Sarcomeres, the fundamental unit of muscle contraction, contain giant polypeptides (>700,000 Da) consisting of multiple immunoglobulin and fibronectin type 3 domains, one or two protein kinase domains, and in some cases, highly elastic regions. C. elegans has 3 such proteins; twitchin (encoded by unc-22), UNC-89 and TTN-1. The function, substrates and mechanism of activation of the kinase domains of these giant proteins are unknown. unc-22 mutants display “twitching” of the animal’s surface, ~1-2 times/s, which originates from the underlying muscle, and also, to varying extents, reduced locomotion and disorganized sarcomeres. We wondered how the kinase domain of twitchin contributes to this phenotype. We converted a highly conserved lysine (K) involved in ATP binding to alanine (A) and showed that this abolishes the kinase activity of recombinant twitchin kinase in vitro towards an artificial substrate. The behavior of the KtoA mutated protein is very similar to that of the wild type protein; crystals have been grown and we plan to obtain a structure and compare it to the published wildtype structure. We used CRISPR/Cas9 to introduce the same mutation in the endogenous unc-22 gene, resulting in expression of a kinase-dead twitchin. This mutant does not twitch, and shows normal sarcomeric structure by immunostaining using a battery of antibodies to sarcomeric proteins. Remarkably, however, a thrashing assay in liquid reveals that unc-22(KtoA) moves ~25% faster than wild type. While crawling on an agar surface unc-22(KtoA) has a ~60% greater velocity than wildtype. Using an optogenetic method, the KtoA mutant shows an abnormal pattern with more overall contraction, and faster rates of contraction and of relaxation. Thus, twitchin kinase activity is required to inhibit the contraction/relaxation cycle. The closest homolog of twitchin in vertebrates is titin, but titin’s kinase domain is an inactive pseudokinase. Twitchin kinase from Aplysia is also active, as is the kinase domain from the twitchin homolog in Drosophila called projectin. We wondered if there was a selective advantage in maintaining kinase activity. To test this hypothesis in C. elegans, we conducted a competitive fitness assay; competing either wild type individuals or unc-22(KtoA) mutants against a GFP-marked tester strain. After 4 generations, the wildtype strain exhibited significantly greater fitness than the unc-22(KtoA) strain. Specifically the wildtype strain increased in frequency ~17% relative to the tester strain, while the unc-22(KtoA) strain decreased ~20% in frequency over 4 generations. Thus, there is a substantial selective advantage in maintaining the kinase activity of twitchin.