The ontogeny of muscle structure and locomotory function in the long-finned squid Doryteuthis pealeii

Muscle fibers can be modified to produce a broad continuum of contractile properties. These modifications, in conjunction with changes in other parts of the motor system, have resulted in muscles that can generate force economically (Twarog, 1967), produce relatively high force over an impressive range of lengths (Hoyle et al., 1965; Lanzavecchia, 1977; Herrel et al., 2002; Woods et al., 2008), shorten with great speed (Rome et al., 1996; Schaeffer et al., 1996; Elemans et al., 2004), act as powerful motors (Lutz and Rome, 1994; Marsh and Olson, 1994), as brakes or shock absorbers (Full et al., 1998), as struts (Roberts et al., 1997; Biewener et al., 1998) or as springs (Tu and Dickinson, 1994). A recognized, but understudied, phenomenon is that muscle fibers can be modified extensively during ontogeny (e.g. Greer-Walker, 1970; Goldspink and Ward, 1979; Anapol and Herring, 1989; Gilly et al., 1991; Kier, 1996; Thompson and Kier, 2006; Etnier et al., 2008). The ultimate causes of ontogenetic muscle modifications are not always apparent from a functional or ecological perspective but they may represent developmental constraints, selection for different levels of performance during ontogeny or both. Understanding the extent to which changes in muscle form and function underlie ontogenetic changes in locomotory behaviors and performance is important to understanding the evolution of musculoskeletal systems and also the ecology of different life stages.

We explored the ontogeny of muscle structure, the expression of isoforms of myosin heavy chain (MHC) and the function of the mantle in squid, with respect to how changes in these parameters affect jetting behavior and performance. Unlike vertebrates and some arthropods, in which striated muscle fibers may be altered during ontogeny via expression of different isoforms of MHC (Bárány, 1967; Goldspink, 1968; Goldspink, 1983; Gauthier et al., 1978; Bandman, 1985; Gondret et al., 1996), changes in the ratio of the alkali light chains (Sweeney et al., 1988) or changes in calcium sensitivity via expression of isoforms of troponinT (Fitzhugh and Marden, 1997), squids appear to modify muscle mechanical properties by altering the lengths of the thick filaments. For example, the thick filaments of the obliquely striated transverse muscles of the arms of the oval squid Sepioteuthis lessoniana increase in length during ontogeny, from 2.2 μm in hatchlings to 6.4μm in adults, while the thick filament lengths of the cross-striated transverse muscles of the tentacles decrease from 2.4 μm to 1.2 μm during the same interval (Kier, 1996). The ontogenetic change in thick filament length of the tentacle muscle is clearly related to changing function (see Kier, 1982) of the tentacles but the functional basis for the arm muscle change is less clear (Kier, 1996).

Production of jet thrust in squids requires the integration of internal (i.e. the components of the motor system, including connective tissues) and external (i.e. jet dynamics) factors. In jetting, contraction of the circular muscles of the mantle pressurizes the mantle cavity and drives water out of the mantle cavity via the funnel (Young, 1938). Mantle contraction rate, the amplitude of mantle cavity pressure, the mass flux of the jet (i.e. the mass of water ejected from the funnel per unit time) and jet velocity are determined largely by the contractile properties of the circular muscles of the mantle, although the muscles of the funnel can adjust the aperture and, therefore, also affect jet velocity. Thus, the contractile properties of the circular muscles directly affect the hydrodynamics of the jet wake, jet thrust and jet propulsive efficiency. Both the rate of mantle contraction and the mass flux of the escape jet, however, change in some squid during growth (Thompson and Kier, 2001a; Thompson and Kier, 2002), and evidence suggests that the contractile properties of the circular muscles may change during ontogeny (Thompson and Kier, 2006).

Two important changes in the circular muscles of the mantle occur during the ontogeny of squids. The first is the relative abundance of the two types of circular muscle fibers. Many squid species possess two types of circular muscle cells: centrally located, mitochondria-poor (CMP) fibers and superficially located, mitochondria-rich (SMR) fibers (Bone et al., 1981; Mommsen et al., 1981) [terminology from Preuss et al. (Preuss et al., 1997)]. Both fiber types are obliquely striated, have a core of mitochondria, a single nucleus and are electrically coupled to adjacent circular muscle fibers, presumably by gap junctions (Marceau, 1905; Young, 1938; Hanson and Lowy, 1957; Kawaguti and Ikemoto, 1957; Millman, 1967; Ward and Wainwright, 1972; Moon and Hulbert, 1975; Bone et al., 1995; Milligan et al., 1997). The SMR circular muscle fibers are hypothesized to provide power for ventilation of the mantle cavity and prolonged, slow-speed jetting (Bone et al., 1981; Mommsen et al., 1981; Bartol, 2001). Conversely, the CMP circular muscle fibers are hypothesized to provide power for high velocity jets, including escape jets (Bone et al., 1981; Mommsen et al., 1981; Gosline et al., 1983; Bartol, 2001). The ratio of SMR:CMP fibers decreases from about 1:1 to 1:10 in the mantle of S. lessoniana and Doryteuthis (formerly Loligo) opalescens during growth from a tiny paralarva to a large adult (Preuss et al., 1997; Thompson and Kier, 2001b). The second important change is that the thick filaments of both the SMR and CMP circular muscles increase 1.5-fold in length, at least during the ontogeny of S. lessoniana (Thompson and Kier, 2006).

The ontogenetic increase in thick filament length may affect the contractile properties of the circular muscles. The shortening velocity of striated muscles depends on the lengths of the thick filaments and sarcomeres, the load on the muscle and the rate of cross-bridge cycling (e.g. Josephson, 1975). Thick filament length is inversely proportional to unloaded shortening velocity and is directly proportional to peak isometric force (e.g. Josephson, 1975). Assuming all else about the fibers is equal, we hypothesize that the circular muscles of paralarval squids produce lower peak isometric stresses and higher unloaded shortening velocities than the circular muscles of adult squids. To better understand how the structure and function of the circular muscles that provide power for jet locomotion change throughout ontogeny in squids, we examined (1) muscle properties using mechanical tests, (2) muscle morphometrics using histological techniques, and (3) the expression of isoforms of MHC using reverse transcription-polymerase chain reaction in paralarvae and, wherever data were lacking from previous studies, adult life stages of the long-finned squid Doryteuthis pealeii.

We performed our experiments on the paralarvae and sexually mature adults of the long-finned squid Doryteuthis (formerly Loligo) pealeii Leseuer. Egg fingers were collected from the waters near Woods Hole, MA, and Walpole, ME (USA) between 1 June and 30 September 2006 and 2007. They were housed in a flow-through seawater system with temperature and salinity ranging from 14 to 18°C and 28 to 32 p.p.t. (parts per thousand), respectively. Copepods, small mysid shrimp and larval brine shrimp were available as prey but feeding by the paralarvae was observed only rarely. Healthy paralarvae (i.e. squid that had normal coloration, that did not rest on the bottom of the tank and that were able to hold their vertical position in the water column via vigorous jetting) were used in experiments within 12 h of hatching. The mean ± s.d. dorsal mantle length (DML) of the paralarvae was 1.6±0.1 mm.

We caught male and female D. pealeii at night from lighted piers in South Bristol and Walpole, ME, USA, in July 2007. We trapped all of the squid with a 4.2 m-diameter cast net and then transported them immediately to the lab in 24 l buckets. Squid were housed in a 1 m × 2m × 0.5 m tank provided with flow-through seawater at 14–18°C and 28–32 p.p.t. salinity. The animals were fed small fish (Clupea spp.) daily and were used within four days of capture. We used only squid that were healthy, had no visible damage to the skin or mantle and swam vigorously. The adults ranged in size from 150 to 167 mm DML.

Paralarvae were anesthetized in cold seawater (3°C) (O'Dor and Shadwick, 1989; Bower et al., 1999), transferred to a drop of ice-cold modified squid saline solution containing (in mmol l⁻¹): NaCl (450), MgCl₂ ·6H₂O (10), Hepes (10), EGTA (10), pH adjusted to 7.8 with 2 mol l⁻¹ NaOH (Milligan et al., 1997), and then killed by impaling the brain with a straight pin. The posterior tip (i.e. about 0.5mm) of the mantle was sliced off. The mantle was then separated from the rest of the body, slit along the dorsal side and unrolled to form an approximately rectangular piece of tissue from what was once a hollow cylinder. The pen was removed and then each slit edge of the mantle was glued with Vetbond (3M, St Paul, MN, USA) to a T-shaped foil clip (Milligan et al., 1997). The clips were attached so that the long axes of the circular muscle fibers were parallel to the long axis of the preparation. The preparation was then transferred to a temperature-controlled bath. The bath was filled with standard squid saline containing (in mmol l⁻¹): NaCl (470), KCl (10), CaCl₂ (10), MgCl₂ ·6H₂O (50), glucose (20), Hepes (10), pH adjusted to 7.8 with 2 mol l⁻¹ NaOH (Milligan et al., 1997).

The muscle preparations were attached at one end to an ASI 400A force transducer (Aurora Scientific Inc., Aurora, ON, Canada) and at the other end to an ASI 322 high-speed length controller (Aurora Scientific Inc.). The length–force relationship of each tissue preparation was determined using supramaximal brief tetanic (2 ms pulses, 50 Hz, 200 ms) and twitch stimulations. Stimuli were provided via platinum foil electrodes that spanned the entire length and width of the preparation. The maximum unloaded shortening velocity (V_max) was determined at 0.9L₀ (where L₀ was the length of the preparation at which isometric tension was highest) using slack tests (Edman, 1979). We chose 0.9L₀ because passive tension was close to zero for all preparations at this length. We performed regular isometric control stimulations to monitor the health of the preparation. If the force produced during isometric contraction at L₀ decreased by more than 10%, we terminated the experiment and discarded all data collected subsequent to the previous control stimulation. We analyzed data from the muscle preparations of 19 paralarvae.

The muscles of the mantle are arranged primarily in two orientations: circumferentially (the circular muscles) and radially (the radial muscles) (Marceau, 1905; Williams, 1909; Young, 1938). Contraction of the circular muscles drives water out of the mantle cavity via the funnel while contraction of the radial muscles helps to refill the mantle cavity with water at the end of the power stroke (Young, 1938). As mentioned earlier, two types of circular muscle have been identified in loliginid and ommastrephid squids: CMP and SMR fibers (Bone et al., 1981; Mommsen et al., 1981) [terminology from Preuss et al. (Preuss et al., 1997)]. As in a previous study of the contractile properties of the CMP and SMR circular muscle fibers of adult D. pealeii (Thompson et al., 2008), we attempted to use a vibratome to cut sheets of circular muscles from the mantles of paralarvae. The paralarval mantles were thin (about 0.06 mm thick) and were much less stiff than the adult mantles, perhaps because they have significantly fewer intramuscular connective tissue fibers than the adults (see Thompson and Kier, 2001b). The result was that the vibratome blade tended to deform the paralarval mantles rather than slice them, and we had no success in isolating the SMR or CMP circular muscle fibers. The paralarval mantle preparations, therefore, contained intact SMR and CMP circular muscle fibers as well as radial fibers. We discuss the limitations of the paralarval intact preparations in the Discussion.

Following the mechanical tests, each preparation was pinned at L₀ in a Sylgard dish. Most of the preparations were fixed for 6–8 h (3% glutaraldehyde, 0.065 mol l⁻¹ phosphate buffer, 0.5% tannic acid and 6% sucrose) and then postfixed for 45 min at 4°C in a 1:1 solution of 2% osmium tetroxide and 2% potassium ferrocyanide in 0.13 mol l⁻¹ cacodylate buffer (Kier, 1985). The tissue was rinsed in chilled 0.13 mol l⁻¹ cacodylate buffer for 15 min, dehydrated in a graded series of acetones and embedded in epoxy resin (Embed 812, Electron Microscopy Sciences, Hatboro, PA, USA). A few of the preparations were fixed in a modified Karnovsky fixative (2% paraformaldehye, 3% glutaraldehyde, 0.065 mol l⁻¹ phosphate buffer and 2% sucrose), rinsed in phosphate buffer, dehydrated in ethanol and then stained in 1% phosphotungstic acid in absolute ethanol prior to embedding in LR White (Electron Microscopy Sciences, Hatfield, PA, USA). As is typical in squid muscle, dense bodies did not stain well in the 3% glutaraldehyde fixative (Kawaguti and Ikemoto, 1957; Ward and Wainwright, 1972; Moon and Hulbert, 1975; Bone et al., 1981; Kier, 1985) but the preservation of the myofilaments, sarcoplasmic reticulum (SR), plasma membrane and mitochondria was excellent. The tissues fixed in the modified Karnovsky and stained in phosphotungstic acid showed excellent preservation and staining of the dense bodies and thick filaments but the preservation and staining of the mitochondria, plasma membrane and SR were poor. For all tissues, transverse sections of the preparations were cut, stained and examined with brightfield and electron microscopy to measure the physiological cross section (pcs) of the circular muscle fibers (for details, see Thompson et al., 2008) and to explore the ultrastructure of the cells.

We prepared the mantle tissue of three adult D. pealeii for histology in order to examine circular muscle cell size, shape and structure in detail, and also to confirm the thick filament lengths reported by Thompson et al. (Thompson et al., 2008). We anesthetized the animals in a 1:1 solution of 7.5%MgCl₂ · 6H₂O:seawater (Messenger et al., 1985), decapitated them and then removed small blocks of mantle tissue from the ventral midline approximately ½ DML from the anterior edge of the mantle. The tissue blocks spanned the thickness of the mantle wall (i.e. they included radial muscle fibers plus both types of circular fibers). They were fixed, post-fixed and then embedded in epoxy resin as described above.

The protocol we followed for measuring thick filament lengths is described in detail elsewhere (Thompson and Kier, 2006; Thompson et al., 2008). Briefly, embedded tissue blocks were sectioned in a plane perpendicular to the longitudinal axis of the mantle (i.e. parallel to the long axes of the circular muscles) using a diamond knife. Thick sections (0.5–1 μm) were cut initially and stained in an aqueous solution of 0.1% Methylene Blue and 0.1% Azure II. Sections were then examined using brightfield microscopy to determine if the long axes of the circular muscle fibers were parallel to the knife edge. Once alignment was achieved, thin sections (silver interference color) were cut, mounted on grids and stained with 2% aqueous uranyl acetate (Bozzola and Russell, 1992) and 0.4% lead citrate (Venable and Coggeshall, 1965). Thin sections were examined with either a Zeiss EM-902 (Oberkochen, Baden-Wurttemberg, Germany) or JEOL JEM-100SX (Tokyo, Japan) transmission electron microscope and photographed.

The electron micrograph negatives were scanned at 4800 dpi (dots per inch), and thick filament lengths measured using ImageJ software (Abramoff et al., 2004). We measured thick filaments from at least five CMP and SMR muscle fibers per preparation, and a total of about 1200 thick filaments from paralarvae and 700 from the three adults. The mean thick filament length per animal was used for the statistical comparisons. Although great care was taken to align the long axes of the circular muscle fibers with the section plane, the thick filament lengths we report here may nevertheless be slight underestimates.

We used reverse transcription-polymerase chain reaction (RT-PCR) to determine if different isoforms of MHC are expressed in the mantle during ontogeny. We anesthetized two adult D. pealeii (155 and 167mm DML) in cold water, decapitated them, quickly excised small blocks of tissue and then placed the blocks into RNAlater™ (Ambion, Inc., Austin, TX, USA). The tissues included the central zone of the mantle (containing CMP and radial muscle fibers but no SMR fibers), portions of the mantle containing all muscle fiber types and the funnel retractor muscle. We also placed 25 whole paralarval mantles (including the fins, skin and small fragments of the ctenidia) in RNAlater™.

The tissues in RNAlater™ were stored at 2°C for several weeks then frozen in liquid nitrogen and pulverized using a chilled mortar and pestle. We then isolated total RNA using a kit (RNAqueous-4PCR kit, Ambion, Inc.) and performed RT-PCR. We used primer sequences designed by Kier and Schachat (Kier and Schachat, 2008) for the RT-PCR. The forward primer was 5′-AGCTTGGCTGGAAAGAAAGATAA-3′; the reverse primer was 5′-CAGCACCGGCAATTTTACCTT-3′. The primers bracketed the putative alternative RNA splice site identified by Matulef et al. (Matulef et al., 1998) in the funnel retractor muscle of D. pealeii. Expression of MHC isoform A or isoform B (sensuMatulef et al., 1998) was determined by the lengths of cDNA products, with the isoform A cDNA being 15 bp longer than that of isoform B. Following 40 cycles of PCR amplification, the products were diluted to 20 μg ml⁻¹ and then resolved by gel electrophoresis using a 4% high-resolution agarose gel in 1× Tris/Borate/EDTA buffer.

For comparison of morphological or muscle mechanics data between two sample populations, we used one-way analysis of variance (ANOVA). For comparisons among multiple groups, we used one-way ANOVA with Tukey's highly significant difference (HSD) post tests. Where indicated in the text, we compared thick filament length and muscle mechanics data from paralarvae collected in this study with similar data from adults reported by Thompson et al. (Thompson et al., 2008) and used here with their permission.

The mantle musculature of paralarval and adult D. pealeii differed notably in several ways. First, when circular muscle cells of paralarvae were cut transverse to their long axes, their shapes were variable (Fig. 1A,B). Although most muscle cells were roughly triangular or ovoid in cross section with a core of mitochondria (typical of squid mantle muscles), numerous muscle cells with crescent or other unusual cross-sectional shapes, and muscle cells with unequal distributions of mitochondria, were present. The frequency of these fibers relative to ‘normal’ circular muscle cells ranged from 1:6 to 1:10 among the 19 paralarvae we surveyed. We did not find crescent-shaped circular fibers or fibers with unusual cross-sectional shapes in the circular muscles of adult squid (Fig. 1C,D).

Second, the circular fibers of paralarvae were smaller than those of adults (Table 1). Among paralarvae, the CMP circular muscle fibers had significantly smaller total cross sectional areas than the SMR fibers (P<0.05), and this difference was also seen in the adults (P<0.0001). Both the SMR and CMP fibers of paralarvae had significantly smaller cross sectional areas than their counterparts in the adults (P<0.0001 for both comparisons).

Third, the general organization of the myofilaments in the cross sections of the circular muscle fibers was different between paralarvae and adults. When sectioned transverse to their long axes, the SMR and CMP muscle fibers of adults showed the banding pattern typical of obliquely striated muscle cells (Figs 1, 2). Briefly, the bands result from alternation of lightly stained regions containing only thin filaments, SR and irregularly spaced dense bodies with more darkly stained regions containing overlapping thick and thin filaments. In the paralarvae, this pattern was not apparent in most of the circular fibers (Fig. 1, Fig. 2A,B). Cross sections of most paralarval circular fibers were composed primarily of regions with both thick and thin filaments, with a few isolated regions composed of either thin filaments or thick filaments only (Fig. 2B). When sectioned parallel to their long axes, the SMR and CMP fibers of both paralarvae and adults exhibited the myofilament arrangements typical of obliquely striated muscle cells but the angle of obliquity of the dense bodies was much greater (i.e. closer to cross striation) in paralarvae than in adults (Figs 3, 4). The angle of obliquity is sensitive to the plane of section (see Rosenbluth, 1965), and thus the mean angle of obliquity must be reported with great caution. Nevertheless, among the hundreds of paralarval and adult muscle cells observed, the means ± s.d. of the 25 lowest angles of obliquity (for SMR and CMP fibers combined) visible in the micrographs were 22.8±2.4 deg. and 4.6±0.8 deg. for paralarvae and adults, respectively. In addition to the difference in angle of obliquity, the dense bodies in some of the paralarval circular fibers did not follow straight trajectories (Fig. 3C). This resulted in adjacent thick filaments being out of the register typically seen in obliquely striated cells. Indeed, some thick filaments within a few circular fibers followed trajectories that were almost perpendicular to the majority of the thick filaments (Fig. 3C).

Fourth, there was a difference in the distribution of SR between paralarvae and adults. In adults, the SR of the circular fibers was well-organized into three zones: a peripheral zone associated with the sarcolemma, a second, central zone surrounding the core of mitochondria, and a third, intramyoplasmic region of flattened tubules located in the plane of the dense bodies that appeared to connect the outer and central zones of SR (Fig. 2). In paralarvae, the outer and central zones of SR were well developed but the intramyoplasmic zone was greatly reduced in some cells or was scattered non-uniformly across the myoplasm (Fig. 2A,B). Whereas in adults the intramyoplasmic SR region was composed primarily of flattened tubules that occasionally swelled into a vesicle, it was composed primarily of large vesicles in the paralarvae, although flattened tubules were present in parts of a small number of paralarval fibers.

Fifth, among paralarvae the thick filaments of the SMR and CMP circular muscle fibers differed slightly in length (P=0.05, Figs 5, 6). During ontogeny, the lengths of the thick filaments of the SMR and CMP fibers increased significantly (P<0.0001 for both comparisons), with the SMR fibers increasing 3.7-fold in length and the CMP fibers 2.5-fold (data for adults from Thompson et al., 2008).

Finally, the thick filament diameter and thin:thick filament ratio changed during ontogeny. Among paralarvae, the CMP thick filaments were significantly (P=0.009) smaller in diameter than the SMR thick filaments. A similar difference was found among adults (P=0.0001) (Table 1). The CMP and SMR thick filaments of paralarvae were significantly smaller in diameter than their counterparts in adults (P=0.037 and P=0.0014 for SMR and CMP fibers, respectively) (Table 1). The thin:thick filament ratio was significantly lower in the CMP fibers of adults compared with the SMR fibers of adults (P=0.05) and with both the CMP and SMR fibers of paralarvae (P<0.05). The thin:thick filament ratios of the SMR and CMP fibers of paralarvae did not differ significantly from each other (P>0.4) (Table 1).

All of the slack tests resulted in tight linear relationships between step length and force recovery time, with the R² values for the least-squares linear regressions of all preparations exceeding 0.98 (not shown). For brief tetanus (2ms pulse width, 50Hz, 200ms duration) at 20°C, the mean maximum unloaded shortening velocity (V_max) of the paralarval preparations was 9.1 L₀ s⁻¹, where L₀ was the preparation length that generated the peak isometric stress (Fig. 7). V_max ranged from 4.3 to 14.8 L₀ s⁻¹.

In the mantle preparations of the paralarvae, the maximum isometric stress in brief tetanus (P₀) was 119±15 mN mm⁻² pcs with a range of 85–160 mN mm⁻² pcs.

The temporal aspects of brief tetani are summarized in Table 2.

Both isoforms of MHC were expressed in the mantles of adults and paralarvae. Isoform ‘A’ (204 bp) composed at least 90% of the RT-PCR product in all animals, based on relative staining intensity of the bands in the agarose gel (Fig. 8).

The fine structure of the circular muscles that provide power for jet locomotion changed in several ways during ontogeny in D. pealeii. These include a difference in myofilament organization, the shapes and distribution of SR and the lengths of the thick filaments. All observed changes may have implications for contractile properties of the circular muscles.

When cut transverse to their long axes, the CMP and SMR circular muscle fibers of adults showed the alternating band pattern typical of obliquely striated muscle cells (e.g. Kawaguti and Ikemoto, 1957; Rosenbluth, 1965; Rosenbluth, 1968; Ward and Wainwright, 1972; Moon and Hulbert, 1975; Lanzavecchia, 1977) but it was absent from virtually all of the CMP and SMR circular fibers of the paralarvae (Figs 1, 2). The pattern visible from the transverse plane of an obliquely striated cell results from the staggered arrangement of the myofilament lattice proteins (see Rosenbluth, 1965). Thus, a change in the banding pattern suggests a difference in the arrangement of the myofilaments.

We propose that the difference in the arrangement of myofilaments between paralarvae and adults results from a greater angle of obliquity of the dense bodies in the circular muscle fibers of paralarvae (Fig. 3). The angle of obliquity increases from 6–12 deg. at rest to 14–18 deg. during contraction in Loligo sp. (Hanson and Lowy, 1957; Rosenbluth, 1965; Millman, 1967) and can change with section plane (Rosenbluth, 1965); thus, making comparisons of angle of obliquity between animals is difficult. Nevertheless, the paralarval fibers consistently exhibited substantially higher angles of obliquity even when sections from highly stretched (prior to fixation) paralarval preparations were compared with highly contracted (during fixation) adult preparations. Higher angles of obliquity make it likely that sections transverse to the long axis of the fiber will primarily contain A-bands (i.e. overlapping thick and thin filaments; Fig. 4). Examining sections that are slightly off-axis or sectioning cells that contracted during the fixation process decreases the probability of seeing the banding typical of adults (Fig. 4).

The effect of this change in the organization of the dense bodies on muscle performance is uncertain because the functional implications of oblique striation are not yet clear. Rosenbluth suggested that oblique striation constrains shortening velocity because the friction associated with shearing of the thick filaments may act as a brake (Rosenbluth, 1968). The higher angle of obliquity in paralarval fibers should reduce this friction and result in higher V_max relative to otherwise identical fibers that have a lower angle of obliquity.

The difference in the arrangement of myofilaments between paralarvae and adults we found is consistent with the photomicrographs of the circular muscle fibers in paralarval and juvenile D. (formerly Loligo) opalescens in Preuss et al. (Preuss et al., 1997), who also noted a difference in the organization of the myofilaments between the SMR fibers and CMP fibers of paralarvae. Preuss and colleagues reported that the myofilaments of paralarvae were ‘poorly organized’ relative to those of adults, and also cautiously suggested that the SMR fibers of paralarval D. opalescens were cross-striated (Preuss et al., 1997). We found no evidence to support the suggestion that either the SMR or CMP fibers of D. pealeii paralarvae are cross-striated (Figs 2, 3). Indeed, the photomicrograph in Preuss et al. (Preuss et al., 1997) of a paralarval SMR circular muscle fiber seems to display the alternating band pattern typical of obliquely striated fibers.

The ontogenetic differences in intramyoplasmic SR organization and shape (i.e. rounded vesicles in paralarvae vs flattened tubules in adults) did not affect the latency between stimulation and the rise of force in the muscle preparations. The delay between stimulation and force rise in the paralarval muscle preparations, 4 ms, was not significantly (P=0.78) shorter than the delay in the CMP (4.4 ms) and SMR (4.6 ms) circular muscle preparations of adult D. pealeii reported by Thompson et al. (Thompson et al., 2008). The paralarval preparations achieved peak isometric force significantly faster than both the CMP and SMR fiber preparations of adults (Thompson et al., 2008) (Table 2) but this probably resulted from the lower absolute forces produced by the paralarval preparations. Indeed, the rate of isometric force rise in the paralarval preparations was substantially lower than in either the adult CMP or SMR preparations (data not shown). This probably occurred because the significantly longer thick filaments of the adult circular fibers permitted activation of more cross bridges per unit time. The paralarval muscle preparations relaxed to 50% P₀ about 3 times more rapidly than the CMP preparations of adults (Table 2). We have not identified a mechanism that explains this difference, although it is possible that the rate of calcium return to the SR is greater in paralarvae. Both the faster times to peak force and the time to 50% P₀ suggest that the paralarvae should be able to produce jet pulses with greater frequency than adults, and this is consistent with the results of Bartol et al. (Bartol et al., 2009a) who found shorter jet pulse duration in paralarval D. pealeii than in adults (see also Anderson and Grosenbaugh, 2005).

Numerous crescent-shaped cells and cells with unequal distributions of mitochondria were present in the mantles of paralarvae but not in those of adults. Bone et al. noted crescent-shaped circular muscle cells in the mantle of Alloteuthis subulata and suggested that these represented early stages of new fibers formed from cell division (Bone et al., 1981). The frequent occurrence of such cells in the mantle of paralarvae in which growth and, presumably, rates of cell division are high (Moltschaniwskj, 1994; Moltschaniwskj, 1995; Moltschaniwskj, 1997), lends support to the hypothesis of Bone et al. (Bone et al., 1981).

There was a significant increase in the lengths of the thick myofilaments of both the CMP and SMR circular muscle fibers (Figs 3, 5) through ontogeny from paralarvae to adults. This increase was similar to that found in the CMP and SMR fibers of another loliginid squid, S. lessoniana (Thompson and Kier, 2006), and suggests that such changes may be common among squids. The ontogenetic increase in the thick filament lengths of the circular muscles of D. pealeii and S. lessoniana are not unique among cephalopod muscles (Kier, 1996), and Kier and colleagues have suggested that loliginid squids modulate muscle contractile properties by altering thick filament length alone (Kier, 1985; Kier, 1996; Kier and Schachat, 1992; Kier and Curtin, 2002; Thompson and Kier, 2006; Kier and Schachat, 2008; Thompson et al., 2008).

The shortening velocity of striated muscles, including the obliquely striated circular muscles, depends on a variety of factors: the lengths of the thick filaments and sarcomeres, the load on the muscle and the rate of cross-bridge cycling (e.g. Bárány, 1967; Josephson, 1975). In general, thick filament length is inversely proportional to unloaded shortening velocity and directly proportional to peak isometric stress (Millman, 1967; Josephson, 1975; Kier and Curtin, 2002).

In adult D. pealeii, there is a strong correlation between thick filament length and contractile properties in the SMR and CMP circular muscle fibers (Thompson et al., 2008). The mean thick filament length of the adult SMR fibers (3.12 μm) is 1.75-fold greater than that of the adult CMP fibers (1.78 μm) (Thompson et al., 2008). Thus, if other aspects of the two fiber types are the same, and if the relationship between thick filament length and isometric stress is linear, the P₀ of the SMR fibers should be 1.75 times greater than the P₀ of the CMP fibers. Thompson et al. (Thompson et al., 2008) found that the P₀ of the adult SMR fibers was approximately 1.5 times greater; correcting for differences in the size of the core of mitochondria between the two fiber types resulted in a P₀ that was about 2 times greater. By similar reasoning, the V_max of the adult SMR fibers should be 1.75 times slower than the CMP fibers, and it was approximately 2 times slower (Thompson et al., 2008).

Assuming that other aspects of the circular muscles do not change during ontogeny, we therefore hypothesized that the circular muscles of paralarvae produce lower peak isometric stresses and higher unloaded shortening velocities than those of adults. Before analyzing this hypothesis in light of our results, we first discuss the limitations of our mechanical testing methods.

The thin, deformable mantles of paralarvae thwarted our attempts to investigate the contractile properties of preparations composed solely of CMP or solely of SMR circular fibers. Thus, we studied mantle preparations that contained both types of circular muscle cells in addition to radial fibers. Electrical stimulation of the paralarval muscle preparations probably resulted in contraction of all muscle fibers, and this complicates the comparisons of muscle contractile properties. Nevertheless, several factors support the validity of our experimental approach. First, the thick filaments of the CMP and SMR fibers of paralarval D. pealeii are roughly similar in length (Figs 3, 5). If, as our results and those from other studies suggest (Kier and Curtin, 2002; Thompson et al., 2008), the contractile properties of cephalopod obliquely striated muscles are determined largely by thick filament length, then the contractile properties of paralarval CMP and SMR circular muscle fibers are likely to be similar. Nonetheless, our experimental design masked any differences in contractile properties between the two fiber types.

Second, the radial muscle fibers are antagonists to the circular fibers and their activity therefore may have decreased both the speed of shortening and the peak isometric stress of the circular muscle fibers. We attempted to estimate the magnitude of this error in paralarvae as follows. We made mantle muscle preparations from two small adult D. pealeii (90 mm and 110 mm DML) that had intact radial fibers and both SMR and CMP circular fibers. We then measured the contractile properties of these preparations using the same methods described earlier in this paper. By comparing the results with those of Thompson et al. (Thompson et al., 2008) for SMR and CMP preparations in which the radial fibers were severed and, therefore, inactivated, we were able to estimate the error introduced by having intact radial muscle fibers in the preparations. The mean V_max from the two intact adult preparations was 4.5±0.7 L₀ s⁻¹, about 12% lower than the mean V_max for the CMP preparations of adult squid (Thompson et al., 2008). The mean P₀ for the adult whole mantle preparations was 306±15 mN mm⁻² pcs, about 8.5% less than the value for the isolated SMR preparations from adults (Thompson et al., 2008). This suggests that our muscle mechanics data underestimate the true V_max and P₀ by 8–12%.

Third, Lowy and Millman measured the contractile properties of the funnel retractor muscles of the octopuses Octopus vulgaris and Eledone moschata and the cuttlefish Sepia officinalis (Lowy and Millman, 1962). The funnel retractor is similar to the mantle in that (1) it is supported by a muscular hydrostat with muscle fibers oriented in two directions – parallel and transverse to the long axis of the muscle; (2) the parallel and transverse fibers are antagonists; and (3) the parallel fibers compose the majority of the muscle, just as the circular fibers compose the majority of the mantle musculature (Kier and Thompson, 2003). Lowy and Millman measured the contractile properties of the parallel fibers in intact funnel retractor muscles and preparations in which they pared the muscle down its long axis, thereby severing the transverse fibers (Lowy and Millman, 1962). They found no differences in the properties of the two types of preparations with the exception that the pared preparations produced higher forces per unit cross sectional area.

Despite these reassurances we nevertheless emphasize that the V_max and P₀ data we report for paralarvae are conservative estimates.

Even with the conservative estimate of unloaded shortening velocity in the paralarvae, the data support the hypothesis that V_max is lower in adult SMR and CMP circular fibers than in the paralarval circular fibers. The mean V_max of the SMR and CMP fibers of adult D. pealeii in brief tetanus was 2.4 and 5.1 L₀ s⁻¹, respectively, where L₀ was the preparation length that yielded the highest force (Thompson et al., 2008). The mean V_max of the adult SMR and CMP preparations was 3.8- and 1.8-fold lower, respectively, than the V_max of the paralarval preparations (9.1 L₀ s⁻¹).

The difference in V_max between paralarvae and adults was less than that predicted based solely on the ontogenetic increase in thick filament length. The thick filaments of adult CMP fibers were 2.5 times longer than those of paralarvae, yet the V_max of adult CMP fibers was only 1.8 times lower (Thompson et al., 2008). Several variables confound this comparison. First, paralarval preparations were composed of intact circular and radial muscle fibers whereas the radial fibers were severed in the adult preparations (Thompson et al., 2008). If the whole mantle preparations of adults described above provide a realistic effect of the radial fibers on V_max, then we need to correct the V_max of paralarvae by 12%. Doing so raises the V_max to 10.2 L₀ s⁻¹ and increases the speed of shortening of the paralarval preparations relative to adult CMP preparations to 2-fold. Second, the SMR fibers of the paralarvae had thick filament lengths that were 1.16 times longer than the CMP fibers (Figs 3, 5). If V_max scales linearly with thick filament length in the paralarvae, the slightly slower SMR fibers may have added a load to the CMP fibers; thus, lowering V_max. Third, the paralarvae had numerous crescent-shaped and other unusually shaped circular muscle fibers, and the ratio of these to ‘normal’ fibers varied from 1:6 to 1:10 among the paralarval preparations. If these cells did not generate as much force per unit cross section as the ‘normal’ circular fibers, the greater the ratio of these cells in a preparation, the greater the load on the ‘normal’ cells and, therefore, the lower the V_max. Indeed, there was greater variation in the paralarval than in the adult preparations, with V_max ranging from 4.3 to 14.8 L₀ s⁻¹ (Fig. 5) in paralarvae. Although undetected damage to the preparations during dissection may account for some of this variation, we believe that most of it reflects natural variation in the contractile properties of the circular muscles. Paralarvae in the lab differed greatly in their swimming performance, and differences in the contractile properties of the mantle musculature, among other factors, contributed to this variation in locomotion.

Some of the paralarvae had a V_max that was nearly 3-fold higher than that of adults. Because thick filament length was relatively uniform among all the paralarvae (Figs 3, 5), this calls into question the notion that thick filament length alone explains the ontogeny of V_max. The passive tension in the paralarval and adult CMP preparations was minimal at L₀ (data not shown). Thus, it is unlikely that elastic recoil during the slack tests contributed to the V_max data we present.

The P₀ we recorded for the paralarval preparations was much lower than that reported by Thompson et al. (Thompson et al., 2008) for the SMR and CMP preparations of adult D. pealeii. It is, however, very similar to the P₀ (131±56 mN mm⁻² psc) of the transverse muscle fibers of the tentacles of adult D. pealeii (Kier and Curtin, 2002). The tentacle muscle fibers are cross-striated (Kier, 1982) and, like the SMR and CMP circular fibers of paralarvae, have short (0.8 μm) thick filaments (Kier and Curtin, 2002).

The dimensions of the myofilaments are but one factor that may affect the contractile properties of the SMR and CMP muscle fibers. The contractile properties of striated muscle fibers may be altered in a variety of ways, including, but not limited to, expression of different isoforms of the myofilament lattice proteins (e.g. Kendrick-Jones et al., 1976; Sweeney et al., 1988; Marden et al., 1998; Schiaffino and Reggiani, 1996; Toniolo et al., 2007). In contrast to mammals, in which nine MHC genes are expressed in skeletal and cardiac muscle and 38 orthologs of MHC have been putatively identified (Maccatrozzo et al., 2007), only a single MHC gene has been identified to date in loliginid squid but it has two splice variants (‘A’ and ‘B’) (Matulef et al., 1998). The splice site occurs along one of the nucleotide binding regions of the protein, so the two splice variants may well differ in their ATPase activity (Matulef et al., 1998).

Both predicted splice variants of the MHC gene were expressed in the mantles of paralarvae and adults. The results of the RT-PCR are consistent with the interpretation that isoform A is expressed in much higher levels than isoform B. Within the mantle of an adult, there is no evidence to suggest different expression patterns of the two isoforms between the central zone alone (i.e. the CMP fiber layer and the radial fibers) and the whole mantle (i.e. the CMP and SMR circular fibers and the radial fibers). Thus, our results are consistent with the hypothesis that the contractile properties of the circular muscles are modulated via changes in thick filament length, and not via changes in the molecular species of MHC present.

In their study of the transverse muscle fibers of the arms and tentacles of D. pealeii, Kier and Schachat found no changes in the expression of MHC between the two appendages (Kier and Schachat, 1992) and, similar to our results, that isoform A is expressed at higher levels (>90%) than isoform B in both appendages (Kier and Schachat, 2008).

As striking as these results are, we nevertheless cannot state firmly that changes in the biochemistry of the circular muscles do not underlie the significant ontogenetic decline in V_max. Work to sequence the genome of D. pealeii is far from complete. Although cephalopods may not have the same diversity of MHC genes as vertebrates and arthropods, additional squid MHC genes may yet be discovered. In addition, expression of different isoforms of other myofilament lattice proteins (e.g. Kendrick-Jones et al., 1976; Sweeney et al., 1988; Lowey et al., 1993; Marden et al., 1998; Schiaffino and Reggiani, 1996; Toniolo et al., 2007) during ontogeny may affect the contractile properties of the circular fibers.

It is important to note that we do not know where each isoform of MHC is expressed in the mantle or any of the other tissues we examined. The isoforms may be co-expressed within a single muscle fiber, as sometimes occurs in the striated muscle fibers of mammals (Staron and Pette, 1987; Schiaffino and Reggiani, 1996) and the nematode Caenorhabditis elegans (Miller et al., 1986), or a single muscle fiber may express only one of the isoforms, as is often the case in the striated fibers of mammals and fish (Goldspink, 1998). We do not know if the two isoforms of MHC confer different contractile properties on the muscles in which they are expressed. It is quite possible that the two isoforms differ in ATPase activity given that the alternative splice site falls along one of the regions of the MHC gene that codes for the putative nucleotide binding region of the protein (Matulef et al., 1998), and changes in the structure of this region can affect ATP turnover rates in the MHC isoforms of scallops (Perreault-Micale et al., 1996). Despite these important gaps in knowledge, the molecular data we present do not illustrate an obvious difference in isoform expression pattern during ontogeny.

If the mechanics of the circular muscles change during ontogeny, an interesting question is raised: is a change in circular muscle mechanics associated with ontogenetic changes in jet dynamics or jet propulsive efficiency? The answer is that yes, indeed, jet locomotion changes during ontogeny in a manner consistent with predictions based on the concurrent changes in mantle muscle structure and function we report. The rationale for this statement is provided below.

Many paralarval squids, including D. pealeii, are planktonic and appear to spend much of their time maintaining vertical position in the water column using vertically directed jets (Boletzky, 1974; Zeidberg and Hamner, 2002). To increase speed, as they would for prey capture or to undergo daily vertical migrations (Zeidberg and Hamner, 2002), paralarval D. pealeii increase the amplitude, but not the rate, of mantle contraction (Bartol et al., 2009a). In other words, paralarvae employ a relatively constant jet velocity across the range of speeds (5–27 DML s⁻¹) considered by Bartol et al. (Bartol et al., 2009a), varying speed by altering the total mass of water expelled per mantle contraction. By contrast, adult D. pealeii increase jetting speed by adjusting both the rate and amplitude of mantle contraction (Anderson and Grosenbaugh, 2005) (J.T.T. and K. R. Taylor, unpublished). Increasing the rate of mantle contraction requires application of greater pressure to the water in the mantle cavity, and this pressure is derived, ultimately, from active stress produced by the circular muscle fibers. The short thick filaments of the paralarval SMR and CMP circular fibers result in significantly lower peak stress production, and this limits the ability of paralarvae to produce high mantle cavity pressures.

Paralarval squid swim in an intermediate Reynolds number (Re) fluid regime in which inertial forces dominate fluid dynamic drag but viscous forces are considerable. Within this environment, they utilize an intermittent or pulsed jet for propulsion. Squids form a jet pulse by first expanding the mantle radially and filling the mantle cavity. Once the mantle cavity is full, the circular muscles contract, thereby increasing the pressure in the mantle cavity and driving water out of the mantle cavity through the funnel. Repetition of this cycle results in a pulsed jet. Studies of mechanically generated pulsed jets at high (>1000) Re issuing into quiescent fluid have demonstrated the important role played by the formation of vortex rings at the leading edge of each jet pulse. Vortex rings provide additional thrust through nozzle exit over-pressure, i.e. fluid pressure above the local ambient pressure during jet ejection, which develops as additional ambient fluid downstream is accelerated by entrainment and added mass effects during ring formation (Krueger and Gharib, 2003; Krueger and Gharib 2005; Choutapalli, 2006). As a result of the additional input through over-pressure, pulsed jets can produce substantially more thrust than equivalent steady jets (Krueger and Gharib, 2005). Recent results have shown that paralarval D. pealeii also develop vortex-ring-like structures during jetting and hence may benefit from enhanced thrust that is useful for combating the drag challenges experienced at intermediate Re (Bartol et al., 2008; Bartol et al., 2009a).

To capitalize on over-pressure benefits provided by pulsing, it is essential to use short pulses. In particular, the ratio of the length of the ejected plug of fluid (L) to the nozzle diameter (D) should be small enough so that isolated vortex rings are formed with each pulse. If L/D is too large, the vortex ring will pinch off from the generating jet (Gharib et al., 1998) and the relative contribution of over-pressure related to the vortex ring formation will decline as the remainder of the pulse is ejected as a relatively steady jet (Krueger and Gharib, 2003). Not only do short jet pulses amplify the thrust benefit of pulsing but recent results from juvenile and adult L. brevis (Bartol et al., 2009b) and ‘Robosquid’ (Nichols et al., 2008) have shown that propulsive efficiency can also be improved by relying on short jet pulses during pulsed jetting. Indeed, paralarval D. pealeii seem to rely almost exclusively on relatively short jet pulses, which appears to be a contributing factor leading to their remarkably high propulsive efficiencies despite the intermediate Re at which they swim (Bartol et al., 2008; Bartol et al., 2009a).

In the context of swimming at intermediate Re and the considerable benefit pulsed jetting with short jet pulses appears to provide, the present results of increased V_max and short circular muscle relaxation times in paralarval D. pealeii, as compared with their adult counterparts, is particularly intriguing. Although we do not know the load on the circular musculature during jetting or how it scales with body size, the significantly higher V_max of the paralarvae undoubtedly contributes to their ability to produce the short jet pulses that are important for maintaining high propulsive efficiency. In addition, the higher V_max and shorter relaxation times may also facilitate the high jet pulsing frequencies that are often observed in paralarvae (I.K.B. and J.T.T., personal observation) (Bartol et al., 2009a), which allow for more continuous swimming and shorter coasting phase durations within each jet cycle. The ability to swim more continuously with short periodic refill periods during jet cycles is an important consideration at intermediate Re where the considerable viscosity inhibits coasting.

High jet propulsive efficiency actually may be more important for paralarvae than older life-history stages because their fins play such a minor role in propulsion relative to the contributions of the highly efficient fins of juveniles and adults (Bartol et al., 2008), requiring paralarvae to rely more heavily on their jet for propulsion. Thus, ultrastructural modifications of the circular muscle fibers over ontogeny may have evolved in response to the need for the rapid mantle contractions and short jet periods that lead to high propulsive efficiency.

We thank the staff of the Darling Marine Center (DMC), especially Mr Tim Miller and Ms Linda Healy, for help in finding and housing squid, and also for lab space and housing for J.T.T. and A.B. We also thank Oluwarotimi Adesina, Stephen Chorney and Karen Ross for help in maintaining squid at the DMC. We thank Dr Robert Jinks for assistance with the electron microscopy and sharing the ultramicrotomes in his lab. The research was supported by National Science Foundation grant IOS-0446081 to J.T.T., I.K.B. and P.S.K., and start-up funds from Franklin & Marshall College.

 
• CMP

  central mitochondria-poor circular muscle fibers

 
• D

  diameter of jet nozzle

 
• DML

  dorsal mantle length

 
• L

  length of the ejected plug of fluid

 
• L₀

  muscle preparation length that produced the peak isometric stress in brief tetanus

 
• MHC

  myosin heavy chain

 
• p

  pressure inside the mantle cavity

 
• P₀

  maximum isometric stress in brief tetanus

 
• pcs

  physiological cross sectional area

 
• r_i

  radius of the inner surface of the mantle

 
• Re

  Reynolds number

 
• RT-PCR

  reverse transcriptase-polymerase chain reaction

 
• SMR

  superficial mitochondria-rich circular muscle fibers

 
• SR

  sarcoplasmic reticulum

 
• t

  thickness of the mantle wall

 
• V_max

  maximum unloaded shortening velocity

 
• σ

  isometric stress produced by the circular muscles