top of page

Cells, Gels, and the Engines of Life


On the surface everything seems to be in order. Virtually all known cellular processes are by now accounted for by well-described mechanisms: ions flow through channels; solutes are transported by pumps; vesicles are moved y motors; etc. But as we shall see as we probe beneath the surface of these solutions, a bewildering level of complexity hints at a situation that could parallel the epicycles.

I propose to step back and regroup. Firm ground needs to be identified. I begin by considering two elements thoughts to be fundamental to cell function: membrane pumps and channel. Pumps transport solutes across the cell boundary against their respective concentration gradients. Channels permit the solutes to trickle back in the opposite direction.

Through a balance between pump-based transport and channel-based leakage, the characteristic partitioning of solutes and ions is though to be established. Thus, potassium concentration is relatively higher inside the cell, and sodium is relatively higher outside.

That pumps and channels existence seems beyond doubt --- or to put it more precisely, the existence of proteins with pump-like or channel-like features cannot be doubted. Genes coding for these proteins have been cloned, and the proteins themselves have been exhaustively studied. There can be not reason why their existence might be challenged.

What I will be considering in this chapter is whether these proteins really mediate ions partitioning. Because a “pump” protein inserted into an artificial membrane can translocate an ion from one side of the membrane to the other, can we be certain that ion partitioning in the living cell necessarily occurs by pumping?


The concept of the cell membrane arose during the era of light microscopy, prior to the time any such membrane could actually be visualized. Biologists of the early 19th century observed that a lump of cytoplasm, described as a “pulpy, homogeneous, gelatinous substance” did not mix with the surrounding solution.

To explain why this gelatinous substance did not dissolve, the idea arose that it must be enveloped by a water-impermeant film. This film could prevent the surrounding solutions from permeating into the cytoplasm and dissolving it. The nature of the membranous film had two suggested variants. Kuhne (1864 ) envisioned it as a layer of coagulated protein, while Schultze (1863 ) imagined it as a layer of condensed cytoplasm. Given the experimental limitations of the era, the nature of the putative film, still not visible, remained uncertain.

The idea of an invisible film was nevertheless attractive to many of the era's scientists, and was increasingly conferred with special attributes. Thus, Theodore Schwann (1839 ) viewed this film as "prior importance to its contents." The membrane grew in significance to become the presumed seat of much of the cell's activity. Yet this view was not accepted by all. Max Schultze, often referred to as the father of modern biology, discounted the evidence for a cytoplasmic film altogether, and instead regarded cells as "membrane less little lumps of protoplasm with a nucleus". In spite of Schultze's prominence, the concept of an enveloping membrane held firm.

The modern idea that the membrane barrier might be semi-permeable came from the plant physiologist Wilhelm Pfeffer. Pfeffer was aware of the ongoing work of Thomas Graham (1861) who had been studying colloids, which are large molecules suspended indefinitely in a liquid medium --- e.g. milk. According to Graham's observations, colloids could not pass through dialysis membranes although water could. To Pfeffer, colloid seemed to resemble thy cytoplasm. If the dialysis membrane were like the cell membrane, Pfeffer reasoned, the cell interior would not dissipate into the surrounding fluid even though the membrane might still be water-permeable. Thus arose the idea of the semi-permeable membrane.

Pfeffer took up the semi-permeable membrane idea and pursued it. He carried out experiments on membrane models made of copper ferrocyanide, which acted much like dialysis membranes in that they could pass water easily but solutes with great difficulty. It was on these experiments that Pfeffer based the modern cell-membrane theory.

The membrane at this stage was presumed permeable to water, but little else.

Although Pfeffer's theory held for some time, it suffered serious setbacks when substances presumed unable to cross the membrane turned out to cross. The first and perhaps most significant of these substance was potassium. Therecognitionm in the early 20th century, that potassium could flow into and out of the cell prompted a fundamental rethinking of theory.


Faced with the needed to explain the pottasium-permeability issue, Boyle and Conway ( 1941 ) proposed an elegant solution: the potassium channel. Since the hydrated potassium ion was known to be smaller than the hydrated sodium ion, Boyle and conway proposed transmembrance channels of critical size --- large enough to pass potassium and its shell of associated water, but small enough to exclude sodium with its shell. The membrane was effectively a sieve that passed small ions, but excluded larger ones.

The Boyle-Conway atomic sieve theory was attractive in that it could also account for several known features of cell behavior without too much difficulty. It explained the accumulation of potassium inside the cell as an attraction to the cell's negatively charged proteins (aka Donnan effect). It explained the cell potential as arising from a charge separation across the membrane (capacitive effect). and it accounted for the changes of cell volume that could be induced by changes of external potassium concentration (osmotic effect). The sieve theory seemed to explain so much in a coherent manner that it was immediately granted an exalted status.

But another problem cropped up, perhaps even more serious than the first. The membrane turned out to be permeable also to sodium. The advent of radioactive sodium made it possible to trace the path of sodium ions, and a cadre of investigators promptly found that sodium did in fact cross the cell boundary. This finding created a problem because the hydrated sodium ion was larger than the channels postulated to accommodate potassium; sodium ions should have been excluded, but they were not. Thus, the atomic-sieve theory collapsed.

Collapse notwithstanding, the transmembrane-channel concept remained appealing. One had to begin somewhere. A channel-based framework could circumvent the sodium problem with separate channels for sodium and potassium: if selectivity were based on some criterion other than size, then distinct channels could suffice. A separate channel for sodium could then account for the observed leakage of sodium ions into the cell.

But leakage of sodium introduced yet another dilemma, of a different nature. Sodium could now pass through the channel, flowing down its concentration gradient and accumulating inside the cell. How then could intracellular sodium remain as low as it is?


The solution was to pump it out. In more-or-less the same manner as a sump-pump removes water that has leaked into your basement, a membrane pump was postulated to rid the cell of the sodium that would otherwise have accumulated inside.

The idea of a membrane pump actually originated before the sodium problem. It began at the turn of the last century with Overton, a prominent physiologist who had advanced the idea that the membrane was made of lipid. Realizing that some solutes could cross an otherwise impermeable lipid membrane, Overton postulated a kind of secretory activity to handle these solutes. Through metabolic energy, the membrane could thus secrete, or pump, certain solutes into or out of the cell.

The pump concept resurfaced some 40 years later, to respond specifically to the sodium-permeability problem. Dean did not have a particular pumping mechanism in mind; in fat, the sodium-pump was put forth as the least objectionable of alternatives. Thus, Dean remarked, "It is safer to assume that there is a pump of unknown mechanism which is doing work at a constant rate excreting sodium as fast as it diffuses into the cell." With this, the sodium pump (later, the Na/K exchange pump) came decidedly into existence.

By the mid 20th century, then, the cell had acquired both channels and pumps. With channels for potassium and sodium, along with pumps to restore ion gradients lost through leakage, the cell's electrophysiology seemed firmly grounded.


Channels and pumps came into being as ad hoc hypotheses needed to patch otherwise flagging theories. The channel arose when a putatively ion-impermeant membrane was found to pass potassium; the channel could pass potassium while properly excluding sodium and other larger hydrated ions. Then, sodium was found to enter the cell and instead of reconsidering the channel concept, a second channel specific for sodium was postulated. Sodium permeability also implied a persistent leak into the cell. To keep gradients from collapsing, sodium pump was postulated.

For the same reason that a sodium pump was needed, it became evident that pumps for other solutes were needed as well. Virtually all of the cell's solutes partition far out of electrochemical equilibrium and therefore need to be pumped. Hydogen-ion pumps, calcium pumps, chloride pumps, and bicarbonate pumps to name a few, soon came into being over and above the postulated sodium/potassium pumps. Yet, even at the height of this intense activity, Glynn and Karish in their classic review had to reluctantly admit that notwithstanding an enormous thrust of experimental work on the subject, still no hypothesis existed to explain how pumps pump.

The channel field exploded similarly. With the advent of the patch-clamp technique in the late 1970s' investigators had gained the capacity to study what appeared to be single ion channels. It seemed for a time that new channels were being identified practically monthly, many of them apparently selective for a particular ion or solute. The number of channels has risen to well over 100. Even water channels have come into being. Elegant work was carried out to try to understand how channels could achieve their vaunted selectivity. At least some channels, it appeared, could pass one ion or solute selectively, while excluding most others. At least, some channels, it appeared, could pass one ion or solute selectively, while excluding most others.


The existence of single ion channels appeared to be confirmed by ground breaking experiments using the patch-clamp technique. In this technique the tip of a micropipette is positioned on the cell surface. Through suction, a patch of membrane is plucked from the cell and remains stuck onto the micropipette orifice. A steady bias voltage is placed across the patch, and the resulting current flow through the patch is measured. This current is not continuous; it occurs as a train of discrete pulses. Because the pulses appear to be quantal in size, each pulse is assumed to correspond to the opening of a single ion channel.

This dazzling result has so revolutionized the field of membrane electrophysiology that the originators of the technique, Erwin Neher and Bert Sakmann, were awarded to Nobel Prize. The observation of discrete events would seem to confirm beyond doubt that the ions flow through discrete channels.

Results from the laboratory of Fred Sachs, on the other hand, make one wonder. Sachs found that when the patch of membrane was replaced by a patch of silicon rubber, the discrete currents did not disappear; they remained essentially indistinguishable from those measured when the membrane was present. Even more surprisingly, the silicon rubber sample showed ion-selectivity features essentially the same as the putative membrane channel.

A similarly troubling observation was made on polymer samples. Current flow through synthetic polymer filters was found to be discrete, just as in silicon rubber. The filters also showed features commonly ascribed to biological channels such as ion selectivity, reversal potential, and gating. Yet, the sample was devoid of any protein or lipid.

In yet another set of experiments, channel-like behavior was observed in pure lipid-bilayer membranes. Following brief exposure to large concentration of lipid vesicles ejected from a pipette tip approximately 0.5 mm distant, these membranes showed typical channel-like fluctuations. Conductance changed in ways usually considered to be indicative of reconstituted protein channels --- including step conductance changes, flickering, ion selectivity, and inactivation. But no channels were present; the membrane contained only lipid.

What are we to do with such observations? It is clear from these three studies that the discrete currents previously taken to confirm the existence of single biological channels seem to be general features of current flow through small samples. The currents presumably arise from some common feature of these specimens that is yet to be determined, but evidently not from single channels since they are absent. The channels may exit --- but the prime evince on which their existence is based is less than conclusive.

A second point to consider is the manner in which the channel achieves its specificity. Channels exist for each one of the cell's ions; additional channels exist for amino acids, peptide, toxins, and sugars, most of these being otherwise unable to cross the lipid bilayer; and there are also channels for water. Thus, a plethora of channels exists, most engineered to be solute specific. How is such specificity achieved?

To explain such exquisite specificity, models of some pomposity have evolved. Most models are sufficiently complex that the solution requires numerical methods. Indeed, calculating the trajectory of a molecule diffusing through a channel during a 100=picosecond time window is the work of a supercomputer. The naive question nevertheless lingers: How is it that small solutes do not pass through large channels?

Textbook depictions of the channel as a hollow tube oversimplify the contemporary view of the channel as a convoluted pathway; and the process of selectivity is thought to rest not on size per se but on some complex interaction between the solute's electric field and structural features of the channel's filter. Also, channel selectivity is not absolute. Nevertheless, the issue of passing only one or a few among a field of numerous possible solutes including many smaller ones remains to be dealt with in a systematic manner. And the issue of non-biological samples producing single-channel currents certainly needs to be evaluated as well. What could all this imply?


Like channels, pumps come in many varieties and most are solute-specific. The number easily exceeds 50. The need for multiple pumps has already been dealt with: unless partitioning between the inside and outside of the cell is in electrochemical equilibrium, pumping is required. Because so few solutes are in equilibrium, one or more pumps are necessary for each solute.

A question that arises is how the cell might pump a solute it has never seen. Antibiotics, for example, remain in high concentration outside the bacterial cell but in low concentration inside. Maintaining the low intracellular concentration implies the need for a pump, and in fact, a tetracycline pump for E.coli has been formally proposed. To cope with substances it has never seen, the cell appears to require pumps over and above those used on a regular basis --- on reserve.

How is it possible? One potion is for existing pumps to adapt themselves to these new substances. But this seems illogical, for if they could adapt so easily why would they have been selective to begin with? An alternative is for the cell to synthesize a new pump each time it encounters a foreign substance. But this option faces the problem of limited space.

A second question is how the cell master the energy required to power all of its pumps. Where might all the ATP come from? Since ions and other solutes cross the membrane continually even in the resting state, pumps must run continuously to counter at these leaks. Pumping does not come free. The sodium pump alone has been estimated, on the basis of oxygen-consumption measurents, to consume 45~50% of all the cell's energy supply (Whittam, 1961). Current textbooks estimate a range of 30~35%.

In sum, pumping faces obstacles of space and energy. The membrane's size is fixed but the number of pumps will inevitably continue to grow. At some stage the demand for space could exceed the supply, and what then? Pumping also requires energy. The Na/K pump alone is estimated to consume an appreciable fraction of the cell's energy supply, and that pump is one of very man. How is the cell to cope with the associated energy requirement?

bottom of page