Understanding muscle from its length

Most physiologists will have measured the tension–length relation of a muscle at some early point in their career and speculated on the mechanisms which cause tension to first increase and subsequently decrease as the muscle is progressively stretched. In the early part of the 20th century it was widely thought that the muscle proteins, actin and myosin, formed continuous filaments across the sarcomere and that contraction involved a uniform decrease in filament length. All that changed with the sliding filament theory of muscle, developed independently by H. E. Huxley & Hanson (1954) and by A. F. Huxley & Niedergerke (1954) and published back-to-back in Nature. The sliding filament theory led to two predictions: (i) that tension was produced only at the region of overlap and that there should be a predictable relation between sarcomere length and the overlap between thick and thin filaments, and (ii) that unloaded shortening velocity should be independent of overlap (Edman, 1979; Reggiani, 2007).

Testing the relation between tension and overlap required accurate measurements of the lengths of the thick and thin filaments and detailed measurements of the tension as a function of sarcomere length. In the 50s and 60s electron micrographs improved dramatically and provided increasingly accurate measurements of the length of the thick and thin filaments. However, measuring muscle tension as a function of sarcomere length proved to be a very difficult problem which required a number of technical developments. First, it was necessary to work with single fibres that had been cleaned of most connective tissue so that passive tension was minimized and sarcomeres could be clearly visualized. This technique had been pioneered by Ramsey & Street (1940) and was already in use in Andrew Huxley's laboratory at University College in London. Huxley & Peachey (1961) had by then shown that in single fibres the sarcomeres were not uniform along a fibre and that sarcomeres were generally shorter at the ends of fibres than in the middle. Consequently, the tension developed by the whole muscle fibre at very long lengths was dominated by the shorter sarcomeres near the ends of the fibre. This intractable problem was solved by the development of the spot-follower sarcomere clamp. This approach, conceptually a development of the squid axon voltage clamp, utilized markers attached to the single fibre in a region where the sarcomeres were particularly uniform. In this way the tension was effectively measured in a central region with almost uniform sarcomeres.

The result was a triumph and is encapsulated in their brilliant figure reproduced below (Fig. 1) (Gordon et al. 1966). It shows the linear decline of tension on the descending limb reaching zero at 3.65 μm which is, within experimental error, the length at which overlap is zero. The plateau and the shoulder at 1.67 μm are all simply explained by the interactions of the thick and thin filaments. This must surely be the most reproduced figure in muscle physiology and a quick survey shows that a version of it is still found in most textbooks of physiology.

Figure 1.

The length–tension relation of frog skeletal muscle A, critical stages in the overlap of thick and thin filaments. B, schematic summary of results. Both panels from Gordon et al. (1966).

The paper remains a delight to read. There are no P values quoted or necessary. It seems clear that the authors built the equipment themselves starting with an up-ended oscilloscope whose spot formed the basis for detecting the position of the markers. It is full of wonderfully arcane technical details (e.g. the lever used to stretch the muscle fibre was made of dried grass stems coated with araldite). There is meticulous attention to the work of earlier studies and detailed discussion of possible errors in measurements and interpretation. Although the classic figure does not show the data points, other figures in the paper show that they cluster around the lines of best fit with almost magnetic precision. The paper has been cited around 1300 times and, interestingly, the citation rate grew slowly in the first 5 years reaching a near-steady level which has subsequently slowly increased over 35 years. The paper must be one of the best contributors to the long half-life of citations for The Journal of Physiology. On the other hand, the paper's citations in the first 2 years, used to calculate the Impact Factor of the journal, would be quite modest demonstrating how classic work can be underestimated by the Impact Factor approach.

The paper provided powerful support for the sliding filament theory because frog muscle could be fully activated and, under these conditions, tension depended only on the geometry of the sliding filaments. Interestingly when skeletal muscle is submaximally activated, as commonly happens under physiological conditions, the shape is different and poorly defined activation-dependent factors come into play (Rack & Westbury, 1969). In cardiac and smooth muscle the issues have proved more complex and are still not fully resolved. An interesting feature is that because muscle has near-constant volume, the distance between thick and thin filaments changes from approximately 9 nm at short lengths to 4 nm at long. It has been proposed in cardiac muscle that this change in the distance over which the cross-bridges have to act is partly the cause of the change in Ca²⁺ sensitivity which is the major contributor to the length dependence of tension in the heart (Fuchs & Martyn, 2005). So, as with many classic studies, the investigators wisely chose a preparation whose performance could be simplified to allow them to uncover a fundamental mechanism.