Site map > Home page > CV > Systems > Consequences of Channelling

Kinetic Consequences of Channelling

This page contains the introduction to the following book chapter: Athel Cornish-Bowden (1997) "Kinetic Consequences of Channelling" pp. 53–70 in Channelling in Intermediary Metabolism, (ed. Loranne Agius and H. Stanley A. Sherratt) Portland Press, London; ISBN 1 875578 075 5. This book is available from Amazon (UK) and from Amazon (USA)

Il est vrai que certaines paroles et certaines cérémonies suffisent pour faire périr un troupeau de moutons, pourvu qu'on y ajoute de l'arsenic. (Voltaire, 1771)

Channeling can become a tremendously powerful tool of information gathering, self-growth and life enrichment. It can help you do everything from choosing a career path that’s right for you to finding the right mate. (Leaflet found in a laundromat, 1989)


Direct transfer of metabolites between consecutive enzymes, or channelling, has been a controversial topic at two different levels. The first is essentially chemical and kinetic, and concerns the strength of the evidence that is adduced in favour of channelling in numerous specific systems. The second is physiological, and concerns the effects, or lack of them, that channelling will have on metabolic systems if it occurs. These two controversies are, of course, independent: one can reasonably believe it to be established beyond doubt that channelling occurs without necessarily accepting that it has any significant physiological consequences; alternatively, one can believe channelling to be of potential physiological importance while still denying the reality of most of the specific examples. The former position is similar to my own, whereas the latter appears to be that of Gutfreund and Chock. In this chapter I shall mainly be concerned with the kinetic consequences of channelling, i.e. with the second controversy, but first I shall briefly review some of the arguments about its existence.

The question of whether channelling actually occurs in the systems where it is reported is especially controversial if the enzymes involved in the channel do not form a stable complex. In such a case the channelling is said to be dynamic, and it implies that diffusional encounter between two macromolecules, an enzyme–substrate complex and a second enzyme, may be faster than diffusion of the free product released by the first enzyme–substrate complex to the second enzyme. This is not impossible, of course, if the concentration of enzyme-bound intermediate is so much higher than that of the free intermediate that collisions between enzymes and enzyme-bound intermediates occur more often than between enzyme and free intermediates despite the faster diffusion of each free intermediate molecule. The corresponding problem does not arise in static channelling, where the enzymes form a stable complex that exists throughout the chemical process, and so no diffusion of macromolecules is required.

There is good evidence evidence that NADH — the most extensively studied candidate for channelling — can indeed be transferred directly from one enzyme active site to another. This evidence comes in particular from the enzyme buffering method designed by Srivastava and Bernhard, which is most easily explained by reference to an example. Suppose one is interested in testing the possibility of direct transfer of NADH from glyceraldehyde 3-phosphate dehydrogenase to lactate dehydrogenase. The first requirement is to measure the kinetic parameters for the lactate dehydrogenase-catalysed oxidation of NADH at some suitable concentration of the other substrate, pyruvate. This is done initially in the absence of glyceraldehyde 3-phosphate dehydrogenase. The parameters allow one to calculate what the rate ought to be at any low concentration of NADH. If such a low concentration is simply achieved by adding less NADH, then one can, of course, check the validity of the calculation directly, and in a properly carried out experiment there should be full agreement between the observed rate and the rate calculated from the kinetic parameters. Suppose, however, that the lower free concentration is achieved by adding enough glyceraldehyde 3-phosphate dehydrogenase to bind almost all of the NADH. In this case there are two possible resultd: if the complex of NADH with glyceraldehyde 3-phosphate dehydrogenase has no reactivity with lactate dehydrogenase the rate of oxidation should be exactly the rate calculated from the free NADH concentration, which can itself be easily calculated from the known dissociation constant of the complex; on the other hand, if the glyceraldehyde 3-phosphate dehydrogenase-NADH complex is a substrate for lactate dehydrogenase the rate will be higher.

In this example the measured rate turned out to be 5-6 times faster than the calculated rate, and corresponding results have now been obtained for 20 or more pairs of enzymes, in many cases with discrepancies of more than 20-fold. Even a 5-6-fold discrepancy is far outside the range of experimental error, which is known from control experiments with pairs of enzymes with the wrong stereochemistry (see below) for direct transfer.

Such results are normally taken to indicate that NADH can indeed be channelled from one member of each such pair to the other. Before accepting this conclusion uncritically, however, we should consider two other possible explanations that do not involve channelling. One is that addition of a very large excess of the buffering enzyme may result in artefactual effects due to its capacity to catalyse the other reaction. In the example considered, the concentration of glyceraldehyde 3-phosphate dehydrogenase was of the order of 106 -fold higher than that of lactate dehydrogenase. As a result, even 0.0005% contamination of the glyceraldehyde 3-phosphate dehydrogenase with lactate dehydrogenase (or a contaminant capable of catalysing the lactate dehydrogenase reaction) would suffice to produce a 6-fold increase in rate over the calculated value. Srivastava and Bernhard were well aware of this danger and took care to eliminate it, and later workers have been equally careful. Nonetheless, it is easy to overlook the problems of contamination that can arise when one enzyme is present in enormous excess over the other and it is important to be conscious of it in any enzyme buffering experiment, especially as it is quite normal for dehydrogenase preparations to be contaminated with other dehydrogenases.

The second possibility is that the glyceraldehyde 3-phosphate dehydrogenase had some activating effect on the lactate dehydrogenase unrelated to the presence of NADH. This type of effect is difficult to eliminate in any one example (because one cannot do kinetic experiments in the absence of substrate), but it stretches credulity to regard it as a general explanation once the enzymes concerned have been classified into pairs that give positive effects and pairs that do not, as I now discuss.

Enzymes that use NADH as substrate show a remarkable stereospecificity that contrasts with the behaviour of most other classes of enzyme. NADH has two non-equivalent H atoms that yield the same NAD+ molecule on oxidation, and any dehydrogenase can in principle show specificity for one or the other; in contrast to most other classes of enzymes, for which all that catalyse the same sort of reaction normally have the same stereospecificity, dehydrogenases are about evenly divided between ones that use the pro-S and ones the use the pro-R H atom of NADH. Benner favoured an explanation in terms of chemistry and evolutionary optimization of the individual enzymes, but an alternative is to suppose that opposite stereospecificity in consecutive enzymes is a precondition for direct transfer of NADH from one active site to another without requiring the molecule to rotate during the transfer. This would imply evolutionary selection for enzymes catalysing consecutive reactions to have opposite stereospecificity as a way to permit the possibility of channelling.

It turns out that there is an excellent correlation between pairs of enzymes that give a positive result in the enzyme buffering test and enzymes with opposite stereospecificity; indeed, all pairs with the same stereospecificity give a null result in the enzyme buffering test. If all of the positive results in the enzyme buffering test were due to an NADH-independent activation of one enzyme by another, one would have to attribute this correlation to chance; it is more reasonable to interpret it as evidence that NADH can indeed be directly transferred between enzymes with opposite stereospecificity but not between enzymes with the same stereospecificity.