Micelles and Membranes

From this point onwards, most of my work was done jointly with Jacqueline Reynolds, my right-hand associate in the laboratory, who quickly became my beloved partner in life as well.

This kind of partnership has become much more common since my generation, but even now partners of this kind, though perhaps sharing the same employer, tend to retain separate identities and separate research interests. In our case, the research objectives were to a large extent shared and this creates obvious problems in the assignment of professional credit. In the text from here on, I shall use the pronouns "we" and "our" in referring to the scientific work published from our laboratory, regardless of whether one alone or both of us appear as authors. There were, of course, usually graduate students or postdoctoral fellows involved in any work that we did, as well as we, the nominal project directors. Moreover, they have all told me (in later years) that Jacqueline was a far more helpful research "advisor" than I had been. (Marc le Maire and Darrell McCaslin, for example, both superb graduate students, ostensibly working under my sole direction as thesis advisor.)

In all of our work on molecular weights of membrane proteins, in particular, I know that my own contribution can only have been peripheral. For example, our innovative development of a method for using the ultracentrifuge to determine the molecular weight of the protein moiety in protein-detergent or pro-tein-lipid complexes without actual knowledge of bound lipid or detergent is in this category [30]. Both the idea itself, which was an extension of the buoyant density principle, using D2O to adjust the solvent density so as to match the density of bound lipid or detergent - as well as the practical execution - were primarily Jacqueline's work.

An example I remember well is our collaborative efforts with Arthur Karlin on the acetylcholine receptor [31], where my job in the partnership was to make frequent trips to the airport (I had a fast car) to collect the samples that had been freshly isolated earlier that day in Karlin's laboratory in New York, and were shipped to us via Eastern Airlines.

Our transition from water-soluble proteins to membrane proteins may not seem to be a very profound change in direction today, but seemed to be at that time. It involved a much broader vista, which incorporated lipids (micelles, membranes) as well as proteins. Phil Handler, when he heard about it, thought I was out of my mind, sacrificing an assured status as a leading physical protein chemist by venturing into this new field with dubious prospects. In a sense Handler's objections justified the change in direction, in that they illustrated the newness of the project, the magnitude of the leap into the unknown. Ventures that are dull and have predictable results do not carry much risk.

To prepare myself for the change, sensing (correctly) that my physico-chemical background would be inadequate, I spent much of the early time in exploring old ground, to organize in my mind what surface chemists and surfactant chemists already knew: the thermodynamic extension of hydrophobicity to micelle formation, size, and shape of micelles, etc. This endeavor resulted in a new book, The Hydrophobic Effect [32]. Just as in the 1950s, when I ventured beyond the teachings of Cohn and Edsall and felt the need to write Physical Chemistry of Macromolecules, so now, again venturing beyond what I already knew, another textbook seemed the surest way to go. (On some later occasion, Walter Kauzmann and I argued about the philosophy of textbook writing. Kauzmann believed in the general rule that you only write a textbook after much experience, when you have become expert enough to be able to vouch personally for the truth of every word. My viewpoint was the opposite, that you use textbook writing to learn a subject, to express a consensus rather than a personal view.)

In the laboratory, the catalyst for getting the new venture going was again the connection to the medical school. This time it was a medical student, John Gwynne, who had opted to spend his third year on a project in basic science and had dropped into my office, to see what might be on offer in our laboratory. Nothing we were doing was likely to be of immediate medical interest, I told him, but I did have a menial job for someone with a brain, which might eventually blossom into some research - and I told him about our desire to expand our work from focussing exclusively on proteins in aqueous solution to proteins which were in their native state in cell membranes and that it was an ambitious step for us for we knew nothing, whatever, about cells and membranes.

Nor could I identify who among my colleagues might profitably be consulted on the subject: many of them could probably qualify as "cell biologists'' of a sort, but surely the practical details of getting cells or membranes from living tissue into laboratory apparatus would be different for each of them: muscles, nerves, the eye, the blood stream, bacteria would all be expected to demand different techniques and specialized apparatus; and what you meant to do with the cells when you got them could make a difference, too - would an electron micro-scopist have different criteria from a physiologist studying transport? Red blood cells seemed particularly attractive to me (in my ignorance), being both abundant and easy to collect. Was this really true? And who in the diverse population of Duke Medical Center would have the answers?

I told John Gwynne that if he could identify appropriate members of the medical faculty and find out how membranes could be prepared free from water-soluble cell contents, we could think of chemical approaches to begin to characterize the membrane proteins. I must admit that I expected a medical student to prefer more precisely defined projects, that did not begin in this way in uncharted waters. I was, therefore, surprised when Gwynne showed up again in a couple of weeks, mission accomplished. Red blood cells were indeed the most promising starting point, and Gwynne was ready to tell us just how red cell "ghosts" (as the membrane preparations were called) were to be made.

We had during the preceding years established that most proteins, when dissolved in the potent denaturing agent guanidinium chloride and reduced to their constituent polypeptide chains by disulfide bond disruption, behave physically as structureless "random coils.'' The constituent amino acids remain covalently linked by peptide bonds, but all effects of hydrogen bonds and other noncovalent structure-forming factors are lost: the physical behavior of the molecule becomes essentially identical to that of a synthetic organic polymer dissolved in an indifferent organic solvent. This means that physical parameters that depend on molecular weight become regular functions of molecular weight. One consequence was that it enabled us to adapt gel exclusion chromatography to molecular weight determination - strictly speaking, molecular length is the effective parameter in chromatographic sorting, but in most applications that will be proportional to molecular weight. The result was a method whereby polypeptides of a mixture could be separated on a column and the molecular weight of all separated fractions could be measured simultaneously [33] - a huge gain in convenience over the analytical ultracentrifuge.

There was good evidence that guanidine hydrochloride would also disrupt noncovalent association between lipids and proteins, i.e., the method should be applicable to proteins from biological membranes (not ordinarily soluble in aqueous media) as well as to those that originate from a cell's cytoplasm. Once John Gwynne could give us an accurate description of how to prepare red cell ghosts, separated from any other protein-containing matter, we had everything ready to go for molecular weight analysis. The most significant aspect of the result that emerged was that a major component of the mixture had an astonishingly high molecular weight, higher than 200,000 [34]. Apart from the heavy chain of myosin, no polypeptide of greater length was known.

About this time, the detergent sodium dodecyl sulphate (SDS) surpassed guanidine hydrochloride as everyone's favorite reagent for membrane disruption cum polypeptide chain segregation by molecular weight - using gel electrophoresis for the latter in place of the simpler gel filtration [35]. We knew that polypeptides in SDS, though thoroughly denatured, did not at all the resemble structureless random coils that they were in guanidine. The physical basis for sorting by molecular weight was, therefore, not immediately obvious, but we quickly worked it out [36], showing in the process that electrophoretic mobility on SDS gels was not as rigorously linked to molecular weights as was the case for gel filtration in guanidine, i.e., there were more likely to be discrepancies when molecular weights derived from SDS gels were compared precisely with values ultimately obtained by ultracentrifugation. (Molecular weights from SDS gels are in fact commonly called "apparent molecular weights.'')

We repeated the analysis of red blood cell membrane proteins in the new system and got similar results, but in SDS we clearly saw two distinct polypeptides around 200,000 molecular weight, which had merged into one broad band in gel filtration [37]. A year later we discovered an even larger new polypeptide, again lipid associated, though not in a membrane: this was the polypeptide constituent of low density lipoprotein (LDL) with a molecular weight of 250,000 [38]. Strangely, this proved to be controversial: the self-appointed authorities on molecular weight in the lipoprotein establishment (often referred to as the "Sardinian mafia'') refused to recognize our measurement, though we repeated and extended it: they were advocating a vastly smaller figure, around 20,000. This controversy was one of very few truly unpleasant events in my experience; in retrospect probably best forgotten. Later genetic analysis showed that LDL can be separated into several different fractions - all of them have principal polypeptide chains with molecular weights above 250,000.

In the process of this work, we became reasonably proficient in handling many practical aspects of amphiphile interactions and in understanding much of the underlying theory, which is of course governed by the quantitative canons of thermodynamics. Detergent micelles and phospholipid vesicles gradually became part of our everyday vocabulary; most of the proteins we handled needed their presence to remain in solution.

Was this article helpful?

0 0

Post a comment