Kir2 Subfamily Underlies Cardiac I_K1

The Kir2 subfamily consists of five members (Kir2.1-Kir2.5), of which only Kir2.1-Kir2.4 are expressed in the mammalian heart. Kir2.5 is expressed in the fish. As shown in Table 4-1, the following genes encode the mammalian cardiac Kir channels: KCNJ2, KCNJ12, KCNJ4, and KCNJ14, for Kir2.1 through Kir2.4, respectively. There is evidence that in the mammalian heart Kir2.1-Kir2.3 isoforms are expressed in cardiac myocytes, and that Kir2.4 is probably only expressed in neuronal cells. It is well established that members of the Kir2 subfamily underlie I_K1, although the subunit composition varies among species and cell types, and channel complexes are likely formed as hetero-tetrameric structures.

Crystal Structure of Kir2 Channels

In recent years, x-ray crystallographic structures of both bacterial and mammalian homologs of several Kir channels have been obtained.⁵ Figure 4-2 highlights some important and common features of Kir channel structure based on the results of work with Kir2.1 and Kir2.2 mammalian channels.⁶ It is now firmly established that Kir channels are tetramers of distinct subunits, each having two transmembrane domains (M1 and M2), relatively small N-terminal, and large C-terminal cytoplasmic domains, and a pore-forming structure between M1 and M2 (see Figure 4-2). The pore structure contains pore helix directed toward the conduction pathway and the characteristic GYG (or GFG) motif, also known as K⁺ channel signature sequence, that contributes to the selectivity filter in all potassium channels. The M1 and M2 transmembrane domains in each subunit are arranged as an antiparallel coiled-coil and make contact with each other. Kir2 channels have a negatively charged amino acid (D172 in Kir2.1) located in approximately the middle of the pore, which has a critical role in the phenomenon of inward rectification discussed later.

There are at least three distinct regions of the intramembrane part of the pore: selectivity filter, a centrally located water filled cavity of approximately 10 Å in diameter, and the narrowing of the pore at the cytoplasmic side as the pore-lining M2 helixes come closer to each other (known as bundle-crossing). M2 helixes also possess a highly conserved glycine residue (a “hinge”; G168 in Kir2.1) that likely contributes to the channel gating by allowing the helixes to bend and change the size of the pore at the bundle crossing.

The so-called slide helix formed by the N-terminus region just preceding the M1 helix is another unique and important regulatory feature of both bacterial and mammalian Kir channels.⁷ This helix intercalates between the inner leaflet of the plasma membrane and cytosol likely because of its amphiphilic nature. The important functional role of the slide helix is highlighted by the fact that many loss-of-function mutations associated with Andersen-Tawil syndrome (LQT7) are located in the slide helix.

A large C-terminal domain provided by four Kir2.x subunits consists primarily of β-strands and potentially strongly interacts with a smaller N-terminal domain. As shown in Figure 4-2, C- terminal domain forms a large intracellular vestibule, approximately 30 Å in length, for easy ion passage and likely provides binding sites for various intracellular agents. The cytoplasmic domain harbors a number of residues (e.g., E224 and E299 in Figure 4-2) known to contribute to inward rectification.

A membrane phospholipid phosphatidylinositol 4,5-bisphosphate (PIP₂) is an important structural and regulatory component of Kir channels.⁸ In the membrane, PIP₂ likely interacts with hydrophobic amino acids on both M1 and M2 transmembrane helices as well as with a well-conserved (among Kir channels) RWR motif located just at the end of the M1 domain (see Figure 4-2). The interaction of PIP₂ with a specific region in the CTD leads to an approximately 6 Å translation of the entire CTD toward the membrane associated with the movement of the M1 helix, which ultimately leads to the opening of the gate.⁶

Mechanism of Polyamine-Induced Rectification

The key experiments conducted in mid-1990s with the first cloned members of the Kir2 subfamily have clearly demonstrated that almost all essential properties of classical strong inward rectification could be explained by a voltage-dependent block of the channel by ubiquitous intracellular organic cations, the polyamines.⁹ Micromolar concentrations of free polyamines (primarily spermine and spermidine) are sufficient to reproduce the degree of rectification observed in native cells. The strength of rectification varies among the different members of the Kir subfamily, and although every Kir channel shows some degree of inward rectification, they can be broadly grouped as either strongly rectifying (Kir2 and Kir3) or weakly rectifying (Kir1 and Kir6) channels. It is well established that the time-dependent activation of strong, inward rectifiers during membrane hyperpolarization reflects the exit of polyamines (primarily sperimine) from the pore.

Rectification Properties Are Related to Electrostatic Interactions in the Cytoplasmic Pore of a Kir Channel

Early work with cloned Kir channels established that strong inward rectification by intracellular polyamines depends on three negatively charged residues located in the second transmembrane domain (D172; the rectification controller; see Figure 4-2) and in the C-terminal tail (E224 and E299; Kir2.1 amino acid [aa] numbering; see Figure 4-2). Positively charged polyamines enter the Kir channel pore to physically occlude it, a process aided by negative charges provided by aspartate and glutamate residues. Both electrophysiologic and structural data are consistent with the idea that channel block occurs in two sequential steps. A weakly voltage-dependent (shallow) blocking step involves entry of polyamine into the Kir2.1 pore at a site provided by a ring of negative charges (E224 and E299) in C-terminus. A more strongly voltage-dependent blocking step reflects the movement of polyamine to its deep binding site near D172 residue. It is also believed that the strong voltage dependence of the polyamine block arises not only from the high valency (z) of polyamines (z≈4 for spermine) but also from a displacement (push) of K⁺ ions through the pore.¹⁰ Among the polyamines, spermine is the most voltage-dependent blocker and also has the highest potency for blocking the channel.

A characteristic feature of Kir2.1 and Kir3 channels is a flexible cytoplasmic pore-facing G-loop (see Figure 4-2) that forms a girdle around the central axis of the Kir channel.¹¹ It was estimated that this girdle constricts the ion permeation pathway to approximately 3 Å. Mutations in the G-loop were shown to disrupt inward rectification. In addition to the previously described E224 and E299 residues, it was shown that A255 and A259 located farther away from the pore axis are also involved in channel rectification. Electrophysiologic experiments using cysteine modifications in the pore region of a mutant Kir6.2 channel (N160D; equivalent to D172 in Kir2.1) showed that spermine binds at a deep site beyond the rectification controller residue D172, a site that is close to the extracellular mouth of the pore. In another study, refinement of the crystal structures of bacterial KirBac1.1 and KirBac3.1 allowed identification of the shallow polyamine binding sites at the cytoplasmic interface between the two subunits.¹² These observations notwithstanding, the precise mechanism involved in rectification is still being worked out, and the exact location of polyamine binding sites in Kir channels is still a controversial issue.

Differential Properties of Kir2.x Subfamily

Consistent with their overall significant sequence homology, all Kir2 channels have the three mentioned residues (D172, E224, and E299) important for rectification at the equivalent positions. Recent studies, however, unexpectedly showed that rectification properties are rather different in the three Kir2 isoforms.¹³ Specifically, the data show that Kir2.2 channels display significantly stronger voltage dependence of rectification than that observed in Kir2.1 and Kir2.3 channels. The voltage dependence of steady-state rectification is different between Kir2.x channels, as are single-channel conductance and kinetics properties of the currents. In particular, polyamine unblock (activation) at negative membrane potentials in Kir2.3 channels is several-fold slower than in Kir2.1 and Kir2.2 channels. Single-channel conductance is smallest in Kir2.3 (approximately 10 pS), medium in Kir.2.1 (approximately 25 pS) and largest in Kir2.2 (approximately 35 pS).