The metabolic demands of swimming behavior influence the evolution of skeletal muscle fiber design in the brachyuran crab family Portunidae

We investigated the influence of intracellular diffusion on muscle fiber design in several swimming and non-swimming brachyuran crabs. Species with sustained swimming behavior had aerobic dark fibers subdivided into small metabolic functional units, creating short diffusion distances necessary to support the high rates of aerobic ATP turnover associated with endurance activity. This dark fiber design was observed in all swimming species including Ovalipes ocellatus, which has apparently evolved swimming behavior independently of other Portunidae. In addition, we observed fiber and subdivision size-dependent differences in organelle distribution. Mitochondria, which rely on oxygen to function, were uniformly distributed in small fibers/subdivisions, but were clustered at the fiber periphery in larger fibers. The inverse pattern was observed for nuclei, which are not oxygen dependent, but rely on the transport of slow diffusing macromolecules. Phylogeneti­ cally independent contrast analysis revealed that these relationships were largely independent of phylogeny. Our results demonstrate cellular responses to diffusion that were necessary for the evolution of swimming and that are likely to be broadly applicable. Introduction The basal muscles that power swimming in the blue crab, Callinectes sapidus, grow hypertrophically during postmetamorphic development, so that fiber diameters increase from\60 lm in juveniles to[600 lm in adults (Boyle et al. 2003). At these fiber sizes, intracellular diffusion distances begin to limit certain processes critical to normal muscle cell function (Hardy et al. 2006; Kinsey et al. 2005). Potential limitations caused by excessive intracellular diffusion dis­ tances have long been hypothesized as one of the primary reasons that cells are relatively small (typically within an order of magnitude of 10 lm) (Koch 1996; Russell et al. 2000; Teissier 1939; Thompson 1917). Rapid intracellular diffusion of oxygen and small metabolites [e.g., ATP, argi­ nine phosphate (AP) and inorganic phosphate (Pi)] is critical to maintaining high rates of aerobic metabolism (Kim et al. 1998; Mainwood and Rakusan 1982), and the diffusion of larger RNA and polypeptide molecules is important for protein turnover (Fusco et al. 2004; Russell and Dix 1992). Thus, diffusion may provide selective pressure for the evo­ lution of basic cellular design and function, and this role may become greater as intracellular diffusion distances and/or aerobic metabolic rates become greater. Basal swimming muscles are composed primarily of two fiber types: light fibers that power anaerobic burst swim­ ming and dark fibers that power aerobic endurance swim­ ming (Tse et al. 1983). The light fibers rely on maximal rates of aerobic metabolism only during post-contractile recovery, which is associated with low ATP turnover rates, while the dark fibers use aerobic processes to support the high rates of ATP turnover associated with sustained contractile activity. We previously demonstrated that dif­ fusion has a considerable impact on cellular organization in the basal swimming muscles of the blue crab and that these effects were distinctly different between the anaerobic light fibers and aerobic dark fibers as a result of their funda­ mentally different metabolic requirements (Boyle et al. 2003; Hardy et al. 2009; Johnson et al. 2004). During hypertrophic growth, the anaerobic light fibers of C. sapidus appear to maintain cellular function by redis­ tributing certain organelles in a way that minimizes intra­ cellular diffusive path lengths. Mitochondria, for example, are homogenously scattered throughout each fiber in juvenile animals so that there are nearly equal numbers of mitochondria in the fiber interior (intermyofibrillar or IM mitochondria) and at the fiber periphery (subsarcolemmal or SS mitochondria). However, during growth, mitochon­ dria begin to cluster near the sarcolemma, and in adults, virtually no mitochondria occur within the fiber core (Boyle et al. 2003; Hardy et al. 2009). This rearrangement effectively reduces transport distances for oxygen from the blood to the mitochondria, but increases travel distances for small, rapidly diffusing metabolites like ATP and AP. Muscle fibers are multi-nucleated cells, and myonuclear distribution also changes during hypertrophic growth, but in a pattern that is the inverse of that seen for mitochondria. In small juvenile fibers, myonuclei are located exclusively at the fiber periphery adjacent to the sarcolemmal mem­ brane (the characteristic pattern in vertebrate fibers), but during fiber growth, nuclei begin to occupy more centrally located positions within the fiber as well. This shift results in reduced intracellular transport distances for the large, slowly diffusing protein and RNA products required by the fiber for turnover of metabolic and contractile machinery (Hardy et al. 2009). The aerobic dark fibers, on the other hand, have to meet much higher ATP demands during steady-state contraction, and high reaction rates can result in a diffusion limiting environment even in fibers with small diameters. To satisfy the opposing demands for hypertrophic growth and short diffusion path lengths, the dark fibers have developed a network of highly perfused, mitochondria-rich subdivisions (Johnson et al. 2004; Tse et al. 1983) that increase in number while maintaining a constant size (*35 lm) during growth (Johnson et al. 2004). In this way, the aerobic fibers preserve an effective metabolic diameter throughout development that is well within the range of cellular dimensions typical of aerobic muscle from other animals. The perfused subdivi­ sions result in greatly reduced diffusion distances and increased oxygen availability to the mitochondria. As such, nuclei and mitochondria do not undergo the ontogenetic shift in organelle distribution observed in anaerobic fibers. In both adult and juvenile animals, nuclei are found exclusively at the subdivision periphery, while mitochondria are found predominantly at the periphery of each subdivision and, at a lower density, between the myofibrils. Reaction–diffusion mathematical models demonstrated that the ontogenetic shifts in both organelle distribution in the light fibers and the subdivision of dark fibers are essential to promote observed rates of aerobic metabolism in C. sapidus muscle (Hardy et al. 2009). Organelle distribution in adult skeletal muscle fibers is a plastic property. Mitochondrial distribution and morphol­ ogy have been shown to vary dramatically in response to factors including temperature (Tyler and Sidell 1984), hypoxia (Hoppeler and Vogt 2001) and exercise (Chilibeck et al. 2002; Howald et al. 1985; Kayar et al. 1986). Like­ wise, nuclei have been reported to realign themselves with newly formed blood vessels in skeletal muscle fibers sub­ ject to chronic stimulation following denervation (Ralston et al. 2006). The processes by which organelles migrate and anchor inside of cells have been studied extensively (Bitoun et al. 2005; Frederick and Shaw 2007; Milner et al. 1996; Ralston et al. 2006; Rube and van der Bliek 2004; Smirnova et al. 1998; Starr 2007; Starr and Han 2002). However, current understanding of the mechanisms that regulate organelle movement is limited, and the signals that result in the relocation of mitochondria or nuclei within an adult (or embryonic) muscle fiber are largely unknown. There are many potential regulatory mechanisms that could dictate the intracellular arrangement of organelles, and these strategies are not necessarily mutually exclusive. The distribution of organelles within a cell may simply be a product of phylogenetic inertia. This term refers to the sta­ bility of a trait that results from the influence of an ancestor on its descendant (for review see Blomberg and Garland 2002). If there is no selective pressure to modify the place­ ment of organelles in a fiber, then an organism will likely share the same distribution as their ancestor. Alternatively, intracellular organelle distribution may be the product of a genetic developmental program (Badrinath and White 2003; van Blerkom 1991). A cell must be able to function over the entire range of sizes it will span in its lifetime. If the terminal cell size or degree of expected hypertrophy is encoded in the genome of an animal, then certain mechanisms may be implemented early in development to prepare each cell for constraints that will surface only after substantial growth has occurred. A third possibility is that the cellular organization of organelles is the direct and immediate product of some prevailing intracellular condition—in particular, diffusion constraints. For example, mitochondrial distribution could be responsive to intracellular oxygen concentrations. During hypertrophic growth, oxygen gradients across the cell may steepen due to increasing fiber size, and mitochondria may shift from areas of low to high oxygen concentration to maintain rates of oxidative phosphorylation sufficient to preserve function in that fiber. The aim of the present study was to investigate cellular organization across a range of fiber sizes spanned by adults of several related, but morphologically and behaviorally divergent, species of Portunid and non-Portunid crabs to examine the relative effects of diffusion and shared com­ mon ancestry on fiber design. The family Portunidae comprises a group of brachyuran crabs well known for their swimming abilities (Fiedler 1930; Judy and Dudley 1970; Spirito 1972). Portunids exhibit a number of char­ acteristic morphological adaptations that have facilitated the evolution of swimming behavior. Most notably, the 5th pereiopods have been modified into flattened, oar-like paddles, and the carapace has been laterally extended and dorsoventrally compressed to increase hydrodynamic effi­ ciency during sideways swimming (Hartnoll 1971). The basal swimming musculature in particularly adept swim­ ming portunids is also generally enlarged and exhibits severe fiber hypertrophy, most likely to fulfill the high power requirements of swimming (Cochran 1935). Within the portunid family, however, there is consid­ erable variation in the extent of these specializations and hence the range of swimming proficiency. In some species, swimming behaviors only occur during brief feeding or escape events, while other species have adopted an entirely pelagic lifestyle (Hartnoll 1971). Carcinus maenus, for example, is particularly interesting because it is one of the only portunids whose 5th pereiopods have retained their original walking leg characteristics. As such, C. maenus is a much weaker swimmer than many of the other more modified portunids. Although this family is popularly referred to as ‘‘swimming crabs,’’ there are representative species from at least 12 other brachyuran families that also exhibit some capacity to swim (reviewed in Hartnoll 1971). Previously, we documented changes in cellular organi­ zation that occurred during ontogenetic muscle fiber hypertrophy in a single Portunid species, C. sapidus (Hardy et al. 2009). From this study, it could not be conclusively determined whether the observed changes in cellular organization were true adaptive responses to changing diffusion constraints associated with swimming behavior. In the current study, we examined the homologous basal muscles from eight different brachyuran crab species (six portunids and two non-portunids)—all in their mature adult form—and measured fiber/subdivision size, as well as mitochondrial and nuclear density and distribution in anaerobic light fibers and aerobic dark fibers. Using 16S rDNA sequences, we generated a phylogeny for these species from which we performed a phylogenetically independent contrast (PIC) analysis (Felsenstein 1985). The PIC analysis determines whether an observed trait is the product of phylogenetic inertia (shared common ancestry), whereas a trait that is found to be independent of phylogeny can be considered an evolutionary adaptation. We used this analysis to discern the relative influence of phylogenetic ancestry and intracellular diffusion limita­ tions on organelle distribution. We hypothesized that all portunids evolved the ability to swim aerobically by sub­ dividing the dark fibers, and that higher aerobic swimming capacity is associated with smaller subdivisions and higher mitochondrial densities. Additionally, we hypothesized that the patterns in cellular design we previously observed during growth in C. sapidus light and dark fibers would be broadly observable across a range of portunid (and non-portunid) species, and that these patterns would be independent of phylogeny. Such an independence from phylogeny would provide further evidence that intracellu­ lar organelle distribution is, in fact, an adaptation to pre­ vailing intracellular diffusion conditions. Materials and methods