The physiological roles of ion channels depend critically on their ability to select one type of an ion from another and allow the selected ion to readily pass through the protein from one side of the cell membrane to the other.  Potassium channels allow at least 10⁶ K⁺ ions/sec to pass while selecting against Na⁺ ions by at least a factor of 1,000.

The crystal structures of several K channels have been solved- the first, in 1998, was the KcsA channel from the bacterium, Spreptomyces lividans (Doyle et al., 1998).  These channels are tetramers and all contain an integral pore of extremely similar structure.  The picture at the right is a side view of the KcsA channel with two subunits removed for clarity.  All K channels have a very similar pore with a narrow "selectivity filter" near the outer entrance and a wide, internal vestibule.

The inner vestibule can change conformation controlling whether or not the channel is open and conducting ions or closed.  The crystallized KcsA appears to be in a closed conformation.  The bacterial MthK Ca²⁺-activated K channel was apparently crystallized in the open conformation as can be seen by the very wide inner vestibule in the figure.

There are two general ways that ion channel proteins achieve selectivity: (1) by selectively binding the preferred ion or (2) by excluding the impermeant ion from the pore.  Na⁺ ions are smaller that K⁺ ions and so might be difficult to exclude from the pore in K channels.  However, ions in aqueous solutions are surrounded by water molecules- the dipolar water molecules are attracted to the charged ions.   Na⁺ ions have a larger water "cloud" than do K⁺ ions.

K channels appear to have solved the selectivity problem by endowing the selectivity filter region with an extremely high (sub micromolar) affinity for K⁺ ions.  Thus, the energy required to (partially) dehydrate a K⁺ ion is compensated by binding in the filter.  No such comfortable fit exists for Na⁺ ions so K⁺ ions successfully out-compete Na⁺ ions for pore occupancy.

The tight binding of K⁺ ions in the pore creates a new problem: how to get a high ion throughput if the permeant ion sticks so tightly.  There are three general ways this could be achieved: 

1. Electrostatic Repulsion.  An ion enters a pore already occupied by a bound ion.  Electrostatic interaction between them repels the "stuck" ion out of the pore.

2. Conformational Change.  An ion enters a pore already occupied by a bound ion.  The binding of the entering ion induces a conformational change in the pore protein and the "stuck" ion is released from the pore.            
3. Using the Stairs.  We often use stairs to help us get to the 3rd floor from the basement.  Most of us can't jump all the way at once: we take the stairs one at a time.  In the same way, an ion can get out of a deep energy well if there are stairs-that is, additional binding sites of lower affinity.  The "climbing stairs" type of ion permeation can be mathematically simulated using "Eyring rate theory" (Eyring, H. 1935. J. Chem. Phys. 3:107-115).  In this approach, ions "hop" from one local energy minimum to the next over intervening energy barriers.  Click here to download a 4-barrier-3-site model for computing ion fluxes using Eyring rate theory.

Note that these are not necessarily mutually exclusive mechanisms.  For example, electrostatic repulsion between K⁺ ions could occur within the selectivity filter but not in the larger inner vestibule.

Our long term goal is to critically test these various ion permeation mechanisms.  To do so requires determining the number, location, and properties of the ion binding sites in the pore of K channels. 

We have found that some K channels can simultaneously accommodate at least four K⁺ ions and so must have at least that many  binding sites (Stampe and Begenisich, 1996).  The early crystallographic work from Rod MacKinnon's lab  (Doyle et al., 1998) identified four K⁺ ion binding sites in the selectivity filter and another in the inner vestibule.   Later analysis from this lab suggested that usually no more than two K⁺ ions could simultaneously occupy the selectivity filter which would imply a maximum simultaneous occupancy of three K ions- at least one less than our evidence demonstrates.  Thus, we predict that there may be more than a single binding site for ions in the inner vestibule- a prediction borne out by recent experiments  (Thompson and Begenisich 2003b- see Publications).  In addition, though not noted in the associated publication (Jiang et al., 2003), the crystallized structure of the KvAP channel reveals two K⁺ ions in the inner vestibule.

With some idea of the number of K⁺ sites within the pore, a more focused approach to understanding the permeation mechanism can be undertaken.  One useful tool for such investigations is the tetraethyl ammonium (TEA) ion.  TEA is about the same size as a hydrated K⁺ ion and carries the same +1 electric charge.  TEA can bind in the inner vestibule near the selectivity filter and can also bind just external to the selectivity filter (but cannot pass through the selectivity filter).  We found that there is no electrostatic repulsion between  TEA ions occupying their two sites flanking the selectivity filter (Thompson and Begenisich, 2000) putting at least some limit to the range of this aspect of permeation.

Other experiments suggests that electrostatic repulsion is likely between TEA ions at their site just external to the selectivity filter and K⁺ ions within the selectivity filter (Thompson and Begenisich, 2003b).  This type of interaction is responsible for the voltage dependence of block by TEA.  TEA ions do not move through the membrane electric field (most of which is probably dropped across the selectivity filter) but, due to their interaction with K⁺ ions in the selectivity filter who DO move through the electric field, external TEA block appears to be voltage dependent.

Additional experiments using TEA allowed us to identify two distinct conformational states of the selectivity filter.  (likely associated with different numbers of ions in the selectivity filter) of the selectivity filter both of which conduct (Thompson and Begenisich, 2005).  These finding agree with the two conformational states seen in crystals of the KcsA channel grown in low or high K⁺ concentrations (Morais-Cabral et al., 2001) but not with the authors' conclusion that the low K-occupancy state is non-conducting.

KcsA Pore Structure

A cut-away, side view of the structure of the bacterial K channel.  The narrow, selectivity filter is on the left.

MthK Pore Structure

In this MthK picture only the pore structure is shown- two of the four subunits have been removed for clarity.  The narrow selectivity filter (left) and wide inner vestibule (right) are apparent.

Ion Selectivity

A (blue) ion channel protein imbedded in a (green) cell membrane.  A pore large enough to accommodate a potassium ion (red sphere) would also allow permeation by the smaller sodium ion (yellow sphere).

Electrostatic Repulsion

In this mechanism an ion entering an occupied pore electrostatically repels the bound ion.

Conformational Change

Here, an entering ion induces a conformational change which frees the stuck ion.

Climbing Stairs

This mechanism preserves the selectivity inherent in a  high affinity site and as long as each step isn't too big, high throughput is also maintained.