Chapter 13. Interaction in membrane assemblies

V.A. Parsegian
Division of Computer Research and Technology
and National Institute of Diabetes, Digestive and Kidney Diseases,
National Institutes of Health, Bethesda, MD 20892, USA

R.P. Rand
Department of Biological Sciences, Brock University,
St Catharines, Ontario, Canada, L2S3A1

1. Introduction

Charting the course of the planets, probing the power of atomic nuclei, and analyzing events at all levels of complexity in between, has always required knowing operative forces. At each level, different kinds of forces emerge to dominate the organization of matter. At each level, they must be measured. Such measurements have been made with lipids that form cell membranes. Unexpected forces have been seen. Now recognized, they should be understood and can be used in rational consideration of membrane organization.

The unexpected comes from the perspective of the macromolecule and of molecular assembly. For them their aqueous milieu becomes more than a medium through which the traditional forces, Van der Waals and electrostatic, act. Water, as it 'hydrates' the participants, becomes a part of the interacting molecular components themselves. Hydration and dehydration reactions make unexpectedly large contributions to the structure and energetics of the system. It is this new perspective that we emphasize in this chapter.

Lipids are interesting first for their morphology or, better, their polymorphism, itself driven by the properties of amphiphiles in an aqueous medium. The many configurations of lipid assemblies that populate the rest of this book have their analogue in natural biological membranes. They form the boundaries defining 'in' and 'out' not only for whole cells or many viruses but also for internal compartments that arrange and carry on the cell's business. Forces within these two-dimensional assemblies can be expected to influence the function of proteins that direct the movement of material or energy between compartments. Membrane proteins behave differently when incorporated in different lipids. Can the strains in packing that drive lipid polymorphism be felt in the proteins that work in these lipids? Only through systematic measurement of packing energetics can one expect to know.

Lipids must also participate in the massive topological rearrangement required for membranes to fuse with one another while maintaining the overall integrity of the compartments they enclose. It is expected that the study of non-bilayer configurations of lipid assemblies would teach us about the energetics of that rearrangement and for that reason we extend our consideration to some of these.

In the larger sense, the field of inquiry rapidly widens beyond lipids alone. The contribution of hydration to the overall energetics of molecular assembly has been recognized for a long time, but it has been difficult to isolate and to measure. The new use of osmotic stress (OS) to measure water's role in lipid assembly [1, 2] is remarkably simple, generally applicable (fig. 1), and is capable of determining whether even extremely weak perturbations of water molecules near surfaces are energetically significant. Forces measured between large polymers like DNA [3] and polysaccharides [4] have also shown that tiny surface perturbations of water, of chemical potential as little different from bulk water as a fraction of a calorie/mole, can result in very large interaction energies when many such water molecules are involved. Further, the cost of removing all the perturbed water, or conversely the benefit of fully hydrating a newly exposed surface, is remarkably similar among these systems and is very high at 1.5-15 kcal/mole per square nanometer of surface area. (ATP hydrolysis yields 7.3 kcal/mole.)

The OS strategy has now been applied to individually functioning molecules. Mitochondrial voltage-dependent anion channels open (hydrate) with increased difficulty in the face of decreased water activity in their vicinity. A measure of that difficulty shows that about 1000 additional water molecules become associated with the open channel [5]. Other channels show similar dependence on water for gating [6, 7], usually to an extent related to their conductance, as if most of the newly associated water is hydrating a newly created aqueous cavity. Further the coupling of channel gating and lipids is seen in the observation that non-bilayer prone lipids modify the conductance-state probabilities of alamethicin channels [8].

Individual proteins in solution, subject to osmotic stress, show clear changes in the numbers of solute-excluding molecules during the transition between conformational states. The numbers are large and their possible roles intriguing.

• Kornblatt and Hoa have measured 10 additional water molecules associated with cytochrome oxidase when cytochrome a binds an electron [9]. These waters are then shed as the electron is passed internally to cytochrome a₃.Far from being 'futile', the water cycle appears necessary for regulating sequential conformational change.
• Colombo, Rau and Parsegian [10] have shown that as four oxygens bind to hemoglobin, sixty additional water molecules also associate with the protein. Sixty is a surprisingly high number and appears to be related to changes in quaternary structure since O₂ binding to myoglobin seems to show no such effect. Is the hydration of 600 angstrom² of newly exposed surface of the oxy conformation energetically significant? If that surface is like others, its full hydration would yield more than 8 kcal/mole Hb, or 2 kcal/mole per heme group.
• Hexokinase releases at least 70 water molecules, and probably many more, in the process of closing a large cleft upon binding glucose [11]. The expulsion of so much water, a molecular mass ten times that of the substrate itself, presumably reflects the dehydration required to confine the phosphorylation to glucose and prevent the futile 'phosphorylation' of water [12]. These three proteins show three qualitatively different solvation reactions on substrate binding; one hydrates, one dehydrates, and one shows a hydration/dehydration cycle.

Our aim in this chapter is, first, to describe how forces can actually be measured, then to survey, tabulate and illustrate the large set of force data now available for lipids. On their own, the fundamental data reviewed here allow some prediction of interactions. Whatever our limited understanding, these data provide good estimates of what it takes to bend membranes and to bring them toward 'contact'. This is a set of facts to motivate quantitative thinking on how cell membranes, and biological macromolecules, do their work. These empirical data are a core of knowledge which still lacks satisfactory theoretical explanation. The theory section deals primarily with ideas proposed to explain bilayer hydration and hydration forces, but also mentions the traditional electrostatic, Van der Waals and steric interactions that are covered in sufficient detail elsewhere.

Fig. 1. In the osmotic stress (OS) strategy water activity is controlled through the concentration of a 'neutral' solute, such as a large polymer. Any aqueous compartment that is inaccessible to that solute, or is made inaccessible by a dialysis membrane, has its water activity controlled by solute concentration. Schematically, imagine the exclusion space indicated by the lightly shaded enclosing lines. In the case of single proteins, the solutes that work are those excluded from the protein's 'hydration shell', in the spirit of the extensive studies of Timasheff and his colleagues [17]. Under such osmotic pressure, p,the components of molecular assemblies get pushed together against their mutual repulsive forces. Isolated proteins undergoing reversible transitions get 'dehydrated' or, kinetically, find it more difficult to get to their more hydrated conformation. The extra, osmotic work is observed as a shift toward the dehydrated state. A measure of the sensitivity of that shift to water activity gives the difference in the number of solute-excluding water molecules associated with each conformation [10].

[Full text] (PDF 707 Kbytes)