D.R. is a 27-year-old man, who presents to the nurse practitioner at the Family Care Clinic complaining of increasing SOB, wheezing, fatigue, cough, stuffy nose, watery eyes, and postnasal

……………………………………………………..

Convergent evolution in mechanical design of lamnid sharks and tunas Jeanine M. Donley1, Chugey A. Sepulveda1, Peter Konstantinidis2, Sven Gemballa2 & Robert E. Shadwick1

1Marine Biology Research Division, Scripps Institution of Oceanography, University of California San Diego, La Jolla, California 92093-0202, USA 2Department of Zoology, University of Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen, Germany ………………………………………………………………………………………………………………………………………………………..

The evolution of ‘thunniform’ body shapes in several different groups of vertebrates, including whales, ichthyosaurs1 and sev- eral species of large pelagic fishes2 supports the view that physical and hydromechanical demands provided important selection pressures to optimize body design for locomotion during ver- tebrate evolution. Recognition of morphological similarities between lamnid sharks (the most well known being the great white and the mako) and tunas has led to a general expectation that they also have converged in their functional design; however, no quantitative data exist on the mechanical performance of the locomotor system in lamnid sharks. Here we examine the swim- ming kinematics, in vivo muscle dynamics and functional mor- phology of the force-transmission system in a lamnid shark, and show that the evolutionary convergence in body shape and mechanical design between the distantly related lamnids and tunas is muchmore than skin deep; it extends to the depths of the myotendinous architecture and the mechanical basis for propul- sive movements. We demonstrate that not only have lamnids and tunas converged to amuch greater extent than previously known, but they have also developed morphological and functional adaptations in their locomotor systems that are unlike virtually all other fishes. During their 400 million years of independent evolution, sharks

and bony fishes have diverged in many fundamental aspects of their anatomy and physiology. However, two groups of dominant open- ocean predators, the lamnid sharks and the tunas, evenwhen looked at superficially, display remarkably similar morphological special- izations related to locomotion3–12 (Fig. 1a). The shared character- istics in these distantly related groups that distinguish them from virtually all other fish have arisen independently, probably as the result of selection for fast and continuous locomotion. Moreover, in both lamnids and tunas, the aerobic (red) musculature that powers cruise swimming is concentrated in a more medial (closer to the

letters to nature

NATURE |VOL 429 | 6 MAY 2004 | www.nature.com/nature 61© 2004 Nature Publishing Group

backbone) and anterior position compared with the relatively uniform and superficial position in other fishes; the body tempera- ture is elevated above that of the surrounding water, facilitated by counter-current heat exchangers associated with the internal red muscle; and themuscle segments (myotomes) are highly elongated3. Recent studies on tunas revealed a host of unique functional

adaptations in their locomotor system that distinguish tunas from

all other bony fishes4,13. Similar investigations on swimming lamnid sharks are lacking because of the difficulty in handling such large and dangerous predators; thus, the dynamic properties of the lamnid locomotor system remain unknown. This paper presents in vivo quantitative measurements of swimming kinematics and muscle dynamics, and analysis of the morphology of the force- transmission system in a lamnid shark.

First, we examined the kinematics of steady swimming in the shortfinmako shark (Isurus oxyrinchus). Lateral displacement of the dorsal midline as a function of body position was calculated from dorsal video images of mako sharks swimming under controlled conditions in a swim tunnel (Fig. 1b; see also Supplementary Video). Descriptions of undulatory swimming modes in fishes are based on the proportion of the body that participates in the lateral thrust-producing movements14, and can be distinguished by differ- ences in patterns of lateral displacement along the body, as shown in Fig. 1c. Compared with other teleosts, tunas exhibit the least undulatory (thunniform) mode of locomotion, in which lateral movements are largely confined to the caudal region where body mass is reduced by tapering. The lateral displacement data pre- sented here show that the lamnid shark kinematically resembles tuna more than other sharks15 or subcarangiform teleosts: the degree of lateral motion along the mako shark’s body from ,0.4L to ,0.8L (where L is total body length) is relatively small, demon- strating that lamnids swim using a thunniform-like mode. The amplitude of lateral motion increases substantially beyond ,0.8L where the body tapers to the narrow caudal peduncle. Lamnids, like tunas, have the least lateral motion in the mid-body region where the bulk of the muscle resides and have a reduced body mass in the caudal region where lateral amplitudes are high, both being features that match predictions for enhancing hydromechanical efficiency of swimming9.

Because lamnids have both internal red muscle and a thunni- form-like swimmingmode, we tested the possibility that shortening of the red muscle would be physically uncoupled from deformation of the adjacent skin, backbone and white muscle, a unique func- tional property of red muscle that has been observed so far only in tunas. Whereas in most teleosts superficial red muscle fibres con- tract sequentially to cause a posteriorly travelling wave of local body bending16–18, the internal position of red muscle in tunas allows them to abandon this pattern of undulation and adopt a novel mechanism to project red muscle action to posterior regions of the body4,19, thereby facilitating thunniform kinematics. In the mako shark we used sonomicrometry to record instantaneous muscle segment length changes during steady swimming as well as during passive, simulated swimming movements induced under anaesthe- sia. The temporal relationship between red and white muscle shortening was measured to determine whether the action of these two muscle masses is synchronized, as in most fish, or uncoupled, as in tunas. During passive swimming movements, peaks in red muscle strain (that is, relative length change) were in phase with peaks in adjacent white muscle strain (Fig. 2a). Thus, when the body bends passively red muscle shortens synchronously with the surrounding white muscle and skin, as one would expect. In contrast, during active swimming the peaks in red muscle strain were delayed relative to peaks in white muscle strain (Fig. 2b). By cross-correlation analysis we determined that the mean phase shift between simultaneous recordings of red andwhitemuscle strainwas 90ms (or,10% of the tail-beat cycle), with one individual as high as 174ms (,17% of tail-beat cycle). The observed phase shift indicates that during steady swimming the red muscle is indeed physically uncoupled from the surrounding tissues and contracts in phase with body bending at a more posterior location.

On the basis of a wave velocity of about 1 l s21 in sharks swimming with a tail-beat frequency of about 1Hz, the red muscle in the mid-body region will be shortening in phase with bending of the backbone 10–17% closer to the tail. A consequence is that

Figure 1 Features of thunniform body shape and patterns of lateral undulation during

steady swimming. a, Thunniform body shape in lamnid sharks25 (right) and tunas26 (left).

Note the highly streamlined fusiform body shape that minimizes pressure drag, stiff,

high-aspect-ratio hydrofoil caudal fin that produces thrust by a hydrodynamic-lift

mechanism, and dorsoventrally flattened and enlarged caudal keel that decreases drag

produced by lateral movement of the peduncle6–8. b, Dorsal image of steady swimming I. oxyrinchus. Scale bar: 10 cm. c, Lateral displacement, relative to the amplitude of lateral motion at the tail tip, versus axial position in the mako shark (n ! 5; red symbols), tuna

(red dashed line27), leopard shark (black solid line15) and subcarangiform teleosts

(black dashed line; modified from ref. 13). This graph emphasizes differences in

displacement in the mid-body where most of the variation among swimming modes

occurs. Comparison of the mako shark and tuna illustrates the similarity in their swimming

mode, where lateral undulations are largely confined to the caudal region, indicated by

shading in dorsal outlines. The reduction in lateral motion in the mid-body region afforded

by their internal red muscle and modified force-transmission system is evident when

comparing the mako shark and tuna to sharks and bony fishes with less-specialized

swimming modes such as the subcarangiforms shown here.

letters to nature

NATURE | VOL 429 | 6 MAY 2004 | www.nature.com/nature62 © 2004 Nature Publishing Group

during short portions of each contraction cycle the red muscle will be lengthening while the adjacent white fibres are shortening, and vice versa. The loose sheath of connective tissue that surrounds the red muscle mass (see Fig. 3d) may facilitate this shearing. These results are remarkably similar to those described in tunas, where the

deep red muscle shortens and transmits force and displacement to more posterior regions of the body rather than effecting local bending4,19. Thus, the same mechanism enables tunas and lamnid sharks to swim with the relatively stiff-bodied thunniform mode of propulsion; even though the bulk of aerobic muscle is located more

Figure 2 Simultaneous recordings of muscle strain (segment length change/mean length) of red (red trace) and adjacent white (grey trace) muscle during passive simulated

swimming movements (a) and active steady swimming (0.5 l s21) (b) in the mako shark.

During active swimming, as verified by red muscle activity (EMG trace), shortening in the

red muscle is delayed relative to the white muscle and is therefore in phase with lateral

motion in more posterior positions.

Figure 3 Collagenous architecture of myosepta of I. oxyrinchus. a, Oblique view focusing

on the hypaxial part between 0.54L and 0.74L (coloured inset). The elongated anterior

pointing cone (AC) of one myoseptum is shown. It intersects with the red musculature

(pale red area) and contains the hypaxial lateral tendon (dark red). The hypaxial lateral

tendon extends between the tip of the anterior pointing cone and the ventral posterior

cone (VPC). The hypaxial lateral tendon of a more anterior myoseptum is also shown

without its myoseptal sheet. The anterior part of this tendon is cut at its intersection with

the transverse plane (blue). b, Excised area of myosepta between the anterior pointing cone and ventral posterior cone flattened out under polarized light. Pathways of

collagenous structures are shown in white. The hypaxial lateral tendon extends between

the red arrows. Dashed line, excision line from vertebral axis; thick white line, excision line

from skin; dotted line, excision line from remaining dorsal part of the myosepta, equivalent

to the dotted line in a; white arrowheads, intersection line of myosepta and loose connective tissue surrounding red muscle. c, d, Transverse sections of left side. Concentric rings of myosepta indicate nesting anterior pointing cones. c, Fresh section, 0.6L, showing red muscle with sections of hypaxial lateral tendons (white) within the red

muscle. d, Histological section at 0.54L. Red muscle is separated from surrounding white

muscle by a sheath of loose connective tissue. Numbers 1–12 indicate anterior portions of

12 hypaxial lateral tendons present in red muscle, whereas numbers 13–24 indicate

posterior portions of 12 additional hypaxial lateral tendons present in white muscle,

meaning that a single tendon covers 24 segments. The inset shows a detailed view of red

muscle and hypaxial lateral tendons (stained orange).

letters to nature

NATURE |VOL 429 | 6 MAY 2004 | www.nature.com/nature 63© 2004 Nature Publishing Group

anteriorly than in less derived species, the lateral motion it produces is primarily focused to the caudal region. Our morphological investigations demonstrate that the anatomi-

cal specializations associated with the force-transmission system are also convergent. We used a new combination of techniques to explore the three-dimensional morphology of the tendinous con- nective tissue linkages (myosepta) that transmit muscular forces to the skin and backbone, and their relationship to the internal red muscle in the mako shark. In principle, the three-dimensional shape of myotomes and their

associatedmyosepta inmako sharks resembles the regular pattern in gnathostome fishes20, which includes a main anterior cone and a dorsal and ventral posterior cone (see Supplementary Fig. 1 for myoseptal parts of gnathostomes andmako sharks). Additionally, in mako sharks two secondary anterior cones are present at the dorsal- and ventral-most part of the myoseptum. The redmuscle is situated in the lower part of the main anterior cone (Fig. 3a; see also Supplementary Fig. 1d, e) where its fibres insert into the collagenous myoseptum. In particular, red muscle fibres insert into the anterior half of a myoseptal tendon (Fig. 3a). This tendon runs from the tip of the main anterior cone through the red muscle towards its end within the white muscle at the ventral posterior cone (Fig. 3a, b). It clearly represents the homologue of the hypaxial lateral tendon in gnathostome fishes (Supplementary Fig. 1). In mako sharks, this tendon is extremely prominent and elongated when compared with other fishes. In fact, myoseptal tendons as long and as distinct as those associated with the red muscle in the mako shark have never been reported in any shark species. We measured tendon lengths as long as 0.19L in the posterior region of the body (Supplementary Fig. 1e). The sonomicrometric results suggest that the action of the red muscle is directed posteriorly along the body by 10–17%. The measured tendon lengths accord well with the values predicted from sonomicrometry, suggesting that the hypaxial lateral tendon is responsible for transmitting red muscle forces posteriorly. The prominent tendons of the posterior body are gradually developed along a rostrocaudal gradient from shorter (,0.06L; Supplementary Fig. 1d) and less distinct hypax- ial lateral tendons of anterior myosepta to longer tendons in the posterior. In tunas, distinct and elongated tendons have also been discov-

ered and have been hypothesized to transfer forces from the red muscle to the axial skeleton, and thus provide the anatomical basis for force transmission from the anterior to the caudal region21. As in mako sharks, the available length measurements and sonomicro- metric data are in good accordance (0.18L experimentally and 0.16L morphologically)22. Although in tunas the primary force-transmit- ting tendons are in the horizontal septum, in the mako shark, as in other sharks23, we found the horizontal septum to be reduced in the posterior half of the body. Instead, the primary linkage to the tail appears to be the hypaxial lateral tendons. Interestingly, although the posterior oblique tendons in tunas and hypaxial lateral tendons in mako sharks provide the same function, they have different anatomical origins. Through distinct evolutionary pathways lamnid sharks and tunas

have converged on the same mechanical design principle, that of having internalized red muscle associated with a highly derived force-transmission system, two features that form the basis for their thunniform swimming mode. Our study shows that not only have the physical demands of the external environment sculpted the body shapes of large pelagic cruisers, but also the internal physiology and morphology of their complex locomotor systems has been fine- tuned over the course of their evolution. A

Methods Shortfinmako sharks (I. oxyrinchus, family Lamnidae) ranging in size from 80 to 112 cm L (total body length) were collected by hook and line off the coast of Southern California and transported to the laboratory facilities at Scripps Institution of Oceanography (SIO) in a

transport chamber equipped with circulating aerated sea water. Once at SIO, the sharks were placed into a large 3,000-l swim tunnel for an acclimation period of several hours before experimentation. All procedures in capture, maintenance and experimentation followed the guidelines of the University of California, San Diego Institutional Animal Care and Use Committee.

In vivo muscle dynamics To examine the dynamics of red andwhitemuscle contractions during swimming, we used electromyography (EMG) and sonomicrometry, a technique for measuring distances in which piezoelectric crystals transmit and receive ultrasonic pulses. Pairs of sonomicrometric crystals were implanted into the deep red and adjacent white muscle to record instantaneous changes in muscle segment length (strain) during active periods of steady swimming as well as during passive, simulated swimming movements induced under anaesthesia. Surgery was performed on anaesthetized individuals partially submerged in a seawater bath according to procedures described previously15. Crystal pairs were implanted approximately 15mm apart along the longitudinal axis of the body and the leads were loosely anchored to the skin with surgical sutures. To verify the passive and active states of the redmuscle, electrical activity was recorded using pairs of electrodes implanted approximately 2mm apart directly bisecting the crystal pairs. After surgery, the sharks were placed into the swim tunnel and allowed to recover before data collection. In the recovery period we recorded red and white muscle strain during passive, simulated swimming movements induced by gentle side-to-side motions of the centre of mass that generated body undulation. Additionally, we recorded and analysed 30–50 consecutive tail-beat cycles for each individual while the shark swam steadily at approximately 0.5 l s21. To measure the relative timing of red and white muscle strain (phase shift), a cross-correlation analysis was performed using waveforms containing approximately ten consecutive tail-beat cycles. Mean phase shift presented in the text represents a mean of five individuals. Sonomicrometric and EMG signals were recorded at 500Hz.

To correlate measurements of local muscle activation and strain with patterns of body bending, five mako sharks were videotaped while swimming against a current of known velocity in the swim tunnel. To synchronize the collection of sonomicrometric, EMG and video recordings, a flashing red diode was recorded in the video sequences and its excitation voltage was recorded with the sonomicrometric and EMG data. Kinematic analysis follows procedures described previously15,24.

Morphology We used a combination of clearing and staining, microdissection, polarized light microscopy, standard histology and computer-based three-dimensional reconstruction to explore the three-dimensional morphology of the tendinous connective tissue linkages (myosepta). A small body segment (0.54–0.55L) of a formalin-fixed mako specimen (65 cm total length) was prepared for standard histology (paraffin embedding; Azan– Domagk staining, slice thickness of 15 mm). The two remaining parts (0–0.54L and 0.55–1.00L) were carefully skinned, stained for cartilage with Alcian blue 8GX (Merck) and then cleared according to a recently described procedure20. Microdissections on the myoseptal system were carried out using fine microsurgery tools. Myosepta or parts of myosepta from all body regions were removed subsequently. Three-dimensional shape of the myosepta was documented by a camera lucida, and tendon lengths and rostrocaudal extensions of complete myosepta were measured in situ. Removed myosepta were photographed under polarized light to visualize the collagen fibre pathways and tendons (Zeiss Stemi 2000C with Fuji X digital camera HC300Z; 1,000 £ 1,450 pixels). Additionally, the distribution of red muscle and its relation to myoseptal cones along the body was examined. A three-dimensional reconstruction was obtained from histological sections. Major landmarks (vertebrae, neural arches, vertical septum, abdominal cavity, tip of main anterior, secondary anterior and ventral posterior cone, sections of tendons, position of red muscle) were digitized and aligned using SurfDriver 3.5.3. Maxon Cinema 4D (Release 6) was used for choosing an adequate perspective and rendering. The obtained three-dimensional view was edited by Adobe Photoshop (final shading, adding of myoseptal shape).

Received 23 January; accepted 25 February 2004; doi:10.1038/nature02435.

1. Motani, R. Scaling effects in caudal fin propulsion and the speed of ichthyosaurs.Nature 415, 309–312

(2002).

2. Lindsey, C. C. in Fish Physiology Vol. VII Locomotion (eds Hoar, W. S. & Randall, D. J.) 1–100

(Academic, New York, 1978).

3. Bernal, D., Dickson, K. A., Shadwick, R. E. & Graham, J. B. Analysis of the evolutionary convergence

for high performance swimming in lamnid sharks and tunas. Comp. Biochem. Physiol. 129, 695–726

(2001).

4. Katz, S. L., Syme, D. A. & Shadwick, R. E. High speed swimming: enhanced power in yellowfin tuna.

Nature 410, 770–771 (2001).

5. Carey, F. G., Teal, J. M., Kanwisher, J. W., Lawson, K. D. & Beckett, J. S. Warm-bodied fish. Am. Zool.

11, 137–143 (1971).

6. Carey, F. G., Casey, J. G., Pratt, H. L., Urquhart, D. & McCosker, J. E. Temperature, heat production

and heat exchange in lamnid sharks. Mem. S. Calif. Acad. Sci. 9, 92–108 (1985).

7. Collette, B. B. in The Physiological Ecology of Tunas (eds Sharp, G. D. & Dizon, A. E.) 7–40 (Academic,

New York, 1978).

8. Graham, J. B. & Dickson, K. A. The evolution of thunniform locomotion and heat conservation in

scombrid fishes: new insights based on the morphology of Allothunnus fallai. Zool. J. Linn. Soc. 129,

419–466 (2000).

9. Lighthill, M. J. Hydromechanics of aquatic animal propulsion—a survey. Annu. Rev. Fluid Mech. 1,

413–446 (1969).

10. Magnuson, J. J. in Fish Physiology Vol. VII Locomotion (eds Hoar, W. S. & Randall, D. J.) 239–313

(Academic, New York, 1978).

letters to nature

NATURE | VOL 429 | 6 MAY 2004 | www.nature.com/nature64 © 2004 Nature Publishing Group

11. Reif, W. E. & Weishampel, D. B. Anatomy and mechanics of the lunate tail in lamnid sharks. Zool. Jb.

Anat. 114, 221–234 (1986).

12. Sfakiotakis,M., Lane, D.M. &Davies, J. B. C. Review of fish swimmingmodes for aquatic locomotion.

J. Ocean. Eng. 24, 237–252 (1999).

13. Altringham, J. D. & Shadwick, R. E. in Tuna: Physiological Ecology and Evolution (eds Block, B. A. &

Stevens, E. D.) 313–344 (Academic, San Diego, 2001).

14. Webb, P. W. Hydrodynamics and energetics of fish propulsion. Bull. Fish. Res. Bd Can. 190, 1–159

(1975).

15. Donley, J. M. & Shadwick, R. E. Steady swimming muscle dynamics in the leopard shark Triakis

semifasciata. J. Exp. Biol. 206, 1117–1126 (2003).

16. Coughlin, D. J., Valdes, L. & Rome, L. C. Muscle length changes during swimming in scup:

sonomicrometry verifies the anatomical high-speed cine technique. J. Exp. Biol. 199, 459–463 (1996).

17. Katz, S. L., Shadwick, R. E. & Rapoport, H. S. Muscle strain histories in swimming milkfish in steady

and sprinting gaits. J. Exp. Biol. 202, 529–541 (1999).

18. Shadwick, R. E., Steffenson, J. F., Katz, S. L. & Knower, T. Muscle dynamics in fish during steady

swimming. Am. Zool. 38, 755–770 (1998).

19. Shadwick, R. E., Katz, S. L., Korsmeyer, K., Knower, T. & Covell, J. W. Muscle dynamics in skipjack

tuna Katsuwonus pelamus: timing of red muscle shortening and body curvature during steady

swimming. J. Exp. Biol. 202, 2139–2150 (1999).

20. Gemballa, S. et al. Evolutionary transformations of myoseptal tendons in gnathostomes. Proc. R. Soc.

Lond. B 270, 1229–1235 (2003).

21. Westneat, M. W., Hoese, W., Pell, C. A. & Wainwright, S. A. The horizontal septum: mechanisms of

force transfer in locomotion of scombrid fishes (Scombridae, Perciformes). J. Morphol. 217, 183–204

(1993).

22. Katz, S. L. Design of heterothermic muscle in fish. J. Exp. Biol. 205, 2251–2266 (2002).

23. Gemballa, S., Hagen, K., Röder, K., Rolf, M. & Treiber, K. Structure and evolution of the horizontal

septum in vertebrates. J. Evol. Biol. 16, 966–975 (2003).

24. Jayne, B. C. & Lauder, G. V. Speed effects on midline kinematics during steady undulatory swimming

in largemouth bass, Micropterus salmoides. J. Exp. Biol. 198, 585–602 (1995).

25. Compagno, L. J. V. in FAO Identification Guide for Fishery Purposes. The Living Marine Resources of the

Western Central Pacific (eds Carpenter, K. E. & Niem, V. H.) 1274–1278 (FAO, Rome, 1998).

26. Joseph, J., Klawe, W. & Murphey, P. Tuna and Billfish—Fish Without A Country 1–69 (Inter-American

Tropical Tuna Commission, La Jolla, California, 1988).

27. Knower, T. Biomechanics of Thunniform Swimming Thesis, Univ. California, San Diego (1998).

Supplementary Information accompanies the paper on www.nature.com/nature.

Acknowledgements We thank A. Biewener, J. Gosline, J. B. Graham, S. Vogel and N. Holland for discussion and reviews. Funding was provided by NSF and UC Regents.

Competing interests statement The authors declare that they have no competing financial interests.

Correspondence and requests for materials should be addressed to J.D. ([email protected]).

i

letters to nature