66Å, respectively. The effective diameter of a water molecule in a hydrated ion is larger than 2Å. Therefore, to pass through the selectivity filter, the K⁺ or Na⁺ ion must remove four water molecules from its first hydration shell, leaving only two (one at the front and another at the back). The smaller Na⁺ ion requires greater dehydration energy than the K⁺ ion, because its nucleus has shorter distance with surrounding water molecules and thus interacting more strongly. As a result, the Na⁺ ion is harder to pass through the potassium channel than the K⁺ ion. The observation that lithium (Li⁺) is excluded from the potassium channel can also be explained by the same energetic consideration. Divalent cations such as Ca²⁺ should also be excluded because their dehydration energies are much higher than monovalent cations.

Figure 3.1. Ion passage through potassium and sodium channels. (A) To pass through the potassium channel, an ion must remove most of its surrounding water molecules, leaving only two - one at the front and another at the back. (B) The selectivity filter of the sodium channel is slightly larger than that of the potassium channel. It may accommodate a Na⁺ ion attached with three water molecules, but not enough for a K⁺ ion attached with three water molecules.

Sodium Channels

In the sodium channel, the Na⁺ ion is more permeable than the K⁺ ion. This is because the selectivity filter of the sodium channel is slightly larger than that of the potassium channel. It is large enough to accommodate a Na⁺ ion attached with three water molecules, but not enough for a K⁺ ion attached with three water molecules. Therefore, to pass through the sodium channel, the Na⁺ ion needs to remove only three, but the K⁺ ion has to remove four, water molecules from its first hydration shell. The required dehydration energy for the K⁺ ion is greater than the Na⁺ ion.

Calcium Channels

In calcium channels, the permeability of monovalent cations (Na⁺ and  K⁺) is about three orders of magnitude smaller than the Ca²⁺ permeability. This ion selectivity does not seem to involve hydration, because Ca²⁺ is more heavily hydrated than Na⁺, and the unhydrated diameters of Ca²⁺ and Na⁺ are almost identical. Then, how could calcium channels select Ca²⁺ over Na⁺?

Although the permeability of monovalent cations in the calcium channel is quite small at normal ionic concentrations, large monovalent cationic current can be observed in the absence of Ca²⁺ and other divalent cations. This suggests that the calcium channel is basically permeable to both divalent and monovalent cations, but the selectivity arises from competition between ions. The calcium channel may contain a negatively charged binding site to facilitate ion conduction. The monovalent cations simply cannot compete with Ca²⁺ for this binding site. This idea has been confirmed experimentally. In the calcium channel, if a negatively charged glutamate residue in the pore-lining region is mutated into a positively charged lysine, the calcium channel becomes more permeable to Na⁺ than Ba²⁺ (reference). Conversely, in the sodium channel, mutation of a pore-lining lysine residue into glutamate transforms the channel from a Na⁺-selective to a Ca²⁺-selective channel (reference).

Channel Structures

K⁺ Channels

There are many types of potassium channels. The one involved in the generation of action potentials is composed of four subunits, each is homologous to the Shaker protein (Fig. 3.2). The hydrophobicity profile indicates that it contains six hydrophobic segments, designated as S1 - S6. These segments are likely to be the transmembrane domains. Other experimental results suggests that the P-region is lining the channel pore. The domain structure of the Shaker protein is shown in Fig. 3.3. 

Figure 3.2. The amino acid sequence of the Shaker protein.

Figure 3.3. The domain structure of the Shaker protein.

Figure 3.4. Schematic drawing of the Shaker potassium channel, which is composed of four Shaker proteins. The channel pore is formed by four P-regions (only three are shown here).

A growing number of channel structures have been determined by x-ray crystallography (web link). These structures were obtained in the absence of membrane environment. Furthermore, an antibody Fab was used to stablize crystal structures. Whether or not the crystal strutures represent native structures remain a question (reference).

Na⁺ and Ca²⁺ Channels

A Na⁺ or Ca²⁺ channel consists of a major pore-forming subunit and possibly other small auxiliary subunits. The major pore-forming subunit is called the α subunit which can be divided into four similar domains. Each domain is analogous to a Shaker protein with six transmembrane segments and a P-region. Thus, an α subunit is sufficient to form an ion channel.

Figure 3.5. The domain structure of the α subunit of a Na⁺ or Ca²⁺ channel.

Synaptic Channels

A large class of ion channels are specifically located at the synapse (the junction between nerve cells). Each synaptic channel consists of five subunits. Figure 3.6 shows the general structure of a nicotinic acetylcholine receptor (nAChR) channel.

Figure 3.6.  Schematic drawing of the nAChR channel which consists of five subunits: α, α, β, γ, and δ.  (a) Side view.  (b) Top view.  (c) The domain structure for each subunit.  Five M2 segments (one from each subunit) form the channel pore as shown in (b).  Activation (opening) of an nAChR channel requires binding of two acetylcholine molecules, one on each α subunit.