Pore Dimensions of Ion Channels: Applications

3. Example application to cholera toxin B subunit pentamer

Cholera toxin is a protein of molecular weight of approximately 84 kD and is responsible for the massive loss of fluids characteristic of the disease by altering the transport of salt and fluids across the intestinal epithelium²². The intact protein is a heterohexamer composed of five B-subunits with a single A-subunit²³. Recently X-ray crystal structures of both the intact heterohexamer¹⁹ and the B₅ homopentamer²² have been determined. Structures of the hexamer show that a ring of B-subunits form a central "donut hole" which is filled with the A2 alpha-helix from the A subunit. However, the B-subunits can form a stable complex which is similar except that the central hole is empty. Krasilnikov and co-workers have shown²¹ that this pentamer is able to form ion channels in artificial membranes at low pH. It remains unclear as to whether the ability of the toxin to form ion channels has any role in its pathological mechanism of action²³.

An examination of the pore dimensions of the crystal structure for the cholera toxin B subunit pentamer was conducted using HOLE. It has been suggested that the A-subunit which must translocate through the membrane as part of its mode of toxicity may do so by passing through the channel formed by the B subunit pentamer²¹ . However, it has been noted that the pore found in the crystal structures is not sufficiently large to allow such a motion²³ and that there may be some rearrangement of the pentamer on incorporation into the membrane. It is therefore interesting to examine whether the crystal structure is compatible with the conductance data available - if it is then such a rearrangement may be excluded.

The structure examined is the 2.2Å resolution structure solved by Merritt et al.¹⁹:

(click to see picture full size)
A view down the channel direction vector of cholera toxin B₅ showing the pore lining helices from the five separate chains

If you have configured your browser to use rasmol then try using this link to the pdb file to interactively view the structure.

The pore dimensions of the channel as calculated by HOLE show a pore which is sufficiently large to be accommodate two water molecules side by side at any place within the channel.

(click to see picture full size)
The pore surface of cholera toxin B₅. If you have a suitable viewer (see http://www.ch.ic.ac.uk/VRML/ for details) then have a look at the vrml version of this figure.

This can also be seen by plotting a graph of pore radius versus distance along the channel coordinate:

(click to see picture full size)

Comparison of the minimum effective pore radius found with that of the porin family is instructive:

System	Minimum effective radius in Å
E. coli porin Ompf¹⁵	3.90
E. coli phosphoporin 1PhoE¹⁵	3.39
E. coli maltoporin¹⁸	2.85
cholera toxin B₅¹⁹	3.24

Both the porins and the maltoporin allow translocation of molecules up to the size of a ketose carbohydrate²⁴. As the minimum radius for cholera toxin B₅ is in the same range it is expected that it would allow the translocation of solutes of around this size. In terms of being able to translocate a protein domain the size of the A-subunit of cholera toxin, this confirms the observation²³ that the crystal structure is incompatible with this, unless the domain was unfolded prior to translocation.

This begs the question; is there a possibility of a conformational rearrangement of the B pentamer, with a widening of the pore, on incorporation into a lipid bilayer as postulated by Zhang²³? If the HOLE prediction of the conductance of the crystal structure was significantly smaller than the experimentally found conductance this would be indicative of such a change. In fact the predicted conductance is consistent with the experimental value ¹¹ so it would appear that a conformational rearrangement is unlikely. This conclusion is further strengthened by comparing the prediction for the expected profile of a PEG addition experiment with that found:

(click to see picture full size)

It can be seen that according to the HOLE calculation the crystal structure is in fact compatible with the effect on conductance data. Thus the possibility of a significant conformational arrangement accompanying the incorporation of cholera toxin B₅ can be seen to be unlikely. This does not necessarily mean that the proposed translocation of the A-subunit through the B pentamer does not occur, only that such a motion would involve a considerable alteration to the structure of either or both the A-subunit and the B pentamer.

Oliver S. Smart (last modified 20/12/96)