2.5.1—
Membrane Pores and Channels

Lipid or oil is a very effective electrical insulator (it is used in high voltage underground electric cables for this purpose) and it is not surprising to find that the electrical resistance of synthetic bilayers made from pure phospholipid is very high, being 10⁷ to 10⁹ ohms cm² . In nature, the current conducted across membranes is carried by ions and we can see that an unmodified lipid bilayer with such a high resistance is a poor material across which to conduct this essential current of electrolyte. It is not surprising to find, therefore that the electrical resistance of natural membranes is very much less than that of bilayers, usually lying in a wide range of 10² to 10⁵ ohms cm² .

As described on p. 55, lipophilic carriers can greatly reduce the resistance (and hence increase the c onductance) of synthetic membranes and may resemble the carrier molecules in membranes, but it is also widely believed that membranes contain water-filled pores which must contribute to their relatively high conductance by allowing water and selected solutes to by-pass the lipid domain of the membrane. Certain antibiotic molecules, such as nystatin, appear to condense cholesterol molecules in both synthetic (Holz & Finkelstein, 1970) and natural membranes (de Kruijff & Demel, 1974) to form pores with a radius of ca. 0.4 nm. The presence of such pores greatly increases the electrical conductance and hydraulic conductivity of the membrane. In passing we might note that many antibiotics have their effect by enormously increasing the passive permeability of cell membranes causing non-resistant cells to lose their contents or to lyse. Interest in these substances stems from the experimental evidence that cell membranes also possess pores of similar size and that the antibiotic merely induces an extreme expresssion of the normal condition.

Although the word 'pore' is often used to describe channels through which solutes and water can move, we should resist the temptation to conclude that all 'pores' are definable structural entities like the ones induced by nystatin (see above). In many instances a 'pore' may be more like a transient imperfection in membrane structure. This latter type may have a certain statistical probability but have no fixed position.

Proteins which traverse the lipid layer may give rise to hydrophilic channels or pores (see Fig. 2.5e). Indirect evidence supporting this idea comes from a study in which red blood cell membranes were exposed by deeply etching frozen cells under vacuum (Pinto da Silva, 1973). Shrinkage of the membrane surface was observed in areas overlying groups of membrane particles possibly due to sublimation of ice from within the embedded particles; for this to have occurred this water would be a free liquid in the thawed condition (cf. bound water p. 35).

The first circumstantial evidence for the existence of pores come from studies by Collander and Barlund (1933) on the permeation of the giant internodal cells of the alga, Chara ceratophylla, by a number of uncharged solutes and water. They found that in almost every case the rate of movement of a substance into a cell depended on its molecular weight and dimensions and on its solubility

in oil relative to water; substances with high oil solubility permeated most rapidly. This strongly suggested that movement across cell membranes involved the movement of the solute out of the water, its solution in lipid and subsequent diffusion through it, and its re-entry into the aqueous phase at the inside face of the membrane. The authors found, however, that several small molecules permeated the membrane far more rapidly (more than 100 times faster in the case of water) than their relative solubility in oil suggested. The upper size limit for molecules which behaved anomalously was a radius of 0.4 nm and it was suggested that the membrane was constructed as a very fine sieve containing pores of 0.4 nm radius through which water (0.25 nm radius) and certain solutes could move. Since this early work a great deal has been done and the equivalent pore radius in many plasma membranes has been confirmed as 0.4 nm (see Solomon & Gary-Bobo, 1972). It must be said, however, that some authorities are reluctant to accept that pores can provide channels for the bulk movement of water and solutes; the arguments for and against pores have been clearly discussed in Oschman et al., (1974).

There has been much discussion about whether pores admit ions, and if so, which ones. Membranes are known to control very precisely the relative rates at which various ions will diffuse across them; potassium ions will diffuse ten to one-hundred times faster than sodium. Ions in solution are hydrated by binding one or more water molecules; it requires a great deal of energy to dehydrate an ion and for this reason we should think of ions in all natural circumstances as being hydrated. Sodium binds 5 water molecules, whereas potassium binds 3, the former is, therefore, the more bulky ion whose diffusion into the narrow water filled pores would be slower than for the smaller hydrated potassium ion. On the other hand the anion chloride, which has only one water molecule in its hydration shell, diffuses much more slowly than either K⁺ or Na⁺ . This is probably due to the fact that 'pores' carry a predominantly negative charge so that cations would be attracted to them, while anions would be repelled—a small number of positively charged pores would handle the flow of anions. Polyvalent cations and anions have much more water in their hydration shells, e.g. Ca²⁺ has 10 and SO₄ ^2– has 8 and it seems likely that these would be totally excluded from the pores.