ABSTRACT
Myoflbrils, the contractile organelles of skeletal muscle, are highly ordered and precisely regulated actomyosin networks. Investigations of myofibril assembly are revealing the cellular mechanisms by which contractile components are arranged and regulated. In order to facilitate this research we have developed formal molecular genetics for myofibrillar proteins of Drosophila flight muscle. Presently, mutations can be used systematically to perturb or eliminate any of the classical myofibrillar proteins within these fibers, and the in vivo consequences can be conveniently evaluated using protein electrophoresis, electron microscopy, or by assaying flight performance. Here we review some recent progress.
INTRODUCTION
Our work aims to further elucidate the assembly and functioning of myofibrils, the contractile organelles of skeletal muscle, and to extend this knowledge to actomyosin networks of the cellular cytoskeleton. Our approach combines molecular genetics with cell biology and physiology. We use mutations to eliminate or perturb particular contractile and cytoskeletal components in the fruit fly, Drosophila melanogaster, then observe the effects on muscle structure and function. The genetic approach allows us to manipulate muscle proteins in vivo, and thus to bridge the gulf that continues to exist between biochemical studies of purified proteins and ultrastructural, biophysical, and immunohistochemical studies of cells, tissues, and organisms. Drosophila, one of the most completely understood metazoans, is an excellent subject for such work because its complex physiology and simple genome are amenable both to genetic and cell biological approaches (refer to Fyrberg and Goldstein, 1990).
Of several types of adult fibers, Drosophila indirect flight muscles are best suited for experimental work. Within the laboratory these muscles are expendable, hence mutations that perturb their structure and function are in many instances not lethal, merely causing the loss of flight. This factor considerably simplifies mutant analyses. Also noteworthy is the fact that myofibrils of these fibers are highly ordered, and thus facilitate discernment of the structural perturbations caused by particular mutations using transmission electron microscopy. Finally, although much smaller than vertebrate muscles, the fibers are large enough to provide material.
Before describing our strategies and observations we should clarify how it is possible to perturb or eliminate contractile proteins of flight muscles without affecting those of essential muscle groups. The most straight-forward approach is to mutate a contractile protein gene or exon that is expressed only within flight muscles. However, one is by no means limited to that strategy, as it is clear that particular missense or regulatory mutations which impede contractile protein accumulation and function are in many instances manifested chiefly in flight muscles (refer to Fyrberg and Beall, 1990, for further discussion).
SUMMING THE PARTS: DISSECTING MYOFIBRIL ASSEMBLY USING NULL ALLELES
Perhaps the most direct method for elaborating the flow of information during myofibril assembly is to eliminate particular components and assess how much of the remaining sarcomeric structure is able to form. These experiments are made possible by a growing number of gene mutations that eliminate the synthesis or accumulation of major myofibrillar components (refer to Fyrberg and Beall, 1990; Fyrberg et al. 1990a). For example, ActBBF1™88, an actin gene allele wherein the codon for tryptophan 79 (TGG) is converted to an opal terminator (TGA), eliminates all actin and thin filaments in flight muscles (Okamoto et al. 1986; Beall et al. 1989). In the absence of actin no thin filaments or Z-discs form, but reasonably well ordered arrays of thick filaments and M-lines nevertheless assemble. In the converse experiment we have used a flight muscle-specific myosin heavy chain allele, Ifm(2)2 to eliminate thick filaments. In this case arrays of thin filaments and Z-lines form. These two experiments demonstrate that thick filament/M-line and thin filamentμ-disc arrays assemble independently, that thick filaments are probably not integral Z-disc components, and further suggest that M-lines and Z-discs are organizing centers for myofibril formation. Analyses of actin and myosin heavy chain, null allele, heterozygotes (flies having one normal allele and one null allele) have demonstrated the importance of filament stoichiometry for normal myofibrillar assembly (Beall et al. 1989).
Analyses of muscles having tropomyosin, troponin-T, and alpha-actinin null alleles are similarly defining the roles of the corresponding proteins in myofibril assembly and maintenance. In the absence of troponin-T or tropomyosin, thin filaments are largely absent from sarcomeres (Fyrberg et al. 19906; our unpublished observations), demonstrating the importance of these proteins in the stabilization of F-actin. Absence of the principal actin filament crosslinking protein, alpha-actinin, leads to progressive muscle degeneration and paralysis in the Drosophila larva, consistent with the notion that these crosslinks are essential for fixing thin filaments in place during contraction, a role similar to that proposed for spectrin in the membrane-associated cytoskeleton.
THE CRUX OF MYOFIBRIL ASSEMBLY: ALIGNING AND INTEGRATING THICK AND THIN FILAMENTS
A central question of myofibril assembly is how the thick and thin filament networks are aligned and integrated. Several hypotheses have been invoked to explain this facet of the assembly process, and genetic experiments have begun to test critically some of these ideas. One notion, based upon electron microscopy of differentiating muscles, is that the initial alignment and integration of thick and thin filaments is mediated in some as yet undetermined fashion by transient arrays of microtubules. Microtubule bundles or arrays can typically be seen in differentiating muscle (Fischman, 1967). Our recent survey of Drosophila myofibrillogenesis (done in collaboration with Mary Reedy of Duke University Medical Center) has revealed that in developing insect muscle microtubules are found closely apposed to, or surrounding, nascent myofibrils (M. C. Reedy, personal communication; refer also to work of Auber and Couteaux, 1963). Antin et al. (1981) previously showed that microtubules can integrate into the myofibrillar lattice in taxol-treated, postmitotic, chicken myoblasts. However, since no specific interaction between a microtubule-associated protein and a myofibrillar component has been demonstrated, a causal relationship between microtubules and myofibrillogenesis has yet to be established. The recent demonstration by Leiss et al. (1988) and Kimble et al. (1989) that a particular isoform of Drosophila tubulin (j83) is transiently expressed in all developing muscle cells has rekindled interest in the issue, and provided a system in which to investigate rigorously the role of transient microtubule arrays in myofibrillogenesis. Preliminary results from Raff and collaborators suggest that certain mutations in this tubulin gene reduce muscle mass and perturb contractile function. Examination of these muscles in the electron microscope should establish whether the corresponding microtubules serve as a nonspecific scaffold or as a specific template for sarcomere formation.
A second hypothesis is that the long elastic protein referred to as titin connects M-lines with Z-discs, and thus facilitates integration of thick and thin filament arrays, also possibly serving as a molecular ruler for establishing thick filament length (Whiting et al. 1989). A related molecule named twitchin has been characterized in nematodes, and the effects of a number of mutant alleles on muscle formation have been evaluated in this organism. On the basis of these results, Benian et al. (1989) concluded that twitchin is involved in regulating the contraction-relaxation cycle and has little, if any, role in myofibril assembly. In order to use genetics to evaluate titin/twitchin function in a second system we have isolated and partially characterized the Drosophila analogue. Using antibodies against insect titin, Belinda Bullard and collaborators (EMBL, Heidelberg) isolated the corresponding cDNA of the giant waterbug, Lethocer- us, from expression libraries (Lakey et al. 1990). By hybridizing this cDNA to Drosophila genomic and cDNA libraries at low stringency we isolated the corresponding genes. Partial sequencing of several cDNAs has confirmed the identity of the Drosophila gene. In situ hybridization of the gene to polytene chromosomes has revealed that the gene is within subdivision 102 of the fourth chromosome, the site of a number of interesting candidate mutations. Presently, we are analyzing a variety of these putative mutants in order to establish which lack titin, and how that deficit affects Drosophila myofibril formation.
DISRUPTING CONTRACTILITY: ANALYSES OF POINT MUTATIONS THAT PRECLUDE NORMAL MYOFIBRILLAR FUNCTION
It is possible, in Drosophila, to mutate any residue of a contractile protein, then to test functioning of the mutant derivatives by introducing them back into the corresponding null mutant. Alternatively, one can characterize a variety of randomly induced point mutations within a contractile protein gene, correlating the molecular changes to the most interesting and informative phenotypes engendered. We are conducting both approaches successfully. Our purposes in these studies are to define better how actin and myosin generate force, and to delineate how force production is regulated by various regulatory proteins.
To understand force generation we are focusing our mutagenesis efforts on actin. Of the several actin mutations analyzed in our laboratory, probably the most informative involved conversion of glycine six and alanine seven to alanine and threonine, respectively. Thick and thin filaments within the perpipheral regions of myofibrils containing this actin are out of register, and one occasionally finds thick filaments flanked by actin filaments having opposite polarities. Electron microscopy of these arrays revealed that crossbridges can form with both of the thin filaments, revealing the extreme torsional flexibility resident within the myosin head or head-tail junction (Reedy et al. 1989). Presently, we are manufacturing a number of additional actin mutations in order to test various structure/function relationships, a project that will benefit considerably from the recent solution of the structure of actin at atomic resolution (Kabsch et al. 1990). One obvious experiment involves evaluating the function of the single methylhistidine residue of actin. We are presently converting it to both arginine and tyrosine. We are also using alanine scanning mutagenesis (Cunningham and Wells, 1989) to perturb regions of actin known to interact with myosin, the hope being that functional testing of such derivatives will ultimately define the nature of the conformational change proposed to accompany the development of tension.
ACKNOWLEDGEMENTS
In order to understand better how the contractionrelaxation cycle is regulated, we have developed molecular genetics for tropomyosin and the three subunits of troponin. The most interesting observation we have made to date is that certain missense mutations in troponin-T and -I assemble into normal myofibrils but subsequently degenerate. These abnormalities appear to be due to aberrant interactions of actin and myosin, as elimination of myosin using a heavy chain null allele prevents the degeneration (Fyrberg et al. 1990a). The most straightforward interpretation of these results is that the mutant troponin subunits foster abnormal interactions of actin and myosin that destabilize the myofibrillar lattices. More detailed functional observations should reveal the precise cause of the myofibrillar degeneration and may ultimately elucidate mechanisms driving the regulated crossbridge cycle.