Myosin Regulation: It helps to be lucky

Sharyn has asked me to give a personal account of the history of some of my previous studies that may be of general interest. I have chosen the discovery of myosin-linked regulation for two reasons. The discovery was quite serendipitous and showed me how important luck is in research. In addition the story is still open-ended and much remains to be understood about how the system operates. Parts of this account were published in the Proceedings of the Yamada Conference XXXIX on Calcium as Cell Signal, April 26-28, 1994 Tokyo, Japan, held in honor of Setsuro Ebashi and Makoto Endo.

The story begins with the arrival of John Kendrick-Jones in 1967 as a postdoctoral fellow to my laboratory at Brandeis. I suggested to him that he should try to isolate native thin filaments. We wanted to provide direct evidence that all the components of the regulatory system, discovered two years previously by Ebashi, were indeed associated with actin in muscle. We had decided to use molluscan adductor muscles because of their long thin filaments and huge thick filaments. This was also advantageous since we were collaborating at the time with Carolyn Cohen on paramyosin assembly. Our first approach, using sucrose density centrifugation, was a failure since both thick and thin filaments sedimented at comparable rates. Then I recalled that myosin filaments form aggregates at low ionic strength, which is the significant step in the isolation of muscle myosin. Indeed, in the absence of calcium and the presence of ATP, the thick filaments sedimented quickly at low centrifugal forces, while the thin filaments, which were detached from dense-body-like structures, remained in the supernatant.

We tested the properties of the thin filaments by hybridization with rabbit myosin, and found that they readily formed actomyosin and stimulated myosin ATPase activity. At this point, however, we had a great surprise: the ATPase activity was independent of calcium. We wondered if we had somehow lost the components we were looking for, but that would be an unlikely and unpleasant explanation, considering our efforts to avoid high ionic strengths and organic solvents to obtain native thin filaments. Then we theorized that calcium regulation in molluscan muscles might differ from those of the vertebrates. Fortunately this more interesting possibility turned out to be correct. The rest was easy. We soon found that, unlike skeletal myosins, molluscan myosins contained a specific calcium binding site which had to be saturated for its function. Molluscan myosins turned out to be regulated molecules. It did not take much imagination to conclude that the light chains functioned as regulatory subunits and that molluscan muscles were regulated by the direct binding of calcium to myosin. To prove this would require some doing, but the simplicity of the system was so appealing that I have concentrated my scientific interest on finding out how this regulation works.

The first task was to prove that the myosin light chains did,indeed, function as regulatory subunits. I chose scallop myosin for these studies because I found that on SDS gels there was only a single low molecular weight band in scallop myosin. That may not have been the best reason, but for an oversimplifier like me, it was obvious that this was the myosin to study. Soon using urea gel-electrophoresis, we established that there were two kinds of light chains, which we called "regulatory" and "essential" light chains. I was then able to show that removal of divalent cations, by EDTA, released regulatory light chains resulting in losses in calcium binding and in the calcium requirement of ATPase activity. These functions were restored upon the rebinding of the light chains in the presence of excess magnesium ions.

At this time Jake Kendrick-Jones returned for a visit from England and Eva Szentkirályi joined these studies. So far, we had shown that the role of the regulatory light chains was to inhibit activity. We did not know where the triggering calcium binding sites were located since the isolated subunits did not bind calcium. We had also observed that subfragment-1 of scallop, in contrast to heavy meromyosin, was not regulated and was fully active in the absence of calcium. I may note here that these experiments could only be performed on scallop myosin, since, as we subsequently showed with William Lehman, regulatory light chains could not be removed in significant degree from other molluscan myosins by EDTA. Later we learned that the reason for this was that scallop regulatory light chains lacked a glutamate and could not form a salt-link with the heavy chains, as regulatory light chains of other myosins did. EDTA treatment removed only one of the two regulatory light chains, even from scallop myosin. The complete dissociation of all of the regulatory light chains was a precondition of later experiments.

The solution of this problem was also a result of a lucky observation. In 1977, with Robert Simmons, we decided to see if regulatory light chains could also be reversibly removed from skinned fiber bundles in order to show their role in tension generation. So Bob decided to spend a fortnight following a Gordon Conference in the Marine Biological Laboratory in Woods Hole where I used to spend my summers. Bob brought along a student strain gauge, and, since we did not have a cooling system we treated the fibers at room temperature with EDTA. We had found that tension development was, indeed, regulated by the regulatory light chain. To our amazement, however,we also found that the removal of the regulatory light chains from the fiber bundles was considerably more complete than from myosin which had been treated with EDTA at 0°C which is, of course, where respectable biochemists work to prevent possible denaturation of proteins. We later learned that the simple explanation for this complete removal lay in the the hydrophobic bonding of the light chains to the heavy chains. This experiment taught us a simple method for the complete removal of regulatory light chains and opened the way for further experimentation.

Naturally, the postdocs and graduate students working with me clarified several aspects of how the system works. Theo Walliman, with the aid of specific antibodies, obtained the first evidence that the essential light chain is also part of the regulatory apparatus. The work of James Sellers, Peter Chantler and Hyockman Kwon has established that the sources of the regulatory light chains had to be regulated myosins, while only molluscan essential light chains were functional. Betsy Goodwin cloned the light chains and studied mutants of the regulatory light chain. László Nyitray sequenced the heavy chain of scallop myosin, work absolutely needed for structure determinations even before knowing the suitability of scallop myosin for such studies. Sylvia Fromherz found that the triggering calcium binding site was not the characteristic EF hand of domain-3 of the essential light chain, rather it was domain-1 that contained a sequence of 5 amino acids uniquely conserved in molluscan essential light chains. Vassilios Kalabokis' cooperativity studies have demonstrated that communication between the nucleotide and the calcium-binding sites requires interaction between the two myosin heads.

An unexpected fringe benefit of scallop myosin was the propensity of its proteolytic fragments to form crystals suitable for high resolution structures as was shown by Carolyn Cohen and her colleagues. This aspect of the research started with the work of Eva Szentkirályi who, in 1984, isolated the light-chain-bearing portion of S1, called the "regulatory domain", nearly 10 years before the development of the lever arm concept. The regulatory domain was crystallized and its structure solved by Xialene Xie, Carolyn's graduate student, whose work showed that all the liganding sites of calcium were contributed by domain-1 of the essential light chain and established how the regulatory light chain stabilized the unusual calcium-binding loop. This work offers a simple explanation as to why all three peptides, the regulatory and the essential light chains, together with the heavy chain, were required for calcium binding. The absence of a key residue, glycine-117, from vertebrate skeletal myosin RLC explained why it was unable to restore calcium binding and calcium sensitivity. In fact, Ágnes Jancsó succeeded in obtaining a "gain of function mutation" by introducing a single glycine residue into the corresponding position of chicken skeletal regulatory light chain. More recently, Anne Houdusse, in Carolyn's laboratory, determined the structure of a new state on an intact, unmodified scallop S1 and was able to define the position of the lever arm in three different states of unattached scallop S1.

The principal message of this story is: "Treasure your unexpected results!" Much of our work was based on surprises. It may be disappointing not to obtain answers to the questions one asks, but that disappointment may force a reexamination of one's preconcieved notions. Initial dismay usually turns into delight because of the new paths one has luckily stumbled upon.

Contributed by Andrew G. Szent-Györgyi

Reference
Szent-Györgyi, A.G., Fromherz, S., Jancsó, A., Nyitray, L. and Kalabokis, V.N. (1995). Regulation of Muscle Contraction by a Calcium-Binding Myosin: Structural and Mutational Studies. In "Calcium as Cell Signal" Proceedings of the Yamada Conference XXXIX, April 26-28, 1994. Maruyama, K., Nonomura, Y. and Kohama, K. editors. Igaku-Shoin, Tokyo-New York.

View links to recent papers on myosin structure and function from the PUBMED database

Return to the Myosin Function Page
Return to the Myosin Home Page

If you use material from the reviews at The Myosin Home Page, we ask that you cite either the Home Page and authors, or the appropriate source publication in your work.

Copyright 2000 All rights reserved